2019-05-29 14:18:09 +00:00
|
|
|
// SPDX-License-Identifier: GPL-2.0-only
|
2014-09-26 07:16:57 +00:00
|
|
|
/* Copyright (c) 2011-2014 PLUMgrid, http://plumgrid.com
|
|
|
|
|
*/
|
|
|
|
|
#include <linux/bpf.h>
|
2021-12-16 02:55:37 +00:00
|
|
|
#include <linux/bpf-cgroup.h>
|
bpf: add initial bpf tracepoints
This work adds a number of tracepoints to paths that are either
considered slow-path or exception-like states, where monitoring or
inspecting them would be desirable.
For bpf(2) syscall, tracepoints have been placed for main commands
when they succeed. In XDP case, tracepoint is for exceptions, that
is, f.e. on abnormal BPF program exit such as unknown or XDP_ABORTED
return code, or when error occurs during XDP_TX action and the packet
could not be forwarded.
Both have been split into separate event headers, and can be further
extended. Worst case, if they unexpectedly should get into our way in
future, they can also removed [1]. Of course, these tracepoints (like
any other) can be analyzed by eBPF itself, etc. Example output:
# ./perf record -a -e bpf:* sleep 10
# ./perf script
sock_example 6197 [005] 283.980322: bpf:bpf_map_create: map type=ARRAY ufd=4 key=4 val=8 max=256 flags=0
sock_example 6197 [005] 283.980721: bpf:bpf_prog_load: prog=a5ea8fa30ea6849c type=SOCKET_FILTER ufd=5
sock_example 6197 [005] 283.988423: bpf:bpf_prog_get_type: prog=a5ea8fa30ea6849c type=SOCKET_FILTER
sock_example 6197 [005] 283.988443: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[06 00 00 00] val=[00 00 00 00 00 00 00 00]
[...]
sock_example 6197 [005] 288.990868: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[01 00 00 00] val=[14 00 00 00 00 00 00 00]
swapper 0 [005] 289.338243: bpf:bpf_prog_put_rcu: prog=a5ea8fa30ea6849c type=SOCKET_FILTER
[1] https://lwn.net/Articles/705270/
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-25 01:28:18 +00:00
|
|
|
#include <linux/bpf_trace.h>
|
2018-05-27 11:24:09 +00:00
|
|
|
#include <linux/bpf_lirc.h>
|
2020-09-29 12:45:51 +00:00
|
|
|
#include <linux/bpf_verifier.h>
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
#include <linux/bsearch.h>
|
2018-04-18 22:56:01 +00:00
|
|
|
#include <linux/btf.h>
|
2014-09-26 07:16:57 +00:00
|
|
|
#include <linux/syscalls.h>
|
|
|
|
|
#include <linux/slab.h>
|
2017-02-08 17:51:30 +00:00
|
|
|
#include <linux/sched/signal.h>
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
#include <linux/vmalloc.h>
|
|
|
|
|
#include <linux/mmzone.h>
|
2014-09-26 07:16:57 +00:00
|
|
|
#include <linux/anon_inodes.h>
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
#include <linux/fdtable.h>
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
#include <linux/file.h>
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
#include <linux/fs.h>
|
2014-09-26 07:17:00 +00:00
|
|
|
#include <linux/license.h>
|
|
|
|
|
#include <linux/filter.h>
|
2016-11-13 18:44:03 +00:00
|
|
|
#include <linux/kernel.h>
|
2017-06-05 19:15:46 +00:00
|
|
|
#include <linux/idr.h>
|
2017-09-27 21:37:52 +00:00
|
|
|
#include <linux/cred.h>
|
|
|
|
|
#include <linux/timekeeping.h>
|
|
|
|
|
#include <linux/ctype.h>
|
2018-05-03 16:04:59 +00:00
|
|
|
#include <linux/nospec.h>
|
2019-12-06 21:49:34 +00:00
|
|
|
#include <linux/audit.h>
|
2019-10-16 03:24:58 +00:00
|
|
|
#include <uapi/linux/btf.h>
|
2020-06-09 04:32:38 +00:00
|
|
|
#include <linux/pgtable.h>
|
2020-03-29 00:43:52 +00:00
|
|
|
#include <linux/bpf_lsm.h>
|
bpf: Implement BPF ring buffer and verifier support for it
This commit adds a new MPSC ring buffer implementation into BPF ecosystem,
which allows multiple CPUs to submit data to a single shared ring buffer. On
the consumption side, only single consumer is assumed.
Motivation
----------
There are two distinctive motivators for this work, which are not satisfied by
existing perf buffer, which prompted creation of a new ring buffer
implementation.
- more efficient memory utilization by sharing ring buffer across CPUs;
- preserving ordering of events that happen sequentially in time, even
across multiple CPUs (e.g., fork/exec/exit events for a task).
These two problems are independent, but perf buffer fails to satisfy both.
Both are a result of a choice to have per-CPU perf ring buffer. Both can be
also solved by having an MPSC implementation of ring buffer. The ordering
problem could technically be solved for perf buffer with some in-kernel
counting, but given the first one requires an MPSC buffer, the same solution
would solve the second problem automatically.
Semantics and APIs
------------------
Single ring buffer is presented to BPF programs as an instance of BPF map of
type BPF_MAP_TYPE_RINGBUF. Two other alternatives considered, but ultimately
rejected.
One way would be to, similar to BPF_MAP_TYPE_PERF_EVENT_ARRAY, make
BPF_MAP_TYPE_RINGBUF could represent an array of ring buffers, but not enforce
"same CPU only" rule. This would be more familiar interface compatible with
existing perf buffer use in BPF, but would fail if application needed more
advanced logic to lookup ring buffer by arbitrary key. HASH_OF_MAPS addresses
this with current approach. Additionally, given the performance of BPF
ringbuf, many use cases would just opt into a simple single ring buffer shared
among all CPUs, for which current approach would be an overkill.
Another approach could introduce a new concept, alongside BPF map, to
represent generic "container" object, which doesn't necessarily have key/value
interface with lookup/update/delete operations. This approach would add a lot
of extra infrastructure that has to be built for observability and verifier
support. It would also add another concept that BPF developers would have to
familiarize themselves with, new syntax in libbpf, etc. But then would really
provide no additional benefits over the approach of using a map.
BPF_MAP_TYPE_RINGBUF doesn't support lookup/update/delete operations, but so
doesn't few other map types (e.g., queue and stack; array doesn't support
delete, etc).
The approach chosen has an advantage of re-using existing BPF map
infrastructure (introspection APIs in kernel, libbpf support, etc), being
familiar concept (no need to teach users a new type of object in BPF program),
and utilizing existing tooling (bpftool). For common scenario of using
a single ring buffer for all CPUs, it's as simple and straightforward, as
would be with a dedicated "container" object. On the other hand, by being
a map, it can be combined with ARRAY_OF_MAPS and HASH_OF_MAPS map-in-maps to
implement a wide variety of topologies, from one ring buffer for each CPU
(e.g., as a replacement for perf buffer use cases), to a complicated
application hashing/sharding of ring buffers (e.g., having a small pool of
ring buffers with hashed task's tgid being a look up key to preserve order,
but reduce contention).
Key and value sizes are enforced to be zero. max_entries is used to specify
the size of ring buffer and has to be a power of 2 value.
There are a bunch of similarities between perf buffer
(BPF_MAP_TYPE_PERF_EVENT_ARRAY) and new BPF ring buffer semantics:
- variable-length records;
- if there is no more space left in ring buffer, reservation fails, no
blocking;
- memory-mappable data area for user-space applications for ease of
consumption and high performance;
- epoll notifications for new incoming data;
- but still the ability to do busy polling for new data to achieve the
lowest latency, if necessary.
BPF ringbuf provides two sets of APIs to BPF programs:
- bpf_ringbuf_output() allows to *copy* data from one place to a ring
buffer, similarly to bpf_perf_event_output();
- bpf_ringbuf_reserve()/bpf_ringbuf_commit()/bpf_ringbuf_discard() APIs
split the whole process into two steps. First, a fixed amount of space is
reserved. If successful, a pointer to a data inside ring buffer data area
is returned, which BPF programs can use similarly to a data inside
array/hash maps. Once ready, this piece of memory is either committed or
discarded. Discard is similar to commit, but makes consumer ignore the
record.
bpf_ringbuf_output() has disadvantage of incurring extra memory copy, because
record has to be prepared in some other place first. But it allows to submit
records of the length that's not known to verifier beforehand. It also closely
matches bpf_perf_event_output(), so will simplify migration significantly.
bpf_ringbuf_reserve() avoids the extra copy of memory by providing a memory
pointer directly to ring buffer memory. In a lot of cases records are larger
than BPF stack space allows, so many programs have use extra per-CPU array as
a temporary heap for preparing sample. bpf_ringbuf_reserve() avoid this needs
completely. But in exchange, it only allows a known constant size of memory to
be reserved, such that verifier can verify that BPF program can't access
memory outside its reserved record space. bpf_ringbuf_output(), while slightly
slower due to extra memory copy, covers some use cases that are not suitable
for bpf_ringbuf_reserve().
The difference between commit and discard is very small. Discard just marks
a record as discarded, and such records are supposed to be ignored by consumer
code. Discard is useful for some advanced use-cases, such as ensuring
all-or-nothing multi-record submission, or emulating temporary malloc()/free()
within single BPF program invocation.
Each reserved record is tracked by verifier through existing
reference-tracking logic, similar to socket ref-tracking. It is thus
impossible to reserve a record, but forget to submit (or discard) it.
bpf_ringbuf_query() helper allows to query various properties of ring buffer.
Currently 4 are supported:
- BPF_RB_AVAIL_DATA returns amount of unconsumed data in ring buffer;
- BPF_RB_RING_SIZE returns the size of ring buffer;
- BPF_RB_CONS_POS/BPF_RB_PROD_POS returns current logical possition of
consumer/producer, respectively.
Returned values are momentarily snapshots of ring buffer state and could be
off by the time helper returns, so this should be used only for
debugging/reporting reasons or for implementing various heuristics, that take
into account highly-changeable nature of some of those characteristics.
One such heuristic might involve more fine-grained control over poll/epoll
notifications about new data availability in ring buffer. Together with
BPF_RB_NO_WAKEUP/BPF_RB_FORCE_WAKEUP flags for output/commit/discard helpers,
it allows BPF program a high degree of control and, e.g., more efficient
batched notifications. Default self-balancing strategy, though, should be
adequate for most applications and will work reliable and efficiently already.
Design and implementation
-------------------------
This reserve/commit schema allows a natural way for multiple producers, either
on different CPUs or even on the same CPU/in the same BPF program, to reserve
independent records and work with them without blocking other producers. This
means that if BPF program was interruped by another BPF program sharing the
same ring buffer, they will both get a record reserved (provided there is
enough space left) and can work with it and submit it independently. This
applies to NMI context as well, except that due to using a spinlock during
reservation, in NMI context, bpf_ringbuf_reserve() might fail to get a lock,
in which case reservation will fail even if ring buffer is not full.
The ring buffer itself internally is implemented as a power-of-2 sized
circular buffer, with two logical and ever-increasing counters (which might
wrap around on 32-bit architectures, that's not a problem):
- consumer counter shows up to which logical position consumer consumed the
data;
- producer counter denotes amount of data reserved by all producers.
Each time a record is reserved, producer that "owns" the record will
successfully advance producer counter. At that point, data is still not yet
ready to be consumed, though. Each record has 8 byte header, which contains
the length of reserved record, as well as two extra bits: busy bit to denote
that record is still being worked on, and discard bit, which might be set at
commit time if record is discarded. In the latter case, consumer is supposed
to skip the record and move on to the next one. Record header also encodes
record's relative offset from the beginning of ring buffer data area (in
pages). This allows bpf_ringbuf_commit()/bpf_ringbuf_discard() to accept only
the pointer to the record itself, without requiring also the pointer to ring
buffer itself. Ring buffer memory location will be restored from record
metadata header. This significantly simplifies verifier, as well as improving
API usability.
Producer counter increments are serialized under spinlock, so there is
a strict ordering between reservations. Commits, on the other hand, are
completely lockless and independent. All records become available to consumer
in the order of reservations, but only after all previous records where
already committed. It is thus possible for slow producers to temporarily hold
off submitted records, that were reserved later.
Reservation/commit/consumer protocol is verified by litmus tests in
Documentation/litmus-test/bpf-rb.
One interesting implementation bit, that significantly simplifies (and thus
speeds up as well) implementation of both producers and consumers is how data
area is mapped twice contiguously back-to-back in the virtual memory. This
allows to not take any special measures for samples that have to wrap around
at the end of the circular buffer data area, because the next page after the
last data page would be first data page again, and thus the sample will still
appear completely contiguous in virtual memory. See comment and a simple ASCII
diagram showing this visually in bpf_ringbuf_area_alloc().
Another feature that distinguishes BPF ringbuf from perf ring buffer is
a self-pacing notifications of new data being availability.
bpf_ringbuf_commit() implementation will send a notification of new record
being available after commit only if consumer has already caught up right up
to the record being committed. If not, consumer still has to catch up and thus
will see new data anyways without needing an extra poll notification.
Benchmarks (see tools/testing/selftests/bpf/benchs/bench_ringbuf.c) show that
this allows to achieve a very high throughput without having to resort to
tricks like "notify only every Nth sample", which are necessary with perf
buffer. For extreme cases, when BPF program wants more manual control of
notifications, commit/discard/output helpers accept BPF_RB_NO_WAKEUP and
BPF_RB_FORCE_WAKEUP flags, which give full control over notifications of data
availability, but require extra caution and diligence in using this API.
Comparison to alternatives
--------------------------
Before considering implementing BPF ring buffer from scratch existing
alternatives in kernel were evaluated, but didn't seem to meet the needs. They
largely fell into few categores:
- per-CPU buffers (perf, ftrace, etc), which don't satisfy two motivations
outlined above (ordering and memory consumption);
- linked list-based implementations; while some were multi-producer designs,
consuming these from user-space would be very complicated and most
probably not performant; memory-mapping contiguous piece of memory is
simpler and more performant for user-space consumers;
- io_uring is SPSC, but also requires fixed-sized elements. Naively turning
SPSC queue into MPSC w/ lock would have subpar performance compared to
locked reserve + lockless commit, as with BPF ring buffer. Fixed sized
elements would be too limiting for BPF programs, given existing BPF
programs heavily rely on variable-sized perf buffer already;
- specialized implementations (like a new printk ring buffer, [0]) with lots
of printk-specific limitations and implications, that didn't seem to fit
well for intended use with BPF programs.
[0] https://lwn.net/Articles/779550/
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200529075424.3139988-2-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2020-05-29 07:54:20 +00:00
|
|
|
#include <linux/poll.h>
|
2022-04-24 21:48:53 +00:00
|
|
|
#include <linux/sort.h>
|
2020-05-31 08:28:36 +00:00
|
|
|
#include <linux/bpf-netns.h>
|
2020-08-27 22:01:11 +00:00
|
|
|
#include <linux/rcupdate_trace.h>
|
2020-12-01 21:58:32 +00:00
|
|
|
#include <linux/memcontrol.h>
|
2022-03-16 12:24:09 +00:00
|
|
|
#include <linux/trace_events.h>
|
2014-09-26 07:16:57 +00:00
|
|
|
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
#include <net/netfilter/nf_bpf_link.h>
|
|
|
|
|
#include <net/netkit.h>
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
#include <net/tcx.h>
|
|
|
|
|
|
2019-11-22 20:07:58 +00:00
|
|
|
#define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
|
|
|
|
|
(map)->map_type == BPF_MAP_TYPE_CGROUP_ARRAY || \
|
|
|
|
|
(map)->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS)
|
|
|
|
|
#define IS_FD_PROG_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PROG_ARRAY)
|
2017-06-28 06:08:34 +00:00
|
|
|
#define IS_FD_HASH(map) ((map)->map_type == BPF_MAP_TYPE_HASH_OF_MAPS)
|
2019-11-22 20:07:58 +00:00
|
|
|
#define IS_FD_MAP(map) (IS_FD_ARRAY(map) || IS_FD_PROG_ARRAY(map) || \
|
|
|
|
|
IS_FD_HASH(map))
|
2017-06-28 06:08:34 +00:00
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
#define BPF_OBJ_FLAG_MASK (BPF_F_RDONLY | BPF_F_WRONLY)
|
|
|
|
|
|
2016-03-08 05:57:13 +00:00
|
|
|
DEFINE_PER_CPU(int, bpf_prog_active);
|
2017-06-05 19:15:46 +00:00
|
|
|
static DEFINE_IDR(prog_idr);
|
|
|
|
|
static DEFINE_SPINLOCK(prog_idr_lock);
|
2017-06-05 19:15:47 +00:00
|
|
|
static DEFINE_IDR(map_idr);
|
|
|
|
|
static DEFINE_SPINLOCK(map_idr_lock);
|
2020-04-29 00:16:06 +00:00
|
|
|
static DEFINE_IDR(link_idr);
|
|
|
|
|
static DEFINE_SPINLOCK(link_idr_lock);
|
2016-03-08 05:57:13 +00:00
|
|
|
|
2021-05-11 20:35:17 +00:00
|
|
|
int sysctl_unprivileged_bpf_disabled __read_mostly =
|
|
|
|
|
IS_BUILTIN(CONFIG_BPF_UNPRIV_DEFAULT_OFF) ? 2 : 0;
|
bpf: enable non-root eBPF programs
In order to let unprivileged users load and execute eBPF programs
teach verifier to prevent pointer leaks.
Verifier will prevent
- any arithmetic on pointers
(except R10+Imm which is used to compute stack addresses)
- comparison of pointers
(except if (map_value_ptr == 0) ... )
- passing pointers to helper functions
- indirectly passing pointers in stack to helper functions
- returning pointer from bpf program
- storing pointers into ctx or maps
Spill/fill of pointers into stack is allowed, but mangling
of pointers stored in the stack or reading them byte by byte is not.
Within bpf programs the pointers do exist, since programs need to
be able to access maps, pass skb pointer to LD_ABS insns, etc
but programs cannot pass such pointer values to the outside
or obfuscate them.
Only allow BPF_PROG_TYPE_SOCKET_FILTER unprivileged programs,
so that socket filters (tcpdump), af_packet (quic acceleration)
and future kcm can use it.
tracing and tc cls/act program types still require root permissions,
since tracing actually needs to be able to see all kernel pointers
and tc is for root only.
For example, the following unprivileged socket filter program is allowed:
int bpf_prog1(struct __sk_buff *skb)
{
u32 index = load_byte(skb, ETH_HLEN + offsetof(struct iphdr, protocol));
u64 *value = bpf_map_lookup_elem(&my_map, &index);
if (value)
*value += skb->len;
return 0;
}
but the following program is not:
int bpf_prog1(struct __sk_buff *skb)
{
u32 index = load_byte(skb, ETH_HLEN + offsetof(struct iphdr, protocol));
u64 *value = bpf_map_lookup_elem(&my_map, &index);
if (value)
*value += (u64) skb;
return 0;
}
since it would leak the kernel address into the map.
Unprivileged socket filter bpf programs have access to the
following helper functions:
- map lookup/update/delete (but they cannot store kernel pointers into them)
- get_random (it's already exposed to unprivileged user space)
- get_smp_processor_id
- tail_call into another socket filter program
- ktime_get_ns
The feature is controlled by sysctl kernel.unprivileged_bpf_disabled.
This toggle defaults to off (0), but can be set true (1). Once true,
bpf programs and maps cannot be accessed from unprivileged process,
and the toggle cannot be set back to false.
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-08 05:23:21 +00:00
|
|
|
|
2017-04-11 13:34:58 +00:00
|
|
|
static const struct bpf_map_ops * const bpf_map_types[] = {
|
2019-11-14 18:57:15 +00:00
|
|
|
#define BPF_PROG_TYPE(_id, _name, prog_ctx_type, kern_ctx_type)
|
2017-04-11 13:34:58 +00:00
|
|
|
#define BPF_MAP_TYPE(_id, _ops) \
|
|
|
|
|
[_id] = &_ops,
|
2020-04-29 00:16:08 +00:00
|
|
|
#define BPF_LINK_TYPE(_id, _name)
|
2017-04-11 13:34:58 +00:00
|
|
|
#include <linux/bpf_types.h>
|
|
|
|
|
#undef BPF_PROG_TYPE
|
|
|
|
|
#undef BPF_MAP_TYPE
|
2020-04-29 00:16:08 +00:00
|
|
|
#undef BPF_LINK_TYPE
|
2017-04-11 13:34:58 +00:00
|
|
|
};
|
2014-09-26 07:16:57 +00:00
|
|
|
|
2017-08-07 18:45:20 +00:00
|
|
|
/*
|
|
|
|
|
* If we're handed a bigger struct than we know of, ensure all the unknown bits
|
|
|
|
|
* are 0 - i.e. new user-space does not rely on any kernel feature extensions
|
|
|
|
|
* we don't know about yet.
|
|
|
|
|
*
|
|
|
|
|
* There is a ToCToU between this function call and the following
|
|
|
|
|
* copy_from_user() call. However, this is not a concern since this function is
|
|
|
|
|
* meant to be a future-proofing of bits.
|
|
|
|
|
*/
|
2021-05-14 00:36:05 +00:00
|
|
|
int bpf_check_uarg_tail_zero(bpfptr_t uaddr,
|
2018-05-22 22:03:31 +00:00
|
|
|
size_t expected_size,
|
|
|
|
|
size_t actual_size)
|
2017-08-07 18:45:19 +00:00
|
|
|
{
|
2020-05-08 04:16:31 +00:00
|
|
|
int res;
|
2017-08-07 18:45:19 +00:00
|
|
|
|
2017-08-07 18:45:20 +00:00
|
|
|
if (unlikely(actual_size > PAGE_SIZE)) /* silly large */
|
|
|
|
|
return -E2BIG;
|
|
|
|
|
|
2017-08-07 18:45:19 +00:00
|
|
|
if (actual_size <= expected_size)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
if (uaddr.is_kernel)
|
|
|
|
|
res = memchr_inv(uaddr.kernel + expected_size, 0,
|
|
|
|
|
actual_size - expected_size) == NULL;
|
|
|
|
|
else
|
|
|
|
|
res = check_zeroed_user(uaddr.user + expected_size,
|
|
|
|
|
actual_size - expected_size);
|
2020-05-08 04:16:31 +00:00
|
|
|
if (res < 0)
|
|
|
|
|
return res;
|
|
|
|
|
return res ? 0 : -E2BIG;
|
2017-08-07 18:45:19 +00:00
|
|
|
}
|
|
|
|
|
|
2018-01-12 04:29:09 +00:00
|
|
|
const struct bpf_map_ops bpf_map_offload_ops = {
|
2020-08-28 01:18:06 +00:00
|
|
|
.map_meta_equal = bpf_map_meta_equal,
|
2018-01-12 04:29:09 +00:00
|
|
|
.map_alloc = bpf_map_offload_map_alloc,
|
|
|
|
|
.map_free = bpf_map_offload_map_free,
|
2018-08-11 23:59:17 +00:00
|
|
|
.map_check_btf = map_check_no_btf,
|
2023-03-05 12:46:14 +00:00
|
|
|
.map_mem_usage = bpf_map_offload_map_mem_usage,
|
2018-01-12 04:29:09 +00:00
|
|
|
};
|
|
|
|
|
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
static void bpf_map_write_active_inc(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
atomic64_inc(&map->writecnt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_map_write_active_dec(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
atomic64_dec(&map->writecnt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool bpf_map_write_active(const struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
return atomic64_read(&map->writecnt) != 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-12-01 21:58:58 +00:00
|
|
|
static u32 bpf_map_value_size(const struct bpf_map *map)
|
2020-01-15 18:43:00 +00:00
|
|
|
{
|
|
|
|
|
if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE)
|
|
|
|
|
return round_up(map->value_size, 8) * num_possible_cpus();
|
|
|
|
|
else if (IS_FD_MAP(map))
|
|
|
|
|
return sizeof(u32);
|
|
|
|
|
else
|
|
|
|
|
return map->value_size;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void maybe_wait_bpf_programs(struct bpf_map *map)
|
|
|
|
|
{
|
2023-12-11 08:34:47 +00:00
|
|
|
/* Wait for any running non-sleepable BPF programs to complete so that
|
|
|
|
|
* userspace, when we return to it, knows that all non-sleepable
|
|
|
|
|
* programs that could be running use the new map value. For sleepable
|
|
|
|
|
* BPF programs, synchronize_rcu_tasks_trace() should be used to wait
|
|
|
|
|
* for the completions of these programs, but considering the waiting
|
|
|
|
|
* time can be very long and userspace may think it will hang forever,
|
|
|
|
|
* so don't handle sleepable BPF programs now.
|
2020-01-15 18:43:00 +00:00
|
|
|
*/
|
|
|
|
|
if (map->map_type == BPF_MAP_TYPE_HASH_OF_MAPS ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS)
|
|
|
|
|
synchronize_rcu();
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-16 07:50:58 +00:00
|
|
|
static int bpf_map_update_value(struct bpf_map *map, struct file *map_file,
|
|
|
|
|
void *key, void *value, __u64 flags)
|
2020-01-15 18:43:00 +00:00
|
|
|
{
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
/* Need to create a kthread, thus must support schedule */
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map)) {
|
2020-01-15 18:43:00 +00:00
|
|
|
return bpf_map_offload_update_elem(map, key, value, flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_CPUMAP ||
|
bpf: Introduce bpf_arena.
Introduce bpf_arena, which is a sparse shared memory region between the bpf
program and user space.
Use cases:
1. User space mmap-s bpf_arena and uses it as a traditional mmap-ed
anonymous region, like memcached or any key/value storage. The bpf
program implements an in-kernel accelerator. XDP prog can search for
a key in bpf_arena and return a value without going to user space.
2. The bpf program builds arbitrary data structures in bpf_arena (hash
tables, rb-trees, sparse arrays), while user space consumes it.
3. bpf_arena is a "heap" of memory from the bpf program's point of view.
The user space may mmap it, but bpf program will not convert pointers
to user base at run-time to improve bpf program speed.
Initially, the kernel vm_area and user vma are not populated. User space
can fault in pages within the range. While servicing a page fault,
bpf_arena logic will insert a new page into the kernel and user vmas. The
bpf program can allocate pages from that region via
bpf_arena_alloc_pages(). This kernel function will insert pages into the
kernel vm_area. The subsequent fault-in from user space will populate that
page into the user vma. The BPF_F_SEGV_ON_FAULT flag at arena creation time
can be used to prevent fault-in from user space. In such a case, if a page
is not allocated by the bpf program and not present in the kernel vm_area,
the user process will segfault. This is useful for use cases 2 and 3 above.
bpf_arena_alloc_pages() is similar to user space mmap(). It allocates pages
either at a specific address within the arena or allocates a range with the
maple tree. bpf_arena_free_pages() is analogous to munmap(), which frees
pages and removes the range from the kernel vm_area and from user process
vmas.
bpf_arena can be used as a bpf program "heap" of up to 4GB. The speed of
bpf program is more important than ease of sharing with user space. This is
use case 3. In such a case, the BPF_F_NO_USER_CONV flag is recommended.
It will tell the verifier to treat the rX = bpf_arena_cast_user(rY)
instruction as a 32-bit move wX = wY, which will improve bpf prog
performance. Otherwise, bpf_arena_cast_user is translated by JIT to
conditionally add the upper 32 bits of user vm_start (if the pointer is not
NULL) to arena pointers before they are stored into memory. This way, user
space sees them as valid 64-bit pointers.
Diff https://github.com/llvm/llvm-project/pull/84410 enables LLVM BPF
backend generate the bpf_addr_space_cast() instruction to cast pointers
between address_space(1) which is reserved for bpf_arena pointers and
default address space zero. All arena pointers in a bpf program written in
C language are tagged as __attribute__((address_space(1))). Hence, clang
provides helpful diagnostics when pointers cross address space. Libbpf and
the kernel support only address_space == 1. All other address space
identifiers are reserved.
rX = bpf_addr_space_cast(rY, /* dst_as */ 1, /* src_as */ 0) tells the
verifier that rX->type = PTR_TO_ARENA. Any further operations on
PTR_TO_ARENA register have to be in the 32-bit domain. The verifier will
mark load/store through PTR_TO_ARENA with PROBE_MEM32. JIT will generate
them as kern_vm_start + 32bit_addr memory accesses. The behavior is similar
to copy_from_kernel_nofault() except that no address checks are necessary.
The address is guaranteed to be in the 4GB range. If the page is not
present, the destination register is zeroed on read, and the operation is
ignored on write.
rX = bpf_addr_space_cast(rY, 0, 1) tells the verifier that rX->type =
unknown scalar. If arena->map_flags has BPF_F_NO_USER_CONV set, then the
verifier converts such cast instructions to mov32. Otherwise, JIT will emit
native code equivalent to:
rX = (u32)rY;
if (rY)
rX |= clear_lo32_bits(arena->user_vm_start); /* replace hi32 bits in rX */
After such conversion, the pointer becomes a valid user pointer within
bpf_arena range. The user process can access data structures created in
bpf_arena without any additional computations. For example, a linked list
built by a bpf program can be walked natively by user space.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Barret Rhoden <brho@google.com>
Link: https://lore.kernel.org/bpf/20240308010812.89848-2-alexei.starovoitov@gmail.com
2024-03-08 01:07:59 +00:00
|
|
|
map->map_type == BPF_MAP_TYPE_ARENA ||
|
2020-01-15 18:43:00 +00:00
|
|
|
map->map_type == BPF_MAP_TYPE_STRUCT_OPS) {
|
|
|
|
|
return map->ops->map_update_elem(map, key, value, flags);
|
2020-08-21 10:29:45 +00:00
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_SOCKHASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_SOCKMAP) {
|
|
|
|
|
return sock_map_update_elem_sys(map, key, value, flags);
|
2020-01-15 18:43:00 +00:00
|
|
|
} else if (IS_FD_PROG_ARRAY(map)) {
|
2022-11-16 07:50:58 +00:00
|
|
|
return bpf_fd_array_map_update_elem(map, map_file, key, value,
|
2020-01-15 18:43:00 +00:00
|
|
|
flags);
|
|
|
|
|
}
|
|
|
|
|
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_disable_instrumentation();
|
2020-01-15 18:43:00 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
|
|
|
|
|
err = bpf_percpu_hash_update(map, key, value, flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY) {
|
|
|
|
|
err = bpf_percpu_array_update(map, key, value, flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE) {
|
|
|
|
|
err = bpf_percpu_cgroup_storage_update(map, key, value,
|
|
|
|
|
flags);
|
|
|
|
|
} else if (IS_FD_ARRAY(map)) {
|
2022-11-16 07:50:58 +00:00
|
|
|
err = bpf_fd_array_map_update_elem(map, map_file, key, value,
|
2020-01-15 18:43:00 +00:00
|
|
|
flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_HASH_OF_MAPS) {
|
2022-11-16 07:50:58 +00:00
|
|
|
err = bpf_fd_htab_map_update_elem(map, map_file, key, value,
|
2020-01-15 18:43:00 +00:00
|
|
|
flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_REUSEPORT_SOCKARRAY) {
|
|
|
|
|
/* rcu_read_lock() is not needed */
|
|
|
|
|
err = bpf_fd_reuseport_array_update_elem(map, key, value,
|
|
|
|
|
flags);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_QUEUE ||
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
map->map_type == BPF_MAP_TYPE_STACK ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_BLOOM_FILTER) {
|
2020-01-15 18:43:00 +00:00
|
|
|
err = map->ops->map_push_elem(map, value, flags);
|
|
|
|
|
} else {
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_update_elem(map, key, value, flags);
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
}
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_enable_instrumentation();
|
2020-01-15 18:43:00 +00:00
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_map_copy_value(struct bpf_map *map, void *key, void *value,
|
|
|
|
|
__u64 flags)
|
|
|
|
|
{
|
|
|
|
|
void *ptr;
|
|
|
|
|
int err;
|
|
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map))
|
2020-01-15 18:43:01 +00:00
|
|
|
return bpf_map_offload_lookup_elem(map, key, value);
|
2020-01-15 18:43:00 +00:00
|
|
|
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_disable_instrumentation();
|
2020-01-15 18:43:00 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
|
|
|
|
|
err = bpf_percpu_hash_copy(map, key, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_PERCPU_ARRAY) {
|
|
|
|
|
err = bpf_percpu_array_copy(map, key, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE) {
|
|
|
|
|
err = bpf_percpu_cgroup_storage_copy(map, key, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_STACK_TRACE) {
|
|
|
|
|
err = bpf_stackmap_copy(map, key, value);
|
|
|
|
|
} else if (IS_FD_ARRAY(map) || IS_FD_PROG_ARRAY(map)) {
|
|
|
|
|
err = bpf_fd_array_map_lookup_elem(map, key, value);
|
|
|
|
|
} else if (IS_FD_HASH(map)) {
|
|
|
|
|
err = bpf_fd_htab_map_lookup_elem(map, key, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_REUSEPORT_SOCKARRAY) {
|
|
|
|
|
err = bpf_fd_reuseport_array_lookup_elem(map, key, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_QUEUE ||
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
map->map_type == BPF_MAP_TYPE_STACK ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_BLOOM_FILTER) {
|
2020-01-15 18:43:00 +00:00
|
|
|
err = map->ops->map_peek_elem(map, value);
|
|
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_STRUCT_OPS) {
|
|
|
|
|
/* struct_ops map requires directly updating "value" */
|
|
|
|
|
err = bpf_struct_ops_map_sys_lookup_elem(map, key, value);
|
|
|
|
|
} else {
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
if (map->ops->map_lookup_elem_sys_only)
|
|
|
|
|
ptr = map->ops->map_lookup_elem_sys_only(map, key);
|
|
|
|
|
else
|
|
|
|
|
ptr = map->ops->map_lookup_elem(map, key);
|
|
|
|
|
if (IS_ERR(ptr)) {
|
|
|
|
|
err = PTR_ERR(ptr);
|
|
|
|
|
} else if (!ptr) {
|
|
|
|
|
err = -ENOENT;
|
|
|
|
|
} else {
|
|
|
|
|
err = 0;
|
|
|
|
|
if (flags & BPF_F_LOCK)
|
|
|
|
|
/* lock 'ptr' and copy everything but lock */
|
|
|
|
|
copy_map_value_locked(map, value, ptr, true);
|
|
|
|
|
else
|
|
|
|
|
copy_map_value(map, value, ptr);
|
2021-07-15 00:54:10 +00:00
|
|
|
/* mask lock and timer, since value wasn't zero inited */
|
|
|
|
|
check_and_init_map_value(map, value);
|
2020-01-15 18:43:00 +00:00
|
|
|
}
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
}
|
|
|
|
|
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_enable_instrumentation();
|
2020-01-15 18:43:00 +00:00
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-12-01 21:58:33 +00:00
|
|
|
/* Please, do not use this function outside from the map creation path
|
|
|
|
|
* (e.g. in map update path) without taking care of setting the active
|
|
|
|
|
* memory cgroup (see at bpf_map_kmalloc_node() for example).
|
|
|
|
|
*/
|
2019-11-20 22:04:44 +00:00
|
|
|
static void *__bpf_map_area_alloc(u64 size, int numa_node, bool mmapable)
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
{
|
bpf: Try harder when allocating memory for large maps
It has been observed that sometimes a higher order memory allocation
for BPF maps fails when there is no obvious memory pressure in a system.
E.g. the map (BPF_MAP_TYPE_LRU_HASH, key=38, value=56, max_elems=524288)
could not be created due to vmalloc unable to allocate 75497472B,
when the system's memory consumption (in MB) was the following:
Total: 3942 Used: 837 (21.24%) Free: 138 Buffers: 239 Cached: 2727
Later analysis [1] by Michal Hocko showed that the vmalloc was not trying
to reclaim memory from the page cache and was failing prematurely due to
__GFP_NORETRY.
Considering dcda9b0471 ("mm, tree wide: replace __GFP_REPEAT by
__GFP_RETRY_MAYFAIL with more useful semantic") and [1], we can replace
__GFP_NORETRY with __GFP_RETRY_MAYFAIL, as it won't invoke OOM killer
and will try harder to fulfil allocation requests.
Unfortunately, replacing the body of the BPF map memory allocation
function with the kvmalloc_node helper function is not an option at
this point in time, given 1) kmalloc is non-optional for higher order
allocations, and 2) passing __GFP_RETRY_MAYFAIL to the kmalloc would
stress the slab allocator too much for large requests.
The change has been tested with the workloads mentioned above and by
observing oom_kill value from /proc/vmstat.
[1]: https://lore.kernel.org/bpf/20190310071318.GW5232@dhcp22.suse.cz/
Signed-off-by: Martynas Pumputis <m@lambda.lt>
Acked-by: Yonghong Song <yhs@fb.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20190318153940.GL8924@dhcp22.suse.cz/
2019-03-18 15:10:26 +00:00
|
|
|
/* We really just want to fail instead of triggering OOM killer
|
|
|
|
|
* under memory pressure, therefore we set __GFP_NORETRY to kmalloc,
|
|
|
|
|
* which is used for lower order allocation requests.
|
|
|
|
|
*
|
|
|
|
|
* It has been observed that higher order allocation requests done by
|
|
|
|
|
* vmalloc with __GFP_NORETRY being set might fail due to not trying
|
|
|
|
|
* to reclaim memory from the page cache, thus we set
|
|
|
|
|
* __GFP_RETRY_MAYFAIL to avoid such situations.
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
*/
|
bpf: Try harder when allocating memory for large maps
It has been observed that sometimes a higher order memory allocation
for BPF maps fails when there is no obvious memory pressure in a system.
E.g. the map (BPF_MAP_TYPE_LRU_HASH, key=38, value=56, max_elems=524288)
could not be created due to vmalloc unable to allocate 75497472B,
when the system's memory consumption (in MB) was the following:
Total: 3942 Used: 837 (21.24%) Free: 138 Buffers: 239 Cached: 2727
Later analysis [1] by Michal Hocko showed that the vmalloc was not trying
to reclaim memory from the page cache and was failing prematurely due to
__GFP_NORETRY.
Considering dcda9b0471 ("mm, tree wide: replace __GFP_REPEAT by
__GFP_RETRY_MAYFAIL with more useful semantic") and [1], we can replace
__GFP_NORETRY with __GFP_RETRY_MAYFAIL, as it won't invoke OOM killer
and will try harder to fulfil allocation requests.
Unfortunately, replacing the body of the BPF map memory allocation
function with the kvmalloc_node helper function is not an option at
this point in time, given 1) kmalloc is non-optional for higher order
allocations, and 2) passing __GFP_RETRY_MAYFAIL to the kmalloc would
stress the slab allocator too much for large requests.
The change has been tested with the workloads mentioned above and by
observing oom_kill value from /proc/vmstat.
[1]: https://lore.kernel.org/bpf/20190310071318.GW5232@dhcp22.suse.cz/
Signed-off-by: Martynas Pumputis <m@lambda.lt>
Acked-by: Yonghong Song <yhs@fb.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20190318153940.GL8924@dhcp22.suse.cz/
2019-03-18 15:10:26 +00:00
|
|
|
|
2023-02-10 15:47:33 +00:00
|
|
|
gfp_t gfp = bpf_memcg_flags(__GFP_NOWARN | __GFP_ZERO);
|
2020-06-02 04:52:02 +00:00
|
|
|
unsigned int flags = 0;
|
|
|
|
|
unsigned long align = 1;
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
void *area;
|
|
|
|
|
|
2019-11-20 22:04:44 +00:00
|
|
|
if (size >= SIZE_MAX)
|
|
|
|
|
return NULL;
|
|
|
|
|
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
/* kmalloc()'ed memory can't be mmap()'ed */
|
2020-06-02 04:52:02 +00:00
|
|
|
if (mmapable) {
|
|
|
|
|
BUG_ON(!PAGE_ALIGNED(size));
|
|
|
|
|
align = SHMLBA;
|
|
|
|
|
flags = VM_USERMAP;
|
|
|
|
|
} else if (size <= (PAGE_SIZE << PAGE_ALLOC_COSTLY_ORDER)) {
|
|
|
|
|
area = kmalloc_node(size, gfp | GFP_USER | __GFP_NORETRY,
|
bpf: Try harder when allocating memory for large maps
It has been observed that sometimes a higher order memory allocation
for BPF maps fails when there is no obvious memory pressure in a system.
E.g. the map (BPF_MAP_TYPE_LRU_HASH, key=38, value=56, max_elems=524288)
could not be created due to vmalloc unable to allocate 75497472B,
when the system's memory consumption (in MB) was the following:
Total: 3942 Used: 837 (21.24%) Free: 138 Buffers: 239 Cached: 2727
Later analysis [1] by Michal Hocko showed that the vmalloc was not trying
to reclaim memory from the page cache and was failing prematurely due to
__GFP_NORETRY.
Considering dcda9b0471 ("mm, tree wide: replace __GFP_REPEAT by
__GFP_RETRY_MAYFAIL with more useful semantic") and [1], we can replace
__GFP_NORETRY with __GFP_RETRY_MAYFAIL, as it won't invoke OOM killer
and will try harder to fulfil allocation requests.
Unfortunately, replacing the body of the BPF map memory allocation
function with the kvmalloc_node helper function is not an option at
this point in time, given 1) kmalloc is non-optional for higher order
allocations, and 2) passing __GFP_RETRY_MAYFAIL to the kmalloc would
stress the slab allocator too much for large requests.
The change has been tested with the workloads mentioned above and by
observing oom_kill value from /proc/vmstat.
[1]: https://lore.kernel.org/bpf/20190310071318.GW5232@dhcp22.suse.cz/
Signed-off-by: Martynas Pumputis <m@lambda.lt>
Acked-by: Yonghong Song <yhs@fb.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20190318153940.GL8924@dhcp22.suse.cz/
2019-03-18 15:10:26 +00:00
|
|
|
numa_node);
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
if (area != NULL)
|
|
|
|
|
return area;
|
|
|
|
|
}
|
2020-06-02 04:52:02 +00:00
|
|
|
|
|
|
|
|
return __vmalloc_node_range(size, align, VMALLOC_START, VMALLOC_END,
|
|
|
|
|
gfp | GFP_KERNEL | __GFP_RETRY_MAYFAIL, PAGE_KERNEL,
|
|
|
|
|
flags, numa_node, __builtin_return_address(0));
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
}
|
|
|
|
|
|
2019-11-20 22:04:44 +00:00
|
|
|
void *bpf_map_area_alloc(u64 size, int numa_node)
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
{
|
|
|
|
|
return __bpf_map_area_alloc(size, numa_node, false);
|
|
|
|
|
}
|
|
|
|
|
|
2019-11-20 22:04:44 +00:00
|
|
|
void *bpf_map_area_mmapable_alloc(u64 size, int numa_node)
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
{
|
|
|
|
|
return __bpf_map_area_alloc(size, numa_node, true);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: don't trigger OOM killer under pressure with map alloc
This patch adds two helpers, bpf_map_area_alloc() and bpf_map_area_free(),
that are to be used for map allocations. Using kmalloc() for very large
allocations can cause excessive work within the page allocator, so i) fall
back earlier to vmalloc() when the attempt is considered costly anyway,
and even more importantly ii) don't trigger OOM killer with any of the
allocators.
Since this is based on a user space request, for example, when creating
maps with element pre-allocation, we really want such requests to fail
instead of killing other user space processes.
Also, don't spam the kernel log with warnings should any of the allocations
fail under pressure. Given that, we can make backend selection in
bpf_map_area_alloc() generic, and convert all maps over to use this API
for spots with potentially large allocation requests.
Note, replacing the one kmalloc_array() is fine as overflow checks happen
earlier in htab_map_alloc(), since it must also protect the multiplication
for vmalloc() should kmalloc_array() fail.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-18 14:14:17 +00:00
|
|
|
void bpf_map_area_free(void *area)
|
|
|
|
|
{
|
|
|
|
|
kvfree(area);
|
|
|
|
|
}
|
|
|
|
|
|
2019-04-09 21:20:04 +00:00
|
|
|
static u32 bpf_map_flags_retain_permanent(u32 flags)
|
|
|
|
|
{
|
|
|
|
|
/* Some map creation flags are not tied to the map object but
|
|
|
|
|
* rather to the map fd instead, so they have no meaning upon
|
|
|
|
|
* map object inspection since multiple file descriptors with
|
|
|
|
|
* different (access) properties can exist here. Thus, given
|
|
|
|
|
* this has zero meaning for the map itself, lets clear these
|
|
|
|
|
* from here.
|
|
|
|
|
*/
|
|
|
|
|
return flags & ~(BPF_F_RDONLY | BPF_F_WRONLY);
|
|
|
|
|
}
|
|
|
|
|
|
2018-01-12 04:29:06 +00:00
|
|
|
void bpf_map_init_from_attr(struct bpf_map *map, union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
map->map_type = attr->map_type;
|
|
|
|
|
map->key_size = attr->key_size;
|
|
|
|
|
map->value_size = attr->value_size;
|
|
|
|
|
map->max_entries = attr->max_entries;
|
2019-04-09 21:20:04 +00:00
|
|
|
map->map_flags = bpf_map_flags_retain_permanent(attr->map_flags);
|
2018-01-12 04:29:06 +00:00
|
|
|
map->numa_node = bpf_map_attr_numa_node(attr);
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
map->map_extra = attr->map_extra;
|
2018-01-12 04:29:06 +00:00
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:47 +00:00
|
|
|
static int bpf_map_alloc_id(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
int id;
|
|
|
|
|
|
2018-03-27 18:53:21 +00:00
|
|
|
idr_preload(GFP_KERNEL);
|
2017-06-05 19:15:47 +00:00
|
|
|
spin_lock_bh(&map_idr_lock);
|
|
|
|
|
id = idr_alloc_cyclic(&map_idr, map, 1, INT_MAX, GFP_ATOMIC);
|
|
|
|
|
if (id > 0)
|
|
|
|
|
map->id = id;
|
|
|
|
|
spin_unlock_bh(&map_idr_lock);
|
2018-03-27 18:53:21 +00:00
|
|
|
idr_preload_end();
|
2017-06-05 19:15:47 +00:00
|
|
|
|
|
|
|
|
if (WARN_ON_ONCE(!id))
|
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
|
|
return id > 0 ? 0 : id;
|
|
|
|
|
}
|
|
|
|
|
|
2023-02-02 14:19:21 +00:00
|
|
|
void bpf_map_free_id(struct bpf_map *map)
|
2017-06-05 19:15:47 +00:00
|
|
|
{
|
2017-09-19 16:15:59 +00:00
|
|
|
unsigned long flags;
|
|
|
|
|
|
2018-01-12 04:29:09 +00:00
|
|
|
/* Offloaded maps are removed from the IDR store when their device
|
|
|
|
|
* disappears - even if someone holds an fd to them they are unusable,
|
|
|
|
|
* the memory is gone, all ops will fail; they are simply waiting for
|
|
|
|
|
* refcnt to drop to be freed.
|
|
|
|
|
*/
|
|
|
|
|
if (!map->id)
|
|
|
|
|
return;
|
|
|
|
|
|
2023-02-02 14:19:21 +00:00
|
|
|
spin_lock_irqsave(&map_idr_lock, flags);
|
2017-06-05 19:15:50 +00:00
|
|
|
|
2017-06-05 19:15:47 +00:00
|
|
|
idr_remove(&map_idr, map->id);
|
2018-01-12 04:29:09 +00:00
|
|
|
map->id = 0;
|
2017-06-05 19:15:50 +00:00
|
|
|
|
2023-02-02 14:19:21 +00:00
|
|
|
spin_unlock_irqrestore(&map_idr_lock, flags);
|
2017-06-05 19:15:47 +00:00
|
|
|
}
|
|
|
|
|
|
2020-12-01 21:58:32 +00:00
|
|
|
#ifdef CONFIG_MEMCG_KMEM
|
|
|
|
|
static void bpf_map_save_memcg(struct bpf_map *map)
|
|
|
|
|
{
|
2022-07-11 16:28:27 +00:00
|
|
|
/* Currently if a map is created by a process belonging to the root
|
|
|
|
|
* memory cgroup, get_obj_cgroup_from_current() will return NULL.
|
|
|
|
|
* So we have to check map->objcg for being NULL each time it's
|
|
|
|
|
* being used.
|
|
|
|
|
*/
|
2023-02-10 15:47:33 +00:00
|
|
|
if (memcg_bpf_enabled())
|
|
|
|
|
map->objcg = get_obj_cgroup_from_current();
|
2020-12-01 21:58:32 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_map_release_memcg(struct bpf_map *map)
|
|
|
|
|
{
|
2022-07-11 16:28:27 +00:00
|
|
|
if (map->objcg)
|
|
|
|
|
obj_cgroup_put(map->objcg);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct mem_cgroup *bpf_map_get_memcg(const struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
if (map->objcg)
|
|
|
|
|
return get_mem_cgroup_from_objcg(map->objcg);
|
|
|
|
|
|
|
|
|
|
return root_mem_cgroup;
|
2020-12-01 21:58:32 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void *bpf_map_kmalloc_node(const struct bpf_map *map, size_t size, gfp_t flags,
|
|
|
|
|
int node)
|
|
|
|
|
{
|
2022-07-11 16:28:27 +00:00
|
|
|
struct mem_cgroup *memcg, *old_memcg;
|
2020-12-01 21:58:32 +00:00
|
|
|
void *ptr;
|
|
|
|
|
|
2022-07-11 16:28:27 +00:00
|
|
|
memcg = bpf_map_get_memcg(map);
|
|
|
|
|
old_memcg = set_active_memcg(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
ptr = kmalloc_node(size, flags | __GFP_ACCOUNT, node);
|
|
|
|
|
set_active_memcg(old_memcg);
|
2022-07-11 16:28:27 +00:00
|
|
|
mem_cgroup_put(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
|
|
|
|
|
return ptr;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void *bpf_map_kzalloc(const struct bpf_map *map, size_t size, gfp_t flags)
|
|
|
|
|
{
|
2022-07-11 16:28:27 +00:00
|
|
|
struct mem_cgroup *memcg, *old_memcg;
|
2020-12-01 21:58:32 +00:00
|
|
|
void *ptr;
|
|
|
|
|
|
2022-07-11 16:28:27 +00:00
|
|
|
memcg = bpf_map_get_memcg(map);
|
|
|
|
|
old_memcg = set_active_memcg(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
ptr = kzalloc(size, flags | __GFP_ACCOUNT);
|
|
|
|
|
set_active_memcg(old_memcg);
|
2022-07-11 16:28:27 +00:00
|
|
|
mem_cgroup_put(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
|
|
|
|
|
return ptr;
|
|
|
|
|
}
|
|
|
|
|
|
2023-02-10 15:47:32 +00:00
|
|
|
void *bpf_map_kvcalloc(struct bpf_map *map, size_t n, size_t size,
|
|
|
|
|
gfp_t flags)
|
|
|
|
|
{
|
|
|
|
|
struct mem_cgroup *memcg, *old_memcg;
|
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
|
|
memcg = bpf_map_get_memcg(map);
|
|
|
|
|
old_memcg = set_active_memcg(memcg);
|
|
|
|
|
ptr = kvcalloc(n, size, flags | __GFP_ACCOUNT);
|
|
|
|
|
set_active_memcg(old_memcg);
|
|
|
|
|
mem_cgroup_put(memcg);
|
|
|
|
|
|
|
|
|
|
return ptr;
|
|
|
|
|
}
|
|
|
|
|
|
2020-12-01 21:58:32 +00:00
|
|
|
void __percpu *bpf_map_alloc_percpu(const struct bpf_map *map, size_t size,
|
|
|
|
|
size_t align, gfp_t flags)
|
|
|
|
|
{
|
2022-07-11 16:28:27 +00:00
|
|
|
struct mem_cgroup *memcg, *old_memcg;
|
2020-12-01 21:58:32 +00:00
|
|
|
void __percpu *ptr;
|
|
|
|
|
|
2022-07-11 16:28:27 +00:00
|
|
|
memcg = bpf_map_get_memcg(map);
|
|
|
|
|
old_memcg = set_active_memcg(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
ptr = __alloc_percpu_gfp(size, align, flags | __GFP_ACCOUNT);
|
|
|
|
|
set_active_memcg(old_memcg);
|
2022-07-11 16:28:27 +00:00
|
|
|
mem_cgroup_put(memcg);
|
2020-12-01 21:58:32 +00:00
|
|
|
|
|
|
|
|
return ptr;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#else
|
|
|
|
|
static void bpf_map_save_memcg(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_map_release_memcg(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
bpf: Introduce bpf_arena.
Introduce bpf_arena, which is a sparse shared memory region between the bpf
program and user space.
Use cases:
1. User space mmap-s bpf_arena and uses it as a traditional mmap-ed
anonymous region, like memcached or any key/value storage. The bpf
program implements an in-kernel accelerator. XDP prog can search for
a key in bpf_arena and return a value without going to user space.
2. The bpf program builds arbitrary data structures in bpf_arena (hash
tables, rb-trees, sparse arrays), while user space consumes it.
3. bpf_arena is a "heap" of memory from the bpf program's point of view.
The user space may mmap it, but bpf program will not convert pointers
to user base at run-time to improve bpf program speed.
Initially, the kernel vm_area and user vma are not populated. User space
can fault in pages within the range. While servicing a page fault,
bpf_arena logic will insert a new page into the kernel and user vmas. The
bpf program can allocate pages from that region via
bpf_arena_alloc_pages(). This kernel function will insert pages into the
kernel vm_area. The subsequent fault-in from user space will populate that
page into the user vma. The BPF_F_SEGV_ON_FAULT flag at arena creation time
can be used to prevent fault-in from user space. In such a case, if a page
is not allocated by the bpf program and not present in the kernel vm_area,
the user process will segfault. This is useful for use cases 2 and 3 above.
bpf_arena_alloc_pages() is similar to user space mmap(). It allocates pages
either at a specific address within the arena or allocates a range with the
maple tree. bpf_arena_free_pages() is analogous to munmap(), which frees
pages and removes the range from the kernel vm_area and from user process
vmas.
bpf_arena can be used as a bpf program "heap" of up to 4GB. The speed of
bpf program is more important than ease of sharing with user space. This is
use case 3. In such a case, the BPF_F_NO_USER_CONV flag is recommended.
It will tell the verifier to treat the rX = bpf_arena_cast_user(rY)
instruction as a 32-bit move wX = wY, which will improve bpf prog
performance. Otherwise, bpf_arena_cast_user is translated by JIT to
conditionally add the upper 32 bits of user vm_start (if the pointer is not
NULL) to arena pointers before they are stored into memory. This way, user
space sees them as valid 64-bit pointers.
Diff https://github.com/llvm/llvm-project/pull/84410 enables LLVM BPF
backend generate the bpf_addr_space_cast() instruction to cast pointers
between address_space(1) which is reserved for bpf_arena pointers and
default address space zero. All arena pointers in a bpf program written in
C language are tagged as __attribute__((address_space(1))). Hence, clang
provides helpful diagnostics when pointers cross address space. Libbpf and
the kernel support only address_space == 1. All other address space
identifiers are reserved.
rX = bpf_addr_space_cast(rY, /* dst_as */ 1, /* src_as */ 0) tells the
verifier that rX->type = PTR_TO_ARENA. Any further operations on
PTR_TO_ARENA register have to be in the 32-bit domain. The verifier will
mark load/store through PTR_TO_ARENA with PROBE_MEM32. JIT will generate
them as kern_vm_start + 32bit_addr memory accesses. The behavior is similar
to copy_from_kernel_nofault() except that no address checks are necessary.
The address is guaranteed to be in the 4GB range. If the page is not
present, the destination register is zeroed on read, and the operation is
ignored on write.
rX = bpf_addr_space_cast(rY, 0, 1) tells the verifier that rX->type =
unknown scalar. If arena->map_flags has BPF_F_NO_USER_CONV set, then the
verifier converts such cast instructions to mov32. Otherwise, JIT will emit
native code equivalent to:
rX = (u32)rY;
if (rY)
rX |= clear_lo32_bits(arena->user_vm_start); /* replace hi32 bits in rX */
After such conversion, the pointer becomes a valid user pointer within
bpf_arena range. The user process can access data structures created in
bpf_arena without any additional computations. For example, a linked list
built by a bpf program can be walked natively by user space.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Barret Rhoden <brho@google.com>
Link: https://lore.kernel.org/bpf/20240308010812.89848-2-alexei.starovoitov@gmail.com
2024-03-08 01:07:59 +00:00
|
|
|
int bpf_map_alloc_pages(const struct bpf_map *map, gfp_t gfp, int nid,
|
|
|
|
|
unsigned long nr_pages, struct page **pages)
|
|
|
|
|
{
|
|
|
|
|
unsigned long i, j;
|
|
|
|
|
struct page *pg;
|
|
|
|
|
int ret = 0;
|
|
|
|
|
#ifdef CONFIG_MEMCG_KMEM
|
|
|
|
|
struct mem_cgroup *memcg, *old_memcg;
|
|
|
|
|
|
|
|
|
|
memcg = bpf_map_get_memcg(map);
|
|
|
|
|
old_memcg = set_active_memcg(memcg);
|
|
|
|
|
#endif
|
|
|
|
|
for (i = 0; i < nr_pages; i++) {
|
|
|
|
|
pg = alloc_pages_node(nid, gfp | __GFP_ACCOUNT, 0);
|
|
|
|
|
|
|
|
|
|
if (pg) {
|
|
|
|
|
pages[i] = pg;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
for (j = 0; j < i; j++)
|
|
|
|
|
__free_page(pages[j]);
|
|
|
|
|
ret = -ENOMEM;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_MEMCG_KMEM
|
|
|
|
|
set_active_memcg(old_memcg);
|
|
|
|
|
mem_cgroup_put(memcg);
|
|
|
|
|
#endif
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
static int btf_field_cmp(const void *a, const void *b)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
{
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
const struct btf_field *f1 = a, *f2 = b;
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (f1->offset < f2->offset)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return -1;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
else if (f1->offset > f2->offset)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return 1;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
struct btf_field *btf_record_find(const struct btf_record *rec, u32 offset,
|
2023-03-09 18:01:08 +00:00
|
|
|
u32 field_mask)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
{
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
struct btf_field *field;
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
|
2023-03-09 18:01:08 +00:00
|
|
|
if (IS_ERR_OR_NULL(rec) || !(rec->field_mask & field_mask))
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
return NULL;
|
|
|
|
|
field = bsearch(&offset, rec->fields, rec->cnt, sizeof(rec->fields[0]), btf_field_cmp);
|
2023-03-09 18:01:08 +00:00
|
|
|
if (!field || !(field->type & field_mask))
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return NULL;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
return field;
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
void btf_record_free(struct btf_record *rec)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (IS_ERR_OR_NULL(rec))
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
for (i = 0; i < rec->cnt; i++) {
|
|
|
|
|
switch (rec->fields[i].type) {
|
|
|
|
|
case BPF_KPTR_UNREF:
|
|
|
|
|
case BPF_KPTR_REF:
|
2023-08-27 15:27:39 +00:00
|
|
|
case BPF_KPTR_PERCPU:
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (rec->fields[i].kptr.module)
|
|
|
|
|
module_put(rec->fields[i].kptr.module);
|
|
|
|
|
btf_put(rec->fields[i].kptr.btf);
|
|
|
|
|
break;
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
case BPF_LIST_HEAD:
|
bpf: Recognize lock and list fields in allocated objects
Allow specifying bpf_spin_lock, bpf_list_head, bpf_list_node fields in a
allocated object.
Also update btf_struct_access to reject direct access to these special
fields.
A bpf_list_head allows implementing map-in-map style use cases, where an
allocated object with bpf_list_head is linked into a list in a map
value. This would require embedding a bpf_list_node, support for which
is also included. The bpf_spin_lock is used to protect the bpf_list_head
and other data.
While we strictly don't require to hold a bpf_spin_lock while touching
the bpf_list_head in such objects, as when have access to it, we have
complete ownership of the object, the locking constraint is still kept
and may be conditionally lifted in the future.
Note that the specification of such types can be done just like map
values, e.g.:
struct bar {
struct bpf_list_node node;
};
struct foo {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(bar, node);
struct bpf_list_node node;
};
struct map_value {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(foo, node);
};
To recognize such types in user BTF, we build a btf_struct_metas array
of metadata items corresponding to each BTF ID. This is done once during
the btf_parse stage to avoid having to do it each time during the
verification process's requirement to inspect the metadata.
Moreover, the computed metadata needs to be passed to some helpers in
future patches which requires allocating them and storing them in the
BTF that is pinned by the program itself, so that valid access can be
assumed to such data during program runtime.
A key thing to note is that once a btf_struct_meta is available for a
type, both the btf_record and btf_field_offs should be available. It is
critical that btf_field_offs is available in case special fields are
present, as we extensively rely on special fields being zeroed out in
map values and allocated objects in later patches. The code ensures that
by bailing out in case of errors and ensuring both are available
together. If the record is not available, the special fields won't be
recognized, so not having both is also fine (in terms of being a
verification error and not a runtime bug).
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:56 +00:00
|
|
|
case BPF_LIST_NODE:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
case BPF_RB_ROOT:
|
|
|
|
|
case BPF_RB_NODE:
|
|
|
|
|
case BPF_SPIN_LOCK:
|
|
|
|
|
case BPF_TIMER:
|
2023-04-15 20:18:04 +00:00
|
|
|
case BPF_REFCOUNT:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
/* Nothing to release */
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
break;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
default:
|
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
|
continue;
|
|
|
|
|
}
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
kfree(rec);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void bpf_map_free_record(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
btf_record_free(map->record);
|
|
|
|
|
map->record = NULL;
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
struct btf_record *btf_record_dup(const struct btf_record *rec)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
{
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
const struct btf_field *fields;
|
|
|
|
|
struct btf_record *new_rec;
|
|
|
|
|
int ret, size, i;
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (IS_ERR_OR_NULL(rec))
|
|
|
|
|
return NULL;
|
|
|
|
|
size = offsetof(struct btf_record, fields[rec->cnt]);
|
|
|
|
|
new_rec = kmemdup(rec, size, GFP_KERNEL | __GFP_NOWARN);
|
|
|
|
|
if (!new_rec)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return ERR_PTR(-ENOMEM);
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
/* Do a deep copy of the btf_record */
|
|
|
|
|
fields = rec->fields;
|
|
|
|
|
new_rec->cnt = 0;
|
|
|
|
|
for (i = 0; i < rec->cnt; i++) {
|
|
|
|
|
switch (fields[i].type) {
|
|
|
|
|
case BPF_KPTR_UNREF:
|
|
|
|
|
case BPF_KPTR_REF:
|
2023-08-27 15:27:39 +00:00
|
|
|
case BPF_KPTR_PERCPU:
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
btf_get(fields[i].kptr.btf);
|
|
|
|
|
if (fields[i].kptr.module && !try_module_get(fields[i].kptr.module)) {
|
|
|
|
|
ret = -ENXIO;
|
|
|
|
|
goto free;
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
break;
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
case BPF_LIST_HEAD:
|
bpf: Recognize lock and list fields in allocated objects
Allow specifying bpf_spin_lock, bpf_list_head, bpf_list_node fields in a
allocated object.
Also update btf_struct_access to reject direct access to these special
fields.
A bpf_list_head allows implementing map-in-map style use cases, where an
allocated object with bpf_list_head is linked into a list in a map
value. This would require embedding a bpf_list_node, support for which
is also included. The bpf_spin_lock is used to protect the bpf_list_head
and other data.
While we strictly don't require to hold a bpf_spin_lock while touching
the bpf_list_head in such objects, as when have access to it, we have
complete ownership of the object, the locking constraint is still kept
and may be conditionally lifted in the future.
Note that the specification of such types can be done just like map
values, e.g.:
struct bar {
struct bpf_list_node node;
};
struct foo {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(bar, node);
struct bpf_list_node node;
};
struct map_value {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(foo, node);
};
To recognize such types in user BTF, we build a btf_struct_metas array
of metadata items corresponding to each BTF ID. This is done once during
the btf_parse stage to avoid having to do it each time during the
verification process's requirement to inspect the metadata.
Moreover, the computed metadata needs to be passed to some helpers in
future patches which requires allocating them and storing them in the
BTF that is pinned by the program itself, so that valid access can be
assumed to such data during program runtime.
A key thing to note is that once a btf_struct_meta is available for a
type, both the btf_record and btf_field_offs should be available. It is
critical that btf_field_offs is available in case special fields are
present, as we extensively rely on special fields being zeroed out in
map values and allocated objects in later patches. The code ensures that
by bailing out in case of errors and ensuring both are available
together. If the record is not available, the special fields won't be
recognized, so not having both is also fine (in terms of being a
verification error and not a runtime bug).
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:56 +00:00
|
|
|
case BPF_LIST_NODE:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
case BPF_RB_ROOT:
|
|
|
|
|
case BPF_RB_NODE:
|
|
|
|
|
case BPF_SPIN_LOCK:
|
|
|
|
|
case BPF_TIMER:
|
2023-04-15 20:18:04 +00:00
|
|
|
case BPF_REFCOUNT:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
/* Nothing to acquire */
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
break;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
default:
|
|
|
|
|
ret = -EFAULT;
|
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
|
goto free;
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
new_rec->cnt++;
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
return new_rec;
|
|
|
|
|
free:
|
|
|
|
|
btf_record_free(new_rec);
|
|
|
|
|
return ERR_PTR(ret);
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
bool btf_record_equal(const struct btf_record *rec_a, const struct btf_record *rec_b)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
{
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
bool a_has_fields = !IS_ERR_OR_NULL(rec_a), b_has_fields = !IS_ERR_OR_NULL(rec_b);
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
int size;
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (!a_has_fields && !b_has_fields)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return true;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (a_has_fields != b_has_fields)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return false;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (rec_a->cnt != rec_b->cnt)
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return false;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
size = offsetof(struct btf_record, fields[rec_a->cnt]);
|
bpf: Add comments for map BTF matching requirement for bpf_list_head
The old behavior of bpf_map_meta_equal was that it compared timer_off
to be equal (but not spin_lock_off, because that was not allowed), and
did memcmp of kptr_off_tab.
Now, we memcmp the btf_record of two bpf_map structs, which has all
fields.
We preserve backwards compat as we kzalloc the array, so if only spin
lock and timer exist in map, we only compare offset while the rest of
unused members in the btf_field struct are zeroed out.
In case of kptr, btf and everything else is of vmlinux or module, so as
long type is same it will match, since kernel btf, module, dtor pointer
will be same across maps.
Now with list_head in the mix, things are a bit complicated. We
implicitly add a requirement that both BTFs are same, because struct
btf_field_list_head has btf and value_rec members.
We obviously shouldn't force BTFs to be equal by default, as that breaks
backwards compatibility.
Currently it is only implicitly required due to list_head matching
struct btf and value_rec member. value_rec points back into a btf_record
stashed in the map BTF (btf member of btf_field_list_head). So that
pointer and btf member has to match exactly.
Document all these subtle details so that things don't break in the
future when touching this code.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-19-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:56:08 +00:00
|
|
|
/* btf_parse_fields uses kzalloc to allocate a btf_record, so unused
|
|
|
|
|
* members are zeroed out. So memcmp is safe to do without worrying
|
|
|
|
|
* about padding/unused fields.
|
|
|
|
|
*
|
|
|
|
|
* While spin_lock, timer, and kptr have no relation to map BTF,
|
|
|
|
|
* list_head metadata is specific to map BTF, the btf and value_rec
|
|
|
|
|
* members in particular. btf is the map BTF, while value_rec points to
|
|
|
|
|
* btf_record in that map BTF.
|
|
|
|
|
*
|
|
|
|
|
* So while by default, we don't rely on the map BTF (which the records
|
|
|
|
|
* were parsed from) matching for both records, which is not backwards
|
|
|
|
|
* compatible, in case list_head is part of it, we implicitly rely on
|
|
|
|
|
* that by way of depending on memcmp succeeding for it.
|
|
|
|
|
*/
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
return !memcmp(rec_a, rec_b, size);
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
}
|
|
|
|
|
|
2022-11-03 19:09:56 +00:00
|
|
|
void bpf_obj_free_timer(const struct btf_record *rec, void *obj)
|
|
|
|
|
{
|
|
|
|
|
if (WARN_ON_ONCE(!btf_record_has_field(rec, BPF_TIMER)))
|
|
|
|
|
return;
|
|
|
|
|
bpf_timer_cancel_and_free(obj + rec->timer_off);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
void bpf_obj_free_fields(const struct btf_record *rec, void *obj)
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
{
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
const struct btf_field *fields;
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
int i;
|
|
|
|
|
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (IS_ERR_OR_NULL(rec))
|
|
|
|
|
return;
|
|
|
|
|
fields = rec->fields;
|
|
|
|
|
for (i = 0; i < rec->cnt; i++) {
|
bpf: Support __kptr to local kptrs
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-10 23:07:41 +00:00
|
|
|
struct btf_struct_meta *pointee_struct_meta;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
const struct btf_field *field = &fields[i];
|
|
|
|
|
void *field_ptr = obj + field->offset;
|
bpf: Support __kptr to local kptrs
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-10 23:07:41 +00:00
|
|
|
void *xchgd_field;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
|
|
|
|
|
switch (fields[i].type) {
|
2022-11-03 19:09:56 +00:00
|
|
|
case BPF_SPIN_LOCK:
|
|
|
|
|
break;
|
|
|
|
|
case BPF_TIMER:
|
|
|
|
|
bpf_timer_cancel_and_free(field_ptr);
|
|
|
|
|
break;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
case BPF_KPTR_UNREF:
|
|
|
|
|
WRITE_ONCE(*(u64 *)field_ptr, 0);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_KPTR_REF:
|
2023-08-27 15:27:39 +00:00
|
|
|
case BPF_KPTR_PERCPU:
|
bpf: Support __kptr to local kptrs
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-10 23:07:41 +00:00
|
|
|
xchgd_field = (void *)xchg((unsigned long *)field_ptr, 0);
|
bpf: Only invoke kptr dtor following non-NULL xchg
When a map value is being freed, we loop over all of the fields of the
corresponding BPF object and issue the appropriate cleanup calls
corresponding to the field's type. If the field is a referenced kptr, we
atomically xchg the value out of the map, and invoke the kptr's
destructor on whatever was there before (or bpf_obj_drop() it if it was
a local kptr).
Currently, we always invoke the destructor (either bpf_obj_drop() or the
kptr's registered destructor) on any KPTR_REF-type field in a map, even
if there wasn't a value in the map. This means that any function serving
as the kptr's KF_RELEASE destructor must always treat the argument as
possibly NULL, as the following can and regularly does happen:
void *xchgd_field;
/* No value was in the map, so xchgd_field is NULL */
xchgd_field = (void *)xchg(unsigned long *field_ptr, 0);
field->kptr.dtor(xchgd_field);
These are odd semantics to impose on KF_RELEASE kfuncs -- BPF programs
are prohibited by the verifier from passing NULL pointers to KF_RELEASE
kfuncs, so it doesn't make sense to require this of BPF programs, but
not the main kernel destructor path. It's also unnecessary to invoke any
cleanup logic for local kptrs. If there is no object there, there's
nothing to drop.
So as to allow KF_RELEASE kfuncs to fully assume that an argument is
non-NULL, this patch updates a KPTR_REF's destructor to only be invoked
when a non-NULL value is xchg'd out of the kptr map field.
Signed-off-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/r/20230325213144.486885-2-void@manifault.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-25 21:31:42 +00:00
|
|
|
if (!xchgd_field)
|
|
|
|
|
break;
|
|
|
|
|
|
bpf: Support __kptr to local kptrs
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-10 23:07:41 +00:00
|
|
|
if (!btf_is_kernel(field->kptr.btf)) {
|
|
|
|
|
pointee_struct_meta = btf_find_struct_meta(field->kptr.btf,
|
|
|
|
|
field->kptr.btf_id);
|
2023-03-13 21:46:41 +00:00
|
|
|
migrate_disable();
|
|
|
|
|
__bpf_obj_drop_impl(xchgd_field, pointee_struct_meta ?
|
2023-10-20 13:32:01 +00:00
|
|
|
pointee_struct_meta->record : NULL,
|
|
|
|
|
fields[i].type == BPF_KPTR_PERCPU);
|
2023-03-13 21:46:41 +00:00
|
|
|
migrate_enable();
|
bpf: Support __kptr to local kptrs
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-03-10 23:07:41 +00:00
|
|
|
} else {
|
|
|
|
|
field->kptr.dtor(xchgd_field);
|
|
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
break;
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
case BPF_LIST_HEAD:
|
|
|
|
|
if (WARN_ON_ONCE(rec->spin_lock_off < 0))
|
|
|
|
|
continue;
|
|
|
|
|
bpf_list_head_free(field, field_ptr, obj + rec->spin_lock_off);
|
|
|
|
|
break;
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
case BPF_RB_ROOT:
|
|
|
|
|
if (WARN_ON_ONCE(rec->spin_lock_off < 0))
|
|
|
|
|
continue;
|
|
|
|
|
bpf_rb_root_free(field, field_ptr, obj + rec->spin_lock_off);
|
|
|
|
|
break;
|
bpf: Recognize lock and list fields in allocated objects
Allow specifying bpf_spin_lock, bpf_list_head, bpf_list_node fields in a
allocated object.
Also update btf_struct_access to reject direct access to these special
fields.
A bpf_list_head allows implementing map-in-map style use cases, where an
allocated object with bpf_list_head is linked into a list in a map
value. This would require embedding a bpf_list_node, support for which
is also included. The bpf_spin_lock is used to protect the bpf_list_head
and other data.
While we strictly don't require to hold a bpf_spin_lock while touching
the bpf_list_head in such objects, as when have access to it, we have
complete ownership of the object, the locking constraint is still kept
and may be conditionally lifted in the future.
Note that the specification of such types can be done just like map
values, e.g.:
struct bar {
struct bpf_list_node node;
};
struct foo {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(bar, node);
struct bpf_list_node node;
};
struct map_value {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(foo, node);
};
To recognize such types in user BTF, we build a btf_struct_metas array
of metadata items corresponding to each BTF ID. This is done once during
the btf_parse stage to avoid having to do it each time during the
verification process's requirement to inspect the metadata.
Moreover, the computed metadata needs to be passed to some helpers in
future patches which requires allocating them and storing them in the
BTF that is pinned by the program itself, so that valid access can be
assumed to such data during program runtime.
A key thing to note is that once a btf_struct_meta is available for a
type, both the btf_record and btf_field_offs should be available. It is
critical that btf_field_offs is available in case special fields are
present, as we extensively rely on special fields being zeroed out in
map values and allocated objects in later patches. The code ensures that
by bailing out in case of errors and ensuring both are available
together. If the record is not available, the special fields won't be
recognized, so not having both is also fine (in terms of being a
verification error and not a runtime bug).
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:56 +00:00
|
|
|
case BPF_LIST_NODE:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
case BPF_RB_NODE:
|
2023-04-15 20:18:04 +00:00
|
|
|
case BPF_REFCOUNT:
|
bpf: Recognize lock and list fields in allocated objects
Allow specifying bpf_spin_lock, bpf_list_head, bpf_list_node fields in a
allocated object.
Also update btf_struct_access to reject direct access to these special
fields.
A bpf_list_head allows implementing map-in-map style use cases, where an
allocated object with bpf_list_head is linked into a list in a map
value. This would require embedding a bpf_list_node, support for which
is also included. The bpf_spin_lock is used to protect the bpf_list_head
and other data.
While we strictly don't require to hold a bpf_spin_lock while touching
the bpf_list_head in such objects, as when have access to it, we have
complete ownership of the object, the locking constraint is still kept
and may be conditionally lifted in the future.
Note that the specification of such types can be done just like map
values, e.g.:
struct bar {
struct bpf_list_node node;
};
struct foo {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(bar, node);
struct bpf_list_node node;
};
struct map_value {
struct bpf_spin_lock lock;
struct bpf_list_head head __contains(foo, node);
};
To recognize such types in user BTF, we build a btf_struct_metas array
of metadata items corresponding to each BTF ID. This is done once during
the btf_parse stage to avoid having to do it each time during the
verification process's requirement to inspect the metadata.
Moreover, the computed metadata needs to be passed to some helpers in
future patches which requires allocating them and storing them in the
BTF that is pinned by the program itself, so that valid access can be
assumed to such data during program runtime.
A key thing to note is that once a btf_struct_meta is available for a
type, both the btf_record and btf_field_offs should be available. It is
critical that btf_field_offs is available in case special fields are
present, as we extensively rely on special fields being zeroed out in
map values and allocated objects in later patches. The code ensures that
by bailing out in case of errors and ensuring both are available
together. If the record is not available, the special fields won't be
recognized, so not having both is also fine (in terms of being a
verification error and not a runtime bug).
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:56 +00:00
|
|
|
break;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
default:
|
|
|
|
|
WARN_ON_ONCE(1);
|
bpf: Wire up freeing of referenced kptr
A destructor kfunc can be defined as void func(type *), where type may
be void or any other pointer type as per convenience.
In this patch, we ensure that the type is sane and capture the function
pointer into off_desc of ptr_off_tab for the specific pointer offset,
with the invariant that the dtor pointer is always set when 'kptr_ref'
tag is applied to the pointer's pointee type, which is indicated by the
flag BPF_MAP_VALUE_OFF_F_REF.
Note that only BTF IDs whose destructor kfunc is registered, thus become
the allowed BTF IDs for embedding as referenced kptr. Hence it serves
the purpose of finding dtor kfunc BTF ID, as well acting as a check
against the whitelist of allowed BTF IDs for this purpose.
Finally, wire up the actual freeing of the referenced pointer if any at
all available offsets, so that no references are leaked after the BPF
map goes away and the BPF program previously moved the ownership a
referenced pointer into it.
The behavior is similar to BPF timers, where bpf_map_{update,delete}_elem
will free any existing referenced kptr. The same case is with LRU map's
bpf_lru_push_free/htab_lru_push_free functions, which are extended to
reset unreferenced and free referenced kptr.
Note that unlike BPF timers, kptr is not reset or freed when map uref
drops to zero.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-8-memxor@gmail.com
2022-04-24 21:48:55 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2014-09-26 07:16:57 +00:00
|
|
|
/* called from workqueue */
|
|
|
|
|
static void bpf_map_free_deferred(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = container_of(work, struct bpf_map, work);
|
bpf: Do btf_record_free outside map_free callback
Since the commit being fixed, we now miss freeing btf_record for local
storage maps which will have a btf_record populated in case they have
bpf_spin_lock element.
This was missed because I made the choice of offloading the job to free
kptr_off_tab (now btf_record) to the map_free callback when adding
support for kptrs.
Revisiting the reason for this decision, there is the possibility that
the btf_record gets used inside map_free callback (e.g. in case of maps
embedding kptrs) to iterate over them and free them, hence doing it
before the map_free callback would be leaking special field memory, and
do invalid memory access. The btf_record keeps module references which
is critical to ensure the dtor call made for referenced kptr is safe to
do.
If doing it after map_free callback, the map area is already freed, so
we cannot access bpf_map structure anymore.
To fix this and prevent such lapses in future, move bpf_map_free_record
out of the map_free callback, and do it after map_free by remembering
the btf_record pointer. There is no need to access bpf_map structure in
that case, and we can avoid missing this case when support for new map
types is added for other special fields.
Since a btf_record and its btf_field_offs are used together, for
consistency delay freeing of field_offs as well. While not a problem
right now, a lot of code assumes that either both record and field_offs
are set or none at once.
Note that in case of map of maps (outer maps), inner_map_meta->record is
only used during verification, not to free fields in map value, hence we
simply keep the bpf_map_free_record call as is in bpf_map_meta_free and
never touch map->inner_map_meta in bpf_map_free_deferred.
Add a comment making note of these details.
Fixes: db559117828d ("bpf: Consolidate spin_lock, timer management into btf_record")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:52 +00:00
|
|
|
struct btf_record *rec = map->record;
|
bpf: Fix a race condition between btf_put() and map_free()
When running `./test_progs -j` in my local vm with latest kernel,
I once hit a kasan error like below:
[ 1887.184724] BUG: KASAN: slab-use-after-free in bpf_rb_root_free+0x1f8/0x2b0
[ 1887.185599] Read of size 4 at addr ffff888106806910 by task kworker/u12:2/2830
[ 1887.186498]
[ 1887.186712] CPU: 3 PID: 2830 Comm: kworker/u12:2 Tainted: G OEL 6.7.0-rc3-00699-g90679706d486-dirty #494
[ 1887.188034] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014
[ 1887.189618] Workqueue: events_unbound bpf_map_free_deferred
[ 1887.190341] Call Trace:
[ 1887.190666] <TASK>
[ 1887.190949] dump_stack_lvl+0xac/0xe0
[ 1887.191423] ? nf_tcp_handle_invalid+0x1b0/0x1b0
[ 1887.192019] ? panic+0x3c0/0x3c0
[ 1887.192449] print_report+0x14f/0x720
[ 1887.192930] ? preempt_count_sub+0x1c/0xd0
[ 1887.193459] ? __virt_addr_valid+0xac/0x120
[ 1887.194004] ? bpf_rb_root_free+0x1f8/0x2b0
[ 1887.194572] kasan_report+0xc3/0x100
[ 1887.195085] ? bpf_rb_root_free+0x1f8/0x2b0
[ 1887.195668] bpf_rb_root_free+0x1f8/0x2b0
[ 1887.196183] ? __bpf_obj_drop_impl+0xb0/0xb0
[ 1887.196736] ? preempt_count_sub+0x1c/0xd0
[ 1887.197270] ? preempt_count_sub+0x1c/0xd0
[ 1887.197802] ? _raw_spin_unlock+0x1f/0x40
[ 1887.198319] bpf_obj_free_fields+0x1d4/0x260
[ 1887.198883] array_map_free+0x1a3/0x260
[ 1887.199380] bpf_map_free_deferred+0x7b/0xe0
[ 1887.199943] process_scheduled_works+0x3a2/0x6c0
[ 1887.200549] worker_thread+0x633/0x890
[ 1887.201047] ? __kthread_parkme+0xd7/0xf0
[ 1887.201574] ? kthread+0x102/0x1d0
[ 1887.202020] kthread+0x1ab/0x1d0
[ 1887.202447] ? pr_cont_work+0x270/0x270
[ 1887.202954] ? kthread_blkcg+0x50/0x50
[ 1887.203444] ret_from_fork+0x34/0x50
[ 1887.203914] ? kthread_blkcg+0x50/0x50
[ 1887.204397] ret_from_fork_asm+0x11/0x20
[ 1887.204913] </TASK>
[ 1887.204913] </TASK>
[ 1887.205209]
[ 1887.205416] Allocated by task 2197:
[ 1887.205881] kasan_set_track+0x3f/0x60
[ 1887.206366] __kasan_kmalloc+0x6e/0x80
[ 1887.206856] __kmalloc+0xac/0x1a0
[ 1887.207293] btf_parse_fields+0xa15/0x1480
[ 1887.207836] btf_parse_struct_metas+0x566/0x670
[ 1887.208387] btf_new_fd+0x294/0x4d0
[ 1887.208851] __sys_bpf+0x4ba/0x600
[ 1887.209292] __x64_sys_bpf+0x41/0x50
[ 1887.209762] do_syscall_64+0x4c/0xf0
[ 1887.210222] entry_SYSCALL_64_after_hwframe+0x63/0x6b
[ 1887.210868]
[ 1887.211074] Freed by task 36:
[ 1887.211460] kasan_set_track+0x3f/0x60
[ 1887.211951] kasan_save_free_info+0x28/0x40
[ 1887.212485] ____kasan_slab_free+0x101/0x180
[ 1887.213027] __kmem_cache_free+0xe4/0x210
[ 1887.213514] btf_free+0x5b/0x130
[ 1887.213918] rcu_core+0x638/0xcc0
[ 1887.214347] __do_softirq+0x114/0x37e
The error happens at bpf_rb_root_free+0x1f8/0x2b0:
00000000000034c0 <bpf_rb_root_free>:
; {
34c0: f3 0f 1e fa endbr64
34c4: e8 00 00 00 00 callq 0x34c9 <bpf_rb_root_free+0x9>
34c9: 55 pushq %rbp
34ca: 48 89 e5 movq %rsp, %rbp
...
; if (rec && rec->refcount_off >= 0 &&
36aa: 4d 85 ed testq %r13, %r13
36ad: 74 a9 je 0x3658 <bpf_rb_root_free+0x198>
36af: 49 8d 7d 10 leaq 0x10(%r13), %rdi
36b3: e8 00 00 00 00 callq 0x36b8 <bpf_rb_root_free+0x1f8>
<==== kasan function
36b8: 45 8b 7d 10 movl 0x10(%r13), %r15d
<==== use-after-free load
36bc: 45 85 ff testl %r15d, %r15d
36bf: 78 8c js 0x364d <bpf_rb_root_free+0x18d>
So the problem is at rec->refcount_off in the above.
I did some source code analysis and find the reason.
CPU A CPU B
bpf_map_put:
...
btf_put with rcu callback
...
bpf_map_free_deferred
with system_unbound_wq
... ... ...
... btf_free_rcu: ...
... ... bpf_map_free_deferred:
... ...
... ---------> btf_struct_metas_free()
... | race condition ...
... ---------> map->ops->map_free()
...
... btf->struct_meta_tab = NULL
In the above, map_free() corresponds to array_map_free() and eventually
calling bpf_rb_root_free() which calls:
...
__bpf_obj_drop_impl(obj, field->graph_root.value_rec, false);
...
Here, 'value_rec' is assigned in btf_check_and_fixup_fields() with following code:
meta = btf_find_struct_meta(btf, btf_id);
if (!meta)
return -EFAULT;
rec->fields[i].graph_root.value_rec = meta->record;
So basically, 'value_rec' is a pointer to the record in struct_metas_tab.
And it is possible that that particular record has been freed by
btf_struct_metas_free() and hence we have a kasan error here.
Actually it is very hard to reproduce the failure with current bpf/bpf-next
code, I only got the above error once. To increase reproducibility, I added
a delay in bpf_map_free_deferred() to delay map->ops->map_free(), which
significantly increased reproducibility.
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 5e43ddd1b83f..aae5b5213e93 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -695,6 +695,7 @@ static void bpf_map_free_deferred(struct work_struct *work)
struct bpf_map *map = container_of(work, struct bpf_map, work);
struct btf_record *rec = map->record;
+ mdelay(100);
security_bpf_map_free(map);
bpf_map_release_memcg(map);
/* implementation dependent freeing */
Hao also provided test cases ([1]) for easily reproducing the above issue.
There are two ways to fix the issue, the v1 of the patch ([2]) moving
btf_put() after map_free callback, and the v5 of the patch ([3]) using
a kptr style fix which tries to get a btf reference during
map_check_btf(). Each approach has its pro and cons. The first approach
delays freeing btf while the second approach needs to acquire reference
depending on context which makes logic not very elegant and may
complicate things with future new data structures. Alexei
suggested in [4] going back to v1 which is what this patch
tries to do.
Rerun './test_progs -j' with the above mdelay() hack for a couple
of times and didn't observe the error for the above rb_root test cases.
Running Hou's test ([1]) is also successful.
[1] https://lore.kernel.org/bpf/20231207141500.917136-1-houtao@huaweicloud.com/
[2] v1: https://lore.kernel.org/bpf/20231204173946.3066377-1-yonghong.song@linux.dev/
[3] v5: https://lore.kernel.org/bpf/20231208041621.2968241-1-yonghong.song@linux.dev/
[4] v4: https://lore.kernel.org/bpf/CAADnVQJ3FiXUhZJwX_81sjZvSYYKCFB3BT6P8D59RS2Gu+0Z7g@mail.gmail.com/
Cc: Hou Tao <houtao@huaweicloud.com>
Fixes: 958cf2e273f0 ("bpf: Introduce bpf_obj_new")
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20231214203815.1469107-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-12-14 20:38:15 +00:00
|
|
|
struct btf *btf = map->btf;
|
2014-09-26 07:16:57 +00:00
|
|
|
|
2017-10-18 20:00:24 +00:00
|
|
|
security_bpf_map_free(map);
|
2020-12-01 21:58:32 +00:00
|
|
|
bpf_map_release_memcg(map);
|
bpf: Do btf_record_free outside map_free callback
Since the commit being fixed, we now miss freeing btf_record for local
storage maps which will have a btf_record populated in case they have
bpf_spin_lock element.
This was missed because I made the choice of offloading the job to free
kptr_off_tab (now btf_record) to the map_free callback when adding
support for kptrs.
Revisiting the reason for this decision, there is the possibility that
the btf_record gets used inside map_free callback (e.g. in case of maps
embedding kptrs) to iterate over them and free them, hence doing it
before the map_free callback would be leaking special field memory, and
do invalid memory access. The btf_record keeps module references which
is critical to ensure the dtor call made for referenced kptr is safe to
do.
If doing it after map_free callback, the map area is already freed, so
we cannot access bpf_map structure anymore.
To fix this and prevent such lapses in future, move bpf_map_free_record
out of the map_free callback, and do it after map_free by remembering
the btf_record pointer. There is no need to access bpf_map structure in
that case, and we can avoid missing this case when support for new map
types is added for other special fields.
Since a btf_record and its btf_field_offs are used together, for
consistency delay freeing of field_offs as well. While not a problem
right now, a lot of code assumes that either both record and field_offs
are set or none at once.
Note that in case of map of maps (outer maps), inner_map_meta->record is
only used during verification, not to free fields in map value, hence we
simply keep the bpf_map_free_record call as is in bpf_map_meta_free and
never touch map->inner_map_meta in bpf_map_free_deferred.
Add a comment making note of these details.
Fixes: db559117828d ("bpf: Consolidate spin_lock, timer management into btf_record")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:52 +00:00
|
|
|
/* implementation dependent freeing */
|
2014-09-26 07:16:57 +00:00
|
|
|
map->ops->map_free(map);
|
bpf: Remove btf_field_offs, use btf_record's fields instead
The btf_field_offs struct contains (offset, size) for btf_record fields,
sorted by offset. btf_field_offs is always used in conjunction with
btf_record, which has btf_field 'fields' array with (offset, type), the
latter of which btf_field_offs' size is derived from via
btf_field_type_size.
This patch adds a size field to struct btf_field and sorts btf_record's
fields by offset, making it possible to get rid of btf_field_offs. Less
data duplication and less code complexity results.
Since btf_field_offs' lifetime closely followed the btf_record used to
populate it, most complexity wins are from removal of initialization
code like:
if (btf_record_successfully_initialized) {
foffs = btf_parse_field_offs(rec);
if (IS_ERR_OR_NULL(foffs))
// free the btf_record and return err
}
Other changes in this patch are pretty mechanical:
* foffs->field_off[i] -> rec->fields[i].offset
* foffs->field_sz[i] -> rec->fields[i].size
* Sort rec->fields in btf_parse_fields before returning
* It's possible that this is necessary independently of other
changes in this patch. btf_record_find in syscall.c expects
btf_record's fields to be sorted by offset, yet there's no
explicit sorting of them before this patch, record's fields are
populated in the order they're read from BTF struct definition.
BTF docs don't say anything about the sortedness of struct fields.
* All functions taking struct btf_field_offs * input now instead take
struct btf_record *. All callsites of these functions already have
access to the correct btf_record.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230415201811.343116-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-04-15 20:18:03 +00:00
|
|
|
/* Delay freeing of btf_record for maps, as map_free
|
bpf: Do btf_record_free outside map_free callback
Since the commit being fixed, we now miss freeing btf_record for local
storage maps which will have a btf_record populated in case they have
bpf_spin_lock element.
This was missed because I made the choice of offloading the job to free
kptr_off_tab (now btf_record) to the map_free callback when adding
support for kptrs.
Revisiting the reason for this decision, there is the possibility that
the btf_record gets used inside map_free callback (e.g. in case of maps
embedding kptrs) to iterate over them and free them, hence doing it
before the map_free callback would be leaking special field memory, and
do invalid memory access. The btf_record keeps module references which
is critical to ensure the dtor call made for referenced kptr is safe to
do.
If doing it after map_free callback, the map area is already freed, so
we cannot access bpf_map structure anymore.
To fix this and prevent such lapses in future, move bpf_map_free_record
out of the map_free callback, and do it after map_free by remembering
the btf_record pointer. There is no need to access bpf_map structure in
that case, and we can avoid missing this case when support for new map
types is added for other special fields.
Since a btf_record and its btf_field_offs are used together, for
consistency delay freeing of field_offs as well. While not a problem
right now, a lot of code assumes that either both record and field_offs
are set or none at once.
Note that in case of map of maps (outer maps), inner_map_meta->record is
only used during verification, not to free fields in map value, hence we
simply keep the bpf_map_free_record call as is in bpf_map_meta_free and
never touch map->inner_map_meta in bpf_map_free_deferred.
Add a comment making note of these details.
Fixes: db559117828d ("bpf: Consolidate spin_lock, timer management into btf_record")
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221118015614.2013203-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-18 01:55:52 +00:00
|
|
|
* callback usually needs access to them. It is better to do it here
|
|
|
|
|
* than require each callback to do the free itself manually.
|
|
|
|
|
*
|
|
|
|
|
* Note that the btf_record stashed in map->inner_map_meta->record was
|
|
|
|
|
* already freed using the map_free callback for map in map case which
|
|
|
|
|
* eventually calls bpf_map_free_meta, since inner_map_meta is only a
|
|
|
|
|
* template bpf_map struct used during verification.
|
|
|
|
|
*/
|
|
|
|
|
btf_record_free(rec);
|
bpf: Fix a race condition between btf_put() and map_free()
When running `./test_progs -j` in my local vm with latest kernel,
I once hit a kasan error like below:
[ 1887.184724] BUG: KASAN: slab-use-after-free in bpf_rb_root_free+0x1f8/0x2b0
[ 1887.185599] Read of size 4 at addr ffff888106806910 by task kworker/u12:2/2830
[ 1887.186498]
[ 1887.186712] CPU: 3 PID: 2830 Comm: kworker/u12:2 Tainted: G OEL 6.7.0-rc3-00699-g90679706d486-dirty #494
[ 1887.188034] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014
[ 1887.189618] Workqueue: events_unbound bpf_map_free_deferred
[ 1887.190341] Call Trace:
[ 1887.190666] <TASK>
[ 1887.190949] dump_stack_lvl+0xac/0xe0
[ 1887.191423] ? nf_tcp_handle_invalid+0x1b0/0x1b0
[ 1887.192019] ? panic+0x3c0/0x3c0
[ 1887.192449] print_report+0x14f/0x720
[ 1887.192930] ? preempt_count_sub+0x1c/0xd0
[ 1887.193459] ? __virt_addr_valid+0xac/0x120
[ 1887.194004] ? bpf_rb_root_free+0x1f8/0x2b0
[ 1887.194572] kasan_report+0xc3/0x100
[ 1887.195085] ? bpf_rb_root_free+0x1f8/0x2b0
[ 1887.195668] bpf_rb_root_free+0x1f8/0x2b0
[ 1887.196183] ? __bpf_obj_drop_impl+0xb0/0xb0
[ 1887.196736] ? preempt_count_sub+0x1c/0xd0
[ 1887.197270] ? preempt_count_sub+0x1c/0xd0
[ 1887.197802] ? _raw_spin_unlock+0x1f/0x40
[ 1887.198319] bpf_obj_free_fields+0x1d4/0x260
[ 1887.198883] array_map_free+0x1a3/0x260
[ 1887.199380] bpf_map_free_deferred+0x7b/0xe0
[ 1887.199943] process_scheduled_works+0x3a2/0x6c0
[ 1887.200549] worker_thread+0x633/0x890
[ 1887.201047] ? __kthread_parkme+0xd7/0xf0
[ 1887.201574] ? kthread+0x102/0x1d0
[ 1887.202020] kthread+0x1ab/0x1d0
[ 1887.202447] ? pr_cont_work+0x270/0x270
[ 1887.202954] ? kthread_blkcg+0x50/0x50
[ 1887.203444] ret_from_fork+0x34/0x50
[ 1887.203914] ? kthread_blkcg+0x50/0x50
[ 1887.204397] ret_from_fork_asm+0x11/0x20
[ 1887.204913] </TASK>
[ 1887.204913] </TASK>
[ 1887.205209]
[ 1887.205416] Allocated by task 2197:
[ 1887.205881] kasan_set_track+0x3f/0x60
[ 1887.206366] __kasan_kmalloc+0x6e/0x80
[ 1887.206856] __kmalloc+0xac/0x1a0
[ 1887.207293] btf_parse_fields+0xa15/0x1480
[ 1887.207836] btf_parse_struct_metas+0x566/0x670
[ 1887.208387] btf_new_fd+0x294/0x4d0
[ 1887.208851] __sys_bpf+0x4ba/0x600
[ 1887.209292] __x64_sys_bpf+0x41/0x50
[ 1887.209762] do_syscall_64+0x4c/0xf0
[ 1887.210222] entry_SYSCALL_64_after_hwframe+0x63/0x6b
[ 1887.210868]
[ 1887.211074] Freed by task 36:
[ 1887.211460] kasan_set_track+0x3f/0x60
[ 1887.211951] kasan_save_free_info+0x28/0x40
[ 1887.212485] ____kasan_slab_free+0x101/0x180
[ 1887.213027] __kmem_cache_free+0xe4/0x210
[ 1887.213514] btf_free+0x5b/0x130
[ 1887.213918] rcu_core+0x638/0xcc0
[ 1887.214347] __do_softirq+0x114/0x37e
The error happens at bpf_rb_root_free+0x1f8/0x2b0:
00000000000034c0 <bpf_rb_root_free>:
; {
34c0: f3 0f 1e fa endbr64
34c4: e8 00 00 00 00 callq 0x34c9 <bpf_rb_root_free+0x9>
34c9: 55 pushq %rbp
34ca: 48 89 e5 movq %rsp, %rbp
...
; if (rec && rec->refcount_off >= 0 &&
36aa: 4d 85 ed testq %r13, %r13
36ad: 74 a9 je 0x3658 <bpf_rb_root_free+0x198>
36af: 49 8d 7d 10 leaq 0x10(%r13), %rdi
36b3: e8 00 00 00 00 callq 0x36b8 <bpf_rb_root_free+0x1f8>
<==== kasan function
36b8: 45 8b 7d 10 movl 0x10(%r13), %r15d
<==== use-after-free load
36bc: 45 85 ff testl %r15d, %r15d
36bf: 78 8c js 0x364d <bpf_rb_root_free+0x18d>
So the problem is at rec->refcount_off in the above.
I did some source code analysis and find the reason.
CPU A CPU B
bpf_map_put:
...
btf_put with rcu callback
...
bpf_map_free_deferred
with system_unbound_wq
... ... ...
... btf_free_rcu: ...
... ... bpf_map_free_deferred:
... ...
... ---------> btf_struct_metas_free()
... | race condition ...
... ---------> map->ops->map_free()
...
... btf->struct_meta_tab = NULL
In the above, map_free() corresponds to array_map_free() and eventually
calling bpf_rb_root_free() which calls:
...
__bpf_obj_drop_impl(obj, field->graph_root.value_rec, false);
...
Here, 'value_rec' is assigned in btf_check_and_fixup_fields() with following code:
meta = btf_find_struct_meta(btf, btf_id);
if (!meta)
return -EFAULT;
rec->fields[i].graph_root.value_rec = meta->record;
So basically, 'value_rec' is a pointer to the record in struct_metas_tab.
And it is possible that that particular record has been freed by
btf_struct_metas_free() and hence we have a kasan error here.
Actually it is very hard to reproduce the failure with current bpf/bpf-next
code, I only got the above error once. To increase reproducibility, I added
a delay in bpf_map_free_deferred() to delay map->ops->map_free(), which
significantly increased reproducibility.
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 5e43ddd1b83f..aae5b5213e93 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -695,6 +695,7 @@ static void bpf_map_free_deferred(struct work_struct *work)
struct bpf_map *map = container_of(work, struct bpf_map, work);
struct btf_record *rec = map->record;
+ mdelay(100);
security_bpf_map_free(map);
bpf_map_release_memcg(map);
/* implementation dependent freeing */
Hao also provided test cases ([1]) for easily reproducing the above issue.
There are two ways to fix the issue, the v1 of the patch ([2]) moving
btf_put() after map_free callback, and the v5 of the patch ([3]) using
a kptr style fix which tries to get a btf reference during
map_check_btf(). Each approach has its pro and cons. The first approach
delays freeing btf while the second approach needs to acquire reference
depending on context which makes logic not very elegant and may
complicate things with future new data structures. Alexei
suggested in [4] going back to v1 which is what this patch
tries to do.
Rerun './test_progs -j' with the above mdelay() hack for a couple
of times and didn't observe the error for the above rb_root test cases.
Running Hou's test ([1]) is also successful.
[1] https://lore.kernel.org/bpf/20231207141500.917136-1-houtao@huaweicloud.com/
[2] v1: https://lore.kernel.org/bpf/20231204173946.3066377-1-yonghong.song@linux.dev/
[3] v5: https://lore.kernel.org/bpf/20231208041621.2968241-1-yonghong.song@linux.dev/
[4] v4: https://lore.kernel.org/bpf/CAADnVQJ3FiXUhZJwX_81sjZvSYYKCFB3BT6P8D59RS2Gu+0Z7g@mail.gmail.com/
Cc: Hou Tao <houtao@huaweicloud.com>
Fixes: 958cf2e273f0 ("bpf: Introduce bpf_obj_new")
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20231214203815.1469107-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-12-14 20:38:15 +00:00
|
|
|
/* Delay freeing of btf for maps, as map_free callback may need
|
|
|
|
|
* struct_meta info which will be freed with btf_put().
|
|
|
|
|
*/
|
|
|
|
|
btf_put(btf);
|
2014-09-26 07:16:57 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
static void bpf_map_put_uref(struct bpf_map *map)
|
|
|
|
|
{
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
if (atomic64_dec_and_test(&map->usercnt)) {
|
2018-04-23 22:39:23 +00:00
|
|
|
if (map->ops->map_release_uref)
|
|
|
|
|
map->ops->map_release_uref(map);
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2023-12-04 14:04:22 +00:00
|
|
|
static void bpf_map_free_in_work(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
INIT_WORK(&map->work, bpf_map_free_deferred);
|
|
|
|
|
/* Avoid spawning kworkers, since they all might contend
|
|
|
|
|
* for the same mutex like slab_mutex.
|
|
|
|
|
*/
|
|
|
|
|
queue_work(system_unbound_wq, &map->work);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_map_free_rcu_gp(struct rcu_head *rcu)
|
|
|
|
|
{
|
|
|
|
|
bpf_map_free_in_work(container_of(rcu, struct bpf_map, rcu));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_map_free_mult_rcu_gp(struct rcu_head *rcu)
|
|
|
|
|
{
|
|
|
|
|
if (rcu_trace_implies_rcu_gp())
|
|
|
|
|
bpf_map_free_rcu_gp(rcu);
|
|
|
|
|
else
|
|
|
|
|
call_rcu(rcu, bpf_map_free_rcu_gp);
|
|
|
|
|
}
|
|
|
|
|
|
2014-09-26 07:16:57 +00:00
|
|
|
/* decrement map refcnt and schedule it for freeing via workqueue
|
2023-02-02 14:19:21 +00:00
|
|
|
* (underlying map implementation ops->map_free() might sleep)
|
2014-09-26 07:16:57 +00:00
|
|
|
*/
|
2023-02-02 14:19:21 +00:00
|
|
|
void bpf_map_put(struct bpf_map *map)
|
2014-09-26 07:16:57 +00:00
|
|
|
{
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
if (atomic64_dec_and_test(&map->refcnt)) {
|
2017-06-05 19:15:48 +00:00
|
|
|
/* bpf_map_free_id() must be called first */
|
2023-02-02 14:19:21 +00:00
|
|
|
bpf_map_free_id(map);
|
2023-12-04 14:04:22 +00:00
|
|
|
|
2023-12-04 14:04:23 +00:00
|
|
|
WARN_ON_ONCE(atomic64_read(&map->sleepable_refcnt));
|
2023-12-04 14:04:22 +00:00
|
|
|
if (READ_ONCE(map->free_after_mult_rcu_gp))
|
|
|
|
|
call_rcu_tasks_trace(&map->rcu, bpf_map_free_mult_rcu_gp);
|
2023-12-04 14:04:23 +00:00
|
|
|
else if (READ_ONCE(map->free_after_rcu_gp))
|
|
|
|
|
call_rcu(&map->rcu, bpf_map_free_rcu_gp);
|
2023-12-04 14:04:22 +00:00
|
|
|
else
|
|
|
|
|
bpf_map_free_in_work(map);
|
2014-09-26 07:16:57 +00:00
|
|
|
}
|
|
|
|
|
}
|
2018-05-04 01:37:09 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_map_put);
|
2017-06-05 19:15:50 +00:00
|
|
|
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
void bpf_map_put_with_uref(struct bpf_map *map)
|
2014-09-26 07:16:57 +00:00
|
|
|
{
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
bpf_map_put_uref(map);
|
2014-09-26 07:16:57 +00:00
|
|
|
bpf_map_put(map);
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_map_release(struct inode *inode, struct file *filp)
|
|
|
|
|
{
|
2016-06-15 20:47:12 +00:00
|
|
|
struct bpf_map *map = filp->private_data;
|
|
|
|
|
|
|
|
|
|
if (map->ops->map_release)
|
|
|
|
|
map->ops->map_release(map, filp);
|
|
|
|
|
|
|
|
|
|
bpf_map_put_with_uref(map);
|
2014-09-26 07:16:57 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
static fmode_t map_get_sys_perms(struct bpf_map *map, struct fd f)
|
|
|
|
|
{
|
|
|
|
|
fmode_t mode = f.file->f_mode;
|
|
|
|
|
|
|
|
|
|
/* Our file permissions may have been overridden by global
|
|
|
|
|
* map permissions facing syscall side.
|
|
|
|
|
*/
|
|
|
|
|
if (READ_ONCE(map->frozen))
|
|
|
|
|
mode &= ~FMODE_CAN_WRITE;
|
|
|
|
|
return mode;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-19 10:56:22 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
2023-03-05 12:45:58 +00:00
|
|
|
/* Show the memory usage of a bpf map */
|
|
|
|
|
static u64 bpf_map_memory_usage(const struct bpf_map *map)
|
2020-12-01 21:58:58 +00:00
|
|
|
{
|
2023-03-05 12:46:15 +00:00
|
|
|
return map->ops->map_mem_usage(map);
|
2020-12-01 21:58:58 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-19 10:56:22 +00:00
|
|
|
static void bpf_map_show_fdinfo(struct seq_file *m, struct file *filp)
|
|
|
|
|
{
|
2022-01-21 10:10:02 +00:00
|
|
|
struct bpf_map *map = filp->private_data;
|
2019-11-22 20:07:56 +00:00
|
|
|
u32 type = 0, jited = 0;
|
2016-11-26 00:28:07 +00:00
|
|
|
|
2022-01-21 10:10:02 +00:00
|
|
|
if (map_type_contains_progs(map)) {
|
|
|
|
|
spin_lock(&map->owner.lock);
|
|
|
|
|
type = map->owner.type;
|
|
|
|
|
jited = map->owner.jited;
|
|
|
|
|
spin_unlock(&map->owner.lock);
|
2016-11-26 00:28:07 +00:00
|
|
|
}
|
2015-11-19 10:56:22 +00:00
|
|
|
|
|
|
|
|
seq_printf(m,
|
|
|
|
|
"map_type:\t%u\n"
|
|
|
|
|
"key_size:\t%u\n"
|
|
|
|
|
"value_size:\t%u\n"
|
2016-03-24 23:30:25 +00:00
|
|
|
"max_entries:\t%u\n"
|
2016-11-26 00:28:07 +00:00
|
|
|
"map_flags:\t%#x\n"
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
"map_extra:\t%#llx\n"
|
2023-03-05 12:45:58 +00:00
|
|
|
"memlock:\t%llu\n"
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
"map_id:\t%u\n"
|
|
|
|
|
"frozen:\t%u\n",
|
2015-11-19 10:56:22 +00:00
|
|
|
map->map_type,
|
|
|
|
|
map->key_size,
|
|
|
|
|
map->value_size,
|
2016-03-24 23:30:25 +00:00
|
|
|
map->max_entries,
|
2016-11-26 00:28:07 +00:00
|
|
|
map->map_flags,
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
(unsigned long long)map->map_extra,
|
2023-03-05 12:45:58 +00:00
|
|
|
bpf_map_memory_usage(map),
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
map->id,
|
|
|
|
|
READ_ONCE(map->frozen));
|
2019-11-22 20:07:56 +00:00
|
|
|
if (type) {
|
|
|
|
|
seq_printf(m, "owner_prog_type:\t%u\n", type);
|
|
|
|
|
seq_printf(m, "owner_jited:\t%u\n", jited);
|
2017-07-02 00:13:28 +00:00
|
|
|
}
|
2015-11-19 10:56:22 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
static ssize_t bpf_dummy_read(struct file *filp, char __user *buf, size_t siz,
|
|
|
|
|
loff_t *ppos)
|
|
|
|
|
{
|
|
|
|
|
/* We need this handler such that alloc_file() enables
|
|
|
|
|
* f_mode with FMODE_CAN_READ.
|
|
|
|
|
*/
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static ssize_t bpf_dummy_write(struct file *filp, const char __user *buf,
|
|
|
|
|
size_t siz, loff_t *ppos)
|
|
|
|
|
{
|
|
|
|
|
/* We need this handler such that alloc_file() enables
|
|
|
|
|
* f_mode with FMODE_CAN_WRITE.
|
|
|
|
|
*/
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
/* called for any extra memory-mapped regions (except initial) */
|
|
|
|
|
static void bpf_map_mmap_open(struct vm_area_struct *vma)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = vma->vm_file->private_data;
|
|
|
|
|
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
if (vma->vm_flags & VM_MAYWRITE)
|
|
|
|
|
bpf_map_write_active_inc(map);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* called for all unmapped memory region (including initial) */
|
|
|
|
|
static void bpf_map_mmap_close(struct vm_area_struct *vma)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = vma->vm_file->private_data;
|
|
|
|
|
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
if (vma->vm_flags & VM_MAYWRITE)
|
|
|
|
|
bpf_map_write_active_dec(map);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static const struct vm_operations_struct bpf_map_default_vmops = {
|
|
|
|
|
.open = bpf_map_mmap_open,
|
|
|
|
|
.close = bpf_map_mmap_close,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int bpf_map_mmap(struct file *filp, struct vm_area_struct *vma)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = filp->private_data;
|
|
|
|
|
int err;
|
|
|
|
|
|
2022-11-03 19:09:56 +00:00
|
|
|
if (!map->ops->map_mmap || !IS_ERR_OR_NULL(map->record))
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
return -ENOTSUPP;
|
|
|
|
|
|
|
|
|
|
if (!(vma->vm_flags & VM_SHARED))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
mutex_lock(&map->freeze_mutex);
|
|
|
|
|
|
2020-05-19 05:38:24 +00:00
|
|
|
if (vma->vm_flags & VM_WRITE) {
|
|
|
|
|
if (map->frozen) {
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
/* map is meant to be read-only, so do not allow mapping as
|
|
|
|
|
* writable, because it's possible to leak a writable page
|
|
|
|
|
* reference and allows user-space to still modify it after
|
|
|
|
|
* freezing, while verifier will assume contents do not change
|
|
|
|
|
*/
|
|
|
|
|
if (map->map_flags & BPF_F_RDONLY_PROG) {
|
|
|
|
|
err = -EACCES;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* set default open/close callbacks */
|
|
|
|
|
vma->vm_ops = &bpf_map_default_vmops;
|
|
|
|
|
vma->vm_private_data = map;
|
2023-01-26 19:37:49 +00:00
|
|
|
vm_flags_clear(vma, VM_MAYEXEC);
|
2020-04-10 20:26:12 +00:00
|
|
|
if (!(vma->vm_flags & VM_WRITE))
|
|
|
|
|
/* disallow re-mapping with PROT_WRITE */
|
2023-01-26 19:37:49 +00:00
|
|
|
vm_flags_clear(vma, VM_MAYWRITE);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
|
|
|
|
|
err = map->ops->map_mmap(map, vma);
|
|
|
|
|
if (err)
|
|
|
|
|
goto out;
|
|
|
|
|
|
2020-04-10 20:26:12 +00:00
|
|
|
if (vma->vm_flags & VM_MAYWRITE)
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_inc(map);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
out:
|
|
|
|
|
mutex_unlock(&map->freeze_mutex);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Implement BPF ring buffer and verifier support for it
This commit adds a new MPSC ring buffer implementation into BPF ecosystem,
which allows multiple CPUs to submit data to a single shared ring buffer. On
the consumption side, only single consumer is assumed.
Motivation
----------
There are two distinctive motivators for this work, which are not satisfied by
existing perf buffer, which prompted creation of a new ring buffer
implementation.
- more efficient memory utilization by sharing ring buffer across CPUs;
- preserving ordering of events that happen sequentially in time, even
across multiple CPUs (e.g., fork/exec/exit events for a task).
These two problems are independent, but perf buffer fails to satisfy both.
Both are a result of a choice to have per-CPU perf ring buffer. Both can be
also solved by having an MPSC implementation of ring buffer. The ordering
problem could technically be solved for perf buffer with some in-kernel
counting, but given the first one requires an MPSC buffer, the same solution
would solve the second problem automatically.
Semantics and APIs
------------------
Single ring buffer is presented to BPF programs as an instance of BPF map of
type BPF_MAP_TYPE_RINGBUF. Two other alternatives considered, but ultimately
rejected.
One way would be to, similar to BPF_MAP_TYPE_PERF_EVENT_ARRAY, make
BPF_MAP_TYPE_RINGBUF could represent an array of ring buffers, but not enforce
"same CPU only" rule. This would be more familiar interface compatible with
existing perf buffer use in BPF, but would fail if application needed more
advanced logic to lookup ring buffer by arbitrary key. HASH_OF_MAPS addresses
this with current approach. Additionally, given the performance of BPF
ringbuf, many use cases would just opt into a simple single ring buffer shared
among all CPUs, for which current approach would be an overkill.
Another approach could introduce a new concept, alongside BPF map, to
represent generic "container" object, which doesn't necessarily have key/value
interface with lookup/update/delete operations. This approach would add a lot
of extra infrastructure that has to be built for observability and verifier
support. It would also add another concept that BPF developers would have to
familiarize themselves with, new syntax in libbpf, etc. But then would really
provide no additional benefits over the approach of using a map.
BPF_MAP_TYPE_RINGBUF doesn't support lookup/update/delete operations, but so
doesn't few other map types (e.g., queue and stack; array doesn't support
delete, etc).
The approach chosen has an advantage of re-using existing BPF map
infrastructure (introspection APIs in kernel, libbpf support, etc), being
familiar concept (no need to teach users a new type of object in BPF program),
and utilizing existing tooling (bpftool). For common scenario of using
a single ring buffer for all CPUs, it's as simple and straightforward, as
would be with a dedicated "container" object. On the other hand, by being
a map, it can be combined with ARRAY_OF_MAPS and HASH_OF_MAPS map-in-maps to
implement a wide variety of topologies, from one ring buffer for each CPU
(e.g., as a replacement for perf buffer use cases), to a complicated
application hashing/sharding of ring buffers (e.g., having a small pool of
ring buffers with hashed task's tgid being a look up key to preserve order,
but reduce contention).
Key and value sizes are enforced to be zero. max_entries is used to specify
the size of ring buffer and has to be a power of 2 value.
There are a bunch of similarities between perf buffer
(BPF_MAP_TYPE_PERF_EVENT_ARRAY) and new BPF ring buffer semantics:
- variable-length records;
- if there is no more space left in ring buffer, reservation fails, no
blocking;
- memory-mappable data area for user-space applications for ease of
consumption and high performance;
- epoll notifications for new incoming data;
- but still the ability to do busy polling for new data to achieve the
lowest latency, if necessary.
BPF ringbuf provides two sets of APIs to BPF programs:
- bpf_ringbuf_output() allows to *copy* data from one place to a ring
buffer, similarly to bpf_perf_event_output();
- bpf_ringbuf_reserve()/bpf_ringbuf_commit()/bpf_ringbuf_discard() APIs
split the whole process into two steps. First, a fixed amount of space is
reserved. If successful, a pointer to a data inside ring buffer data area
is returned, which BPF programs can use similarly to a data inside
array/hash maps. Once ready, this piece of memory is either committed or
discarded. Discard is similar to commit, but makes consumer ignore the
record.
bpf_ringbuf_output() has disadvantage of incurring extra memory copy, because
record has to be prepared in some other place first. But it allows to submit
records of the length that's not known to verifier beforehand. It also closely
matches bpf_perf_event_output(), so will simplify migration significantly.
bpf_ringbuf_reserve() avoids the extra copy of memory by providing a memory
pointer directly to ring buffer memory. In a lot of cases records are larger
than BPF stack space allows, so many programs have use extra per-CPU array as
a temporary heap for preparing sample. bpf_ringbuf_reserve() avoid this needs
completely. But in exchange, it only allows a known constant size of memory to
be reserved, such that verifier can verify that BPF program can't access
memory outside its reserved record space. bpf_ringbuf_output(), while slightly
slower due to extra memory copy, covers some use cases that are not suitable
for bpf_ringbuf_reserve().
The difference between commit and discard is very small. Discard just marks
a record as discarded, and such records are supposed to be ignored by consumer
code. Discard is useful for some advanced use-cases, such as ensuring
all-or-nothing multi-record submission, or emulating temporary malloc()/free()
within single BPF program invocation.
Each reserved record is tracked by verifier through existing
reference-tracking logic, similar to socket ref-tracking. It is thus
impossible to reserve a record, but forget to submit (or discard) it.
bpf_ringbuf_query() helper allows to query various properties of ring buffer.
Currently 4 are supported:
- BPF_RB_AVAIL_DATA returns amount of unconsumed data in ring buffer;
- BPF_RB_RING_SIZE returns the size of ring buffer;
- BPF_RB_CONS_POS/BPF_RB_PROD_POS returns current logical possition of
consumer/producer, respectively.
Returned values are momentarily snapshots of ring buffer state and could be
off by the time helper returns, so this should be used only for
debugging/reporting reasons or for implementing various heuristics, that take
into account highly-changeable nature of some of those characteristics.
One such heuristic might involve more fine-grained control over poll/epoll
notifications about new data availability in ring buffer. Together with
BPF_RB_NO_WAKEUP/BPF_RB_FORCE_WAKEUP flags for output/commit/discard helpers,
it allows BPF program a high degree of control and, e.g., more efficient
batched notifications. Default self-balancing strategy, though, should be
adequate for most applications and will work reliable and efficiently already.
Design and implementation
-------------------------
This reserve/commit schema allows a natural way for multiple producers, either
on different CPUs or even on the same CPU/in the same BPF program, to reserve
independent records and work with them without blocking other producers. This
means that if BPF program was interruped by another BPF program sharing the
same ring buffer, they will both get a record reserved (provided there is
enough space left) and can work with it and submit it independently. This
applies to NMI context as well, except that due to using a spinlock during
reservation, in NMI context, bpf_ringbuf_reserve() might fail to get a lock,
in which case reservation will fail even if ring buffer is not full.
The ring buffer itself internally is implemented as a power-of-2 sized
circular buffer, with two logical and ever-increasing counters (which might
wrap around on 32-bit architectures, that's not a problem):
- consumer counter shows up to which logical position consumer consumed the
data;
- producer counter denotes amount of data reserved by all producers.
Each time a record is reserved, producer that "owns" the record will
successfully advance producer counter. At that point, data is still not yet
ready to be consumed, though. Each record has 8 byte header, which contains
the length of reserved record, as well as two extra bits: busy bit to denote
that record is still being worked on, and discard bit, which might be set at
commit time if record is discarded. In the latter case, consumer is supposed
to skip the record and move on to the next one. Record header also encodes
record's relative offset from the beginning of ring buffer data area (in
pages). This allows bpf_ringbuf_commit()/bpf_ringbuf_discard() to accept only
the pointer to the record itself, without requiring also the pointer to ring
buffer itself. Ring buffer memory location will be restored from record
metadata header. This significantly simplifies verifier, as well as improving
API usability.
Producer counter increments are serialized under spinlock, so there is
a strict ordering between reservations. Commits, on the other hand, are
completely lockless and independent. All records become available to consumer
in the order of reservations, but only after all previous records where
already committed. It is thus possible for slow producers to temporarily hold
off submitted records, that were reserved later.
Reservation/commit/consumer protocol is verified by litmus tests in
Documentation/litmus-test/bpf-rb.
One interesting implementation bit, that significantly simplifies (and thus
speeds up as well) implementation of both producers and consumers is how data
area is mapped twice contiguously back-to-back in the virtual memory. This
allows to not take any special measures for samples that have to wrap around
at the end of the circular buffer data area, because the next page after the
last data page would be first data page again, and thus the sample will still
appear completely contiguous in virtual memory. See comment and a simple ASCII
diagram showing this visually in bpf_ringbuf_area_alloc().
Another feature that distinguishes BPF ringbuf from perf ring buffer is
a self-pacing notifications of new data being availability.
bpf_ringbuf_commit() implementation will send a notification of new record
being available after commit only if consumer has already caught up right up
to the record being committed. If not, consumer still has to catch up and thus
will see new data anyways without needing an extra poll notification.
Benchmarks (see tools/testing/selftests/bpf/benchs/bench_ringbuf.c) show that
this allows to achieve a very high throughput without having to resort to
tricks like "notify only every Nth sample", which are necessary with perf
buffer. For extreme cases, when BPF program wants more manual control of
notifications, commit/discard/output helpers accept BPF_RB_NO_WAKEUP and
BPF_RB_FORCE_WAKEUP flags, which give full control over notifications of data
availability, but require extra caution and diligence in using this API.
Comparison to alternatives
--------------------------
Before considering implementing BPF ring buffer from scratch existing
alternatives in kernel were evaluated, but didn't seem to meet the needs. They
largely fell into few categores:
- per-CPU buffers (perf, ftrace, etc), which don't satisfy two motivations
outlined above (ordering and memory consumption);
- linked list-based implementations; while some were multi-producer designs,
consuming these from user-space would be very complicated and most
probably not performant; memory-mapping contiguous piece of memory is
simpler and more performant for user-space consumers;
- io_uring is SPSC, but also requires fixed-sized elements. Naively turning
SPSC queue into MPSC w/ lock would have subpar performance compared to
locked reserve + lockless commit, as with BPF ring buffer. Fixed sized
elements would be too limiting for BPF programs, given existing BPF
programs heavily rely on variable-sized perf buffer already;
- specialized implementations (like a new printk ring buffer, [0]) with lots
of printk-specific limitations and implications, that didn't seem to fit
well for intended use with BPF programs.
[0] https://lwn.net/Articles/779550/
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200529075424.3139988-2-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2020-05-29 07:54:20 +00:00
|
|
|
static __poll_t bpf_map_poll(struct file *filp, struct poll_table_struct *pts)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = filp->private_data;
|
|
|
|
|
|
|
|
|
|
if (map->ops->map_poll)
|
|
|
|
|
return map->ops->map_poll(map, filp, pts);
|
|
|
|
|
|
|
|
|
|
return EPOLLERR;
|
|
|
|
|
}
|
|
|
|
|
|
2024-03-07 03:12:25 +00:00
|
|
|
static unsigned long bpf_get_unmapped_area(struct file *filp, unsigned long addr,
|
|
|
|
|
unsigned long len, unsigned long pgoff,
|
|
|
|
|
unsigned long flags)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map = filp->private_data;
|
|
|
|
|
|
|
|
|
|
if (map->ops->map_get_unmapped_area)
|
|
|
|
|
return map->ops->map_get_unmapped_area(filp, addr, len, pgoff, flags);
|
|
|
|
|
#ifdef CONFIG_MMU
|
|
|
|
|
return current->mm->get_unmapped_area(filp, addr, len, pgoff, flags);
|
|
|
|
|
#else
|
|
|
|
|
return addr;
|
|
|
|
|
#endif
|
|
|
|
|
}
|
|
|
|
|
|
2017-10-18 20:00:26 +00:00
|
|
|
const struct file_operations bpf_map_fops = {
|
2015-11-19 10:56:22 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
.show_fdinfo = bpf_map_show_fdinfo,
|
|
|
|
|
#endif
|
|
|
|
|
.release = bpf_map_release,
|
2017-10-18 20:00:22 +00:00
|
|
|
.read = bpf_dummy_read,
|
|
|
|
|
.write = bpf_dummy_write,
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
.mmap = bpf_map_mmap,
|
bpf: Implement BPF ring buffer and verifier support for it
This commit adds a new MPSC ring buffer implementation into BPF ecosystem,
which allows multiple CPUs to submit data to a single shared ring buffer. On
the consumption side, only single consumer is assumed.
Motivation
----------
There are two distinctive motivators for this work, which are not satisfied by
existing perf buffer, which prompted creation of a new ring buffer
implementation.
- more efficient memory utilization by sharing ring buffer across CPUs;
- preserving ordering of events that happen sequentially in time, even
across multiple CPUs (e.g., fork/exec/exit events for a task).
These two problems are independent, but perf buffer fails to satisfy both.
Both are a result of a choice to have per-CPU perf ring buffer. Both can be
also solved by having an MPSC implementation of ring buffer. The ordering
problem could technically be solved for perf buffer with some in-kernel
counting, but given the first one requires an MPSC buffer, the same solution
would solve the second problem automatically.
Semantics and APIs
------------------
Single ring buffer is presented to BPF programs as an instance of BPF map of
type BPF_MAP_TYPE_RINGBUF. Two other alternatives considered, but ultimately
rejected.
One way would be to, similar to BPF_MAP_TYPE_PERF_EVENT_ARRAY, make
BPF_MAP_TYPE_RINGBUF could represent an array of ring buffers, but not enforce
"same CPU only" rule. This would be more familiar interface compatible with
existing perf buffer use in BPF, but would fail if application needed more
advanced logic to lookup ring buffer by arbitrary key. HASH_OF_MAPS addresses
this with current approach. Additionally, given the performance of BPF
ringbuf, many use cases would just opt into a simple single ring buffer shared
among all CPUs, for which current approach would be an overkill.
Another approach could introduce a new concept, alongside BPF map, to
represent generic "container" object, which doesn't necessarily have key/value
interface with lookup/update/delete operations. This approach would add a lot
of extra infrastructure that has to be built for observability and verifier
support. It would also add another concept that BPF developers would have to
familiarize themselves with, new syntax in libbpf, etc. But then would really
provide no additional benefits over the approach of using a map.
BPF_MAP_TYPE_RINGBUF doesn't support lookup/update/delete operations, but so
doesn't few other map types (e.g., queue and stack; array doesn't support
delete, etc).
The approach chosen has an advantage of re-using existing BPF map
infrastructure (introspection APIs in kernel, libbpf support, etc), being
familiar concept (no need to teach users a new type of object in BPF program),
and utilizing existing tooling (bpftool). For common scenario of using
a single ring buffer for all CPUs, it's as simple and straightforward, as
would be with a dedicated "container" object. On the other hand, by being
a map, it can be combined with ARRAY_OF_MAPS and HASH_OF_MAPS map-in-maps to
implement a wide variety of topologies, from one ring buffer for each CPU
(e.g., as a replacement for perf buffer use cases), to a complicated
application hashing/sharding of ring buffers (e.g., having a small pool of
ring buffers with hashed task's tgid being a look up key to preserve order,
but reduce contention).
Key and value sizes are enforced to be zero. max_entries is used to specify
the size of ring buffer and has to be a power of 2 value.
There are a bunch of similarities between perf buffer
(BPF_MAP_TYPE_PERF_EVENT_ARRAY) and new BPF ring buffer semantics:
- variable-length records;
- if there is no more space left in ring buffer, reservation fails, no
blocking;
- memory-mappable data area for user-space applications for ease of
consumption and high performance;
- epoll notifications for new incoming data;
- but still the ability to do busy polling for new data to achieve the
lowest latency, if necessary.
BPF ringbuf provides two sets of APIs to BPF programs:
- bpf_ringbuf_output() allows to *copy* data from one place to a ring
buffer, similarly to bpf_perf_event_output();
- bpf_ringbuf_reserve()/bpf_ringbuf_commit()/bpf_ringbuf_discard() APIs
split the whole process into two steps. First, a fixed amount of space is
reserved. If successful, a pointer to a data inside ring buffer data area
is returned, which BPF programs can use similarly to a data inside
array/hash maps. Once ready, this piece of memory is either committed or
discarded. Discard is similar to commit, but makes consumer ignore the
record.
bpf_ringbuf_output() has disadvantage of incurring extra memory copy, because
record has to be prepared in some other place first. But it allows to submit
records of the length that's not known to verifier beforehand. It also closely
matches bpf_perf_event_output(), so will simplify migration significantly.
bpf_ringbuf_reserve() avoids the extra copy of memory by providing a memory
pointer directly to ring buffer memory. In a lot of cases records are larger
than BPF stack space allows, so many programs have use extra per-CPU array as
a temporary heap for preparing sample. bpf_ringbuf_reserve() avoid this needs
completely. But in exchange, it only allows a known constant size of memory to
be reserved, such that verifier can verify that BPF program can't access
memory outside its reserved record space. bpf_ringbuf_output(), while slightly
slower due to extra memory copy, covers some use cases that are not suitable
for bpf_ringbuf_reserve().
The difference between commit and discard is very small. Discard just marks
a record as discarded, and such records are supposed to be ignored by consumer
code. Discard is useful for some advanced use-cases, such as ensuring
all-or-nothing multi-record submission, or emulating temporary malloc()/free()
within single BPF program invocation.
Each reserved record is tracked by verifier through existing
reference-tracking logic, similar to socket ref-tracking. It is thus
impossible to reserve a record, but forget to submit (or discard) it.
bpf_ringbuf_query() helper allows to query various properties of ring buffer.
Currently 4 are supported:
- BPF_RB_AVAIL_DATA returns amount of unconsumed data in ring buffer;
- BPF_RB_RING_SIZE returns the size of ring buffer;
- BPF_RB_CONS_POS/BPF_RB_PROD_POS returns current logical possition of
consumer/producer, respectively.
Returned values are momentarily snapshots of ring buffer state and could be
off by the time helper returns, so this should be used only for
debugging/reporting reasons or for implementing various heuristics, that take
into account highly-changeable nature of some of those characteristics.
One such heuristic might involve more fine-grained control over poll/epoll
notifications about new data availability in ring buffer. Together with
BPF_RB_NO_WAKEUP/BPF_RB_FORCE_WAKEUP flags for output/commit/discard helpers,
it allows BPF program a high degree of control and, e.g., more efficient
batched notifications. Default self-balancing strategy, though, should be
adequate for most applications and will work reliable and efficiently already.
Design and implementation
-------------------------
This reserve/commit schema allows a natural way for multiple producers, either
on different CPUs or even on the same CPU/in the same BPF program, to reserve
independent records and work with them without blocking other producers. This
means that if BPF program was interruped by another BPF program sharing the
same ring buffer, they will both get a record reserved (provided there is
enough space left) and can work with it and submit it independently. This
applies to NMI context as well, except that due to using a spinlock during
reservation, in NMI context, bpf_ringbuf_reserve() might fail to get a lock,
in which case reservation will fail even if ring buffer is not full.
The ring buffer itself internally is implemented as a power-of-2 sized
circular buffer, with two logical and ever-increasing counters (which might
wrap around on 32-bit architectures, that's not a problem):
- consumer counter shows up to which logical position consumer consumed the
data;
- producer counter denotes amount of data reserved by all producers.
Each time a record is reserved, producer that "owns" the record will
successfully advance producer counter. At that point, data is still not yet
ready to be consumed, though. Each record has 8 byte header, which contains
the length of reserved record, as well as two extra bits: busy bit to denote
that record is still being worked on, and discard bit, which might be set at
commit time if record is discarded. In the latter case, consumer is supposed
to skip the record and move on to the next one. Record header also encodes
record's relative offset from the beginning of ring buffer data area (in
pages). This allows bpf_ringbuf_commit()/bpf_ringbuf_discard() to accept only
the pointer to the record itself, without requiring also the pointer to ring
buffer itself. Ring buffer memory location will be restored from record
metadata header. This significantly simplifies verifier, as well as improving
API usability.
Producer counter increments are serialized under spinlock, so there is
a strict ordering between reservations. Commits, on the other hand, are
completely lockless and independent. All records become available to consumer
in the order of reservations, but only after all previous records where
already committed. It is thus possible for slow producers to temporarily hold
off submitted records, that were reserved later.
Reservation/commit/consumer protocol is verified by litmus tests in
Documentation/litmus-test/bpf-rb.
One interesting implementation bit, that significantly simplifies (and thus
speeds up as well) implementation of both producers and consumers is how data
area is mapped twice contiguously back-to-back in the virtual memory. This
allows to not take any special measures for samples that have to wrap around
at the end of the circular buffer data area, because the next page after the
last data page would be first data page again, and thus the sample will still
appear completely contiguous in virtual memory. See comment and a simple ASCII
diagram showing this visually in bpf_ringbuf_area_alloc().
Another feature that distinguishes BPF ringbuf from perf ring buffer is
a self-pacing notifications of new data being availability.
bpf_ringbuf_commit() implementation will send a notification of new record
being available after commit only if consumer has already caught up right up
to the record being committed. If not, consumer still has to catch up and thus
will see new data anyways without needing an extra poll notification.
Benchmarks (see tools/testing/selftests/bpf/benchs/bench_ringbuf.c) show that
this allows to achieve a very high throughput without having to resort to
tricks like "notify only every Nth sample", which are necessary with perf
buffer. For extreme cases, when BPF program wants more manual control of
notifications, commit/discard/output helpers accept BPF_RB_NO_WAKEUP and
BPF_RB_FORCE_WAKEUP flags, which give full control over notifications of data
availability, but require extra caution and diligence in using this API.
Comparison to alternatives
--------------------------
Before considering implementing BPF ring buffer from scratch existing
alternatives in kernel were evaluated, but didn't seem to meet the needs. They
largely fell into few categores:
- per-CPU buffers (perf, ftrace, etc), which don't satisfy two motivations
outlined above (ordering and memory consumption);
- linked list-based implementations; while some were multi-producer designs,
consuming these from user-space would be very complicated and most
probably not performant; memory-mapping contiguous piece of memory is
simpler and more performant for user-space consumers;
- io_uring is SPSC, but also requires fixed-sized elements. Naively turning
SPSC queue into MPSC w/ lock would have subpar performance compared to
locked reserve + lockless commit, as with BPF ring buffer. Fixed sized
elements would be too limiting for BPF programs, given existing BPF
programs heavily rely on variable-sized perf buffer already;
- specialized implementations (like a new printk ring buffer, [0]) with lots
of printk-specific limitations and implications, that didn't seem to fit
well for intended use with BPF programs.
[0] https://lwn.net/Articles/779550/
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200529075424.3139988-2-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2020-05-29 07:54:20 +00:00
|
|
|
.poll = bpf_map_poll,
|
2024-03-07 03:12:25 +00:00
|
|
|
.get_unmapped_area = bpf_get_unmapped_area,
|
2014-09-26 07:16:57 +00:00
|
|
|
};
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
int bpf_map_new_fd(struct bpf_map *map, int flags)
|
2015-10-29 13:58:06 +00:00
|
|
|
{
|
2017-10-18 20:00:24 +00:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
ret = security_bpf_map(map, OPEN_FMODE(flags));
|
|
|
|
|
if (ret < 0)
|
|
|
|
|
return ret;
|
|
|
|
|
|
2015-10-29 13:58:06 +00:00
|
|
|
return anon_inode_getfd("bpf-map", &bpf_map_fops, map,
|
2017-10-18 20:00:22 +00:00
|
|
|
flags | O_CLOEXEC);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int bpf_get_file_flag(int flags)
|
|
|
|
|
{
|
|
|
|
|
if ((flags & BPF_F_RDONLY) && (flags & BPF_F_WRONLY))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (flags & BPF_F_RDONLY)
|
|
|
|
|
return O_RDONLY;
|
|
|
|
|
if (flags & BPF_F_WRONLY)
|
|
|
|
|
return O_WRONLY;
|
|
|
|
|
return O_RDWR;
|
2015-10-29 13:58:06 +00:00
|
|
|
}
|
|
|
|
|
|
2014-09-26 07:16:57 +00:00
|
|
|
/* helper macro to check that unused fields 'union bpf_attr' are zero */
|
|
|
|
|
#define CHECK_ATTR(CMD) \
|
|
|
|
|
memchr_inv((void *) &attr->CMD##_LAST_FIELD + \
|
|
|
|
|
sizeof(attr->CMD##_LAST_FIELD), 0, \
|
|
|
|
|
sizeof(*attr) - \
|
|
|
|
|
offsetof(union bpf_attr, CMD##_LAST_FIELD) - \
|
|
|
|
|
sizeof(attr->CMD##_LAST_FIELD)) != NULL
|
|
|
|
|
|
2020-03-14 01:02:09 +00:00
|
|
|
/* dst and src must have at least "size" number of bytes.
|
|
|
|
|
* Return strlen on success and < 0 on error.
|
2017-09-27 21:37:52 +00:00
|
|
|
*/
|
2020-03-14 01:02:09 +00:00
|
|
|
int bpf_obj_name_cpy(char *dst, const char *src, unsigned int size)
|
2017-09-27 21:37:52 +00:00
|
|
|
{
|
2020-03-14 01:02:09 +00:00
|
|
|
const char *end = src + size;
|
|
|
|
|
const char *orig_src = src;
|
2017-09-27 21:37:52 +00:00
|
|
|
|
2020-03-14 01:02:09 +00:00
|
|
|
memset(dst, 0, size);
|
2019-04-09 21:20:07 +00:00
|
|
|
/* Copy all isalnum(), '_' and '.' chars. */
|
2017-09-27 21:37:52 +00:00
|
|
|
while (src < end && *src) {
|
2019-04-09 21:20:07 +00:00
|
|
|
if (!isalnum(*src) &&
|
|
|
|
|
*src != '_' && *src != '.')
|
2017-09-27 21:37:52 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
*dst++ = *src++;
|
|
|
|
|
}
|
|
|
|
|
|
2020-03-14 01:02:09 +00:00
|
|
|
/* No '\0' found in "size" number of bytes */
|
2017-09-27 21:37:52 +00:00
|
|
|
if (src == end)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-03-14 01:02:09 +00:00
|
|
|
return src - orig_src;
|
2017-09-27 21:37:52 +00:00
|
|
|
}
|
|
|
|
|
|
2018-08-11 23:59:17 +00:00
|
|
|
int map_check_no_btf(const struct bpf_map *map,
|
2018-12-10 23:43:00 +00:00
|
|
|
const struct btf *btf,
|
2018-08-11 23:59:17 +00:00
|
|
|
const struct btf_type *key_type,
|
|
|
|
|
const struct btf_type *value_type)
|
|
|
|
|
{
|
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
2024-01-24 02:21:01 +00:00
|
|
|
static int map_check_btf(struct bpf_map *map, struct bpf_token *token,
|
|
|
|
|
const struct btf *btf, u32 btf_key_id, u32 btf_value_id)
|
2018-08-11 23:59:17 +00:00
|
|
|
{
|
|
|
|
|
const struct btf_type *key_type, *value_type;
|
|
|
|
|
u32 key_size, value_size;
|
|
|
|
|
int ret = 0;
|
|
|
|
|
|
2019-04-09 21:20:10 +00:00
|
|
|
/* Some maps allow key to be unspecified. */
|
|
|
|
|
if (btf_key_id) {
|
|
|
|
|
key_type = btf_type_id_size(btf, &btf_key_id, &key_size);
|
|
|
|
|
if (!key_type || key_size != map->key_size)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
} else {
|
|
|
|
|
key_type = btf_type_by_id(btf, 0);
|
|
|
|
|
if (!map->ops->map_check_btf)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
2018-08-11 23:59:17 +00:00
|
|
|
|
|
|
|
|
value_type = btf_type_id_size(btf, &btf_value_id, &value_size);
|
|
|
|
|
if (!value_type || value_size != map->value_size)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
map->record = btf_parse_fields(btf, value_type,
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
BPF_SPIN_LOCK | BPF_TIMER | BPF_KPTR | BPF_LIST_HEAD |
|
2023-04-15 20:18:04 +00:00
|
|
|
BPF_RB_ROOT | BPF_REFCOUNT,
|
2022-11-03 19:09:56 +00:00
|
|
|
map->value_size);
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (!IS_ERR_OR_NULL(map->record)) {
|
|
|
|
|
int i;
|
|
|
|
|
|
2024-01-24 02:21:01 +00:00
|
|
|
if (!bpf_token_capable(token, CAP_BPF)) {
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
ret = -EPERM;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
|
|
|
|
if (map->map_flags & (BPF_F_RDONLY_PROG | BPF_F_WRONLY_PROG)) {
|
|
|
|
|
ret = -EACCES;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
for (i = 0; i < sizeof(map->record->field_mask) * 8; i++) {
|
|
|
|
|
switch (map->record->field_mask & (1 << i)) {
|
|
|
|
|
case 0:
|
|
|
|
|
continue;
|
2022-11-03 19:09:56 +00:00
|
|
|
case BPF_SPIN_LOCK:
|
|
|
|
|
if (map->map_type != BPF_MAP_TYPE_HASH &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_ARRAY &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_CGROUP_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_SK_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_INODE_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_TASK_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_CGRP_STORAGE) {
|
|
|
|
|
ret = -EOPNOTSUPP;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case BPF_TIMER:
|
|
|
|
|
if (map->map_type != BPF_MAP_TYPE_HASH &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_LRU_HASH &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_ARRAY) {
|
2022-11-18 01:55:51 +00:00
|
|
|
ret = -EOPNOTSUPP;
|
2022-11-03 19:09:56 +00:00
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
|
|
|
|
break;
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
case BPF_KPTR_UNREF:
|
|
|
|
|
case BPF_KPTR_REF:
|
2023-08-27 15:27:39 +00:00
|
|
|
case BPF_KPTR_PERCPU:
|
2023-04-15 20:18:04 +00:00
|
|
|
case BPF_REFCOUNT:
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
if (map->map_type != BPF_MAP_TYPE_HASH &&
|
2023-02-25 15:40:08 +00:00
|
|
|
map->map_type != BPF_MAP_TYPE_PERCPU_HASH &&
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
map->map_type != BPF_MAP_TYPE_LRU_HASH &&
|
2023-02-25 15:40:08 +00:00
|
|
|
map->map_type != BPF_MAP_TYPE_LRU_PERCPU_HASH &&
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
map->map_type != BPF_MAP_TYPE_ARRAY &&
|
2023-02-25 15:40:09 +00:00
|
|
|
map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_SK_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_INODE_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_TASK_STORAGE &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_CGRP_STORAGE) {
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
ret = -EOPNOTSUPP;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case BPF_LIST_HEAD:
|
bpf: Add basic bpf_rb_{root,node} support
This patch adds special BPF_RB_{ROOT,NODE} btf_field_types similar to
BPF_LIST_{HEAD,NODE}, adds the necessary plumbing to detect the new
types, and adds bpf_rb_root_free function for freeing bpf_rb_root in
map_values.
structs bpf_rb_root and bpf_rb_node are opaque types meant to
obscure structs rb_root_cached rb_node, respectively.
btf_struct_access will prevent BPF programs from touching these special
fields automatically now that they're recognized.
btf_check_and_fixup_fields now groups list_head and rb_root together as
"graph root" fields and {list,rb}_node as "graph node", and does same
ownership cycle checking as before. Note that this function does _not_
prevent ownership type mixups (e.g. rb_root owning list_node) - that's
handled by btf_parse_graph_root.
After this patch, a bpf program can have a struct bpf_rb_root in a
map_value, but not add anything to nor do anything useful with it.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230214004017.2534011-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-02-14 00:40:10 +00:00
|
|
|
case BPF_RB_ROOT:
|
bpf: Support bpf_list_head in map values
Add the support on the map side to parse, recognize, verify, and build
metadata table for a new special field of the type struct bpf_list_head.
To parameterize the bpf_list_head for a certain value type and the
list_node member it will accept in that value type, we use BTF
declaration tags.
The definition of bpf_list_head in a map value will be done as follows:
struct foo {
struct bpf_list_node node;
int data;
};
struct map_value {
struct bpf_list_head head __contains(foo, node);
};
Then, the bpf_list_head only allows adding to the list 'head' using the
bpf_list_node 'node' for the type struct foo.
The 'contains' annotation is a BTF declaration tag composed of four
parts, "contains:name:node" where the name is then used to look up the
type in the map BTF, with its kind hardcoded to BTF_KIND_STRUCT during
the lookup. The node defines name of the member in this type that has
the type struct bpf_list_node, which is actually used for linking into
the linked list. For now, 'kind' part is hardcoded as struct.
This allows building intrusive linked lists in BPF, using container_of
to obtain pointer to entry, while being completely type safe from the
perspective of the verifier. The verifier knows exactly the type of the
nodes, and knows that list helpers return that type at some fixed offset
where the bpf_list_node member used for this list exists. The verifier
also uses this information to disallow adding types that are not
accepted by a certain list.
For now, no elements can be added to such lists. Support for that is
coming in future patches, hence draining and freeing items is done with
a TODO that will be resolved in a future patch.
Note that the bpf_list_head_free function moves the list out to a local
variable under the lock and releases it, doing the actual draining of
the list items outside the lock. While this helps with not holding the
lock for too long pessimizing other concurrent list operations, it is
also necessary for deadlock prevention: unless every function called in
the critical section would be notrace, a fentry/fexit program could
attach and call bpf_map_update_elem again on the map, leading to the
same lock being acquired if the key matches and lead to a deadlock.
While this requires some special effort on part of the BPF programmer to
trigger and is highly unlikely to occur in practice, it is always better
if we can avoid such a condition.
While notrace would prevent this, doing the draining outside the lock
has advantages of its own, hence it is used to also fix the deadlock
related problem.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221114191547.1694267-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-14 19:15:25 +00:00
|
|
|
if (map->map_type != BPF_MAP_TYPE_HASH &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_LRU_HASH &&
|
|
|
|
|
map->map_type != BPF_MAP_TYPE_ARRAY) {
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
ret = -EOPNOTSUPP;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
/* Fail if map_type checks are missing for a field type */
|
|
|
|
|
ret = -EOPNOTSUPP;
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-18 01:55:57 +00:00
|
|
|
ret = btf_check_and_fixup_fields(btf, map->record);
|
|
|
|
|
if (ret < 0)
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
if (map->ops->map_check_btf) {
|
2018-12-10 23:43:00 +00:00
|
|
|
ret = map->ops->map_check_btf(map, btf, key_type, value_type);
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
if (ret < 0)
|
|
|
|
|
goto free_map_tab;
|
|
|
|
|
}
|
2018-08-11 23:59:17 +00:00
|
|
|
|
bpf: Allow storing unreferenced kptr in map
This commit introduces a new pointer type 'kptr' which can be embedded
in a map value to hold a PTR_TO_BTF_ID stored by a BPF program during
its invocation. When storing such a kptr, BPF program's PTR_TO_BTF_ID
register must have the same type as in the map value's BTF, and loading
a kptr marks the destination register as PTR_TO_BTF_ID with the correct
kernel BTF and BTF ID.
Such kptr are unreferenced, i.e. by the time another invocation of the
BPF program loads this pointer, the object which the pointer points to
may not longer exist. Since PTR_TO_BTF_ID loads (using BPF_LDX) are
patched to PROBE_MEM loads by the verifier, it would safe to allow user
to still access such invalid pointer, but passing such pointers into
BPF helpers and kfuncs should not be permitted. A future patch in this
series will close this gap.
The flexibility offered by allowing programs to dereference such invalid
pointers while being safe at runtime frees the verifier from doing
complex lifetime tracking. As long as the user may ensure that the
object remains valid, it can ensure data read by it from the kernel
object is valid.
The user indicates that a certain pointer must be treated as kptr
capable of accepting stores of PTR_TO_BTF_ID of a certain type, by using
a BTF type tag 'kptr' on the pointed to type of the pointer. Then, this
information is recorded in the object BTF which will be passed into the
kernel by way of map's BTF information. The name and kind from the map
value BTF is used to look up the in-kernel type, and the actual BTF and
BTF ID is recorded in the map struct in a new kptr_off_tab member. For
now, only storing pointers to structs is permitted.
An example of this specification is shown below:
#define __kptr __attribute__((btf_type_tag("kptr")))
struct map_value {
...
struct task_struct __kptr *task;
...
};
Then, in a BPF program, user may store PTR_TO_BTF_ID with the type
task_struct into the map, and then load it later.
Note that the destination register is marked PTR_TO_BTF_ID_OR_NULL, as
the verifier cannot know whether the value is NULL or not statically, it
must treat all potential loads at that map value offset as loading a
possibly NULL pointer.
Only BPF_LDX, BPF_STX, and BPF_ST (with insn->imm = 0 to denote NULL)
are allowed instructions that can access such a pointer. On BPF_LDX, the
destination register is updated to be a PTR_TO_BTF_ID, and on BPF_STX,
it is checked whether the source register type is a PTR_TO_BTF_ID with
same BTF type as specified in the map BTF. The access size must always
be BPF_DW.
For the map in map support, the kptr_off_tab for outer map is copied
from the inner map's kptr_off_tab. It was chosen to do a deep copy
instead of introducing a refcount to kptr_off_tab, because the copy only
needs to be done when paramterizing using inner_map_fd in the map in map
case, hence would be unnecessary for all other users.
It is not permitted to use MAP_FREEZE command and mmap for BPF map
having kptrs, similar to the bpf_timer case. A kptr also requires that
BPF program has both read and write access to the map (hence both
BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG are disallowed).
Note that check_map_access must be called from both
check_helper_mem_access and for the BPF instructions, hence the kptr
check must distinguish between ACCESS_DIRECT and ACCESS_HELPER, and
reject ACCESS_HELPER cases. We rename stack_access_src to bpf_access_src
and reuse it for this purpose.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220424214901.2743946-2-memxor@gmail.com
2022-04-24 21:48:49 +00:00
|
|
|
return ret;
|
|
|
|
|
free_map_tab:
|
bpf: Refactor kptr_off_tab into btf_record
To prepare the BPF verifier to handle special fields in both map values
and program allocated types coming from program BTF, we need to refactor
the kptr_off_tab handling code into something more generic and reusable
across both cases to avoid code duplication.
Later patches also require passing this data to helpers at runtime, so
that they can work on user defined types, initialize them, destruct
them, etc.
The main observation is that both map values and such allocated types
point to a type in program BTF, hence they can be handled similarly. We
can prepare a field metadata table for both cases and store them in
struct bpf_map or struct btf depending on the use case.
Hence, refactor the code into generic btf_record and btf_field member
structs. The btf_record represents the fields of a specific btf_type in
user BTF. The cnt indicates the number of special fields we successfully
recognized, and field_mask is a bitmask of fields that were found, to
enable quick determination of availability of a certain field.
Subsequently, refactor the rest of the code to work with these generic
types, remove assumptions about kptr and kptr_off_tab, rename variables
to more meaningful names, etc.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20221103191013.1236066-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2022-11-03 19:09:55 +00:00
|
|
|
bpf_map_free_record(map);
|
2018-08-11 23:59:17 +00:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2024-01-24 02:20:58 +00:00
|
|
|
static bool bpf_net_capable(void)
|
|
|
|
|
{
|
|
|
|
|
return capable(CAP_NET_ADMIN) || capable(CAP_SYS_ADMIN);
|
|
|
|
|
}
|
|
|
|
|
|
2024-01-24 02:21:01 +00:00
|
|
|
#define BPF_MAP_CREATE_LAST_FIELD map_token_fd
|
2014-09-26 07:16:57 +00:00
|
|
|
/* called via syscall */
|
|
|
|
|
static int map_create(union bpf_attr *attr)
|
|
|
|
|
{
|
2023-06-13 22:35:31 +00:00
|
|
|
const struct bpf_map_ops *ops;
|
2024-01-24 02:21:01 +00:00
|
|
|
struct bpf_token *token = NULL;
|
2017-08-18 18:28:00 +00:00
|
|
|
int numa_node = bpf_map_attr_numa_node(attr);
|
2023-06-13 22:35:31 +00:00
|
|
|
u32 map_type = attr->map_type;
|
2014-09-26 07:16:57 +00:00
|
|
|
struct bpf_map *map;
|
2024-01-24 02:21:01 +00:00
|
|
|
bool token_flag;
|
2017-10-18 20:00:22 +00:00
|
|
|
int f_flags;
|
2014-09-26 07:16:57 +00:00
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
err = CHECK_ATTR(BPF_MAP_CREATE);
|
|
|
|
|
if (err)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2024-01-24 02:21:01 +00:00
|
|
|
/* check BPF_F_TOKEN_FD flag, remember if it's set, and then clear it
|
|
|
|
|
* to avoid per-map type checks tripping on unknown flag
|
|
|
|
|
*/
|
|
|
|
|
token_flag = attr->map_flags & BPF_F_TOKEN_FD;
|
|
|
|
|
attr->map_flags &= ~BPF_F_TOKEN_FD;
|
|
|
|
|
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
if (attr->btf_vmlinux_value_type_id) {
|
|
|
|
|
if (attr->map_type != BPF_MAP_TYPE_STRUCT_OPS ||
|
|
|
|
|
attr->btf_key_type_id || attr->btf_value_type_id)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
} else if (attr->btf_key_type_id && !attr->btf_value_type_id) {
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
if (attr->map_type != BPF_MAP_TYPE_BLOOM_FILTER &&
|
bpf: Introduce bpf_arena.
Introduce bpf_arena, which is a sparse shared memory region between the bpf
program and user space.
Use cases:
1. User space mmap-s bpf_arena and uses it as a traditional mmap-ed
anonymous region, like memcached or any key/value storage. The bpf
program implements an in-kernel accelerator. XDP prog can search for
a key in bpf_arena and return a value without going to user space.
2. The bpf program builds arbitrary data structures in bpf_arena (hash
tables, rb-trees, sparse arrays), while user space consumes it.
3. bpf_arena is a "heap" of memory from the bpf program's point of view.
The user space may mmap it, but bpf program will not convert pointers
to user base at run-time to improve bpf program speed.
Initially, the kernel vm_area and user vma are not populated. User space
can fault in pages within the range. While servicing a page fault,
bpf_arena logic will insert a new page into the kernel and user vmas. The
bpf program can allocate pages from that region via
bpf_arena_alloc_pages(). This kernel function will insert pages into the
kernel vm_area. The subsequent fault-in from user space will populate that
page into the user vma. The BPF_F_SEGV_ON_FAULT flag at arena creation time
can be used to prevent fault-in from user space. In such a case, if a page
is not allocated by the bpf program and not present in the kernel vm_area,
the user process will segfault. This is useful for use cases 2 and 3 above.
bpf_arena_alloc_pages() is similar to user space mmap(). It allocates pages
either at a specific address within the arena or allocates a range with the
maple tree. bpf_arena_free_pages() is analogous to munmap(), which frees
pages and removes the range from the kernel vm_area and from user process
vmas.
bpf_arena can be used as a bpf program "heap" of up to 4GB. The speed of
bpf program is more important than ease of sharing with user space. This is
use case 3. In such a case, the BPF_F_NO_USER_CONV flag is recommended.
It will tell the verifier to treat the rX = bpf_arena_cast_user(rY)
instruction as a 32-bit move wX = wY, which will improve bpf prog
performance. Otherwise, bpf_arena_cast_user is translated by JIT to
conditionally add the upper 32 bits of user vm_start (if the pointer is not
NULL) to arena pointers before they are stored into memory. This way, user
space sees them as valid 64-bit pointers.
Diff https://github.com/llvm/llvm-project/pull/84410 enables LLVM BPF
backend generate the bpf_addr_space_cast() instruction to cast pointers
between address_space(1) which is reserved for bpf_arena pointers and
default address space zero. All arena pointers in a bpf program written in
C language are tagged as __attribute__((address_space(1))). Hence, clang
provides helpful diagnostics when pointers cross address space. Libbpf and
the kernel support only address_space == 1. All other address space
identifiers are reserved.
rX = bpf_addr_space_cast(rY, /* dst_as */ 1, /* src_as */ 0) tells the
verifier that rX->type = PTR_TO_ARENA. Any further operations on
PTR_TO_ARENA register have to be in the 32-bit domain. The verifier will
mark load/store through PTR_TO_ARENA with PROBE_MEM32. JIT will generate
them as kern_vm_start + 32bit_addr memory accesses. The behavior is similar
to copy_from_kernel_nofault() except that no address checks are necessary.
The address is guaranteed to be in the 4GB range. If the page is not
present, the destination register is zeroed on read, and the operation is
ignored on write.
rX = bpf_addr_space_cast(rY, 0, 1) tells the verifier that rX->type =
unknown scalar. If arena->map_flags has BPF_F_NO_USER_CONV set, then the
verifier converts such cast instructions to mov32. Otherwise, JIT will emit
native code equivalent to:
rX = (u32)rY;
if (rY)
rX |= clear_lo32_bits(arena->user_vm_start); /* replace hi32 bits in rX */
After such conversion, the pointer becomes a valid user pointer within
bpf_arena range. The user process can access data structures created in
bpf_arena without any additional computations. For example, a linked list
built by a bpf program can be walked natively by user space.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Barret Rhoden <brho@google.com>
Link: https://lore.kernel.org/bpf/20240308010812.89848-2-alexei.starovoitov@gmail.com
2024-03-08 01:07:59 +00:00
|
|
|
attr->map_type != BPF_MAP_TYPE_ARENA &&
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
attr->map_extra != 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
f_flags = bpf_get_file_flag(attr->map_flags);
|
|
|
|
|
if (f_flags < 0)
|
|
|
|
|
return f_flags;
|
|
|
|
|
|
2017-08-18 18:28:00 +00:00
|
|
|
if (numa_node != NUMA_NO_NODE &&
|
2017-09-05 05:41:02 +00:00
|
|
|
((unsigned int)numa_node >= nr_node_ids ||
|
|
|
|
|
!node_online(numa_node)))
|
2017-08-18 18:28:00 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
2014-09-26 07:16:57 +00:00
|
|
|
/* find map type and init map: hashtable vs rbtree vs bloom vs ... */
|
2023-06-13 22:35:31 +00:00
|
|
|
map_type = attr->map_type;
|
|
|
|
|
if (map_type >= ARRAY_SIZE(bpf_map_types))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
map_type = array_index_nospec(map_type, ARRAY_SIZE(bpf_map_types));
|
|
|
|
|
ops = bpf_map_types[map_type];
|
|
|
|
|
if (!ops)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (ops->map_alloc_check) {
|
|
|
|
|
err = ops->map_alloc_check(attr);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
if (attr->map_ifindex)
|
|
|
|
|
ops = &bpf_map_offload_ops;
|
|
|
|
|
if (!ops->map_mem_usage)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2024-01-24 02:21:01 +00:00
|
|
|
if (token_flag) {
|
|
|
|
|
token = bpf_token_get_from_fd(attr->map_token_fd);
|
|
|
|
|
if (IS_ERR(token))
|
|
|
|
|
return PTR_ERR(token);
|
|
|
|
|
|
|
|
|
|
/* if current token doesn't grant map creation permissions,
|
|
|
|
|
* then we can't use this token, so ignore it and rely on
|
|
|
|
|
* system-wide capabilities checks
|
|
|
|
|
*/
|
|
|
|
|
if (!bpf_token_allow_cmd(token, BPF_MAP_CREATE) ||
|
|
|
|
|
!bpf_token_allow_map_type(token, attr->map_type)) {
|
|
|
|
|
bpf_token_put(token);
|
|
|
|
|
token = NULL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
|
2023-06-13 22:35:30 +00:00
|
|
|
/* Intent here is for unprivileged_bpf_disabled to block BPF map
|
|
|
|
|
* creation for unprivileged users; other actions depend
|
|
|
|
|
* on fd availability and access to bpffs, so are dependent on
|
|
|
|
|
* object creation success. Even with unprivileged BPF disabled,
|
|
|
|
|
* capability checks are still carried out.
|
|
|
|
|
*/
|
2024-01-24 02:21:01 +00:00
|
|
|
if (sysctl_unprivileged_bpf_disabled && !bpf_token_capable(token, CAP_BPF))
|
|
|
|
|
goto put_token;
|
2023-06-13 22:35:30 +00:00
|
|
|
|
2023-06-13 22:35:32 +00:00
|
|
|
/* check privileged map type permissions */
|
|
|
|
|
switch (map_type) {
|
|
|
|
|
case BPF_MAP_TYPE_ARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_PERCPU_ARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_PROG_ARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_PERF_EVENT_ARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_CGROUP_ARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_ARRAY_OF_MAPS:
|
|
|
|
|
case BPF_MAP_TYPE_HASH:
|
|
|
|
|
case BPF_MAP_TYPE_PERCPU_HASH:
|
|
|
|
|
case BPF_MAP_TYPE_HASH_OF_MAPS:
|
|
|
|
|
case BPF_MAP_TYPE_RINGBUF:
|
|
|
|
|
case BPF_MAP_TYPE_USER_RINGBUF:
|
|
|
|
|
case BPF_MAP_TYPE_CGROUP_STORAGE:
|
|
|
|
|
case BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE:
|
|
|
|
|
/* unprivileged */
|
|
|
|
|
break;
|
|
|
|
|
case BPF_MAP_TYPE_SK_STORAGE:
|
|
|
|
|
case BPF_MAP_TYPE_INODE_STORAGE:
|
|
|
|
|
case BPF_MAP_TYPE_TASK_STORAGE:
|
|
|
|
|
case BPF_MAP_TYPE_CGRP_STORAGE:
|
|
|
|
|
case BPF_MAP_TYPE_BLOOM_FILTER:
|
|
|
|
|
case BPF_MAP_TYPE_LPM_TRIE:
|
|
|
|
|
case BPF_MAP_TYPE_REUSEPORT_SOCKARRAY:
|
|
|
|
|
case BPF_MAP_TYPE_STACK_TRACE:
|
|
|
|
|
case BPF_MAP_TYPE_QUEUE:
|
|
|
|
|
case BPF_MAP_TYPE_STACK:
|
|
|
|
|
case BPF_MAP_TYPE_LRU_HASH:
|
|
|
|
|
case BPF_MAP_TYPE_LRU_PERCPU_HASH:
|
|
|
|
|
case BPF_MAP_TYPE_STRUCT_OPS:
|
|
|
|
|
case BPF_MAP_TYPE_CPUMAP:
|
bpf: Introduce bpf_arena.
Introduce bpf_arena, which is a sparse shared memory region between the bpf
program and user space.
Use cases:
1. User space mmap-s bpf_arena and uses it as a traditional mmap-ed
anonymous region, like memcached or any key/value storage. The bpf
program implements an in-kernel accelerator. XDP prog can search for
a key in bpf_arena and return a value without going to user space.
2. The bpf program builds arbitrary data structures in bpf_arena (hash
tables, rb-trees, sparse arrays), while user space consumes it.
3. bpf_arena is a "heap" of memory from the bpf program's point of view.
The user space may mmap it, but bpf program will not convert pointers
to user base at run-time to improve bpf program speed.
Initially, the kernel vm_area and user vma are not populated. User space
can fault in pages within the range. While servicing a page fault,
bpf_arena logic will insert a new page into the kernel and user vmas. The
bpf program can allocate pages from that region via
bpf_arena_alloc_pages(). This kernel function will insert pages into the
kernel vm_area. The subsequent fault-in from user space will populate that
page into the user vma. The BPF_F_SEGV_ON_FAULT flag at arena creation time
can be used to prevent fault-in from user space. In such a case, if a page
is not allocated by the bpf program and not present in the kernel vm_area,
the user process will segfault. This is useful for use cases 2 and 3 above.
bpf_arena_alloc_pages() is similar to user space mmap(). It allocates pages
either at a specific address within the arena or allocates a range with the
maple tree. bpf_arena_free_pages() is analogous to munmap(), which frees
pages and removes the range from the kernel vm_area and from user process
vmas.
bpf_arena can be used as a bpf program "heap" of up to 4GB. The speed of
bpf program is more important than ease of sharing with user space. This is
use case 3. In such a case, the BPF_F_NO_USER_CONV flag is recommended.
It will tell the verifier to treat the rX = bpf_arena_cast_user(rY)
instruction as a 32-bit move wX = wY, which will improve bpf prog
performance. Otherwise, bpf_arena_cast_user is translated by JIT to
conditionally add the upper 32 bits of user vm_start (if the pointer is not
NULL) to arena pointers before they are stored into memory. This way, user
space sees them as valid 64-bit pointers.
Diff https://github.com/llvm/llvm-project/pull/84410 enables LLVM BPF
backend generate the bpf_addr_space_cast() instruction to cast pointers
between address_space(1) which is reserved for bpf_arena pointers and
default address space zero. All arena pointers in a bpf program written in
C language are tagged as __attribute__((address_space(1))). Hence, clang
provides helpful diagnostics when pointers cross address space. Libbpf and
the kernel support only address_space == 1. All other address space
identifiers are reserved.
rX = bpf_addr_space_cast(rY, /* dst_as */ 1, /* src_as */ 0) tells the
verifier that rX->type = PTR_TO_ARENA. Any further operations on
PTR_TO_ARENA register have to be in the 32-bit domain. The verifier will
mark load/store through PTR_TO_ARENA with PROBE_MEM32. JIT will generate
them as kern_vm_start + 32bit_addr memory accesses. The behavior is similar
to copy_from_kernel_nofault() except that no address checks are necessary.
The address is guaranteed to be in the 4GB range. If the page is not
present, the destination register is zeroed on read, and the operation is
ignored on write.
rX = bpf_addr_space_cast(rY, 0, 1) tells the verifier that rX->type =
unknown scalar. If arena->map_flags has BPF_F_NO_USER_CONV set, then the
verifier converts such cast instructions to mov32. Otherwise, JIT will emit
native code equivalent to:
rX = (u32)rY;
if (rY)
rX |= clear_lo32_bits(arena->user_vm_start); /* replace hi32 bits in rX */
After such conversion, the pointer becomes a valid user pointer within
bpf_arena range. The user process can access data structures created in
bpf_arena without any additional computations. For example, a linked list
built by a bpf program can be walked natively by user space.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Barret Rhoden <brho@google.com>
Link: https://lore.kernel.org/bpf/20240308010812.89848-2-alexei.starovoitov@gmail.com
2024-03-08 01:07:59 +00:00
|
|
|
case BPF_MAP_TYPE_ARENA:
|
2024-01-24 02:21:01 +00:00
|
|
|
if (!bpf_token_capable(token, CAP_BPF))
|
|
|
|
|
goto put_token;
|
2023-06-13 22:35:32 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_MAP_TYPE_SOCKMAP:
|
|
|
|
|
case BPF_MAP_TYPE_SOCKHASH:
|
|
|
|
|
case BPF_MAP_TYPE_DEVMAP:
|
|
|
|
|
case BPF_MAP_TYPE_DEVMAP_HASH:
|
|
|
|
|
case BPF_MAP_TYPE_XSKMAP:
|
2024-01-24 02:21:01 +00:00
|
|
|
if (!bpf_token_capable(token, CAP_NET_ADMIN))
|
|
|
|
|
goto put_token;
|
2023-06-13 22:35:32 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
WARN(1, "unsupported map type %d", map_type);
|
2024-01-24 02:21:01 +00:00
|
|
|
goto put_token;
|
2023-06-13 22:35:32 +00:00
|
|
|
}
|
|
|
|
|
|
2023-06-13 22:35:31 +00:00
|
|
|
map = ops->map_alloc(attr);
|
2024-01-24 02:21:01 +00:00
|
|
|
if (IS_ERR(map)) {
|
|
|
|
|
err = PTR_ERR(map);
|
|
|
|
|
goto put_token;
|
|
|
|
|
}
|
2023-06-13 22:35:31 +00:00
|
|
|
map->ops = ops;
|
|
|
|
|
map->map_type = map_type;
|
2014-09-26 07:16:57 +00:00
|
|
|
|
2020-03-14 01:02:09 +00:00
|
|
|
err = bpf_obj_name_cpy(map->name, attr->map_name,
|
|
|
|
|
sizeof(attr->map_name));
|
|
|
|
|
if (err < 0)
|
2019-05-30 01:03:58 +00:00
|
|
|
goto free_map;
|
2017-09-27 21:37:53 +00:00
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
atomic64_set(&map->refcnt, 1);
|
|
|
|
|
atomic64_set(&map->usercnt, 1);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
mutex_init(&map->freeze_mutex);
|
2022-01-21 10:10:02 +00:00
|
|
|
spin_lock_init(&map->owner.lock);
|
2014-09-26 07:16:57 +00:00
|
|
|
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
if (attr->btf_key_type_id || attr->btf_value_type_id ||
|
|
|
|
|
/* Even the map's value is a kernel's struct,
|
|
|
|
|
* the bpf_prog.o must have BTF to begin with
|
|
|
|
|
* to figure out the corresponding kernel's
|
|
|
|
|
* counter part. Thus, attr->btf_fd has
|
|
|
|
|
* to be valid also.
|
|
|
|
|
*/
|
|
|
|
|
attr->btf_vmlinux_value_type_id) {
|
2018-04-18 22:56:03 +00:00
|
|
|
struct btf *btf;
|
|
|
|
|
|
|
|
|
|
btf = btf_get_by_fd(attr->btf_fd);
|
|
|
|
|
if (IS_ERR(btf)) {
|
|
|
|
|
err = PTR_ERR(btf);
|
2019-05-30 01:03:58 +00:00
|
|
|
goto free_map;
|
2018-04-18 22:56:03 +00:00
|
|
|
}
|
2021-03-07 22:52:48 +00:00
|
|
|
if (btf_is_kernel(btf)) {
|
|
|
|
|
btf_put(btf);
|
|
|
|
|
err = -EACCES;
|
|
|
|
|
goto free_map;
|
|
|
|
|
}
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
map->btf = btf;
|
2018-04-18 22:56:03 +00:00
|
|
|
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
if (attr->btf_value_type_id) {
|
2024-01-24 02:21:01 +00:00
|
|
|
err = map_check_btf(map, token, btf, attr->btf_key_type_id,
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
attr->btf_value_type_id);
|
|
|
|
|
if (err)
|
|
|
|
|
goto free_map;
|
2018-04-18 22:56:03 +00:00
|
|
|
}
|
|
|
|
|
|
2018-05-22 21:57:21 +00:00
|
|
|
map->btf_key_type_id = attr->btf_key_type_id;
|
|
|
|
|
map->btf_value_type_id = attr->btf_value_type_id;
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
map->btf_vmlinux_value_type_id =
|
|
|
|
|
attr->btf_vmlinux_value_type_id;
|
2018-04-18 22:56:03 +00:00
|
|
|
}
|
|
|
|
|
|
bpf,lsm: Refactor bpf_map_alloc/bpf_map_free LSM hooks
Similarly to bpf_prog_alloc LSM hook, rename and extend bpf_map_alloc
hook into bpf_map_create, taking not just struct bpf_map, but also
bpf_attr and bpf_token, to give a fuller context to LSMs.
Unlike bpf_prog_alloc, there is no need to move the hook around, as it
currently is firing right before allocating BPF map ID and FD, which
seems to be a sweet spot.
But like bpf_prog_alloc/bpf_prog_free combo, make sure that bpf_map_free
LSM hook is called even if bpf_map_create hook returned error, as if few
LSMs are combined together it could be that one LSM successfully
allocated security blob for its needs, while subsequent LSM rejected BPF
map creation. The former LSM would still need to free up LSM blob, so we
need to ensure security_bpf_map_free() is called regardless of the
outcome.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-11-andrii@kernel.org
2024-01-24 02:21:07 +00:00
|
|
|
err = security_bpf_map_create(map, attr, token);
|
2022-04-24 21:48:53 +00:00
|
|
|
if (err)
|
bpf,lsm: Refactor bpf_map_alloc/bpf_map_free LSM hooks
Similarly to bpf_prog_alloc LSM hook, rename and extend bpf_map_alloc
hook into bpf_map_create, taking not just struct bpf_map, but also
bpf_attr and bpf_token, to give a fuller context to LSMs.
Unlike bpf_prog_alloc, there is no need to move the hook around, as it
currently is firing right before allocating BPF map ID and FD, which
seems to be a sweet spot.
But like bpf_prog_alloc/bpf_prog_free combo, make sure that bpf_map_free
LSM hook is called even if bpf_map_create hook returned error, as if few
LSMs are combined together it could be that one LSM successfully
allocated security blob for its needs, while subsequent LSM rejected BPF
map creation. The former LSM would still need to free up LSM blob, so we
need to ensure security_bpf_map_free() is called regardless of the
outcome.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-11-andrii@kernel.org
2024-01-24 02:21:07 +00:00
|
|
|
goto free_map_sec;
|
2022-04-24 21:48:53 +00:00
|
|
|
|
2017-06-05 19:15:47 +00:00
|
|
|
err = bpf_map_alloc_id(map);
|
|
|
|
|
if (err)
|
2019-05-30 01:03:58 +00:00
|
|
|
goto free_map_sec;
|
2017-06-05 19:15:47 +00:00
|
|
|
|
2020-12-01 21:58:32 +00:00
|
|
|
bpf_map_save_memcg(map);
|
2024-01-24 02:21:01 +00:00
|
|
|
bpf_token_put(token);
|
2020-12-01 21:58:32 +00:00
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
err = bpf_map_new_fd(map, f_flags);
|
2017-06-05 19:15:50 +00:00
|
|
|
if (err < 0) {
|
|
|
|
|
/* failed to allocate fd.
|
2019-02-27 14:36:25 +00:00
|
|
|
* bpf_map_put_with_uref() is needed because the above
|
2017-06-05 19:15:50 +00:00
|
|
|
* bpf_map_alloc_id() has published the map
|
|
|
|
|
* to the userspace and the userspace may
|
|
|
|
|
* have refcnt-ed it through BPF_MAP_GET_FD_BY_ID.
|
|
|
|
|
*/
|
2019-02-27 14:36:25 +00:00
|
|
|
bpf_map_put_with_uref(map);
|
2017-06-05 19:15:50 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
2014-09-26 07:16:57 +00:00
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
|
2017-10-18 20:00:24 +00:00
|
|
|
free_map_sec:
|
|
|
|
|
security_bpf_map_free(map);
|
2019-05-30 01:03:58 +00:00
|
|
|
free_map:
|
2018-04-18 22:56:03 +00:00
|
|
|
btf_put(map->btf);
|
2014-09-26 07:16:57 +00:00
|
|
|
map->ops->map_free(map);
|
2024-01-24 02:21:01 +00:00
|
|
|
put_token:
|
|
|
|
|
bpf_token_put(token);
|
2014-09-26 07:16:57 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
/* if error is returned, fd is released.
|
|
|
|
|
* On success caller should complete fd access with matching fdput()
|
|
|
|
|
*/
|
2015-10-29 13:58:07 +00:00
|
|
|
struct bpf_map *__bpf_map_get(struct fd f)
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
{
|
|
|
|
|
if (!f.file)
|
|
|
|
|
return ERR_PTR(-EBADF);
|
|
|
|
|
if (f.file->f_op != &bpf_map_fops) {
|
|
|
|
|
fdput(f);
|
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
}
|
|
|
|
|
|
2015-10-29 13:58:07 +00:00
|
|
|
return f.file->private_data;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
void bpf_map_inc(struct bpf_map *map)
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
{
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
atomic64_inc(&map->refcnt);
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
}
|
2018-05-04 01:37:09 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_map_inc);
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
void bpf_map_inc_with_uref(struct bpf_map *map)
|
|
|
|
|
{
|
|
|
|
|
atomic64_inc(&map->refcnt);
|
|
|
|
|
atomic64_inc(&map->usercnt);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(bpf_map_inc_with_uref);
|
|
|
|
|
|
2020-02-25 23:04:21 +00:00
|
|
|
struct bpf_map *bpf_map_get(u32 ufd)
|
|
|
|
|
{
|
|
|
|
|
struct fd f = fdget(ufd);
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
|
|
|
|
|
map = __bpf_map_get(f);
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return map;
|
|
|
|
|
|
|
|
|
|
bpf_map_inc(map);
|
|
|
|
|
fdput(f);
|
|
|
|
|
|
|
|
|
|
return map;
|
|
|
|
|
}
|
2022-02-09 23:19:57 +00:00
|
|
|
EXPORT_SYMBOL(bpf_map_get);
|
2020-02-25 23:04:21 +00:00
|
|
|
|
bpf: fix clearing on persistent program array maps
Currently, when having map file descriptors pointing to program arrays,
there's still the issue that we unconditionally flush program array
contents via bpf_fd_array_map_clear() in bpf_map_release(). This happens
when such a file descriptor is released and is independent of the map's
refcount.
Having this flush independent of the refcount is for a reason: there
can be arbitrary complex dependency chains among tail calls, also circular
ones (direct or indirect, nesting limit determined during runtime), and
we need to make sure that the map drops all references to eBPF programs
it holds, so that the map's refcount can eventually drop to zero and
initiate its freeing. Btw, a walk of the whole dependency graph would
not be possible for various reasons, one being complexity and another
one inconsistency, i.e. new programs can be added to parts of the graph
at any time, so there's no guaranteed consistent state for the time of
such a walk.
Now, the program array pinning itself works, but the issue is that each
derived file descriptor on close would nevertheless call unconditionally
into bpf_fd_array_map_clear(). Instead, keep track of users and postpone
this flush until the last reference to a user is dropped. As this only
concerns a subset of references (f.e. a prog array could hold a program
that itself has reference on the prog array holding it, etc), we need to
track them separately.
Short analysis on the refcounting: on map creation time usercnt will be
one, so there's no change in behaviour for bpf_map_release(), if unpinned.
If we already fail in map_create(), we are immediately freed, and no
file descriptor has been made public yet. In bpf_obj_pin_user(), we need
to probe for a possible map in bpf_fd_probe_obj() already with a usercnt
reference, so before we drop the reference on the fd with fdput().
Therefore, if actual pinning fails, we need to drop that reference again
in bpf_any_put(), otherwise we keep holding it. When last reference
drops on the inode, the bpf_any_put() in bpf_evict_inode() will take
care of dropping the usercnt again. In the bpf_obj_get_user() case, the
bpf_any_get() will grab a reference on the usercnt, still at a time when
we have the reference on the path. Should we later on fail to grab a new
file descriptor, bpf_any_put() will drop it, otherwise we hold it until
bpf_map_release() time.
Joint work with Alexei.
Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-11-24 20:28:15 +00:00
|
|
|
struct bpf_map *bpf_map_get_with_uref(u32 ufd)
|
2015-10-29 13:58:07 +00:00
|
|
|
{
|
|
|
|
|
struct fd f = fdget(ufd);
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
|
|
|
|
|
map = __bpf_map_get(f);
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return map;
|
|
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
bpf_map_inc_with_uref(map);
|
2015-10-29 13:58:07 +00:00
|
|
|
fdput(f);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
|
|
|
|
return map;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-23 03:23:58 +00:00
|
|
|
/* map_idr_lock should have been held or the map should have been
|
|
|
|
|
* protected by rcu read lock.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_map *__bpf_map_inc_not_zero(struct bpf_map *map, bool uref)
|
2017-06-05 19:15:50 +00:00
|
|
|
{
|
|
|
|
|
int refold;
|
|
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
refold = atomic64_fetch_add_unless(&map->refcnt, 1, 0);
|
2017-06-05 19:15:50 +00:00
|
|
|
if (!refold)
|
|
|
|
|
return ERR_PTR(-ENOENT);
|
|
|
|
|
if (uref)
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
atomic64_inc(&map->usercnt);
|
2017-06-05 19:15:50 +00:00
|
|
|
|
|
|
|
|
return map;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
struct bpf_map *bpf_map_inc_not_zero(struct bpf_map *map)
|
2019-08-14 17:37:48 +00:00
|
|
|
{
|
|
|
|
|
spin_lock_bh(&map_idr_lock);
|
bpf: Switch bpf_map ref counter to atomic64_t so bpf_map_inc() never fails
92117d8443bc ("bpf: fix refcnt overflow") turned refcounting of bpf_map into
potentially failing operation, when refcount reaches BPF_MAX_REFCNT limit
(32k). Due to using 32-bit counter, it's possible in practice to overflow
refcounter and make it wrap around to 0, causing erroneous map free, while
there are still references to it, causing use-after-free problems.
But having a failing refcounting operations are problematic in some cases. One
example is mmap() interface. After establishing initial memory-mapping, user
is allowed to arbitrarily map/remap/unmap parts of mapped memory, arbitrarily
splitting it into multiple non-contiguous regions. All this happening without
any control from the users of mmap subsystem. Rather mmap subsystem sends
notifications to original creator of memory mapping through open/close
callbacks, which are optionally specified during initial memory mapping
creation. These callbacks are used to maintain accurate refcount for bpf_map
(see next patch in this series). The problem is that open() callback is not
supposed to fail, because memory-mapped resource is set up and properly
referenced. This is posing a problem for using memory-mapping with BPF maps.
One solution to this is to maintain separate refcount for just memory-mappings
and do single bpf_map_inc/bpf_map_put when it goes from/to zero, respectively.
There are similar use cases in current work on tcp-bpf, necessitating extra
counter as well. This seems like a rather unfortunate and ugly solution that
doesn't scale well to various new use cases.
Another approach to solve this is to use non-failing refcount_t type, which
uses 32-bit counter internally, but, once reaching overflow state at UINT_MAX,
stays there. This utlimately causes memory leak, but prevents use after free.
But given refcounting is not the most performance-critical operation with BPF
maps (it's not used from running BPF program code), we can also just switch to
64-bit counter that can't overflow in practice, potentially disadvantaging
32-bit platforms a tiny bit. This simplifies semantics and allows above
described scenarios to not worry about failing refcount increment operation.
In terms of struct bpf_map size, we are still good and use the same amount of
space:
BEFORE (3 cache lines, 8 bytes of padding at the end):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic_t refcnt __attribute__((__aligned__(64))); /* 128 4 */
atomic_t usercnt; /* 132 4 */
struct work_struct work; /* 136 32 */
char name[16]; /* 168 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 146, holes: 1, sum holes: 38 */
/* padding: 8 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
AFTER (same 3 cache lines, no extra padding now):
struct bpf_map {
const struct bpf_map_ops * ops __attribute__((__aligned__(64))); /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
int spin_lock_off; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
u32 btf_key_type_id; /* 56 4 */
u32 btf_value_type_id; /* 60 4 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct btf * btf; /* 64 8 */
struct bpf_map_memory memory; /* 72 16 */
bool unpriv_array; /* 88 1 */
bool frozen; /* 89 1 */
/* XXX 38 bytes hole, try to pack */
/* --- cacheline 2 boundary (128 bytes) --- */
atomic64_t refcnt __attribute__((__aligned__(64))); /* 128 8 */
atomic64_t usercnt; /* 136 8 */
struct work_struct work; /* 144 32 */
char name[16]; /* 176 16 */
/* size: 192, cachelines: 3, members: 21 */
/* sum members: 154, holes: 1, sum holes: 38 */
/* forced alignments: 2, forced holes: 1, sum forced holes: 38 */
} __attribute__((__aligned__(64)));
This patch, while modifying all users of bpf_map_inc, also cleans up its
interface to match bpf_map_put with separate operations for bpf_map_inc and
bpf_map_inc_with_uref (to match bpf_map_put and bpf_map_put_with_uref,
respectively). Also, given there are no users of bpf_map_inc_not_zero
specifying uref=true, remove uref flag and default to uref=false internally.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-2-andriin@fb.com
2019-11-17 17:28:02 +00:00
|
|
|
map = __bpf_map_inc_not_zero(map, false);
|
2019-08-14 17:37:48 +00:00
|
|
|
spin_unlock_bh(&map_idr_lock);
|
|
|
|
|
|
|
|
|
|
return map;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(bpf_map_inc_not_zero);
|
|
|
|
|
|
2016-03-10 02:56:49 +00:00
|
|
|
int __weak bpf_stackmap_copy(struct bpf_map *map, void *key, void *value)
|
|
|
|
|
{
|
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
2018-10-18 13:16:14 +00:00
|
|
|
static void *__bpf_copy_key(void __user *ukey, u64 key_size)
|
|
|
|
|
{
|
|
|
|
|
if (key_size)
|
2021-08-18 23:52:16 +00:00
|
|
|
return vmemdup_user(ukey, key_size);
|
2018-10-18 13:16:14 +00:00
|
|
|
|
|
|
|
|
if (ukey)
|
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
static void *___bpf_copy_key(bpfptr_t ukey, u64 key_size)
|
|
|
|
|
{
|
|
|
|
|
if (key_size)
|
2021-08-18 23:52:16 +00:00
|
|
|
return kvmemdup_bpfptr(ukey, key_size);
|
2021-05-14 00:36:05 +00:00
|
|
|
|
|
|
|
|
if (!bpfptr_is_null(ukey))
|
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
/* last field in 'union bpf_attr' used by this command */
|
2019-01-31 23:40:09 +00:00
|
|
|
#define BPF_MAP_LOOKUP_ELEM_LAST_FIELD flags
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
|
|
|
|
static int map_lookup_elem(union bpf_attr *attr)
|
|
|
|
|
{
|
2016-11-13 18:44:03 +00:00
|
|
|
void __user *ukey = u64_to_user_ptr(attr->key);
|
|
|
|
|
void __user *uvalue = u64_to_user_ptr(attr->value);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
2020-01-15 18:43:00 +00:00
|
|
|
void *key, *value;
|
bpf: add lookup/update support for per-cpu hash and array maps
The functions bpf_map_lookup_elem(map, key, value) and
bpf_map_update_elem(map, key, value, flags) need to get/set
values from all-cpus for per-cpu hash and array maps,
so that user space can aggregate/update them as necessary.
Example of single counter aggregation in user space:
unsigned int nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
long values[nr_cpus];
long value = 0;
bpf_lookup_elem(fd, key, values);
for (i = 0; i < nr_cpus; i++)
value += values[i];
The user space must provide round_up(value_size, 8) * nr_cpus
array to get/set values, since kernel will use 'long' copy
of per-cpu values to try to copy good counters atomically.
It's a best-effort, since bpf programs and user space are racing
to access the same memory.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-02 06:39:55 +00:00
|
|
|
u32 value_size;
|
2015-09-08 16:00:09 +00:00
|
|
|
struct fd f;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_LOOKUP_ELEM))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2019-01-31 23:40:09 +00:00
|
|
|
if (attr->flags & ~BPF_F_LOCK)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2015-09-08 16:00:09 +00:00
|
|
|
f = fdget(ufd);
|
2015-10-29 13:58:07 +00:00
|
|
|
map = __bpf_map_get(f);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) {
|
2017-10-18 20:00:22 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2019-01-31 23:40:09 +00:00
|
|
|
if ((attr->flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK)) {
|
2019-01-31 23:40:09 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2018-10-18 13:16:14 +00:00
|
|
|
key = __bpf_copy_key(ukey, map->key_size);
|
2017-05-13 22:43:00 +00:00
|
|
|
if (IS_ERR(key)) {
|
|
|
|
|
err = PTR_ERR(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
goto err_put;
|
2017-05-13 22:43:00 +00:00
|
|
|
}
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
2020-01-15 18:43:00 +00:00
|
|
|
value_size = bpf_map_value_size(map);
|
bpf: add lookup/update support for per-cpu hash and array maps
The functions bpf_map_lookup_elem(map, key, value) and
bpf_map_update_elem(map, key, value, flags) need to get/set
values from all-cpus for per-cpu hash and array maps,
so that user space can aggregate/update them as necessary.
Example of single counter aggregation in user space:
unsigned int nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
long values[nr_cpus];
long value = 0;
bpf_lookup_elem(fd, key, values);
for (i = 0; i < nr_cpus; i++)
value += values[i];
The user space must provide round_up(value_size, 8) * nr_cpus
array to get/set values, since kernel will use 'long' copy
of per-cpu values to try to copy good counters atomically.
It's a best-effort, since bpf programs and user space are racing
to access the same memory.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-02 06:39:55 +00:00
|
|
|
|
2015-01-23 01:11:08 +00:00
|
|
|
err = -ENOMEM;
|
2021-08-18 23:52:15 +00:00
|
|
|
value = kvmalloc(value_size, GFP_USER | __GFP_NOWARN);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (!value)
|
2015-01-23 01:11:08 +00:00
|
|
|
goto free_key;
|
|
|
|
|
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_BLOOM_FILTER) {
|
|
|
|
|
if (copy_from_user(value, uvalue, value_size))
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
else
|
|
|
|
|
err = bpf_map_copy_value(map, key, value, attr->flags);
|
|
|
|
|
goto free_value;
|
|
|
|
|
}
|
|
|
|
|
|
2020-01-15 18:43:00 +00:00
|
|
|
err = bpf_map_copy_value(map, key, value, attr->flags);
|
bpf: add lookup/update support for per-cpu hash and array maps
The functions bpf_map_lookup_elem(map, key, value) and
bpf_map_update_elem(map, key, value, flags) need to get/set
values from all-cpus for per-cpu hash and array maps,
so that user space can aggregate/update them as necessary.
Example of single counter aggregation in user space:
unsigned int nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
long values[nr_cpus];
long value = 0;
bpf_lookup_elem(fd, key, values);
for (i = 0; i < nr_cpus; i++)
value += values[i];
The user space must provide round_up(value_size, 8) * nr_cpus
array to get/set values, since kernel will use 'long' copy
of per-cpu values to try to copy good counters atomically.
It's a best-effort, since bpf programs and user space are racing
to access the same memory.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-02 06:39:55 +00:00
|
|
|
if (err)
|
2015-01-23 01:11:08 +00:00
|
|
|
goto free_value;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
|
|
|
|
err = -EFAULT;
|
bpf: add lookup/update support for per-cpu hash and array maps
The functions bpf_map_lookup_elem(map, key, value) and
bpf_map_update_elem(map, key, value, flags) need to get/set
values from all-cpus for per-cpu hash and array maps,
so that user space can aggregate/update them as necessary.
Example of single counter aggregation in user space:
unsigned int nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
long values[nr_cpus];
long value = 0;
bpf_lookup_elem(fd, key, values);
for (i = 0; i < nr_cpus; i++)
value += values[i];
The user space must provide round_up(value_size, 8) * nr_cpus
array to get/set values, since kernel will use 'long' copy
of per-cpu values to try to copy good counters atomically.
It's a best-effort, since bpf programs and user space are racing
to access the same memory.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-02 06:39:55 +00:00
|
|
|
if (copy_to_user(uvalue, value, value_size) != 0)
|
2015-01-23 01:11:08 +00:00
|
|
|
goto free_value;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
|
|
2015-01-23 01:11:08 +00:00
|
|
|
free_value:
|
2021-08-18 23:52:15 +00:00
|
|
|
kvfree(value);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
free_key:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
err_put:
|
|
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2018-10-12 10:54:27 +00:00
|
|
|
|
bpf: add 'flags' attribute to BPF_MAP_UPDATE_ELEM command
the current meaning of BPF_MAP_UPDATE_ELEM syscall command is:
either update existing map element or create a new one.
Initially the plan was to add a new command to handle the case of
'create new element if it didn't exist', but 'flags' style looks
cleaner and overall diff is much smaller (more code reused), so add 'flags'
attribute to BPF_MAP_UPDATE_ELEM command with the following meaning:
#define BPF_ANY 0 /* create new element or update existing */
#define BPF_NOEXIST 1 /* create new element if it didn't exist */
#define BPF_EXIST 2 /* update existing element */
bpf_update_elem(fd, key, value, BPF_NOEXIST) call can fail with EEXIST
if element already exists.
bpf_update_elem(fd, key, value, BPF_EXIST) can fail with ENOENT
if element doesn't exist.
Userspace will call it as:
int bpf_update_elem(int fd, void *key, void *value, __u64 flags)
{
union bpf_attr attr = {
.map_fd = fd,
.key = ptr_to_u64(key),
.value = ptr_to_u64(value),
.flags = flags;
};
return bpf(BPF_MAP_UPDATE_ELEM, &attr, sizeof(attr));
}
First two bits of 'flags' are used to encode style of bpf_update_elem() command.
Bits 2-63 are reserved for future use.
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-11-14 01:36:44 +00:00
|
|
|
#define BPF_MAP_UPDATE_ELEM_LAST_FIELD flags
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
static int map_update_elem(union bpf_attr *attr, bpfptr_t uattr)
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
{
|
2021-05-14 00:36:05 +00:00
|
|
|
bpfptr_t ukey = make_bpfptr(attr->key, uattr.is_kernel);
|
|
|
|
|
bpfptr_t uvalue = make_bpfptr(attr->value, uattr.is_kernel);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
void *key, *value;
|
bpf: add lookup/update support for per-cpu hash and array maps
The functions bpf_map_lookup_elem(map, key, value) and
bpf_map_update_elem(map, key, value, flags) need to get/set
values from all-cpus for per-cpu hash and array maps,
so that user space can aggregate/update them as necessary.
Example of single counter aggregation in user space:
unsigned int nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
long values[nr_cpus];
long value = 0;
bpf_lookup_elem(fd, key, values);
for (i = 0; i < nr_cpus; i++)
value += values[i];
The user space must provide round_up(value_size, 8) * nr_cpus
array to get/set values, since kernel will use 'long' copy
of per-cpu values to try to copy good counters atomically.
It's a best-effort, since bpf programs and user space are racing
to access the same memory.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-02 06:39:55 +00:00
|
|
|
u32 value_size;
|
2015-09-08 16:00:09 +00:00
|
|
|
struct fd f;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_UPDATE_ELEM))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2015-09-08 16:00:09 +00:00
|
|
|
f = fdget(ufd);
|
2015-10-29 13:58:07 +00:00
|
|
|
map = __bpf_map_get(f);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_inc(map);
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) {
|
2017-10-18 20:00:22 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2019-01-31 23:40:09 +00:00
|
|
|
if ((attr->flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK)) {
|
2019-01-31 23:40:09 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
key = ___bpf_copy_key(ukey, map->key_size);
|
2017-05-13 22:43:00 +00:00
|
|
|
if (IS_ERR(key)) {
|
|
|
|
|
err = PTR_ERR(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
goto err_put;
|
2017-05-13 22:43:00 +00:00
|
|
|
}
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
2021-08-18 23:52:15 +00:00
|
|
|
value_size = bpf_map_value_size(map);
|
2022-09-13 08:40:33 +00:00
|
|
|
value = kvmemdup_bpfptr(uvalue, value_size);
|
|
|
|
|
if (IS_ERR(value)) {
|
|
|
|
|
err = PTR_ERR(value);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
goto free_key;
|
2022-09-13 08:40:33 +00:00
|
|
|
}
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
2022-11-16 07:50:58 +00:00
|
|
|
err = bpf_map_update_value(map, f.file, key, value, attr->flags);
|
2023-12-08 10:23:52 +00:00
|
|
|
if (!err)
|
|
|
|
|
maybe_wait_bpf_programs(map);
|
2017-10-16 10:19:28 +00:00
|
|
|
|
2021-08-18 23:52:15 +00:00
|
|
|
kvfree(value);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
free_key:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
err_put:
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_dec(map);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define BPF_MAP_DELETE_ELEM_LAST_FIELD key
|
|
|
|
|
|
2022-08-24 13:40:36 +00:00
|
|
|
static int map_delete_elem(union bpf_attr *attr, bpfptr_t uattr)
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
{
|
2022-08-24 13:40:36 +00:00
|
|
|
bpfptr_t ukey = make_bpfptr(attr->key, uattr.is_kernel);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
2015-09-08 16:00:09 +00:00
|
|
|
struct fd f;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
void *key;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_DELETE_ELEM))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2015-09-08 16:00:09 +00:00
|
|
|
f = fdget(ufd);
|
2015-10-29 13:58:07 +00:00
|
|
|
map = __bpf_map_get(f);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_inc(map);
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) {
|
2017-10-18 20:00:22 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2022-08-24 13:40:36 +00:00
|
|
|
key = ___bpf_copy_key(ukey, map->key_size);
|
2017-05-13 22:43:00 +00:00
|
|
|
if (IS_ERR(key)) {
|
|
|
|
|
err = PTR_ERR(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
goto err_put;
|
2017-05-13 22:43:00 +00:00
|
|
|
}
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map)) {
|
2018-01-12 04:29:09 +00:00
|
|
|
err = bpf_map_offload_delete_elem(map, key);
|
|
|
|
|
goto out;
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
} else if (IS_FD_PROG_ARRAY(map) ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_STRUCT_OPS) {
|
|
|
|
|
/* These maps require sleepable context */
|
2019-11-22 20:07:58 +00:00
|
|
|
err = map->ops->map_delete_elem(map, key);
|
|
|
|
|
goto out;
|
2018-01-12 04:29:09 +00:00
|
|
|
}
|
|
|
|
|
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_disable_instrumentation();
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_delete_elem(map, key);
|
|
|
|
|
rcu_read_unlock();
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_enable_instrumentation();
|
2023-12-08 10:23:52 +00:00
|
|
|
if (!err)
|
|
|
|
|
maybe_wait_bpf_programs(map);
|
2018-01-12 04:29:09 +00:00
|
|
|
out:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
err_put:
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_dec(map);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* last field in 'union bpf_attr' used by this command */
|
|
|
|
|
#define BPF_MAP_GET_NEXT_KEY_LAST_FIELD next_key
|
|
|
|
|
|
|
|
|
|
static int map_get_next_key(union bpf_attr *attr)
|
|
|
|
|
{
|
2016-11-13 18:44:03 +00:00
|
|
|
void __user *ukey = u64_to_user_ptr(attr->key);
|
|
|
|
|
void __user *unext_key = u64_to_user_ptr(attr->next_key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
void *key, *next_key;
|
2015-09-08 16:00:09 +00:00
|
|
|
struct fd f;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_GET_NEXT_KEY))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2015-09-08 16:00:09 +00:00
|
|
|
f = fdget(ufd);
|
2015-10-29 13:58:07 +00:00
|
|
|
map = __bpf_map_get(f);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) {
|
2017-10-18 20:00:22 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2017-04-25 02:00:37 +00:00
|
|
|
if (ukey) {
|
2018-10-18 13:16:14 +00:00
|
|
|
key = __bpf_copy_key(ukey, map->key_size);
|
2017-05-13 22:43:00 +00:00
|
|
|
if (IS_ERR(key)) {
|
|
|
|
|
err = PTR_ERR(key);
|
2017-04-25 02:00:37 +00:00
|
|
|
goto err_put;
|
2017-05-13 22:43:00 +00:00
|
|
|
}
|
2017-04-25 02:00:37 +00:00
|
|
|
} else {
|
|
|
|
|
key = NULL;
|
|
|
|
|
}
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
|
|
|
|
|
err = -ENOMEM;
|
2021-08-18 23:52:16 +00:00
|
|
|
next_key = kvmalloc(map->key_size, GFP_USER);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (!next_key)
|
|
|
|
|
goto free_key;
|
|
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map)) {
|
2018-01-12 04:29:09 +00:00
|
|
|
err = bpf_map_offload_get_next_key(map, key, next_key);
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_get_next_key(map, key, next_key);
|
|
|
|
|
rcu_read_unlock();
|
2018-01-12 04:29:09 +00:00
|
|
|
out:
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
if (err)
|
|
|
|
|
goto free_next_key;
|
|
|
|
|
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
if (copy_to_user(unext_key, next_key, map->key_size) != 0)
|
|
|
|
|
goto free_next_key;
|
|
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
|
|
|
|
|
|
free_next_key:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(next_key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
free_key:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
err_put:
|
|
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-01-15 18:43:02 +00:00
|
|
|
int generic_map_delete_batch(struct bpf_map *map,
|
|
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
void __user *keys = u64_to_user_ptr(attr->batch.keys);
|
|
|
|
|
u32 cp, max_count;
|
|
|
|
|
int err = 0;
|
|
|
|
|
void *key;
|
|
|
|
|
|
|
|
|
|
if (attr->batch.elem_flags & ~BPF_F_LOCK)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if ((attr->batch.elem_flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK)) {
|
2020-01-15 18:43:02 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
max_count = attr->batch.count;
|
|
|
|
|
if (!max_count)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2023-12-08 10:23:53 +00:00
|
|
|
if (put_user(0, &uattr->batch.count))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
2021-08-18 23:52:16 +00:00
|
|
|
key = kvmalloc(map->key_size, GFP_USER | __GFP_NOWARN);
|
2020-01-19 19:40:40 +00:00
|
|
|
if (!key)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
2020-01-15 18:43:02 +00:00
|
|
|
for (cp = 0; cp < max_count; cp++) {
|
2020-01-19 19:40:40 +00:00
|
|
|
err = -EFAULT;
|
|
|
|
|
if (copy_from_user(key, keys + cp * map->key_size,
|
|
|
|
|
map->key_size))
|
2020-01-15 18:43:02 +00:00
|
|
|
break;
|
|
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map)) {
|
2020-01-15 18:43:02 +00:00
|
|
|
err = bpf_map_offload_delete_elem(map, key);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_disable_instrumentation();
|
2020-01-15 18:43:02 +00:00
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_delete_elem(map, key);
|
|
|
|
|
rcu_read_unlock();
|
2020-02-24 14:01:49 +00:00
|
|
|
bpf_enable_instrumentation();
|
2020-01-15 18:43:02 +00:00
|
|
|
if (err)
|
|
|
|
|
break;
|
2022-02-17 18:19:02 +00:00
|
|
|
cond_resched();
|
2020-01-15 18:43:02 +00:00
|
|
|
}
|
|
|
|
|
if (copy_to_user(&uattr->batch.count, &cp, sizeof(cp)))
|
|
|
|
|
err = -EFAULT;
|
2020-01-19 19:40:40 +00:00
|
|
|
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
2022-02-18 18:18:01 +00:00
|
|
|
|
2020-01-15 18:43:02 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2022-11-16 07:50:58 +00:00
|
|
|
int generic_map_update_batch(struct bpf_map *map, struct file *map_file,
|
2020-01-15 18:43:02 +00:00
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
void __user *values = u64_to_user_ptr(attr->batch.values);
|
|
|
|
|
void __user *keys = u64_to_user_ptr(attr->batch.keys);
|
|
|
|
|
u32 value_size, cp, max_count;
|
|
|
|
|
void *key, *value;
|
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
|
|
if (attr->batch.elem_flags & ~BPF_F_LOCK)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if ((attr->batch.elem_flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK)) {
|
2020-01-15 18:43:02 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
value_size = bpf_map_value_size(map);
|
|
|
|
|
|
|
|
|
|
max_count = attr->batch.count;
|
|
|
|
|
if (!max_count)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2023-12-08 10:23:53 +00:00
|
|
|
if (put_user(0, &uattr->batch.count))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
2021-08-18 23:52:16 +00:00
|
|
|
key = kvmalloc(map->key_size, GFP_USER | __GFP_NOWARN);
|
2020-01-19 19:40:40 +00:00
|
|
|
if (!key)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
2021-08-18 23:52:15 +00:00
|
|
|
value = kvmalloc(value_size, GFP_USER | __GFP_NOWARN);
|
2020-01-19 19:40:40 +00:00
|
|
|
if (!value) {
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
2020-01-15 18:43:02 +00:00
|
|
|
return -ENOMEM;
|
2020-01-19 19:40:40 +00:00
|
|
|
}
|
2020-01-15 18:43:02 +00:00
|
|
|
|
|
|
|
|
for (cp = 0; cp < max_count; cp++) {
|
|
|
|
|
err = -EFAULT;
|
2020-01-19 19:40:40 +00:00
|
|
|
if (copy_from_user(key, keys + cp * map->key_size,
|
|
|
|
|
map->key_size) ||
|
|
|
|
|
copy_from_user(value, values + cp * value_size, value_size))
|
2020-01-15 18:43:02 +00:00
|
|
|
break;
|
|
|
|
|
|
2022-11-16 07:50:58 +00:00
|
|
|
err = bpf_map_update_value(map, map_file, key, value,
|
2020-01-15 18:43:02 +00:00
|
|
|
attr->batch.elem_flags);
|
|
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
|
break;
|
2022-02-17 18:19:02 +00:00
|
|
|
cond_resched();
|
2020-01-15 18:43:02 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (copy_to_user(&uattr->batch.count, &cp, sizeof(cp)))
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
|
2021-08-18 23:52:15 +00:00
|
|
|
kvfree(value);
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
2023-12-08 10:23:50 +00:00
|
|
|
|
2020-01-15 18:43:02 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-01-15 18:43:01 +00:00
|
|
|
#define MAP_LOOKUP_RETRIES 3
|
|
|
|
|
|
|
|
|
|
int generic_map_lookup_batch(struct bpf_map *map,
|
|
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
void __user *uobatch = u64_to_user_ptr(attr->batch.out_batch);
|
|
|
|
|
void __user *ubatch = u64_to_user_ptr(attr->batch.in_batch);
|
|
|
|
|
void __user *values = u64_to_user_ptr(attr->batch.values);
|
|
|
|
|
void __user *keys = u64_to_user_ptr(attr->batch.keys);
|
|
|
|
|
void *buf, *buf_prevkey, *prev_key, *key, *value;
|
|
|
|
|
int err, retry = MAP_LOOKUP_RETRIES;
|
|
|
|
|
u32 value_size, cp, max_count;
|
|
|
|
|
|
|
|
|
|
if (attr->batch.elem_flags & ~BPF_F_LOCK)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if ((attr->batch.elem_flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK))
|
2020-01-15 18:43:01 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
value_size = bpf_map_value_size(map);
|
|
|
|
|
|
|
|
|
|
max_count = attr->batch.count;
|
|
|
|
|
if (!max_count)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
if (put_user(0, &uattr->batch.count))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
2021-08-18 23:52:16 +00:00
|
|
|
buf_prevkey = kvmalloc(map->key_size, GFP_USER | __GFP_NOWARN);
|
2020-01-15 18:43:01 +00:00
|
|
|
if (!buf_prevkey)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
2021-08-18 23:52:15 +00:00
|
|
|
buf = kvmalloc(map->key_size + value_size, GFP_USER | __GFP_NOWARN);
|
2020-01-15 18:43:01 +00:00
|
|
|
if (!buf) {
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(buf_prevkey);
|
2020-01-15 18:43:01 +00:00
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
prev_key = NULL;
|
|
|
|
|
if (ubatch && copy_from_user(buf_prevkey, ubatch, map->key_size))
|
|
|
|
|
goto free_buf;
|
|
|
|
|
key = buf;
|
|
|
|
|
value = key + map->key_size;
|
|
|
|
|
if (ubatch)
|
|
|
|
|
prev_key = buf_prevkey;
|
|
|
|
|
|
|
|
|
|
for (cp = 0; cp < max_count;) {
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_get_next_key(map, prev_key, key);
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
if (err)
|
|
|
|
|
break;
|
|
|
|
|
err = bpf_map_copy_value(map, key, value,
|
|
|
|
|
attr->batch.elem_flags);
|
|
|
|
|
|
|
|
|
|
if (err == -ENOENT) {
|
|
|
|
|
if (retry) {
|
|
|
|
|
retry--;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
err = -EINTR;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
|
goto free_buf;
|
|
|
|
|
|
|
|
|
|
if (copy_to_user(keys + cp * map->key_size, key,
|
|
|
|
|
map->key_size)) {
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
goto free_buf;
|
|
|
|
|
}
|
|
|
|
|
if (copy_to_user(values + cp * value_size, value, value_size)) {
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
goto free_buf;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!prev_key)
|
|
|
|
|
prev_key = buf_prevkey;
|
|
|
|
|
|
|
|
|
|
swap(prev_key, key);
|
|
|
|
|
retry = MAP_LOOKUP_RETRIES;
|
|
|
|
|
cp++;
|
2022-02-17 18:19:02 +00:00
|
|
|
cond_resched();
|
2020-01-15 18:43:01 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (err == -EFAULT)
|
|
|
|
|
goto free_buf;
|
|
|
|
|
|
|
|
|
|
if ((copy_to_user(&uattr->batch.count, &cp, sizeof(cp)) ||
|
|
|
|
|
(cp && copy_to_user(uobatch, prev_key, map->key_size))))
|
|
|
|
|
err = -EFAULT;
|
|
|
|
|
|
|
|
|
|
free_buf:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(buf_prevkey);
|
2021-08-18 23:52:15 +00:00
|
|
|
kvfree(buf);
|
2020-01-15 18:43:01 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-11 21:00:04 +00:00
|
|
|
#define BPF_MAP_LOOKUP_AND_DELETE_ELEM_LAST_FIELD flags
|
2018-10-18 13:16:30 +00:00
|
|
|
|
|
|
|
|
static int map_lookup_and_delete_elem(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
void __user *ukey = u64_to_user_ptr(attr->key);
|
|
|
|
|
void __user *uvalue = u64_to_user_ptr(attr->value);
|
|
|
|
|
int ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
2018-10-19 20:52:38 +00:00
|
|
|
void *key, *value;
|
2018-10-18 13:16:30 +00:00
|
|
|
u32 value_size;
|
|
|
|
|
struct fd f;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_LOOKUP_AND_DELETE_ELEM))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2021-05-11 21:00:04 +00:00
|
|
|
if (attr->flags & ~BPF_F_LOCK)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2018-10-18 13:16:30 +00:00
|
|
|
f = fdget(ufd);
|
|
|
|
|
map = __bpf_map_get(f);
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_inc(map);
|
2020-05-27 18:56:59 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ) ||
|
|
|
|
|
!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) {
|
2018-10-18 13:16:30 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-11 21:00:04 +00:00
|
|
|
if (attr->flags &&
|
|
|
|
|
(map->map_type == BPF_MAP_TYPE_QUEUE ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_STACK)) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if ((attr->flags & BPF_F_LOCK) &&
|
2022-11-03 19:09:56 +00:00
|
|
|
!btf_record_has_field(map->record, BPF_SPIN_LOCK)) {
|
2021-05-11 21:00:04 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2018-10-18 13:16:30 +00:00
|
|
|
key = __bpf_copy_key(ukey, map->key_size);
|
|
|
|
|
if (IS_ERR(key)) {
|
|
|
|
|
err = PTR_ERR(key);
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-11 21:00:04 +00:00
|
|
|
value_size = bpf_map_value_size(map);
|
2018-10-18 13:16:30 +00:00
|
|
|
|
|
|
|
|
err = -ENOMEM;
|
2021-08-18 23:52:15 +00:00
|
|
|
value = kvmalloc(value_size, GFP_USER | __GFP_NOWARN);
|
2018-10-18 13:16:30 +00:00
|
|
|
if (!value)
|
|
|
|
|
goto free_key;
|
|
|
|
|
|
2021-05-11 21:00:04 +00:00
|
|
|
err = -ENOTSUPP;
|
2018-10-18 13:16:30 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_QUEUE ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_STACK) {
|
|
|
|
|
err = map->ops->map_pop_elem(map, value);
|
2021-05-11 21:00:04 +00:00
|
|
|
} else if (map->map_type == BPF_MAP_TYPE_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_LRU_HASH ||
|
|
|
|
|
map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
|
2023-01-19 22:15:21 +00:00
|
|
|
if (!bpf_map_is_offloaded(map)) {
|
2021-05-11 21:00:04 +00:00
|
|
|
bpf_disable_instrumentation();
|
|
|
|
|
rcu_read_lock();
|
|
|
|
|
err = map->ops->map_lookup_and_delete_elem(map, key, value, attr->flags);
|
|
|
|
|
rcu_read_unlock();
|
|
|
|
|
bpf_enable_instrumentation();
|
|
|
|
|
}
|
2018-10-18 13:16:30 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
|
goto free_value;
|
|
|
|
|
|
2020-04-30 08:18:51 +00:00
|
|
|
if (copy_to_user(uvalue, value, value_size) != 0) {
|
|
|
|
|
err = -EFAULT;
|
2018-10-18 13:16:30 +00:00
|
|
|
goto free_value;
|
2020-04-30 08:18:51 +00:00
|
|
|
}
|
2018-10-18 13:16:30 +00:00
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
|
|
|
|
|
|
free_value:
|
2021-08-18 23:52:15 +00:00
|
|
|
kvfree(value);
|
2018-10-18 13:16:30 +00:00
|
|
|
free_key:
|
2021-08-18 23:52:16 +00:00
|
|
|
kvfree(key);
|
2018-10-18 13:16:30 +00:00
|
|
|
err_put:
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_dec(map);
|
2018-10-18 13:16:30 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
#define BPF_MAP_FREEZE_LAST_FIELD map_fd
|
|
|
|
|
|
|
|
|
|
static int map_freeze(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
int err = 0, ufd = attr->map_fd;
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
struct fd f;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_FREEZE))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
f = fdget(ufd);
|
|
|
|
|
map = __bpf_map_get(f);
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
|
2022-11-03 19:09:56 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_STRUCT_OPS || !IS_ERR_OR_NULL(map->record)) {
|
2020-03-05 01:34:54 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
2023-05-24 22:54:19 +00:00
|
|
|
if (!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) {
|
2023-05-26 10:13:56 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return -EPERM;
|
2023-05-24 22:54:19 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
mutex_lock(&map->freeze_mutex);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
if (bpf_map_write_active(map)) {
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
err = -EBUSY;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
if (READ_ONCE(map->frozen)) {
|
|
|
|
|
err = -EBUSY;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
WRITE_ONCE(map->frozen, true);
|
|
|
|
|
err_put:
|
bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY
Add ability to memory-map contents of BPF array map. This is extremely useful
for working with BPF global data from userspace programs. It allows to avoid
typical bpf_map_{lookup,update}_elem operations, improving both performance
and usability.
There had to be special considerations for map freezing, to avoid having
writable memory view into a frozen map. To solve this issue, map freezing and
mmap-ing is happening under mutex now:
- if map is already frozen, no writable mapping is allowed;
- if map has writable memory mappings active (accounted in map->writecnt),
map freezing will keep failing with -EBUSY;
- once number of writable memory mappings drops to zero, map freezing can be
performed again.
Only non-per-CPU plain arrays are supported right now. Maps with spinlocks
can't be memory mapped either.
For BPF_F_MMAPABLE array, memory allocation has to be done through vmalloc()
to be mmap()'able. We also need to make sure that array data memory is
page-sized and page-aligned, so we over-allocate memory in such a way that
struct bpf_array is at the end of a single page of memory with array->value
being aligned with the start of the second page. On deallocation we need to
accomodate this memory arrangement to free vmalloc()'ed memory correctly.
One important consideration regarding how memory-mapping subsystem functions.
Memory-mapping subsystem provides few optional callbacks, among them open()
and close(). close() is called for each memory region that is unmapped, so
that users can decrease their reference counters and free up resources, if
necessary. open() is *almost* symmetrical: it's called for each memory region
that is being mapped, **except** the very first one. So bpf_map_mmap does
initial refcnt bump, while open() will do any extra ones after that. Thus
number of close() calls is equal to number of open() calls plus one more.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/bpf/20191117172806.2195367-4-andriin@fb.com
2019-11-17 17:28:04 +00:00
|
|
|
mutex_unlock(&map->freeze_mutex);
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2017-10-16 23:40:53 +00:00
|
|
|
static const struct bpf_prog_ops * const bpf_prog_types[] = {
|
2019-11-14 18:57:15 +00:00
|
|
|
#define BPF_PROG_TYPE(_id, _name, prog_ctx_type, kern_ctx_type) \
|
2017-10-16 23:40:53 +00:00
|
|
|
[_id] = & _name ## _prog_ops,
|
|
|
|
|
#define BPF_MAP_TYPE(_id, _ops)
|
2020-04-29 00:16:08 +00:00
|
|
|
#define BPF_LINK_TYPE(_id, _name)
|
2017-10-16 23:40:53 +00:00
|
|
|
#include <linux/bpf_types.h>
|
|
|
|
|
#undef BPF_PROG_TYPE
|
|
|
|
|
#undef BPF_MAP_TYPE
|
2020-04-29 00:16:08 +00:00
|
|
|
#undef BPF_LINK_TYPE
|
2017-10-16 23:40:53 +00:00
|
|
|
};
|
|
|
|
|
|
2014-09-26 07:17:00 +00:00
|
|
|
static int find_prog_type(enum bpf_prog_type type, struct bpf_prog *prog)
|
|
|
|
|
{
|
2018-05-04 00:13:57 +00:00
|
|
|
const struct bpf_prog_ops *ops;
|
|
|
|
|
|
|
|
|
|
if (type >= ARRAY_SIZE(bpf_prog_types))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
type = array_index_nospec(type, ARRAY_SIZE(bpf_prog_types));
|
|
|
|
|
ops = bpf_prog_types[type];
|
|
|
|
|
if (!ops)
|
2017-04-11 13:34:57 +00:00
|
|
|
return -EINVAL;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (!bpf_prog_is_offloaded(prog->aux))
|
2018-05-04 00:13:57 +00:00
|
|
|
prog->aux->ops = ops;
|
2017-11-03 20:56:17 +00:00
|
|
|
else
|
|
|
|
|
prog->aux->ops = &bpf_offload_prog_ops;
|
2017-04-11 13:34:57 +00:00
|
|
|
prog->type = type;
|
|
|
|
|
return 0;
|
2014-09-26 07:17:00 +00:00
|
|
|
}
|
|
|
|
|
|
2019-12-06 21:49:34 +00:00
|
|
|
enum bpf_audit {
|
|
|
|
|
BPF_AUDIT_LOAD,
|
|
|
|
|
BPF_AUDIT_UNLOAD,
|
|
|
|
|
BPF_AUDIT_MAX,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static const char * const bpf_audit_str[BPF_AUDIT_MAX] = {
|
|
|
|
|
[BPF_AUDIT_LOAD] = "LOAD",
|
|
|
|
|
[BPF_AUDIT_UNLOAD] = "UNLOAD",
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static void bpf_audit_prog(const struct bpf_prog *prog, unsigned int op)
|
|
|
|
|
{
|
|
|
|
|
struct audit_context *ctx = NULL;
|
|
|
|
|
struct audit_buffer *ab;
|
|
|
|
|
|
|
|
|
|
if (WARN_ON_ONCE(op >= BPF_AUDIT_MAX))
|
|
|
|
|
return;
|
|
|
|
|
if (audit_enabled == AUDIT_OFF)
|
|
|
|
|
return;
|
bpf: restore the ebpf program ID for BPF_AUDIT_UNLOAD and PERF_BPF_EVENT_PROG_UNLOAD
When changing the ebpf program put() routines to support being called
from within IRQ context the program ID was reset to zero prior to
calling the perf event and audit UNLOAD record generators, which
resulted in problems as the ebpf program ID was bogus (always zero).
This patch addresses this problem by removing an unnecessary call to
bpf_prog_free_id() in __bpf_prog_offload_destroy() and adjusting
__bpf_prog_put() to only call bpf_prog_free_id() after audit and perf
have finished their bpf program unload tasks in
bpf_prog_put_deferred(). For the record, no one can determine, or
remember, why it was necessary to free the program ID, and remove it
from the IDR, prior to executing bpf_prog_put_deferred();
regardless, both Stanislav and Alexei agree that the approach in this
patch should be safe.
It is worth noting that when moving the bpf_prog_free_id() call, the
do_idr_lock parameter was forced to true as the ebpf devs determined
this was the correct as the do_idr_lock should always be true. The
do_idr_lock parameter will be removed in a follow-up patch, but it
was kept here to keep the patch small in an effort to ease any stable
backports.
I also modified the bpf_audit_prog() logic used to associate the
AUDIT_BPF record with other associated records, e.g. @ctx != NULL.
Instead of keying off the operation, it now keys off the execution
context, e.g. '!in_irg && !irqs_disabled()', which is much more
appropriate and should help better connect the UNLOAD operations with
the associated audit state (other audit records).
Cc: stable@vger.kernel.org
Fixes: d809e134be7a ("bpf: Prepare bpf_prog_put() to be called from irq context.")
Reported-by: Burn Alting <burn.alting@iinet.net.au>
Reported-by: Jiri Olsa <olsajiri@gmail.com>
Suggested-by: Stanislav Fomichev <sdf@google.com>
Suggested-by: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Link: https://lore.kernel.org/r/20230106154400.74211-1-paul@paul-moore.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-01-06 15:43:59 +00:00
|
|
|
if (!in_irq() && !irqs_disabled())
|
2019-12-06 21:49:34 +00:00
|
|
|
ctx = audit_context();
|
|
|
|
|
ab = audit_log_start(ctx, GFP_ATOMIC, AUDIT_BPF);
|
|
|
|
|
if (unlikely(!ab))
|
|
|
|
|
return;
|
|
|
|
|
audit_log_format(ab, "prog-id=%u op=%s",
|
|
|
|
|
prog->aux->id, bpf_audit_str[op]);
|
|
|
|
|
audit_log_end(ab);
|
|
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:46 +00:00
|
|
|
static int bpf_prog_alloc_id(struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
int id;
|
|
|
|
|
|
2018-03-27 18:53:21 +00:00
|
|
|
idr_preload(GFP_KERNEL);
|
2017-06-05 19:15:46 +00:00
|
|
|
spin_lock_bh(&prog_idr_lock);
|
|
|
|
|
id = idr_alloc_cyclic(&prog_idr, prog, 1, INT_MAX, GFP_ATOMIC);
|
|
|
|
|
if (id > 0)
|
|
|
|
|
prog->aux->id = id;
|
|
|
|
|
spin_unlock_bh(&prog_idr_lock);
|
2018-03-27 18:53:21 +00:00
|
|
|
idr_preload_end();
|
2017-06-05 19:15:46 +00:00
|
|
|
|
|
|
|
|
/* id is in [1, INT_MAX) */
|
|
|
|
|
if (WARN_ON_ONCE(!id))
|
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
|
|
return id > 0 ? 0 : id;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-06 15:44:00 +00:00
|
|
|
void bpf_prog_free_id(struct bpf_prog *prog)
|
2017-06-05 19:15:46 +00:00
|
|
|
{
|
2021-07-15 00:54:07 +00:00
|
|
|
unsigned long flags;
|
|
|
|
|
|
2017-12-28 02:39:07 +00:00
|
|
|
/* cBPF to eBPF migrations are currently not in the idr store.
|
|
|
|
|
* Offloaded programs are removed from the store when their device
|
|
|
|
|
* disappears - even if someone grabs an fd to them they are unusable,
|
|
|
|
|
* simply waiting for refcnt to drop to be freed.
|
|
|
|
|
*/
|
2017-06-05 19:15:46 +00:00
|
|
|
if (!prog->aux->id)
|
|
|
|
|
return;
|
|
|
|
|
|
2023-01-06 15:44:00 +00:00
|
|
|
spin_lock_irqsave(&prog_idr_lock, flags);
|
2017-06-05 19:15:46 +00:00
|
|
|
idr_remove(&prog_idr, prog->aux->id);
|
2017-12-28 02:39:07 +00:00
|
|
|
prog->aux->id = 0;
|
2023-01-06 15:44:00 +00:00
|
|
|
spin_unlock_irqrestore(&prog_idr_lock, flags);
|
2017-06-05 19:15:46 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: generally move prog destruction to RCU deferral
Jann Horn reported following analysis that could potentially result
in a very hard to trigger (if not impossible) UAF race, to quote his
event timeline:
- Set up a process with threads T1, T2 and T3
- Let T1 set up a socket filter F1 that invokes another filter F2
through a BPF map [tail call]
- Let T1 trigger the socket filter via a unix domain socket write,
don't wait for completion
- Let T2 call PERF_EVENT_IOC_SET_BPF with F2, don't wait for completion
- Now T2 should be behind bpf_prog_get(), but before bpf_prog_put()
- Let T3 close the file descriptor for F2, dropping the reference
count of F2 to 2
- At this point, T1 should have looked up F2 from the map, but not
finished executing it
- Let T3 remove F2 from the BPF map, dropping the reference count of
F2 to 1
- Now T2 should call bpf_prog_put() (wrong BPF program type), dropping
the reference count of F2 to 0 and scheduling bpf_prog_free_deferred()
via schedule_work()
- At this point, the BPF program could be freed
- BPF execution is still running in a freed BPF program
While at PERF_EVENT_IOC_SET_BPF time it's only guaranteed that the perf
event fd we're doing the syscall on doesn't disappear from underneath us
for whole syscall time, it may not be the case for the bpf fd used as
an argument only after we did the put. It needs to be a valid fd pointing
to a BPF program at the time of the call to make the bpf_prog_get() and
while T2 gets preempted, F2 must have dropped reference to 1 on the other
CPU. The fput() from the close() in T3 should also add additionally delay
to the reference drop via exit_task_work() when bpf_prog_release() gets
called as well as scheduling bpf_prog_free_deferred().
That said, it makes nevertheless sense to move the BPF prog destruction
generally after RCU grace period to guarantee that such scenario above,
but also others as recently fixed in ceb56070359b ("bpf, perf: delay release
of BPF prog after grace period") with regards to tail calls won't happen.
Integrating bpf_prog_free_deferred() directly into the RCU callback is
not allowed since the invocation might happen from either softirq or
process context, so we're not permitted to block. Reviewing all bpf_prog_put()
invocations from eBPF side (note, cBPF -> eBPF progs don't use this for
their destruction) with call_rcu() look good to me.
Since we don't know whether at the time of attaching the program, we're
already part of a tail call map, we need to use RCU variant. However, due
to this, there won't be severely more stress on the RCU callback queue:
situations with above bpf_prog_get() and bpf_prog_put() combo in practice
normally won't lead to releases, but even if they would, enough effort/
cycles have to be put into loading a BPF program into the kernel already.
Reported-by: Jann Horn <jannh@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-06-30 15:24:43 +00:00
|
|
|
static void __bpf_prog_put_rcu(struct rcu_head *rcu)
|
2015-05-29 02:26:02 +00:00
|
|
|
{
|
|
|
|
|
struct bpf_prog_aux *aux = container_of(rcu, struct bpf_prog_aux, rcu);
|
|
|
|
|
|
2019-10-22 21:30:38 +00:00
|
|
|
kvfree(aux->func_info);
|
2019-11-14 18:57:16 +00:00
|
|
|
kfree(aux->func_info_aux);
|
2020-12-01 21:58:59 +00:00
|
|
|
free_uid(aux->user);
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
security_bpf_prog_free(aux->prog);
|
2015-05-29 02:26:02 +00:00
|
|
|
bpf_prog_free(aux->prog);
|
|
|
|
|
}
|
|
|
|
|
|
2019-10-22 13:57:23 +00:00
|
|
|
static void __bpf_prog_put_noref(struct bpf_prog *prog, bool deferred)
|
|
|
|
|
{
|
|
|
|
|
bpf_prog_kallsyms_del_all(prog);
|
|
|
|
|
btf_put(prog->aux->btf);
|
2023-03-10 07:40:59 +00:00
|
|
|
module_put(prog->aux->mod);
|
2021-03-25 01:51:30 +00:00
|
|
|
kvfree(prog->aux->jited_linfo);
|
|
|
|
|
kvfree(prog->aux->linfo);
|
bpf: Support bpf program calling kernel function
This patch adds support to BPF verifier to allow bpf program calling
kernel function directly.
The use case included in this set is to allow bpf-tcp-cc to directly
call some tcp-cc helper functions (e.g. "tcp_cong_avoid_ai()"). Those
functions have already been used by some kernel tcp-cc implementations.
This set will also allow the bpf-tcp-cc program to directly call the
kernel tcp-cc implementation, For example, a bpf_dctcp may only want to
implement its own dctcp_cwnd_event() and reuse other dctcp_*() directly
from the kernel tcp_dctcp.c instead of reimplementing (or
copy-and-pasting) them.
The tcp-cc kernel functions mentioned above will be white listed
for the struct_ops bpf-tcp-cc programs to use in a later patch.
The white listed functions are not bounded to a fixed ABI contract.
Those functions have already been used by the existing kernel tcp-cc.
If any of them has changed, both in-tree and out-of-tree kernel tcp-cc
implementations have to be changed. The same goes for the struct_ops
bpf-tcp-cc programs which have to be adjusted accordingly.
This patch is to make the required changes in the bpf verifier.
First change is in btf.c, it adds a case in "btf_check_func_arg_match()".
When the passed in "btf->kernel_btf == true", it means matching the
verifier regs' states with a kernel function. This will handle the
PTR_TO_BTF_ID reg. It also maps PTR_TO_SOCK_COMMON, PTR_TO_SOCKET,
and PTR_TO_TCP_SOCK to its kernel's btf_id.
In the later libbpf patch, the insn calling a kernel function will
look like:
insn->code == (BPF_JMP | BPF_CALL)
insn->src_reg == BPF_PSEUDO_KFUNC_CALL /* <- new in this patch */
insn->imm == func_btf_id /* btf_id of the running kernel */
[ For the future calling function-in-kernel-module support, an array
of module btf_fds can be passed at the load time and insn->off
can be used to index into this array. ]
At the early stage of verifier, the verifier will collect all kernel
function calls into "struct bpf_kfunc_desc". Those
descriptors are stored in "prog->aux->kfunc_tab" and will
be available to the JIT. Since this "add" operation is similar
to the current "add_subprog()" and looking for the same insn->code,
they are done together in the new "add_subprog_and_kfunc()".
In the "do_check()" stage, the new "check_kfunc_call()" is added
to verify the kernel function call instruction:
1. Ensure the kernel function can be used by a particular BPF_PROG_TYPE.
A new bpf_verifier_ops "check_kfunc_call" is added to do that.
The bpf-tcp-cc struct_ops program will implement this function in
a later patch.
2. Call "btf_check_kfunc_args_match()" to ensure the regs can be
used as the args of a kernel function.
3. Mark the regs' type, subreg_def, and zext_dst.
At the later do_misc_fixups() stage, the new fixup_kfunc_call()
will replace the insn->imm with the function address (relative
to __bpf_call_base). If needed, the jit can find the btf_func_model
by calling the new bpf_jit_find_kfunc_model(prog, insn).
With the imm set to the function address, "bpftool prog dump xlated"
will be able to display the kernel function calls the same way as
it displays other bpf helper calls.
gpl_compatible program is required to call kernel function.
This feature currently requires JIT.
The verifier selftests are adjusted because of the changes in
the verbose log in add_subprog_and_kfunc().
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210325015142.1544736-1-kafai@fb.com
2021-03-25 01:51:42 +00:00
|
|
|
kfree(prog->aux->kfunc_tab);
|
bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier
Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.
This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.
Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.
Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org
2020-12-03 20:46:29 +00:00
|
|
|
if (prog->aux->attach_btf)
|
|
|
|
|
btf_put(prog->aux->attach_btf);
|
2019-10-22 13:57:23 +00:00
|
|
|
|
2020-08-27 22:01:11 +00:00
|
|
|
if (deferred) {
|
2024-03-09 00:47:39 +00:00
|
|
|
if (prog->sleepable)
|
2020-08-27 22:01:11 +00:00
|
|
|
call_rcu_tasks_trace(&prog->aux->rcu, __bpf_prog_put_rcu);
|
|
|
|
|
else
|
|
|
|
|
call_rcu(&prog->aux->rcu, __bpf_prog_put_rcu);
|
|
|
|
|
} else {
|
2019-10-22 13:57:23 +00:00
|
|
|
__bpf_prog_put_rcu(&prog->aux->rcu);
|
2020-08-27 22:01:11 +00:00
|
|
|
}
|
2019-10-22 13:57:23 +00:00
|
|
|
}
|
|
|
|
|
|
2021-07-15 00:54:07 +00:00
|
|
|
static void bpf_prog_put_deferred(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog_aux *aux;
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
|
|
|
|
|
aux = container_of(work, struct bpf_prog_aux, work);
|
|
|
|
|
prog = aux->prog;
|
|
|
|
|
perf_event_bpf_event(prog, PERF_BPF_EVENT_PROG_UNLOAD, 0);
|
|
|
|
|
bpf_audit_prog(prog, BPF_AUDIT_UNLOAD);
|
2023-01-06 15:44:00 +00:00
|
|
|
bpf_prog_free_id(prog);
|
2021-07-15 00:54:07 +00:00
|
|
|
__bpf_prog_put_noref(prog, true);
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-06 15:44:00 +00:00
|
|
|
static void __bpf_prog_put(struct bpf_prog *prog)
|
2014-09-26 07:17:00 +00:00
|
|
|
{
|
2021-07-15 00:54:07 +00:00
|
|
|
struct bpf_prog_aux *aux = prog->aux;
|
|
|
|
|
|
|
|
|
|
if (atomic64_dec_and_test(&aux->refcnt)) {
|
|
|
|
|
if (in_irq() || irqs_disabled()) {
|
|
|
|
|
INIT_WORK(&aux->work, bpf_prog_put_deferred);
|
|
|
|
|
schedule_work(&aux->work);
|
|
|
|
|
} else {
|
|
|
|
|
bpf_prog_put_deferred(&aux->work);
|
|
|
|
|
}
|
bpf: add initial bpf tracepoints
This work adds a number of tracepoints to paths that are either
considered slow-path or exception-like states, where monitoring or
inspecting them would be desirable.
For bpf(2) syscall, tracepoints have been placed for main commands
when they succeed. In XDP case, tracepoint is for exceptions, that
is, f.e. on abnormal BPF program exit such as unknown or XDP_ABORTED
return code, or when error occurs during XDP_TX action and the packet
could not be forwarded.
Both have been split into separate event headers, and can be further
extended. Worst case, if they unexpectedly should get into our way in
future, they can also removed [1]. Of course, these tracepoints (like
any other) can be analyzed by eBPF itself, etc. Example output:
# ./perf record -a -e bpf:* sleep 10
# ./perf script
sock_example 6197 [005] 283.980322: bpf:bpf_map_create: map type=ARRAY ufd=4 key=4 val=8 max=256 flags=0
sock_example 6197 [005] 283.980721: bpf:bpf_prog_load: prog=a5ea8fa30ea6849c type=SOCKET_FILTER ufd=5
sock_example 6197 [005] 283.988423: bpf:bpf_prog_get_type: prog=a5ea8fa30ea6849c type=SOCKET_FILTER
sock_example 6197 [005] 283.988443: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[06 00 00 00] val=[00 00 00 00 00 00 00 00]
[...]
sock_example 6197 [005] 288.990868: bpf:bpf_map_lookup_elem: map type=ARRAY ufd=4 key=[01 00 00 00] val=[14 00 00 00 00 00 00 00]
swapper 0 [005] 289.338243: bpf:bpf_prog_put_rcu: prog=a5ea8fa30ea6849c type=SOCKET_FILTER
[1] https://lwn.net/Articles/705270/
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-01-25 01:28:18 +00:00
|
|
|
}
|
2014-09-26 07:17:00 +00:00
|
|
|
}
|
2017-06-05 19:15:49 +00:00
|
|
|
|
|
|
|
|
void bpf_prog_put(struct bpf_prog *prog)
|
|
|
|
|
{
|
2023-01-06 15:44:00 +00:00
|
|
|
__bpf_prog_put(prog);
|
2017-06-05 19:15:49 +00:00
|
|
|
}
|
cls_bpf: add initial eBPF support for programmable classifiers
This work extends the "classic" BPF programmable tc classifier by
extending its scope also to native eBPF code!
This allows for user space to implement own custom, 'safe' C like
classifiers (or whatever other frontend language LLVM et al may
provide in future), that can then be compiled with the LLVM eBPF
backend to an eBPF elf file. The result of this can be loaded into
the kernel via iproute2's tc. In the kernel, they can be JITed on
major archs and thus run in native performance.
Simple, minimal toy example to demonstrate the workflow:
#include <linux/ip.h>
#include <linux/if_ether.h>
#include <linux/bpf.h>
#include "tc_bpf_api.h"
__section("classify")
int cls_main(struct sk_buff *skb)
{
return (0x800 << 16) | load_byte(skb, ETH_HLEN + __builtin_offsetof(struct iphdr, tos));
}
char __license[] __section("license") = "GPL";
The classifier can then be compiled into eBPF opcodes and loaded
via tc, for example:
clang -O2 -emit-llvm -c cls.c -o - | llc -march=bpf -filetype=obj -o cls.o
tc filter add dev em1 parent 1: bpf cls.o [...]
As it has been demonstrated, the scope can even reach up to a fully
fledged flow dissector (similarly as in samples/bpf/sockex2_kern.c).
For tc, maps are allowed to be used, but from kernel context only,
in other words, eBPF code can keep state across filter invocations.
In future, we perhaps may reattach from a different application to
those maps e.g., to read out collected statistics/state.
Similarly as in socket filters, we may extend functionality for eBPF
classifiers over time depending on the use cases. For that purpose,
cls_bpf programs are using BPF_PROG_TYPE_SCHED_CLS program type, so
we can allow additional functions/accessors (e.g. an ABI compatible
offset translation to skb fields/metadata). For an initial cls_bpf
support, we allow the same set of helper functions as eBPF socket
filters, but we could diverge at some point in time w/o problem.
I was wondering whether cls_bpf and act_bpf could share C programs,
I can imagine that at some point, we introduce i) further common
handlers for both (or even beyond their scope), and/or if truly needed
ii) some restricted function space for each of them. Both can be
abstracted easily through struct bpf_verifier_ops in future.
The context of cls_bpf versus act_bpf is slightly different though:
a cls_bpf program will return a specific classid whereas act_bpf a
drop/non-drop return code, latter may also in future mangle skbs.
That said, we can surely have a "classify" and "action" section in
a single object file, or considered mentioned constraint add a
possibility of a shared section.
The workflow for getting native eBPF running from tc [1] is as
follows: for f_bpf, I've added a slightly modified ELF parser code
from Alexei's kernel sample, which reads out the LLVM compiled
object, sets up maps (and dynamically fixes up map fds) if any, and
loads the eBPF instructions all centrally through the bpf syscall.
The resulting fd from the loaded program itself is being passed down
to cls_bpf, which looks up struct bpf_prog from the fd store, and
holds reference, so that it stays available also after tc program
lifetime. On tc filter destruction, it will then drop its reference.
Moreover, I've also added the optional possibility to annotate an
eBPF filter with a name (e.g. path to object file, or something
else if preferred) so that when tc dumps currently installed filters,
some more context can be given to an admin for a given instance (as
opposed to just the file descriptor number).
Last but not least, bpf_prog_get() and bpf_prog_put() needed to be
exported, so that eBPF can be used from cls_bpf built as a module.
Thanks to 60a3b2253c41 ("net: bpf: make eBPF interpreter images
read-only") I think this is of no concern since anything wanting to
alter eBPF opcode after verification stage would crash the kernel.
[1] http://git.breakpoint.cc/cgit/dborkman/iproute2.git/log/?h=ebpf
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-03-01 11:31:48 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_put);
|
2014-09-26 07:17:00 +00:00
|
|
|
|
|
|
|
|
static int bpf_prog_release(struct inode *inode, struct file *filp)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog = filp->private_data;
|
|
|
|
|
|
bpf: generally move prog destruction to RCU deferral
Jann Horn reported following analysis that could potentially result
in a very hard to trigger (if not impossible) UAF race, to quote his
event timeline:
- Set up a process with threads T1, T2 and T3
- Let T1 set up a socket filter F1 that invokes another filter F2
through a BPF map [tail call]
- Let T1 trigger the socket filter via a unix domain socket write,
don't wait for completion
- Let T2 call PERF_EVENT_IOC_SET_BPF with F2, don't wait for completion
- Now T2 should be behind bpf_prog_get(), but before bpf_prog_put()
- Let T3 close the file descriptor for F2, dropping the reference
count of F2 to 2
- At this point, T1 should have looked up F2 from the map, but not
finished executing it
- Let T3 remove F2 from the BPF map, dropping the reference count of
F2 to 1
- Now T2 should call bpf_prog_put() (wrong BPF program type), dropping
the reference count of F2 to 0 and scheduling bpf_prog_free_deferred()
via schedule_work()
- At this point, the BPF program could be freed
- BPF execution is still running in a freed BPF program
While at PERF_EVENT_IOC_SET_BPF time it's only guaranteed that the perf
event fd we're doing the syscall on doesn't disappear from underneath us
for whole syscall time, it may not be the case for the bpf fd used as
an argument only after we did the put. It needs to be a valid fd pointing
to a BPF program at the time of the call to make the bpf_prog_get() and
while T2 gets preempted, F2 must have dropped reference to 1 on the other
CPU. The fput() from the close() in T3 should also add additionally delay
to the reference drop via exit_task_work() when bpf_prog_release() gets
called as well as scheduling bpf_prog_free_deferred().
That said, it makes nevertheless sense to move the BPF prog destruction
generally after RCU grace period to guarantee that such scenario above,
but also others as recently fixed in ceb56070359b ("bpf, perf: delay release
of BPF prog after grace period") with regards to tail calls won't happen.
Integrating bpf_prog_free_deferred() directly into the RCU callback is
not allowed since the invocation might happen from either softirq or
process context, so we're not permitted to block. Reviewing all bpf_prog_put()
invocations from eBPF side (note, cBPF -> eBPF progs don't use this for
their destruction) with call_rcu() look good to me.
Since we don't know whether at the time of attaching the program, we're
already part of a tail call map, we need to use RCU variant. However, due
to this, there won't be severely more stress on the RCU callback queue:
situations with above bpf_prog_get() and bpf_prog_put() combo in practice
normally won't lead to releases, but even if they would, enough effort/
cycles have to be put into loading a BPF program into the kernel already.
Reported-by: Jann Horn <jannh@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-06-30 15:24:43 +00:00
|
|
|
bpf_prog_put(prog);
|
2014-09-26 07:17:00 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2021-10-26 21:41:33 +00:00
|
|
|
struct bpf_prog_kstats {
|
|
|
|
|
u64 nsecs;
|
|
|
|
|
u64 cnt;
|
|
|
|
|
u64 misses;
|
|
|
|
|
};
|
|
|
|
|
|
2022-09-16 07:19:14 +00:00
|
|
|
void notrace bpf_prog_inc_misses_counter(struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog_stats *stats;
|
|
|
|
|
unsigned int flags;
|
|
|
|
|
|
|
|
|
|
stats = this_cpu_ptr(prog->stats);
|
|
|
|
|
flags = u64_stats_update_begin_irqsave(&stats->syncp);
|
|
|
|
|
u64_stats_inc(&stats->misses);
|
|
|
|
|
u64_stats_update_end_irqrestore(&stats->syncp, flags);
|
|
|
|
|
}
|
|
|
|
|
|
2019-02-25 22:28:39 +00:00
|
|
|
static void bpf_prog_get_stats(const struct bpf_prog *prog,
|
2021-10-26 21:41:33 +00:00
|
|
|
struct bpf_prog_kstats *stats)
|
2019-02-25 22:28:39 +00:00
|
|
|
{
|
2021-02-10 03:36:31 +00:00
|
|
|
u64 nsecs = 0, cnt = 0, misses = 0;
|
2019-02-25 22:28:39 +00:00
|
|
|
int cpu;
|
|
|
|
|
|
|
|
|
|
for_each_possible_cpu(cpu) {
|
|
|
|
|
const struct bpf_prog_stats *st;
|
|
|
|
|
unsigned int start;
|
2021-02-10 03:36:31 +00:00
|
|
|
u64 tnsecs, tcnt, tmisses;
|
2019-02-25 22:28:39 +00:00
|
|
|
|
2021-02-10 03:36:26 +00:00
|
|
|
st = per_cpu_ptr(prog->stats, cpu);
|
2019-02-25 22:28:39 +00:00
|
|
|
do {
|
2022-10-26 12:31:10 +00:00
|
|
|
start = u64_stats_fetch_begin(&st->syncp);
|
2021-10-26 21:41:33 +00:00
|
|
|
tnsecs = u64_stats_read(&st->nsecs);
|
|
|
|
|
tcnt = u64_stats_read(&st->cnt);
|
|
|
|
|
tmisses = u64_stats_read(&st->misses);
|
2022-10-26 12:31:10 +00:00
|
|
|
} while (u64_stats_fetch_retry(&st->syncp, start));
|
2019-02-25 22:28:39 +00:00
|
|
|
nsecs += tnsecs;
|
|
|
|
|
cnt += tcnt;
|
2021-02-10 03:36:31 +00:00
|
|
|
misses += tmisses;
|
2019-02-25 22:28:39 +00:00
|
|
|
}
|
|
|
|
|
stats->nsecs = nsecs;
|
|
|
|
|
stats->cnt = cnt;
|
2021-02-10 03:36:31 +00:00
|
|
|
stats->misses = misses;
|
2019-02-25 22:28:39 +00:00
|
|
|
}
|
|
|
|
|
|
2016-12-04 22:19:41 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
static void bpf_prog_show_fdinfo(struct seq_file *m, struct file *filp)
|
|
|
|
|
{
|
|
|
|
|
const struct bpf_prog *prog = filp->private_data;
|
2017-01-13 22:38:15 +00:00
|
|
|
char prog_tag[sizeof(prog->tag) * 2 + 1] = { };
|
2021-10-26 21:41:33 +00:00
|
|
|
struct bpf_prog_kstats stats;
|
2016-12-04 22:19:41 +00:00
|
|
|
|
2019-02-25 22:28:39 +00:00
|
|
|
bpf_prog_get_stats(prog, &stats);
|
2017-01-13 22:38:15 +00:00
|
|
|
bin2hex(prog_tag, prog->tag, sizeof(prog->tag));
|
2016-12-04 22:19:41 +00:00
|
|
|
seq_printf(m,
|
|
|
|
|
"prog_type:\t%u\n"
|
|
|
|
|
"prog_jited:\t%u\n"
|
2017-01-13 22:38:15 +00:00
|
|
|
"prog_tag:\t%s\n"
|
2018-06-02 21:06:34 +00:00
|
|
|
"memlock:\t%llu\n"
|
2019-02-25 22:28:39 +00:00
|
|
|
"prog_id:\t%u\n"
|
|
|
|
|
"run_time_ns:\t%llu\n"
|
2021-02-10 03:36:31 +00:00
|
|
|
"run_cnt:\t%llu\n"
|
2021-10-20 07:48:17 +00:00
|
|
|
"recursion_misses:\t%llu\n"
|
|
|
|
|
"verified_insns:\t%u\n",
|
2016-12-04 22:19:41 +00:00
|
|
|
prog->type,
|
|
|
|
|
prog->jited,
|
2017-01-13 22:38:15 +00:00
|
|
|
prog_tag,
|
2018-06-02 21:06:34 +00:00
|
|
|
prog->pages * 1ULL << PAGE_SHIFT,
|
2019-02-25 22:28:39 +00:00
|
|
|
prog->aux->id,
|
|
|
|
|
stats.nsecs,
|
2021-02-10 03:36:31 +00:00
|
|
|
stats.cnt,
|
2021-10-20 07:48:17 +00:00
|
|
|
stats.misses,
|
|
|
|
|
prog->aux->verified_insns);
|
2016-12-04 22:19:41 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
2017-10-18 20:00:26 +00:00
|
|
|
const struct file_operations bpf_prog_fops = {
|
2016-12-04 22:19:41 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
.show_fdinfo = bpf_prog_show_fdinfo,
|
|
|
|
|
#endif
|
|
|
|
|
.release = bpf_prog_release,
|
2017-10-18 20:00:22 +00:00
|
|
|
.read = bpf_dummy_read,
|
|
|
|
|
.write = bpf_dummy_write,
|
2014-09-26 07:17:00 +00:00
|
|
|
};
|
|
|
|
|
|
2015-10-29 13:58:09 +00:00
|
|
|
int bpf_prog_new_fd(struct bpf_prog *prog)
|
2015-10-29 13:58:06 +00:00
|
|
|
{
|
2017-10-18 20:00:24 +00:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
ret = security_bpf_prog(prog);
|
|
|
|
|
if (ret < 0)
|
|
|
|
|
return ret;
|
|
|
|
|
|
2015-10-29 13:58:06 +00:00
|
|
|
return anon_inode_getfd("bpf-prog", &bpf_prog_fops, prog,
|
|
|
|
|
O_RDWR | O_CLOEXEC);
|
|
|
|
|
}
|
|
|
|
|
|
2016-06-30 15:24:44 +00:00
|
|
|
static struct bpf_prog *____bpf_prog_get(struct fd f)
|
2014-09-26 07:17:00 +00:00
|
|
|
{
|
|
|
|
|
if (!f.file)
|
|
|
|
|
return ERR_PTR(-EBADF);
|
|
|
|
|
if (f.file->f_op != &bpf_prog_fops) {
|
|
|
|
|
fdput(f);
|
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
}
|
|
|
|
|
|
2015-10-29 13:58:07 +00:00
|
|
|
return f.file->private_data;
|
2014-09-26 07:17:00 +00:00
|
|
|
}
|
|
|
|
|
|
2019-11-17 17:28:03 +00:00
|
|
|
void bpf_prog_add(struct bpf_prog *prog, int i)
|
2016-04-28 01:56:20 +00:00
|
|
|
{
|
2019-11-17 17:28:03 +00:00
|
|
|
atomic64_add(i, &prog->aux->refcnt);
|
2016-04-28 01:56:20 +00:00
|
|
|
}
|
2016-07-19 19:16:46 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_add);
|
|
|
|
|
|
2016-11-09 21:02:34 +00:00
|
|
|
void bpf_prog_sub(struct bpf_prog *prog, int i)
|
|
|
|
|
{
|
|
|
|
|
/* Only to be used for undoing previous bpf_prog_add() in some
|
|
|
|
|
* error path. We still know that another entity in our call
|
|
|
|
|
* path holds a reference to the program, thus atomic_sub() can
|
|
|
|
|
* be safely used in such cases!
|
|
|
|
|
*/
|
2019-11-17 17:28:03 +00:00
|
|
|
WARN_ON(atomic64_sub_return(i, &prog->aux->refcnt) == 0);
|
2016-11-09 21:02:34 +00:00
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_sub);
|
|
|
|
|
|
2019-11-17 17:28:03 +00:00
|
|
|
void bpf_prog_inc(struct bpf_prog *prog)
|
2016-07-19 19:16:46 +00:00
|
|
|
{
|
2019-11-17 17:28:03 +00:00
|
|
|
atomic64_inc(&prog->aux->refcnt);
|
2016-07-19 19:16:46 +00:00
|
|
|
}
|
2016-11-19 00:45:00 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_inc);
|
2016-04-28 01:56:20 +00:00
|
|
|
|
2017-06-05 19:15:49 +00:00
|
|
|
/* prog_idr_lock should have been held */
|
2017-08-16 05:32:22 +00:00
|
|
|
struct bpf_prog *bpf_prog_inc_not_zero(struct bpf_prog *prog)
|
2017-06-05 19:15:49 +00:00
|
|
|
{
|
|
|
|
|
int refold;
|
|
|
|
|
|
2019-11-17 17:28:03 +00:00
|
|
|
refold = atomic64_fetch_add_unless(&prog->aux->refcnt, 1, 0);
|
2017-06-05 19:15:49 +00:00
|
|
|
|
|
|
|
|
if (!refold)
|
|
|
|
|
return ERR_PTR(-ENOENT);
|
|
|
|
|
|
|
|
|
|
return prog;
|
|
|
|
|
}
|
2017-08-16 05:32:22 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_inc_not_zero);
|
2017-06-05 19:15:49 +00:00
|
|
|
|
2017-12-03 01:20:38 +00:00
|
|
|
bool bpf_prog_get_ok(struct bpf_prog *prog,
|
2017-11-20 23:21:54 +00:00
|
|
|
enum bpf_prog_type *attach_type, bool attach_drv)
|
2017-11-03 20:56:20 +00:00
|
|
|
{
|
2017-11-20 23:21:54 +00:00
|
|
|
/* not an attachment, just a refcount inc, always allow */
|
|
|
|
|
if (!attach_type)
|
|
|
|
|
return true;
|
2017-11-03 20:56:20 +00:00
|
|
|
|
|
|
|
|
if (prog->type != *attach_type)
|
|
|
|
|
return false;
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_prog_is_offloaded(prog->aux) && !attach_drv)
|
2017-11-03 20:56:20 +00:00
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct bpf_prog *__bpf_prog_get(u32 ufd, enum bpf_prog_type *attach_type,
|
2017-11-20 23:21:54 +00:00
|
|
|
bool attach_drv)
|
2014-09-26 07:17:00 +00:00
|
|
|
{
|
|
|
|
|
struct fd f = fdget(ufd);
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
|
2016-06-30 15:24:44 +00:00
|
|
|
prog = ____bpf_prog_get(f);
|
2014-09-26 07:17:00 +00:00
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return prog;
|
2017-11-20 23:21:54 +00:00
|
|
|
if (!bpf_prog_get_ok(prog, attach_type, attach_drv)) {
|
2016-06-30 15:24:44 +00:00
|
|
|
prog = ERR_PTR(-EINVAL);
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2019-11-17 17:28:03 +00:00
|
|
|
bpf_prog_inc(prog);
|
2016-06-30 15:24:44 +00:00
|
|
|
out:
|
2014-09-26 07:17:00 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return prog;
|
|
|
|
|
}
|
2016-06-30 15:24:44 +00:00
|
|
|
|
|
|
|
|
struct bpf_prog *bpf_prog_get(u32 ufd)
|
|
|
|
|
{
|
2017-11-20 23:21:54 +00:00
|
|
|
return __bpf_prog_get(ufd, NULL, false);
|
2016-06-30 15:24:44 +00:00
|
|
|
}
|
|
|
|
|
|
2017-11-03 20:56:20 +00:00
|
|
|
struct bpf_prog *bpf_prog_get_type_dev(u32 ufd, enum bpf_prog_type type,
|
2017-11-20 23:21:54 +00:00
|
|
|
bool attach_drv)
|
2017-11-03 20:56:20 +00:00
|
|
|
{
|
2018-04-29 02:56:37 +00:00
|
|
|
return __bpf_prog_get(ufd, &type, attach_drv);
|
2017-11-03 20:56:20 +00:00
|
|
|
}
|
2017-11-03 20:56:21 +00:00
|
|
|
EXPORT_SYMBOL_GPL(bpf_prog_get_type_dev);
|
2017-11-03 20:56:20 +00:00
|
|
|
|
2018-03-30 22:08:07 +00:00
|
|
|
/* Initially all BPF programs could be loaded w/o specifying
|
|
|
|
|
* expected_attach_type. Later for some of them specifying expected_attach_type
|
|
|
|
|
* at load time became required so that program could be validated properly.
|
|
|
|
|
* Programs of types that are allowed to be loaded both w/ and w/o (for
|
|
|
|
|
* backward compatibility) expected_attach_type, should have the default attach
|
|
|
|
|
* type assigned to expected_attach_type for the latter case, so that it can be
|
|
|
|
|
* validated later at attach time.
|
|
|
|
|
*
|
|
|
|
|
* bpf_prog_load_fixup_attach_type() sets expected_attach_type in @attr if
|
|
|
|
|
* prog type requires it but has some attach types that have to be backward
|
|
|
|
|
* compatible.
|
|
|
|
|
*/
|
|
|
|
|
static void bpf_prog_load_fixup_attach_type(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
switch (attr->prog_type) {
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
/* Unfortunately BPF_ATTACH_TYPE_UNSPEC enumeration doesn't
|
|
|
|
|
* exist so checking for non-zero is the way to go here.
|
|
|
|
|
*/
|
|
|
|
|
if (!attr->expected_attach_type)
|
|
|
|
|
attr->expected_attach_type =
|
|
|
|
|
BPF_CGROUP_INET_SOCK_CREATE;
|
|
|
|
|
break;
|
bpf: Support socket migration by eBPF.
This patch introduces a new bpf_attach_type for BPF_PROG_TYPE_SK_REUSEPORT
to check if the attached eBPF program is capable of migrating sockets. When
the eBPF program is attached, we run it for socket migration if the
expected_attach_type is BPF_SK_REUSEPORT_SELECT_OR_MIGRATE or
net.ipv4.tcp_migrate_req is enabled.
Currently, the expected_attach_type is not enforced for the
BPF_PROG_TYPE_SK_REUSEPORT type of program. Thus, this commit follows the
earlier idea in the commit aac3fc320d94 ("bpf: Post-hooks for sys_bind") to
fix up the zero expected_attach_type in bpf_prog_load_fixup_attach_type().
Moreover, this patch adds a new field (migrating_sk) to sk_reuseport_md to
select a new listener based on the child socket. migrating_sk varies
depending on if it is migrating a request in the accept queue or during
3WHS.
- accept_queue : sock (ESTABLISHED/SYN_RECV)
- 3WHS : request_sock (NEW_SYN_RECV)
In the eBPF program, we can select a new listener by
BPF_FUNC_sk_select_reuseport(). Also, we can cancel migration by returning
SK_DROP. This feature is useful when listeners have different settings at
the socket API level or when we want to free resources as soon as possible.
- SK_PASS with selected_sk, select it as a new listener
- SK_PASS with selected_sk NULL, fallbacks to the random selection
- SK_DROP, cancel the migration.
There is a noteworthy point. We select a listening socket in three places,
but we do not have struct skb at closing a listener or retransmitting a
SYN+ACK. On the other hand, some helper functions do not expect skb is NULL
(e.g. skb_header_pointer() in BPF_FUNC_skb_load_bytes(), skb_tail_pointer()
in BPF_FUNC_skb_load_bytes_relative()). So we allocate an empty skb
temporarily before running the eBPF program.
Suggested-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/netdev/20201123003828.xjpjdtk4ygl6tg6h@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/netdev/20201203042402.6cskdlit5f3mw4ru@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/netdev/20201209030903.hhow5r53l6fmozjn@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/bpf/20210612123224.12525-10-kuniyu@amazon.co.jp
2021-06-12 12:32:22 +00:00
|
|
|
case BPF_PROG_TYPE_SK_REUSEPORT:
|
|
|
|
|
if (!attr->expected_attach_type)
|
|
|
|
|
attr->expected_attach_type =
|
|
|
|
|
BPF_SK_REUSEPORT_SELECT;
|
|
|
|
|
break;
|
2018-03-30 22:08:07 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-03-30 22:08:00 +00:00
|
|
|
static int
|
2019-10-16 03:24:58 +00:00
|
|
|
bpf_prog_load_check_attach(enum bpf_prog_type prog_type,
|
|
|
|
|
enum bpf_attach_type expected_attach_type,
|
2020-12-03 20:46:30 +00:00
|
|
|
struct btf *attach_btf, u32 btf_id,
|
|
|
|
|
struct bpf_prog *dst_prog)
|
2018-03-30 22:08:00 +00:00
|
|
|
{
|
2020-01-09 00:35:03 +00:00
|
|
|
if (btf_id) {
|
2019-10-18 06:09:33 +00:00
|
|
|
if (btf_id > BTF_MAX_TYPE)
|
|
|
|
|
return -EINVAL;
|
2020-01-09 00:35:03 +00:00
|
|
|
|
2020-12-03 20:46:30 +00:00
|
|
|
if (!attach_btf && !dst_prog)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-01-09 00:35:03 +00:00
|
|
|
switch (prog_type) {
|
|
|
|
|
case BPF_PROG_TYPE_TRACING:
|
2020-03-29 00:43:52 +00:00
|
|
|
case BPF_PROG_TYPE_LSM:
|
2020-01-09 00:35:03 +00:00
|
|
|
case BPF_PROG_TYPE_STRUCT_OPS:
|
2020-01-21 00:53:46 +00:00
|
|
|
case BPF_PROG_TYPE_EXT:
|
2020-01-09 00:35:03 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
2019-10-18 06:09:33 +00:00
|
|
|
return -EINVAL;
|
2020-01-09 00:35:03 +00:00
|
|
|
}
|
2019-10-18 06:09:33 +00:00
|
|
|
}
|
|
|
|
|
|
2020-12-03 20:46:30 +00:00
|
|
|
if (attach_btf && (!btf_id || dst_prog))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (dst_prog && prog_type != BPF_PROG_TYPE_TRACING &&
|
2020-01-21 00:53:46 +00:00
|
|
|
prog_type != BPF_PROG_TYPE_EXT)
|
2020-01-09 00:35:03 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
2018-03-30 22:08:02 +00:00
|
|
|
switch (prog_type) {
|
2018-03-30 22:08:07 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
switch (expected_attach_type) {
|
|
|
|
|
case BPF_CGROUP_INET_SOCK_CREATE:
|
2020-07-06 23:01:25 +00:00
|
|
|
case BPF_CGROUP_INET_SOCK_RELEASE:
|
2018-03-30 22:08:07 +00:00
|
|
|
case BPF_CGROUP_INET4_POST_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_POST_BIND:
|
|
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
2018-03-30 22:08:02 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
switch (expected_attach_type) {
|
|
|
|
|
case BPF_CGROUP_INET4_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_BIND:
|
2018-03-30 22:08:05 +00:00
|
|
|
case BPF_CGROUP_INET4_CONNECT:
|
|
|
|
|
case BPF_CGROUP_INET6_CONNECT:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_CONNECT:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETPEERNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETPEERNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETPEERNAME:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETSOCKNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETSOCKNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETSOCKNAME:
|
2018-05-25 15:55:23 +00:00
|
|
|
case BPF_CGROUP_UDP4_SENDMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_SENDMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_SENDMSG:
|
bpf: fix unconnected udp hooks
Intention of cgroup bind/connect/sendmsg BPF hooks is to act transparently
to applications as also stated in original motivation in 7828f20e3779 ("Merge
branch 'bpf-cgroup-bind-connect'"). When recently integrating the latter
two hooks into Cilium to enable host based load-balancing with Kubernetes,
I ran into the issue that pods couldn't start up as DNS got broken. Kubernetes
typically sets up DNS as a service and is thus subject to load-balancing.
Upon further debugging, it turns out that the cgroupv2 sendmsg BPF hooks API
is currently insufficient and thus not usable as-is for standard applications
shipped with most distros. To break down the issue we ran into with a simple
example:
# cat /etc/resolv.conf
nameserver 147.75.207.207
nameserver 147.75.207.208
For the purpose of a simple test, we set up above IPs as service IPs and
transparently redirect traffic to a different DNS backend server for that
node:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
The attached BPF program is basically selecting one of the backends if the
service IP/port matches on the cgroup hook. DNS breaks here, because the
hooks are not transparent enough to applications which have built-in msg_name
address checks:
# nslookup 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
;; connection timed out; no servers could be reached
# dig 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; connection timed out; no servers could be reached
For comparison, if none of the service IPs is used, and we tell nslookup
to use 8.8.8.8 directly it works just fine, of course:
# nslookup 1.1.1.1 8.8.8.8
1.1.1.1.in-addr.arpa name = one.one.one.one.
In order to fix this and thus act more transparent to the application,
this needs reverse translation on recvmsg() side. A minimal fix for this
API is to add similar recvmsg() hooks behind the BPF cgroups static key
such that the program can track state and replace the current sockaddr_in{,6}
with the original service IP. From BPF side, this basically tracks the
service tuple plus socket cookie in an LRU map where the reverse NAT can
then be retrieved via map value as one example. Side-note: the BPF cgroups
static key should be converted to a per-hook static key in future.
Same example after this fix:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
Lookups work fine now:
# nslookup 1.1.1.1
1.1.1.1.in-addr.arpa name = one.one.one.one.
Authoritative answers can be found from:
# dig 1.1.1.1
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 51550
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 512
;; QUESTION SECTION:
;1.1.1.1. IN A
;; AUTHORITY SECTION:
. 23426 IN SOA a.root-servers.net. nstld.verisign-grs.com. 2019052001 1800 900 604800 86400
;; Query time: 17 msec
;; SERVER: 147.75.207.207#53(147.75.207.207)
;; WHEN: Tue May 21 12:59:38 UTC 2019
;; MSG SIZE rcvd: 111
And from an actual packet level it shows that we're using the back end
server when talking via 147.75.207.20{7,8} front end:
# tcpdump -i any udp
[...]
12:59:52.698732 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.698735 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
[...]
In order to be flexible and to have same semantics as in sendmsg BPF
programs, we only allow return codes in [1,1] range. In the sendmsg case
the program is called if msg->msg_name is present which can be the case
in both, connected and unconnected UDP.
The former only relies on the sockaddr_in{,6} passed via connect(2) if
passed msg->msg_name was NULL. Therefore, on recvmsg side, we act in similar
way to call into the BPF program whenever a non-NULL msg->msg_name was
passed independent of sk->sk_state being TCP_ESTABLISHED or not. Note
that for TCP case, the msg->msg_name is ignored in the regular recvmsg
path and therefore not relevant.
For the case of ip{,v6}_recv_error() paths, picked up via MSG_ERRQUEUE,
the hook is not called. This is intentional as it aligns with the same
semantics as in case of TCP cgroup BPF hooks right now. This might be
better addressed in future through a different bpf_attach_type such
that this case can be distinguished from the regular recvmsg paths,
for example.
Fixes: 1cedee13d25a ("bpf: Hooks for sys_sendmsg")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrey Ignatov <rdna@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Martynas Pumputis <m@lambda.lt>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-06 23:48:57 +00:00
|
|
|
case BPF_CGROUP_UDP4_RECVMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_RECVMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_RECVMSG:
|
2018-03-30 22:08:02 +00:00
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
2019-05-28 23:59:36 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
|
|
|
|
switch (expected_attach_type) {
|
|
|
|
|
case BPF_CGROUP_INET_INGRESS:
|
|
|
|
|
case BPF_CGROUP_INET_EGRESS:
|
|
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
bpf: implement getsockopt and setsockopt hooks
Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and
BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks.
BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before
passing them down to the kernel or bypass kernel completely.
BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that
kernel returns.
Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure.
The buffer memory is pre-allocated (because I don't think there is
a precedent for working with __user memory from bpf). This might be
slow to do for each {s,g}etsockopt call, that's why I've added
__cgroup_bpf_prog_array_is_empty that exits early if there is nothing
attached to a cgroup. Note, however, that there is a race between
__cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup
program layout might have changed; this should not be a problem
because in general there is a race between multiple calls to
{s,g}etsocktop and user adding/removing bpf progs from a cgroup.
The return code of the BPF program is handled as follows:
* 0: EPERM
* 1: success, continue with next BPF program in the cgroup chain
v9:
* allow overwriting setsockopt arguments (Alexei Starovoitov):
* use set_fs (same as kernel_setsockopt)
* buffer is always kzalloc'd (no small on-stack buffer)
v8:
* use s32 for optlen (Andrii Nakryiko)
v7:
* return only 0 or 1 (Alexei Starovoitov)
* always run all progs (Alexei Starovoitov)
* use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov)
(decided to use optval=-1 instead, optval=0 might be a valid input)
* call getsockopt hook after kernel handlers (Alexei Starovoitov)
v6:
* rework cgroup chaining; stop as soon as bpf program returns
0 or 2; see patch with the documentation for the details
* drop Andrii's and Martin's Acked-by (not sure they are comfortable
with the new state of things)
v5:
* skip copy_to_user() and put_user() when ret == 0 (Martin Lau)
v4:
* don't export bpf_sk_fullsock helper (Martin Lau)
* size != sizeof(__u64) for uapi pointers (Martin Lau)
* offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau)
v3:
* typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko)
* reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii
Nakryiko)
* use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau)
* use BPF_FIELD_SIZEOF() for consistency (Martin Lau)
* new CG_SOCKOPT_ACCESS macro to wrap repeated parts
v2:
* moved bpf_sockopt_kern fields around to remove a hole (Martin Lau)
* aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau)
* bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau)
* added [0,2] return code check to verifier (Martin Lau)
* dropped unused buf[64] from the stack (Martin Lau)
* use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau)
* dropped bpf_target_off from ctx rewrites (Martin Lau)
* use return code for kernel bypass (Martin Lau & Andrii Nakryiko)
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Martin Lau <kafai@fb.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-27 20:38:47 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
switch (expected_attach_type) {
|
|
|
|
|
case BPF_CGROUP_SETSOCKOPT:
|
|
|
|
|
case BPF_CGROUP_GETSOCKOPT:
|
|
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 10:35:23 +00:00
|
|
|
case BPF_PROG_TYPE_SK_LOOKUP:
|
|
|
|
|
if (expected_attach_type == BPF_SK_LOOKUP)
|
|
|
|
|
return 0;
|
|
|
|
|
return -EINVAL;
|
bpf: Support socket migration by eBPF.
This patch introduces a new bpf_attach_type for BPF_PROG_TYPE_SK_REUSEPORT
to check if the attached eBPF program is capable of migrating sockets. When
the eBPF program is attached, we run it for socket migration if the
expected_attach_type is BPF_SK_REUSEPORT_SELECT_OR_MIGRATE or
net.ipv4.tcp_migrate_req is enabled.
Currently, the expected_attach_type is not enforced for the
BPF_PROG_TYPE_SK_REUSEPORT type of program. Thus, this commit follows the
earlier idea in the commit aac3fc320d94 ("bpf: Post-hooks for sys_bind") to
fix up the zero expected_attach_type in bpf_prog_load_fixup_attach_type().
Moreover, this patch adds a new field (migrating_sk) to sk_reuseport_md to
select a new listener based on the child socket. migrating_sk varies
depending on if it is migrating a request in the accept queue or during
3WHS.
- accept_queue : sock (ESTABLISHED/SYN_RECV)
- 3WHS : request_sock (NEW_SYN_RECV)
In the eBPF program, we can select a new listener by
BPF_FUNC_sk_select_reuseport(). Also, we can cancel migration by returning
SK_DROP. This feature is useful when listeners have different settings at
the socket API level or when we want to free resources as soon as possible.
- SK_PASS with selected_sk, select it as a new listener
- SK_PASS with selected_sk NULL, fallbacks to the random selection
- SK_DROP, cancel the migration.
There is a noteworthy point. We select a listening socket in three places,
but we do not have struct skb at closing a listener or retransmitting a
SYN+ACK. On the other hand, some helper functions do not expect skb is NULL
(e.g. skb_header_pointer() in BPF_FUNC_skb_load_bytes(), skb_tail_pointer()
in BPF_FUNC_skb_load_bytes_relative()). So we allocate an empty skb
temporarily before running the eBPF program.
Suggested-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/netdev/20201123003828.xjpjdtk4ygl6tg6h@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/netdev/20201203042402.6cskdlit5f3mw4ru@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/netdev/20201209030903.hhow5r53l6fmozjn@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/bpf/20210612123224.12525-10-kuniyu@amazon.co.jp
2021-06-12 12:32:22 +00:00
|
|
|
case BPF_PROG_TYPE_SK_REUSEPORT:
|
|
|
|
|
switch (expected_attach_type) {
|
|
|
|
|
case BPF_SK_REUSEPORT_SELECT:
|
|
|
|
|
case BPF_SK_REUSEPORT_SELECT_OR_MIGRATE:
|
|
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
2023-06-05 13:14:45 +00:00
|
|
|
case BPF_PROG_TYPE_NETFILTER:
|
|
|
|
|
if (expected_attach_type == BPF_NETFILTER)
|
|
|
|
|
return 0;
|
|
|
|
|
return -EINVAL;
|
2021-05-14 00:36:03 +00:00
|
|
|
case BPF_PROG_TYPE_SYSCALL:
|
2020-01-21 00:53:46 +00:00
|
|
|
case BPF_PROG_TYPE_EXT:
|
|
|
|
|
if (expected_attach_type)
|
|
|
|
|
return -EINVAL;
|
2020-08-23 22:36:59 +00:00
|
|
|
fallthrough;
|
2018-03-30 22:08:02 +00:00
|
|
|
default:
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2018-03-30 22:08:00 +00:00
|
|
|
}
|
|
|
|
|
|
2020-05-13 23:03:54 +00:00
|
|
|
static bool is_net_admin_prog_type(enum bpf_prog_type prog_type)
|
|
|
|
|
{
|
|
|
|
|
switch (prog_type) {
|
|
|
|
|
case BPF_PROG_TYPE_SCHED_CLS:
|
|
|
|
|
case BPF_PROG_TYPE_SCHED_ACT:
|
|
|
|
|
case BPF_PROG_TYPE_XDP:
|
|
|
|
|
case BPF_PROG_TYPE_LWT_IN:
|
|
|
|
|
case BPF_PROG_TYPE_LWT_OUT:
|
|
|
|
|
case BPF_PROG_TYPE_LWT_XMIT:
|
|
|
|
|
case BPF_PROG_TYPE_LWT_SEG6LOCAL:
|
|
|
|
|
case BPF_PROG_TYPE_SK_SKB:
|
|
|
|
|
case BPF_PROG_TYPE_SK_MSG:
|
|
|
|
|
case BPF_PROG_TYPE_FLOW_DISSECTOR:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_DEVICE:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SYSCTL:
|
|
|
|
|
case BPF_PROG_TYPE_SOCK_OPS:
|
|
|
|
|
case BPF_PROG_TYPE_EXT: /* extends any prog */
|
2023-04-21 17:02:54 +00:00
|
|
|
case BPF_PROG_TYPE_NETFILTER:
|
2020-05-13 23:03:54 +00:00
|
|
|
return true;
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
|
|
|
|
/* always unpriv */
|
|
|
|
|
case BPF_PROG_TYPE_SK_REUSEPORT:
|
|
|
|
|
/* equivalent to SOCKET_FILTER. need CAP_BPF only */
|
|
|
|
|
default:
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool is_perfmon_prog_type(enum bpf_prog_type prog_type)
|
|
|
|
|
{
|
|
|
|
|
switch (prog_type) {
|
|
|
|
|
case BPF_PROG_TYPE_KPROBE:
|
|
|
|
|
case BPF_PROG_TYPE_TRACEPOINT:
|
|
|
|
|
case BPF_PROG_TYPE_PERF_EVENT:
|
|
|
|
|
case BPF_PROG_TYPE_RAW_TRACEPOINT:
|
|
|
|
|
case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE:
|
|
|
|
|
case BPF_PROG_TYPE_TRACING:
|
|
|
|
|
case BPF_PROG_TYPE_LSM:
|
|
|
|
|
case BPF_PROG_TYPE_STRUCT_OPS: /* has access to struct sock */
|
|
|
|
|
case BPF_PROG_TYPE_EXT: /* extends any prog */
|
|
|
|
|
return true;
|
|
|
|
|
default:
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2014-09-26 07:17:00 +00:00
|
|
|
/* last field in 'union bpf_attr' used by this command */
|
2024-01-24 02:21:03 +00:00
|
|
|
#define BPF_PROG_LOAD_LAST_FIELD prog_token_fd
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2023-04-06 23:41:58 +00:00
|
|
|
static int bpf_prog_load(union bpf_attr *attr, bpfptr_t uattr, u32 uattr_size)
|
2014-09-26 07:17:00 +00:00
|
|
|
{
|
|
|
|
|
enum bpf_prog_type type = attr->prog_type;
|
2020-12-03 20:46:30 +00:00
|
|
|
struct bpf_prog *prog, *dst_prog = NULL;
|
|
|
|
|
struct btf *attach_btf = NULL;
|
2024-01-24 02:21:03 +00:00
|
|
|
struct bpf_token *token = NULL;
|
|
|
|
|
bool bpf_cap;
|
2014-09-26 07:17:00 +00:00
|
|
|
int err;
|
|
|
|
|
char license[128];
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_LOAD))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2019-05-24 22:25:16 +00:00
|
|
|
if (attr->prog_flags & ~(BPF_F_STRICT_ALIGNMENT |
|
|
|
|
|
BPF_F_ANY_ALIGNMENT |
|
2019-08-23 05:52:12 +00:00
|
|
|
BPF_F_TEST_STATE_FREQ |
|
2020-08-27 22:01:11 +00:00
|
|
|
BPF_F_SLEEPABLE |
|
2022-01-21 10:09:52 +00:00
|
|
|
BPF_F_TEST_RND_HI32 |
|
2023-01-19 22:15:24 +00:00
|
|
|
BPF_F_XDP_HAS_FRAGS |
|
bpf: add register bounds sanity checks and sanitization
Add simple sanity checks that validate well-formed ranges (min <= max)
across u64, s64, u32, and s32 ranges. Also for cases when the value is
constant (either 64-bit or 32-bit), we validate that ranges and tnums
are in agreement.
These bounds checks are performed at the end of BPF_ALU/BPF_ALU64
operations, on conditional jumps, and for LDX instructions (where subreg
zero/sign extension is probably the most important to check). This
covers most of the interesting cases.
Also, we validate the sanity of the return register when manually
adjusting it for some special helpers.
By default, sanity violation will trigger a warning in verifier log and
resetting register bounds to "unbounded" ones. But to aid development
and debugging, BPF_F_TEST_SANITY_STRICT flag is added, which will
trigger hard failure of verification with -EFAULT on register bounds
violations. This allows selftests to catch such issues. veristat will
also gain a CLI option to enable this behavior.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Link: https://lore.kernel.org/r/20231112010609.848406-5-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-11-12 01:06:00 +00:00
|
|
|
BPF_F_XDP_DEV_BOUND_ONLY |
|
2024-01-24 02:21:03 +00:00
|
|
|
BPF_F_TEST_REG_INVARIANTS |
|
|
|
|
|
BPF_F_TOKEN_FD))
|
2017-05-10 18:38:07 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
2024-01-24 02:21:03 +00:00
|
|
|
bpf_prog_load_fixup_attach_type(attr);
|
|
|
|
|
|
|
|
|
|
if (attr->prog_flags & BPF_F_TOKEN_FD) {
|
|
|
|
|
token = bpf_token_get_from_fd(attr->prog_token_fd);
|
|
|
|
|
if (IS_ERR(token))
|
|
|
|
|
return PTR_ERR(token);
|
|
|
|
|
/* if current token doesn't grant prog loading permissions,
|
|
|
|
|
* then we can't use this token, so ignore it and rely on
|
|
|
|
|
* system-wide capabilities checks
|
|
|
|
|
*/
|
|
|
|
|
if (!bpf_token_allow_cmd(token, BPF_PROG_LOAD) ||
|
|
|
|
|
!bpf_token_allow_prog_type(token, attr->prog_type,
|
|
|
|
|
attr->expected_attach_type)) {
|
|
|
|
|
bpf_token_put(token);
|
|
|
|
|
token = NULL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bpf_cap = bpf_token_capable(token, CAP_BPF);
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
|
2018-12-01 05:08:14 +00:00
|
|
|
if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
|
|
|
|
|
(attr->prog_flags & BPF_F_ANY_ALIGNMENT) &&
|
2024-01-24 02:21:03 +00:00
|
|
|
!bpf_cap)
|
|
|
|
|
goto put_token;
|
2018-12-01 05:08:14 +00:00
|
|
|
|
2023-06-13 22:35:30 +00:00
|
|
|
/* Intent here is for unprivileged_bpf_disabled to block BPF program
|
|
|
|
|
* creation for unprivileged users; other actions depend
|
|
|
|
|
* on fd availability and access to bpffs, so are dependent on
|
|
|
|
|
* object creation success. Even with unprivileged BPF disabled,
|
|
|
|
|
* capability checks are still carried out for these
|
|
|
|
|
* and other operations.
|
|
|
|
|
*/
|
2024-01-24 02:21:03 +00:00
|
|
|
if (sysctl_unprivileged_bpf_disabled && !bpf_cap)
|
|
|
|
|
goto put_token;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2019-04-02 04:27:45 +00:00
|
|
|
if (attr->insn_cnt == 0 ||
|
2024-01-24 02:21:03 +00:00
|
|
|
attr->insn_cnt > (bpf_cap ? BPF_COMPLEXITY_LIMIT_INSNS : BPF_MAXINSNS)) {
|
|
|
|
|
err = -E2BIG;
|
|
|
|
|
goto put_token;
|
|
|
|
|
}
|
2017-06-01 01:16:00 +00:00
|
|
|
if (type != BPF_PROG_TYPE_SOCKET_FILTER &&
|
|
|
|
|
type != BPF_PROG_TYPE_CGROUP_SKB &&
|
2024-01-24 02:21:03 +00:00
|
|
|
!bpf_cap)
|
|
|
|
|
goto put_token;
|
2020-05-13 23:03:54 +00:00
|
|
|
|
2024-01-24 02:21:03 +00:00
|
|
|
if (is_net_admin_prog_type(type) && !bpf_token_capable(token, CAP_NET_ADMIN))
|
|
|
|
|
goto put_token;
|
|
|
|
|
if (is_perfmon_prog_type(type) && !bpf_token_capable(token, CAP_PERFMON))
|
|
|
|
|
goto put_token;
|
bpf: enable non-root eBPF programs
In order to let unprivileged users load and execute eBPF programs
teach verifier to prevent pointer leaks.
Verifier will prevent
- any arithmetic on pointers
(except R10+Imm which is used to compute stack addresses)
- comparison of pointers
(except if (map_value_ptr == 0) ... )
- passing pointers to helper functions
- indirectly passing pointers in stack to helper functions
- returning pointer from bpf program
- storing pointers into ctx or maps
Spill/fill of pointers into stack is allowed, but mangling
of pointers stored in the stack or reading them byte by byte is not.
Within bpf programs the pointers do exist, since programs need to
be able to access maps, pass skb pointer to LD_ABS insns, etc
but programs cannot pass such pointer values to the outside
or obfuscate them.
Only allow BPF_PROG_TYPE_SOCKET_FILTER unprivileged programs,
so that socket filters (tcpdump), af_packet (quic acceleration)
and future kcm can use it.
tracing and tc cls/act program types still require root permissions,
since tracing actually needs to be able to see all kernel pointers
and tc is for root only.
For example, the following unprivileged socket filter program is allowed:
int bpf_prog1(struct __sk_buff *skb)
{
u32 index = load_byte(skb, ETH_HLEN + offsetof(struct iphdr, protocol));
u64 *value = bpf_map_lookup_elem(&my_map, &index);
if (value)
*value += skb->len;
return 0;
}
but the following program is not:
int bpf_prog1(struct __sk_buff *skb)
{
u32 index = load_byte(skb, ETH_HLEN + offsetof(struct iphdr, protocol));
u64 *value = bpf_map_lookup_elem(&my_map, &index);
if (value)
*value += (u64) skb;
return 0;
}
since it would leak the kernel address into the map.
Unprivileged socket filter bpf programs have access to the
following helper functions:
- map lookup/update/delete (but they cannot store kernel pointers into them)
- get_random (it's already exposed to unprivileged user space)
- get_smp_processor_id
- tail_call into another socket filter program
- ktime_get_ns
The feature is controlled by sysctl kernel.unprivileged_bpf_disabled.
This toggle defaults to off (0), but can be set true (1). Once true,
bpf programs and maps cannot be accessed from unprivileged process,
and the toggle cannot be set back to false.
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-08 05:23:21 +00:00
|
|
|
|
2020-12-03 20:46:30 +00:00
|
|
|
/* attach_prog_fd/attach_btf_obj_fd can specify fd of either bpf_prog
|
|
|
|
|
* or btf, we need to check which one it is
|
|
|
|
|
*/
|
|
|
|
|
if (attr->attach_prog_fd) {
|
|
|
|
|
dst_prog = bpf_prog_get(attr->attach_prog_fd);
|
|
|
|
|
if (IS_ERR(dst_prog)) {
|
|
|
|
|
dst_prog = NULL;
|
|
|
|
|
attach_btf = btf_get_by_fd(attr->attach_btf_obj_fd);
|
2024-01-24 02:21:03 +00:00
|
|
|
if (IS_ERR(attach_btf)) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto put_token;
|
|
|
|
|
}
|
2020-12-03 20:46:30 +00:00
|
|
|
if (!btf_is_kernel(attach_btf)) {
|
2020-12-08 06:43:26 +00:00
|
|
|
/* attaching through specifying bpf_prog's BTF
|
|
|
|
|
* objects directly might be supported eventually
|
|
|
|
|
*/
|
2020-12-03 20:46:30 +00:00
|
|
|
btf_put(attach_btf);
|
2024-01-24 02:21:03 +00:00
|
|
|
err = -ENOTSUPP;
|
|
|
|
|
goto put_token;
|
2020-12-03 20:46:30 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
} else if (attr->attach_btf_id) {
|
|
|
|
|
/* fall back to vmlinux BTF, if BTF type ID is specified */
|
|
|
|
|
attach_btf = bpf_get_btf_vmlinux();
|
2024-01-24 02:21:03 +00:00
|
|
|
if (IS_ERR(attach_btf)) {
|
|
|
|
|
err = PTR_ERR(attach_btf);
|
|
|
|
|
goto put_token;
|
|
|
|
|
}
|
|
|
|
|
if (!attach_btf) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto put_token;
|
|
|
|
|
}
|
2020-12-03 20:46:30 +00:00
|
|
|
btf_get(attach_btf);
|
|
|
|
|
}
|
|
|
|
|
|
2019-10-16 03:24:58 +00:00
|
|
|
if (bpf_prog_load_check_attach(type, attr->expected_attach_type,
|
2020-12-03 20:46:30 +00:00
|
|
|
attach_btf, attr->attach_btf_id,
|
|
|
|
|
dst_prog)) {
|
|
|
|
|
if (dst_prog)
|
|
|
|
|
bpf_prog_put(dst_prog);
|
|
|
|
|
if (attach_btf)
|
|
|
|
|
btf_put(attach_btf);
|
2024-01-24 02:21:03 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto put_token;
|
2020-12-03 20:46:30 +00:00
|
|
|
}
|
2018-03-30 22:08:00 +00:00
|
|
|
|
2014-09-26 07:17:00 +00:00
|
|
|
/* plain bpf_prog allocation */
|
|
|
|
|
prog = bpf_prog_alloc(bpf_prog_size(attr->insn_cnt), GFP_USER);
|
2020-12-03 20:46:30 +00:00
|
|
|
if (!prog) {
|
|
|
|
|
if (dst_prog)
|
|
|
|
|
bpf_prog_put(dst_prog);
|
|
|
|
|
if (attach_btf)
|
|
|
|
|
btf_put(attach_btf);
|
2024-01-24 02:21:03 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto put_token;
|
2020-12-03 20:46:30 +00:00
|
|
|
}
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2018-03-30 22:08:00 +00:00
|
|
|
prog->expected_attach_type = attr->expected_attach_type;
|
2024-03-09 00:47:39 +00:00
|
|
|
prog->sleepable = !!(attr->prog_flags & BPF_F_SLEEPABLE);
|
2020-12-03 20:46:30 +00:00
|
|
|
prog->aux->attach_btf = attach_btf;
|
2019-10-16 03:24:58 +00:00
|
|
|
prog->aux->attach_btf_id = attr->attach_btf_id;
|
2020-12-03 20:46:30 +00:00
|
|
|
prog->aux->dst_prog = dst_prog;
|
2023-01-19 22:15:24 +00:00
|
|
|
prog->aux->dev_bound = !!attr->prog_ifindex;
|
2022-01-21 10:09:52 +00:00
|
|
|
prog->aux->xdp_has_frags = attr->prog_flags & BPF_F_XDP_HAS_FRAGS;
|
2017-12-28 02:39:04 +00:00
|
|
|
|
2024-01-24 02:21:03 +00:00
|
|
|
/* move token into prog->aux, reuse taken refcnt */
|
|
|
|
|
prog->aux->token = token;
|
|
|
|
|
token = NULL;
|
|
|
|
|
|
2020-12-01 21:58:59 +00:00
|
|
|
prog->aux->user = get_current_user();
|
2014-09-26 07:17:00 +00:00
|
|
|
prog->len = attr->insn_cnt;
|
|
|
|
|
|
|
|
|
|
err = -EFAULT;
|
2021-05-14 00:36:05 +00:00
|
|
|
if (copy_from_bpfptr(prog->insns,
|
|
|
|
|
make_bpfptr(attr->insns, uattr.is_kernel),
|
|
|
|
|
bpf_prog_insn_size(prog)) != 0)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
2023-06-13 22:35:33 +00:00
|
|
|
/* copy eBPF program license from user space */
|
|
|
|
|
if (strncpy_from_bpfptr(license,
|
|
|
|
|
make_bpfptr(attr->license, uattr.is_kernel),
|
|
|
|
|
sizeof(license) - 1) < 0)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
2023-06-13 22:35:33 +00:00
|
|
|
license[sizeof(license) - 1] = 0;
|
|
|
|
|
|
|
|
|
|
/* eBPF programs must be GPL compatible to use GPL-ed functions */
|
|
|
|
|
prog->gpl_compatible = license_is_gpl_compatible(license) ? 1 : 0;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
|
|
|
|
prog->orig_prog = NULL;
|
2015-09-29 23:41:50 +00:00
|
|
|
prog->jited = 0;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2019-11-17 17:28:03 +00:00
|
|
|
atomic64_set(&prog->aux->refcnt, 1);
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2017-12-28 02:39:04 +00:00
|
|
|
if (bpf_prog_is_dev_bound(prog->aux)) {
|
2023-01-19 22:15:24 +00:00
|
|
|
err = bpf_prog_dev_bound_init(prog, attr);
|
2017-11-03 20:56:17 +00:00
|
|
|
if (err)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
2017-11-03 20:56:17 +00:00
|
|
|
}
|
|
|
|
|
|
2023-01-19 22:15:27 +00:00
|
|
|
if (type == BPF_PROG_TYPE_EXT && dst_prog &&
|
|
|
|
|
bpf_prog_is_dev_bound(dst_prog->aux)) {
|
|
|
|
|
err = bpf_prog_dev_bound_inherit(prog, dst_prog);
|
2017-11-03 20:56:17 +00:00
|
|
|
if (err)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
2017-11-03 20:56:17 +00:00
|
|
|
}
|
|
|
|
|
|
2024-01-03 19:05:44 +00:00
|
|
|
/*
|
|
|
|
|
* Bookkeeping for managing the program attachment chain.
|
|
|
|
|
*
|
|
|
|
|
* It might be tempting to set attach_tracing_prog flag at the attachment
|
|
|
|
|
* time, but this will not prevent from loading bunch of tracing prog
|
|
|
|
|
* first, then attach them one to another.
|
|
|
|
|
*
|
|
|
|
|
* The flag attach_tracing_prog is set for the whole program lifecycle, and
|
|
|
|
|
* doesn't have to be cleared in bpf_tracing_link_release, since tracing
|
|
|
|
|
* programs cannot change attachment target.
|
|
|
|
|
*/
|
|
|
|
|
if (type == BPF_PROG_TYPE_TRACING && dst_prog &&
|
|
|
|
|
dst_prog->type == BPF_PROG_TYPE_TRACING) {
|
|
|
|
|
prog->aux->attach_tracing_prog = true;
|
|
|
|
|
}
|
|
|
|
|
|
2014-09-26 07:17:00 +00:00
|
|
|
/* find program type: socket_filter vs tracing_filter */
|
|
|
|
|
err = find_prog_type(type, prog);
|
|
|
|
|
if (err < 0)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2019-06-21 20:32:48 +00:00
|
|
|
prog->aux->load_time = ktime_get_boottime_ns();
|
2020-03-14 01:02:09 +00:00
|
|
|
err = bpf_obj_name_cpy(prog->aux->name, attr->prog_name,
|
|
|
|
|
sizeof(attr->prog_name));
|
|
|
|
|
if (err < 0)
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
goto free_prog;
|
|
|
|
|
|
|
|
|
|
err = security_bpf_prog_load(prog, attr, token);
|
|
|
|
|
if (err)
|
2020-12-01 21:58:59 +00:00
|
|
|
goto free_prog_sec;
|
2017-09-27 21:37:52 +00:00
|
|
|
|
2014-09-26 07:17:00 +00:00
|
|
|
/* run eBPF verifier */
|
2023-04-06 23:41:58 +00:00
|
|
|
err = bpf_check(&prog, attr, uattr, uattr_size);
|
2014-09-26 07:17:00 +00:00
|
|
|
if (err < 0)
|
|
|
|
|
goto free_used_maps;
|
|
|
|
|
|
bpf: reject any prog that failed read-only lock
We currently lock any JITed image as read-only via bpf_jit_binary_lock_ro()
as well as the BPF image as read-only through bpf_prog_lock_ro(). In
the case any of these would fail we throw a WARN_ON_ONCE() in order to
yell loudly to the log. Perhaps, to some extend, this may be comparable
to an allocation where __GFP_NOWARN is explicitly not set.
Added via 65869a47f348 ("bpf: improve read-only handling"), this behavior
is slightly different compared to any of the other in-kernel set_memory_ro()
users who do not check the return code of set_memory_ro() and friends /at
all/ (e.g. in the case of module_enable_ro() / module_disable_ro()). Given
in BPF this is mandatory hardening step, we want to know whether there
are any issues that would leave both BPF data writable. So it happens
that syzkaller enabled fault injection and it triggered memory allocation
failure deep inside x86's change_page_attr_set_clr() which was triggered
from set_memory_ro().
Now, there are two options: i) leaving everything as is, and ii) reworking
the image locking code in order to have a final checkpoint out of the
central bpf_prog_select_runtime() which probes whether any of the calls
during prog setup weren't successful, and then bailing out with an error.
Option ii) is a better approach since this additional paranoia avoids
altogether leaving any potential W+X pages from BPF side in the system.
Therefore, lets be strict about it, and reject programs in such unlikely
occasion. While testing I noticed also that one bpf_prog_lock_ro()
call was missing on the outer dummy prog in case of calls, e.g. in the
destructor we call bpf_prog_free_deferred() on the main prog where we
try to bpf_prog_unlock_free() the program, and since we go via
bpf_prog_select_runtime() do that as well.
Reported-by: syzbot+3b889862e65a98317058@syzkaller.appspotmail.com
Reported-by: syzbot+9e762b52dd17e616a7a5@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-06-15 00:30:48 +00:00
|
|
|
prog = bpf_prog_select_runtime(prog, &err);
|
bpf: allow bpf programs to tail-call other bpf programs
introduce bpf_tail_call(ctx, &jmp_table, index) helper function
which can be used from BPF programs like:
int bpf_prog(struct pt_regs *ctx)
{
...
bpf_tail_call(ctx, &jmp_table, index);
...
}
that is roughly equivalent to:
int bpf_prog(struct pt_regs *ctx)
{
...
if (jmp_table[index])
return (*jmp_table[index])(ctx);
...
}
The important detail that it's not a normal call, but a tail call.
The kernel stack is precious, so this helper reuses the current
stack frame and jumps into another BPF program without adding
extra call frame.
It's trivially done in interpreter and a bit trickier in JITs.
In case of x64 JIT the bigger part of generated assembler prologue
is common for all programs, so it is simply skipped while jumping.
Other JITs can do similar prologue-skipping optimization or
do stack unwind before jumping into the next program.
bpf_tail_call() arguments:
ctx - context pointer
jmp_table - one of BPF_MAP_TYPE_PROG_ARRAY maps used as the jump table
index - index in the jump table
Since all BPF programs are idenitified by file descriptor, user space
need to populate the jmp_table with FDs of other BPF programs.
If jmp_table[index] is empty the bpf_tail_call() doesn't jump anywhere
and program execution continues as normal.
New BPF_MAP_TYPE_PROG_ARRAY map type is introduced so that user space can
populate this jmp_table array with FDs of other bpf programs.
Programs can share the same jmp_table array or use multiple jmp_tables.
The chain of tail calls can form unpredictable dynamic loops therefore
tail_call_cnt is used to limit the number of calls and currently is set to 32.
Use cases:
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
==========
- simplify complex programs by splitting them into a sequence of small programs
- dispatch routine
For tracing and future seccomp the program may be triggered on all system
calls, but processing of syscall arguments will be different. It's more
efficient to implement them as:
int syscall_entry(struct seccomp_data *ctx)
{
bpf_tail_call(ctx, &syscall_jmp_table, ctx->nr /* syscall number */);
... default: process unknown syscall ...
}
int sys_write_event(struct seccomp_data *ctx) {...}
int sys_read_event(struct seccomp_data *ctx) {...}
syscall_jmp_table[__NR_write] = sys_write_event;
syscall_jmp_table[__NR_read] = sys_read_event;
For networking the program may call into different parsers depending on
packet format, like:
int packet_parser(struct __sk_buff *skb)
{
... parse L2, L3 here ...
__u8 ipproto = load_byte(skb, ... offsetof(struct iphdr, protocol));
bpf_tail_call(skb, &ipproto_jmp_table, ipproto);
... default: process unknown protocol ...
}
int parse_tcp(struct __sk_buff *skb) {...}
int parse_udp(struct __sk_buff *skb) {...}
ipproto_jmp_table[IPPROTO_TCP] = parse_tcp;
ipproto_jmp_table[IPPROTO_UDP] = parse_udp;
- for TC use case, bpf_tail_call() allows to implement reclassify-like logic
- bpf_map_update_elem/delete calls into BPF_MAP_TYPE_PROG_ARRAY jump table
are atomic, so user space can build chains of BPF programs on the fly
Implementation details:
=======================
- high performance of bpf_tail_call() is the goal.
It could have been implemented without JIT changes as a wrapper on top of
BPF_PROG_RUN() macro, but with two downsides:
. all programs would have to pay performance penalty for this feature and
tail call itself would be slower, since mandatory stack unwind, return,
stack allocate would be done for every tailcall.
. tailcall would be limited to programs running preempt_disabled, since
generic 'void *ctx' doesn't have room for 'tail_call_cnt' and it would
need to be either global per_cpu variable accessed by helper and by wrapper
or global variable protected by locks.
In this implementation x64 JIT bypasses stack unwind and jumps into the
callee program after prologue.
- bpf_prog_array_compatible() ensures that prog_type of callee and caller
are the same and JITed/non-JITed flag is the same, since calling JITed
program from non-JITed is invalid, since stack frames are different.
Similarly calling kprobe type program from socket type program is invalid.
- jump table is implemented as BPF_MAP_TYPE_PROG_ARRAY to reuse 'map'
abstraction, its user space API and all of verifier logic.
It's in the existing arraymap.c file, since several functions are
shared with regular array map.
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-19 23:59:03 +00:00
|
|
|
if (err < 0)
|
|
|
|
|
goto free_used_maps;
|
2014-09-26 07:17:00 +00:00
|
|
|
|
2017-06-05 19:15:46 +00:00
|
|
|
err = bpf_prog_alloc_id(prog);
|
|
|
|
|
if (err)
|
|
|
|
|
goto free_used_maps;
|
|
|
|
|
|
bpf: fix use after free in prog symbol exposure
syzkaller managed to trigger the warning in bpf_jit_free() which checks via
bpf_prog_kallsyms_verify_off() for potentially unlinked JITed BPF progs
in kallsyms, and subsequently trips over GPF when walking kallsyms entries:
[...]
8021q: adding VLAN 0 to HW filter on device batadv0
8021q: adding VLAN 0 to HW filter on device batadv0
WARNING: CPU: 0 PID: 9869 at kernel/bpf/core.c:810 bpf_jit_free+0x1e8/0x2a0
Kernel panic - not syncing: panic_on_warn set ...
CPU: 0 PID: 9869 Comm: kworker/0:7 Not tainted 5.0.0-rc8+ #1
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: events bpf_prog_free_deferred
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x113/0x167 lib/dump_stack.c:113
panic+0x212/0x40b kernel/panic.c:214
__warn.cold.8+0x1b/0x38 kernel/panic.c:571
report_bug+0x1a4/0x200 lib/bug.c:186
fixup_bug arch/x86/kernel/traps.c:178 [inline]
do_error_trap+0x11b/0x200 arch/x86/kernel/traps.c:271
do_invalid_op+0x36/0x40 arch/x86/kernel/traps.c:290
invalid_op+0x14/0x20 arch/x86/entry/entry_64.S:973
RIP: 0010:bpf_jit_free+0x1e8/0x2a0
Code: 02 4c 89 e2 83 e2 07 38 d0 7f 08 84 c0 0f 85 86 00 00 00 48 ba 00 02 00 00 00 00 ad de 0f b6 43 02 49 39 d6 0f 84 5f fe ff ff <0f> 0b e9 58 fe ff ff 48 b8 00 00 00 00 00 fc ff df 4c 89 e2 48 c1
RSP: 0018:ffff888092f67cd8 EFLAGS: 00010202
RAX: 0000000000000007 RBX: ffffc90001947000 RCX: ffffffff816e9d88
RDX: dead000000000200 RSI: 0000000000000008 RDI: ffff88808769f7f0
RBP: ffff888092f67d00 R08: fffffbfff1394059 R09: fffffbfff1394058
R10: fffffbfff1394058 R11: ffffffff89ca02c7 R12: ffffc90001947002
R13: ffffc90001947020 R14: ffffffff881eca80 R15: ffff88808769f7e8
BUG: unable to handle kernel paging request at fffffbfff400d000
#PF error: [normal kernel read fault]
PGD 21ffee067 P4D 21ffee067 PUD 21ffed067 PMD 9f942067 PTE 0
Oops: 0000 [#1] PREEMPT SMP KASAN
CPU: 0 PID: 9869 Comm: kworker/0:7 Not tainted 5.0.0-rc8+ #1
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: events bpf_prog_free_deferred
RIP: 0010:bpf_get_prog_addr_region kernel/bpf/core.c:495 [inline]
RIP: 0010:bpf_tree_comp kernel/bpf/core.c:558 [inline]
RIP: 0010:__lt_find include/linux/rbtree_latch.h:115 [inline]
RIP: 0010:latch_tree_find include/linux/rbtree_latch.h:208 [inline]
RIP: 0010:bpf_prog_kallsyms_find+0x107/0x2e0 kernel/bpf/core.c:632
Code: 00 f0 ff ff 44 38 c8 7f 08 84 c0 0f 85 fa 00 00 00 41 f6 45 02 01 75 02 0f 0b 48 39 da 0f 82 92 00 00 00 48 89 d8 48 c1 e8 03 <42> 0f b6 04 30 84 c0 74 08 3c 03 0f 8e 45 01 00 00 8b 03 48 c1 e0
[...]
Upon further debugging, it turns out that whenever we trigger this
issue, the kallsyms removal in bpf_prog_ksym_node_del() was /skipped/
but yet bpf_jit_free() reported that the entry is /in use/.
Problem is that symbol exposure via bpf_prog_kallsyms_add() but also
perf_event_bpf_event() were done /after/ bpf_prog_new_fd(). Once the
fd is exposed to the public, a parallel close request came in right
before we attempted to do the bpf_prog_kallsyms_add().
Given at this time the prog reference count is one, we start to rip
everything underneath us via bpf_prog_release() -> bpf_prog_put().
The memory is eventually released via deferred free, so we're seeing
that bpf_jit_free() has a kallsym entry because we added it from
bpf_prog_load() but /after/ bpf_prog_put() from the remote CPU.
Therefore, move both notifications /before/ we install the fd. The
issue was never seen between bpf_prog_alloc_id() and bpf_prog_new_fd()
because upon bpf_prog_get_fd_by_id() we'll take another reference to
the BPF prog, so we're still holding the original reference from the
bpf_prog_load().
Fixes: 6ee52e2a3fe4 ("perf, bpf: Introduce PERF_RECORD_BPF_EVENT")
Fixes: 74451e66d516 ("bpf: make jited programs visible in traces")
Reported-by: syzbot+bd3bba6ff3fcea7a6ec6@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Song Liu <songliubraving@fb.com>
2019-08-23 20:14:23 +00:00
|
|
|
/* Upon success of bpf_prog_alloc_id(), the BPF prog is
|
|
|
|
|
* effectively publicly exposed. However, retrieving via
|
|
|
|
|
* bpf_prog_get_fd_by_id() will take another reference,
|
|
|
|
|
* therefore it cannot be gone underneath us.
|
|
|
|
|
*
|
|
|
|
|
* Only for the time /after/ successful bpf_prog_new_fd()
|
|
|
|
|
* and before returning to userspace, we might just hold
|
|
|
|
|
* one reference and any parallel close on that fd could
|
|
|
|
|
* rip everything out. Hence, below notifications must
|
|
|
|
|
* happen before bpf_prog_new_fd().
|
|
|
|
|
*
|
|
|
|
|
* Also, any failure handling from this point onwards must
|
|
|
|
|
* be using bpf_prog_put() given the program is exposed.
|
|
|
|
|
*/
|
bpf: make jited programs visible in traces
Long standing issue with JITed programs is that stack traces from
function tracing check whether a given address is kernel code
through {__,}kernel_text_address(), which checks for code in core
kernel, modules and dynamically allocated ftrace trampolines. But
what is still missing is BPF JITed programs (interpreted programs
are not an issue as __bpf_prog_run() will be attributed to them),
thus when a stack trace is triggered, the code walking the stack
won't see any of the JITed ones. The same for address correlation
done from user space via reading /proc/kallsyms. This is read by
tools like perf, but the latter is also useful for permanent live
tracing with eBPF itself in combination with stack maps when other
eBPF types are part of the callchain. See offwaketime example on
dumping stack from a map.
This work tries to tackle that issue by making the addresses and
symbols known to the kernel. The lookup from *kernel_text_address()
is implemented through a latched RB tree that can be read under
RCU in fast-path that is also shared for symbol/size/offset lookup
for a specific given address in kallsyms. The slow-path iteration
through all symbols in the seq file done via RCU list, which holds
a tiny fraction of all exported ksyms, usually below 0.1 percent.
Function symbols are exported as bpf_prog_<tag>, in order to aide
debugging and attribution. This facility is currently enabled for
root-only when bpf_jit_kallsyms is set to 1, and disabled if hardening
is active in any mode. The rationale behind this is that still a lot
of systems ship with world read permissions on kallsyms thus addresses
should not get suddenly exposed for them. If that situation gets
much better in future, we always have the option to change the
default on this. Likewise, unprivileged programs are not allowed
to add entries there either, but that is less of a concern as most
such programs types relevant in this context are for root-only anyway.
If enabled, call graphs and stack traces will then show a correct
attribution; one example is illustrated below, where the trace is
now visible in tooling such as perf script --kallsyms=/proc/kallsyms
and friends.
Before:
7fff8166889d bpf_clone_redirect+0x80007f0020ed (/lib/modules/4.9.0-rc8+/build/vmlinux)
f5d80 __sendmsg_nocancel+0xffff006451f1a007 (/usr/lib64/libc-2.18.so)
After:
7fff816688b7 bpf_clone_redirect+0x80007f002107 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fffa0575728 bpf_prog_33c45a467c9e061a+0x8000600020fb (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fffa07ef1fc cls_bpf_classify+0x8000600020dc (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff81678b68 tc_classify+0x80007f002078 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8164d40b __netif_receive_skb_core+0x80007f0025fb (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8164d718 __netif_receive_skb+0x80007f002018 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8164e565 process_backlog+0x80007f002095 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8164dc71 net_rx_action+0x80007f002231 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff81767461 __softirqentry_text_start+0x80007f0020d1 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff817658ac do_softirq_own_stack+0x80007f00201c (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff810a2c20 do_softirq+0x80007f002050 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff810a2cb5 __local_bh_enable_ip+0x80007f002085 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8168d452 ip_finish_output2+0x80007f002152 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8168ea3d ip_finish_output+0x80007f00217d (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff8168f2af ip_output+0x80007f00203f (/lib/modules/4.9.0-rc8+/build/vmlinux)
[...]
7fff81005854 do_syscall_64+0x80007f002054 (/lib/modules/4.9.0-rc8+/build/vmlinux)
7fff817649eb return_from_SYSCALL_64+0x80007f002000 (/lib/modules/4.9.0-rc8+/build/vmlinux)
f5d80 __sendmsg_nocancel+0xffff01c484812007 (/usr/lib64/libc-2.18.so)
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-02-16 21:24:50 +00:00
|
|
|
bpf_prog_kallsyms_add(prog);
|
2019-01-17 16:15:15 +00:00
|
|
|
perf_event_bpf_event(prog, PERF_BPF_EVENT_PROG_LOAD, 0);
|
2019-12-06 21:49:34 +00:00
|
|
|
bpf_audit_prog(prog, BPF_AUDIT_LOAD);
|
bpf: fix use after free in prog symbol exposure
syzkaller managed to trigger the warning in bpf_jit_free() which checks via
bpf_prog_kallsyms_verify_off() for potentially unlinked JITed BPF progs
in kallsyms, and subsequently trips over GPF when walking kallsyms entries:
[...]
8021q: adding VLAN 0 to HW filter on device batadv0
8021q: adding VLAN 0 to HW filter on device batadv0
WARNING: CPU: 0 PID: 9869 at kernel/bpf/core.c:810 bpf_jit_free+0x1e8/0x2a0
Kernel panic - not syncing: panic_on_warn set ...
CPU: 0 PID: 9869 Comm: kworker/0:7 Not tainted 5.0.0-rc8+ #1
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: events bpf_prog_free_deferred
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x113/0x167 lib/dump_stack.c:113
panic+0x212/0x40b kernel/panic.c:214
__warn.cold.8+0x1b/0x38 kernel/panic.c:571
report_bug+0x1a4/0x200 lib/bug.c:186
fixup_bug arch/x86/kernel/traps.c:178 [inline]
do_error_trap+0x11b/0x200 arch/x86/kernel/traps.c:271
do_invalid_op+0x36/0x40 arch/x86/kernel/traps.c:290
invalid_op+0x14/0x20 arch/x86/entry/entry_64.S:973
RIP: 0010:bpf_jit_free+0x1e8/0x2a0
Code: 02 4c 89 e2 83 e2 07 38 d0 7f 08 84 c0 0f 85 86 00 00 00 48 ba 00 02 00 00 00 00 ad de 0f b6 43 02 49 39 d6 0f 84 5f fe ff ff <0f> 0b e9 58 fe ff ff 48 b8 00 00 00 00 00 fc ff df 4c 89 e2 48 c1
RSP: 0018:ffff888092f67cd8 EFLAGS: 00010202
RAX: 0000000000000007 RBX: ffffc90001947000 RCX: ffffffff816e9d88
RDX: dead000000000200 RSI: 0000000000000008 RDI: ffff88808769f7f0
RBP: ffff888092f67d00 R08: fffffbfff1394059 R09: fffffbfff1394058
R10: fffffbfff1394058 R11: ffffffff89ca02c7 R12: ffffc90001947002
R13: ffffc90001947020 R14: ffffffff881eca80 R15: ffff88808769f7e8
BUG: unable to handle kernel paging request at fffffbfff400d000
#PF error: [normal kernel read fault]
PGD 21ffee067 P4D 21ffee067 PUD 21ffed067 PMD 9f942067 PTE 0
Oops: 0000 [#1] PREEMPT SMP KASAN
CPU: 0 PID: 9869 Comm: kworker/0:7 Not tainted 5.0.0-rc8+ #1
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: events bpf_prog_free_deferred
RIP: 0010:bpf_get_prog_addr_region kernel/bpf/core.c:495 [inline]
RIP: 0010:bpf_tree_comp kernel/bpf/core.c:558 [inline]
RIP: 0010:__lt_find include/linux/rbtree_latch.h:115 [inline]
RIP: 0010:latch_tree_find include/linux/rbtree_latch.h:208 [inline]
RIP: 0010:bpf_prog_kallsyms_find+0x107/0x2e0 kernel/bpf/core.c:632
Code: 00 f0 ff ff 44 38 c8 7f 08 84 c0 0f 85 fa 00 00 00 41 f6 45 02 01 75 02 0f 0b 48 39 da 0f 82 92 00 00 00 48 89 d8 48 c1 e8 03 <42> 0f b6 04 30 84 c0 74 08 3c 03 0f 8e 45 01 00 00 8b 03 48 c1 e0
[...]
Upon further debugging, it turns out that whenever we trigger this
issue, the kallsyms removal in bpf_prog_ksym_node_del() was /skipped/
but yet bpf_jit_free() reported that the entry is /in use/.
Problem is that symbol exposure via bpf_prog_kallsyms_add() but also
perf_event_bpf_event() were done /after/ bpf_prog_new_fd(). Once the
fd is exposed to the public, a parallel close request came in right
before we attempted to do the bpf_prog_kallsyms_add().
Given at this time the prog reference count is one, we start to rip
everything underneath us via bpf_prog_release() -> bpf_prog_put().
The memory is eventually released via deferred free, so we're seeing
that bpf_jit_free() has a kallsym entry because we added it from
bpf_prog_load() but /after/ bpf_prog_put() from the remote CPU.
Therefore, move both notifications /before/ we install the fd. The
issue was never seen between bpf_prog_alloc_id() and bpf_prog_new_fd()
because upon bpf_prog_get_fd_by_id() we'll take another reference to
the BPF prog, so we're still holding the original reference from the
bpf_prog_load().
Fixes: 6ee52e2a3fe4 ("perf, bpf: Introduce PERF_RECORD_BPF_EVENT")
Fixes: 74451e66d516 ("bpf: make jited programs visible in traces")
Reported-by: syzbot+bd3bba6ff3fcea7a6ec6@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Song Liu <songliubraving@fb.com>
2019-08-23 20:14:23 +00:00
|
|
|
|
|
|
|
|
err = bpf_prog_new_fd(prog);
|
|
|
|
|
if (err < 0)
|
|
|
|
|
bpf_prog_put(prog);
|
2014-09-26 07:17:00 +00:00
|
|
|
return err;
|
|
|
|
|
|
|
|
|
|
free_used_maps:
|
2019-10-22 13:57:23 +00:00
|
|
|
/* In case we have subprogs, we need to wait for a grace
|
|
|
|
|
* period before we can tear down JIT memory since symbols
|
|
|
|
|
* are already exposed under kallsyms.
|
|
|
|
|
*/
|
bpf: Implement support for adding hidden subprogs
Introduce support in the verifier for generating a subprogram and
include it as part of a BPF program dynamically after the do_check phase
is complete. The first user will be the next patch which generates
default exception callbacks if none are set for the program. The phase
of invocation will be do_misc_fixups. Note that this is an internal
verifier function, and should be used with instruction blocks which
uphold the invariants stated in check_subprogs.
Since these subprogs are always appended to the end of the instruction
sequence of the program, it becomes relatively inexpensive to do the
related adjustments to the subprog_info of the program. Only the fake
exit subprogram is shifted forward, making room for our new subprog.
This is useful to insert a new subprogram, get it JITed, and obtain its
function pointer. The next patch will use this functionality to insert a
default exception callback which will be invoked after unwinding the
stack.
Note that these added subprograms are invisible to userspace, and never
reported in BPF_OBJ_GET_INFO_BY_ID etc. For now, only a single
subprogram is supported, but more can be easily supported in the future.
To this end, two function counts are introduced now, the existing
func_cnt, and real_func_cnt, the latter including hidden programs. This
allows us to conver the JIT code to use the real_func_cnt for management
of resources while syscall path continues working with existing
func_cnt.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20230912233214.1518551-4-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-09-12 23:32:00 +00:00
|
|
|
__bpf_prog_put_noref(prog, prog->aux->real_func_cnt);
|
2019-10-22 13:57:23 +00:00
|
|
|
return err;
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
|
2017-10-18 20:00:24 +00:00
|
|
|
free_prog_sec:
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
security_bpf_prog_free(prog);
|
2023-12-19 15:37:35 +00:00
|
|
|
free_prog:
|
bpf,lsm: Refactor bpf_prog_alloc/bpf_prog_free LSM hooks
Based on upstream discussion ([0]), rework existing
bpf_prog_alloc_security LSM hook. Rename it to bpf_prog_load and instead
of passing bpf_prog_aux, pass proper bpf_prog pointer for a full BPF
program struct. Also, we pass bpf_attr union with all the user-provided
arguments for BPF_PROG_LOAD command. This will give LSMs as much
information as we can basically provide.
The hook is also BPF token-aware now, and optional bpf_token struct is
passed as a third argument. bpf_prog_load LSM hook is called after
a bunch of sanity checks were performed, bpf_prog and bpf_prog_aux were
allocated and filled out, but right before performing full-fledged BPF
verification step.
bpf_prog_free LSM hook is now accepting struct bpf_prog argument, for
consistency. SELinux code is adjusted to all new names, types, and
signatures.
Note, given that bpf_prog_load (previously bpf_prog_alloc) hook can be
used by some LSMs to allocate extra security blob, but also by other
LSMs to reject BPF program loading, we need to make sure that
bpf_prog_free LSM hook is called after bpf_prog_load/bpf_prog_alloc one
*even* if the hook itself returned error. If we don't do that, we run
the risk of leaking memory. This seems to be possible today when
combining SELinux and BPF LSM, as one example, depending on their
relative ordering.
Also, for BPF LSM setup, add bpf_prog_load and bpf_prog_free to
sleepable LSM hooks list, as they are both executed in sleepable
context. Also drop bpf_prog_load hook from untrusted, as there is no
issue with refcount or anything else anymore, that originally forced us
to add it to untrusted list in c0c852dd1876 ("bpf: Do not mark certain LSM
hook arguments as trusted"). We now trigger this hook much later and it
should not be an issue anymore.
[0] https://lore.kernel.org/bpf/9fe88aef7deabbe87d3fc38c4aea3c69.paul@paul-moore.com/
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-10-andrii@kernel.org
2024-01-24 02:21:06 +00:00
|
|
|
free_uid(prog->aux->user);
|
bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier
Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.
This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.
Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.
Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org
2020-12-03 20:46:29 +00:00
|
|
|
if (prog->aux->attach_btf)
|
|
|
|
|
btf_put(prog->aux->attach_btf);
|
2014-09-26 07:17:00 +00:00
|
|
|
bpf_prog_free(prog);
|
2024-01-24 02:21:03 +00:00
|
|
|
put_token:
|
|
|
|
|
bpf_token_put(token);
|
2014-09-26 07:17:00 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
#define BPF_OBJ_LAST_FIELD path_fd
|
2015-10-29 13:58:09 +00:00
|
|
|
|
|
|
|
|
static int bpf_obj_pin(const union bpf_attr *attr)
|
|
|
|
|
{
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
int path_fd;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_OBJ) || attr->file_flags & ~BPF_F_PATH_FD)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
/* path_fd has to be accompanied by BPF_F_PATH_FD flag */
|
|
|
|
|
if (!(attr->file_flags & BPF_F_PATH_FD) && attr->path_fd)
|
2015-10-29 13:58:09 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
path_fd = attr->file_flags & BPF_F_PATH_FD ? attr->path_fd : AT_FDCWD;
|
|
|
|
|
return bpf_obj_pin_user(attr->bpf_fd, path_fd,
|
|
|
|
|
u64_to_user_ptr(attr->pathname));
|
2015-10-29 13:58:09 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_obj_get(const union bpf_attr *attr)
|
|
|
|
|
{
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
int path_fd;
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
if (CHECK_ATTR(BPF_OBJ) || attr->bpf_fd != 0 ||
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
attr->file_flags & ~(BPF_OBJ_FLAG_MASK | BPF_F_PATH_FD))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
/* path_fd has to be accompanied by BPF_F_PATH_FD flag */
|
|
|
|
|
if (!(attr->file_flags & BPF_F_PATH_FD) && attr->path_fd)
|
2015-10-29 13:58:09 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
Current UAPI of BPF_OBJ_PIN and BPF_OBJ_GET commands of bpf() syscall
forces users to specify pinning location as a string-based absolute or
relative (to current working directory) path. This has various
implications related to security (e.g., symlink-based attacks), forces
BPF FS to be exposed in the file system, which can cause races with
other applications.
One of the feedbacks we got from folks working with containers heavily
was that inability to use purely FD-based location specification was an
unfortunate limitation and hindrance for BPF_OBJ_PIN and BPF_OBJ_GET
commands. This patch closes this oversight, adding path_fd field to
BPF_OBJ_PIN and BPF_OBJ_GET UAPI, following conventions established by
*at() syscalls for dirfd + pathname combinations.
This now allows interesting possibilities like working with detached BPF
FS mount (e.g., to perform multiple pinnings without running a risk of
someone interfering with them), and generally making pinning/getting
more secure and not prone to any races and/or security attacks.
This is demonstrated by a selftest added in subsequent patch that takes
advantage of new mount APIs (fsopen, fsconfig, fsmount) to demonstrate
creating detached BPF FS mount, pinning, and then getting BPF map out of
it, all while never exposing this private instance of BPF FS to outside
worlds.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20230523170013.728457-4-andrii@kernel.org
2023-05-15 23:48:06 +00:00
|
|
|
path_fd = attr->file_flags & BPF_F_PATH_FD ? attr->path_fd : AT_FDCWD;
|
|
|
|
|
return bpf_obj_get_user(path_fd, u64_to_user_ptr(attr->pathname),
|
2017-10-18 20:00:22 +00:00
|
|
|
attr->file_flags);
|
2015-10-29 13:58:09 +00:00
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:08 +00:00
|
|
|
void bpf_link_init(struct bpf_link *link, enum bpf_link_type type,
|
2020-04-29 00:16:06 +00:00
|
|
|
const struct bpf_link_ops *ops, struct bpf_prog *prog)
|
2019-11-14 18:57:04 +00:00
|
|
|
{
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
atomic64_set(&link->refcnt, 1);
|
2020-04-29 00:16:08 +00:00
|
|
|
link->type = type;
|
2020-04-29 00:16:06 +00:00
|
|
|
link->id = 0;
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
link->ops = ops;
|
|
|
|
|
link->prog = prog;
|
|
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
static void bpf_link_free_id(int id)
|
|
|
|
|
{
|
|
|
|
|
if (!id)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
spin_lock_bh(&link_idr_lock);
|
|
|
|
|
idr_remove(&link_idr, id);
|
|
|
|
|
spin_unlock_bh(&link_idr_lock);
|
|
|
|
|
}
|
|
|
|
|
|
2020-03-13 00:21:28 +00:00
|
|
|
/* Clean up bpf_link and corresponding anon_inode file and FD. After
|
|
|
|
|
* anon_inode is created, bpf_link can't be just kfree()'d due to deferred
|
2023-08-09 08:34:15 +00:00
|
|
|
* anon_inode's release() call. This helper marks bpf_link as
|
2020-04-29 00:16:06 +00:00
|
|
|
* defunct, releases anon_inode file and puts reserved FD. bpf_prog's refcnt
|
|
|
|
|
* is not decremented, it's the responsibility of a calling code that failed
|
|
|
|
|
* to complete bpf_link initialization.
|
2023-08-09 08:34:15 +00:00
|
|
|
* This helper eventually calls link's dealloc callback, but does not call
|
|
|
|
|
* link's release callback.
|
2020-03-13 00:21:28 +00:00
|
|
|
*/
|
2020-04-29 00:16:06 +00:00
|
|
|
void bpf_link_cleanup(struct bpf_link_primer *primer)
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
{
|
2020-04-29 00:16:06 +00:00
|
|
|
primer->link->prog = NULL;
|
|
|
|
|
bpf_link_free_id(primer->id);
|
|
|
|
|
fput(primer->file);
|
|
|
|
|
put_unused_fd(primer->fd);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
void bpf_link_inc(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
atomic64_inc(&link->refcnt);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: support deferring bpf_link dealloc to after RCU grace period
BPF link for some program types is passed as a "context" which can be
used by those BPF programs to look up additional information. E.g., for
multi-kprobes and multi-uprobes, link is used to fetch BPF cookie values.
Because of this runtime dependency, when bpf_link refcnt drops to zero
there could still be active BPF programs running accessing link data.
This patch adds generic support to defer bpf_link dealloc callback to
after RCU GP, if requested. This is done by exposing two different
deallocation callbacks, one synchronous and one deferred. If deferred
one is provided, bpf_link_free() will schedule dealloc_deferred()
callback to happen after RCU GP.
BPF is using two flavors of RCU: "classic" non-sleepable one and RCU
tasks trace one. The latter is used when sleepable BPF programs are
used. bpf_link_free() accommodates that by checking underlying BPF
program's sleepable flag, and goes either through normal RCU GP only for
non-sleepable, or through RCU tasks trace GP *and* then normal RCU GP
(taking into account rcu_trace_implies_rcu_gp() optimization), if BPF
program is sleepable.
We use this for multi-kprobe and multi-uprobe links, which dereference
link during program run. We also preventively switch raw_tp link to use
deferred dealloc callback, as upcoming changes in bpf-next tree expose
raw_tp link data (specifically, cookie value) to BPF program at runtime
as well.
Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
Fixes: 89ae89f53d20 ("bpf: Add multi uprobe link")
Reported-by: syzbot+981935d9485a560bfbcb@syzkaller.appspotmail.com
Reported-by: syzbot+2cb5a6c573e98db598cc@syzkaller.appspotmail.com
Reported-by: syzbot+62d8b26793e8a2bd0516@syzkaller.appspotmail.com
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240328052426.3042617-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-28 05:24:26 +00:00
|
|
|
static void bpf_link_defer_dealloc_rcu_gp(struct rcu_head *rcu)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link = container_of(rcu, struct bpf_link, rcu);
|
|
|
|
|
|
|
|
|
|
/* free bpf_link and its containing memory */
|
|
|
|
|
link->ops->dealloc_deferred(link);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_link_defer_dealloc_mult_rcu_gp(struct rcu_head *rcu)
|
|
|
|
|
{
|
|
|
|
|
if (rcu_trace_implies_rcu_gp())
|
|
|
|
|
bpf_link_defer_dealloc_rcu_gp(rcu);
|
|
|
|
|
else
|
|
|
|
|
call_rcu(rcu, bpf_link_defer_dealloc_rcu_gp);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
/* bpf_link_free is guaranteed to be called from process context */
|
|
|
|
|
static void bpf_link_free(struct bpf_link *link)
|
|
|
|
|
{
|
bpf: support deferring bpf_link dealloc to after RCU grace period
BPF link for some program types is passed as a "context" which can be
used by those BPF programs to look up additional information. E.g., for
multi-kprobes and multi-uprobes, link is used to fetch BPF cookie values.
Because of this runtime dependency, when bpf_link refcnt drops to zero
there could still be active BPF programs running accessing link data.
This patch adds generic support to defer bpf_link dealloc callback to
after RCU GP, if requested. This is done by exposing two different
deallocation callbacks, one synchronous and one deferred. If deferred
one is provided, bpf_link_free() will schedule dealloc_deferred()
callback to happen after RCU GP.
BPF is using two flavors of RCU: "classic" non-sleepable one and RCU
tasks trace one. The latter is used when sleepable BPF programs are
used. bpf_link_free() accommodates that by checking underlying BPF
program's sleepable flag, and goes either through normal RCU GP only for
non-sleepable, or through RCU tasks trace GP *and* then normal RCU GP
(taking into account rcu_trace_implies_rcu_gp() optimization), if BPF
program is sleepable.
We use this for multi-kprobe and multi-uprobe links, which dereference
link during program run. We also preventively switch raw_tp link to use
deferred dealloc callback, as upcoming changes in bpf-next tree expose
raw_tp link data (specifically, cookie value) to BPF program at runtime
as well.
Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
Fixes: 89ae89f53d20 ("bpf: Add multi uprobe link")
Reported-by: syzbot+981935d9485a560bfbcb@syzkaller.appspotmail.com
Reported-by: syzbot+2cb5a6c573e98db598cc@syzkaller.appspotmail.com
Reported-by: syzbot+62d8b26793e8a2bd0516@syzkaller.appspotmail.com
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240328052426.3042617-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-28 05:24:26 +00:00
|
|
|
bool sleepable = false;
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
bpf_link_free_id(link->id);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
if (link->prog) {
|
bpf: support deferring bpf_link dealloc to after RCU grace period
BPF link for some program types is passed as a "context" which can be
used by those BPF programs to look up additional information. E.g., for
multi-kprobes and multi-uprobes, link is used to fetch BPF cookie values.
Because of this runtime dependency, when bpf_link refcnt drops to zero
there could still be active BPF programs running accessing link data.
This patch adds generic support to defer bpf_link dealloc callback to
after RCU GP, if requested. This is done by exposing two different
deallocation callbacks, one synchronous and one deferred. If deferred
one is provided, bpf_link_free() will schedule dealloc_deferred()
callback to happen after RCU GP.
BPF is using two flavors of RCU: "classic" non-sleepable one and RCU
tasks trace one. The latter is used when sleepable BPF programs are
used. bpf_link_free() accommodates that by checking underlying BPF
program's sleepable flag, and goes either through normal RCU GP only for
non-sleepable, or through RCU tasks trace GP *and* then normal RCU GP
(taking into account rcu_trace_implies_rcu_gp() optimization), if BPF
program is sleepable.
We use this for multi-kprobe and multi-uprobe links, which dereference
link during program run. We also preventively switch raw_tp link to use
deferred dealloc callback, as upcoming changes in bpf-next tree expose
raw_tp link data (specifically, cookie value) to BPF program at runtime
as well.
Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
Fixes: 89ae89f53d20 ("bpf: Add multi uprobe link")
Reported-by: syzbot+981935d9485a560bfbcb@syzkaller.appspotmail.com
Reported-by: syzbot+2cb5a6c573e98db598cc@syzkaller.appspotmail.com
Reported-by: syzbot+62d8b26793e8a2bd0516@syzkaller.appspotmail.com
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240328052426.3042617-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-28 05:24:26 +00:00
|
|
|
sleepable = link->prog->sleepable;
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
/* detach BPF program, clean up used resources */
|
|
|
|
|
link->ops->release(link);
|
|
|
|
|
bpf_prog_put(link->prog);
|
|
|
|
|
}
|
bpf: support deferring bpf_link dealloc to after RCU grace period
BPF link for some program types is passed as a "context" which can be
used by those BPF programs to look up additional information. E.g., for
multi-kprobes and multi-uprobes, link is used to fetch BPF cookie values.
Because of this runtime dependency, when bpf_link refcnt drops to zero
there could still be active BPF programs running accessing link data.
This patch adds generic support to defer bpf_link dealloc callback to
after RCU GP, if requested. This is done by exposing two different
deallocation callbacks, one synchronous and one deferred. If deferred
one is provided, bpf_link_free() will schedule dealloc_deferred()
callback to happen after RCU GP.
BPF is using two flavors of RCU: "classic" non-sleepable one and RCU
tasks trace one. The latter is used when sleepable BPF programs are
used. bpf_link_free() accommodates that by checking underlying BPF
program's sleepable flag, and goes either through normal RCU GP only for
non-sleepable, or through RCU tasks trace GP *and* then normal RCU GP
(taking into account rcu_trace_implies_rcu_gp() optimization), if BPF
program is sleepable.
We use this for multi-kprobe and multi-uprobe links, which dereference
link during program run. We also preventively switch raw_tp link to use
deferred dealloc callback, as upcoming changes in bpf-next tree expose
raw_tp link data (specifically, cookie value) to BPF program at runtime
as well.
Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
Fixes: 89ae89f53d20 ("bpf: Add multi uprobe link")
Reported-by: syzbot+981935d9485a560bfbcb@syzkaller.appspotmail.com
Reported-by: syzbot+2cb5a6c573e98db598cc@syzkaller.appspotmail.com
Reported-by: syzbot+62d8b26793e8a2bd0516@syzkaller.appspotmail.com
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240328052426.3042617-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-28 05:24:26 +00:00
|
|
|
if (link->ops->dealloc_deferred) {
|
|
|
|
|
/* schedule BPF link deallocation; if underlying BPF program
|
|
|
|
|
* is sleepable, we need to first wait for RCU tasks trace
|
|
|
|
|
* sync, then go through "classic" RCU grace period
|
|
|
|
|
*/
|
|
|
|
|
if (sleepable)
|
|
|
|
|
call_rcu_tasks_trace(&link->rcu, bpf_link_defer_dealloc_mult_rcu_gp);
|
|
|
|
|
else
|
|
|
|
|
call_rcu(&link->rcu, bpf_link_defer_dealloc_rcu_gp);
|
|
|
|
|
}
|
|
|
|
|
if (link->ops->dealloc)
|
|
|
|
|
link->ops->dealloc(link);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_link_put_deferred(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link = container_of(work, struct bpf_link, work);
|
|
|
|
|
|
|
|
|
|
bpf_link_free(link);
|
|
|
|
|
}
|
|
|
|
|
|
2023-06-14 08:34:30 +00:00
|
|
|
/* bpf_link_put might be called from atomic context. It needs to be called
|
|
|
|
|
* from sleepable context in order to acquire sleeping locks during the process.
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
*/
|
|
|
|
|
void bpf_link_put(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
if (!atomic64_dec_and_test(&link->refcnt))
|
|
|
|
|
return;
|
|
|
|
|
|
2023-06-14 08:34:30 +00:00
|
|
|
INIT_WORK(&link->work, bpf_link_put_deferred);
|
|
|
|
|
schedule_work(&link->work);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
}
|
2022-02-09 23:20:01 +00:00
|
|
|
EXPORT_SYMBOL(bpf_link_put);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
2023-06-14 08:34:30 +00:00
|
|
|
static void bpf_link_put_direct(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
if (!atomic64_dec_and_test(&link->refcnt))
|
|
|
|
|
return;
|
|
|
|
|
bpf_link_free(link);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
static int bpf_link_release(struct inode *inode, struct file *filp)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link = filp->private_data;
|
|
|
|
|
|
2023-06-14 08:34:30 +00:00
|
|
|
bpf_link_put_direct(link);
|
2019-11-14 18:57:04 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
2020-04-29 00:16:08 +00:00
|
|
|
#define BPF_PROG_TYPE(_id, _name, prog_ctx_type, kern_ctx_type)
|
|
|
|
|
#define BPF_MAP_TYPE(_id, _ops)
|
|
|
|
|
#define BPF_LINK_TYPE(_id, _name) [_id] = #_name,
|
|
|
|
|
static const char *bpf_link_type_strs[] = {
|
|
|
|
|
[BPF_LINK_TYPE_UNSPEC] = "<invalid>",
|
|
|
|
|
#include <linux/bpf_types.h>
|
|
|
|
|
};
|
|
|
|
|
#undef BPF_PROG_TYPE
|
|
|
|
|
#undef BPF_MAP_TYPE
|
|
|
|
|
#undef BPF_LINK_TYPE
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
|
|
|
|
static void bpf_link_show_fdinfo(struct seq_file *m, struct file *filp)
|
|
|
|
|
{
|
|
|
|
|
const struct bpf_link *link = filp->private_data;
|
|
|
|
|
const struct bpf_prog *prog = link->prog;
|
|
|
|
|
char prog_tag[sizeof(prog->tag) * 2 + 1] = { };
|
|
|
|
|
|
|
|
|
|
seq_printf(m,
|
|
|
|
|
"link_type:\t%s\n"
|
bpf: Create links for BPF struct_ops maps.
Make bpf_link support struct_ops. Previously, struct_ops were always
used alone without any associated links. Upon updating its value, a
struct_ops would be activated automatically. Yet other BPF program
types required to make a bpf_link with their instances before they
could become active. Now, however, you can create an inactive
struct_ops, and create a link to activate it later.
With bpf_links, struct_ops has a behavior similar to other BPF program
types. You can pin/unpin them from their links and the struct_ops will
be deactivated when its link is removed while previously need someone
to delete the value for it to be deactivated.
bpf_links are responsible for registering their associated
struct_ops. You can only use a struct_ops that has the BPF_F_LINK flag
set to create a bpf_link, while a structs without this flag behaves in
the same manner as before and is registered upon updating its value.
The BPF_LINK_TYPE_STRUCT_OPS serves a dual purpose. Not only is it
used to craft the links for BPF struct_ops programs, but also to
create links for BPF struct_ops them-self. Since the links of BPF
struct_ops programs are only used to create trampolines internally,
they are never seen in other contexts. Thus, they can be reused for
struct_ops themself.
To maintain a reference to the map supporting this link, we add
bpf_struct_ops_link as an additional type. The pointer of the map is
RCU and won't be necessary until later in the patchset.
Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Link: https://lore.kernel.org/r/20230323032405.3735486-4-kuifeng@meta.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-03-23 03:24:00 +00:00
|
|
|
"link_id:\t%u\n",
|
2020-04-29 00:16:08 +00:00
|
|
|
bpf_link_type_strs[link->type],
|
bpf: Create links for BPF struct_ops maps.
Make bpf_link support struct_ops. Previously, struct_ops were always
used alone without any associated links. Upon updating its value, a
struct_ops would be activated automatically. Yet other BPF program
types required to make a bpf_link with their instances before they
could become active. Now, however, you can create an inactive
struct_ops, and create a link to activate it later.
With bpf_links, struct_ops has a behavior similar to other BPF program
types. You can pin/unpin them from their links and the struct_ops will
be deactivated when its link is removed while previously need someone
to delete the value for it to be deactivated.
bpf_links are responsible for registering their associated
struct_ops. You can only use a struct_ops that has the BPF_F_LINK flag
set to create a bpf_link, while a structs without this flag behaves in
the same manner as before and is registered upon updating its value.
The BPF_LINK_TYPE_STRUCT_OPS serves a dual purpose. Not only is it
used to craft the links for BPF struct_ops programs, but also to
create links for BPF struct_ops them-self. Since the links of BPF
struct_ops programs are only used to create trampolines internally,
they are never seen in other contexts. Thus, they can be reused for
struct_ops themself.
To maintain a reference to the map supporting this link, we add
bpf_struct_ops_link as an additional type. The pointer of the map is
RCU and won't be necessary until later in the patchset.
Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Link: https://lore.kernel.org/r/20230323032405.3735486-4-kuifeng@meta.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-03-23 03:24:00 +00:00
|
|
|
link->id);
|
|
|
|
|
if (prog) {
|
|
|
|
|
bin2hex(prog_tag, prog->tag, sizeof(prog->tag));
|
|
|
|
|
seq_printf(m,
|
|
|
|
|
"prog_tag:\t%s\n"
|
|
|
|
|
"prog_id:\t%u\n",
|
|
|
|
|
prog_tag,
|
|
|
|
|
prog->aux->id);
|
|
|
|
|
}
|
2020-04-29 00:16:08 +00:00
|
|
|
if (link->ops->show_fdinfo)
|
|
|
|
|
link->ops->show_fdinfo(link, m);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
2020-04-23 02:32:40 +00:00
|
|
|
static const struct file_operations bpf_link_fops = {
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
.show_fdinfo = bpf_link_show_fdinfo,
|
|
|
|
|
#endif
|
|
|
|
|
.release = bpf_link_release,
|
2019-11-14 18:57:04 +00:00
|
|
|
.read = bpf_dummy_read,
|
|
|
|
|
.write = bpf_dummy_write,
|
|
|
|
|
};
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
static int bpf_link_alloc_id(struct bpf_link *link)
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
{
|
2020-04-29 00:16:06 +00:00
|
|
|
int id;
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
idr_preload(GFP_KERNEL);
|
|
|
|
|
spin_lock_bh(&link_idr_lock);
|
|
|
|
|
id = idr_alloc_cyclic(&link_idr, link, 1, INT_MAX, GFP_ATOMIC);
|
|
|
|
|
spin_unlock_bh(&link_idr_lock);
|
|
|
|
|
idr_preload_end();
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
return id;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Prepare bpf_link to be exposed to user-space by allocating anon_inode file,
|
|
|
|
|
* reserving unused FD and allocating ID from link_idr. This is to be paired
|
|
|
|
|
* with bpf_link_settle() to install FD and ID and expose bpf_link to
|
|
|
|
|
* user-space, if bpf_link is successfully attached. If not, bpf_link and
|
|
|
|
|
* pre-allocated resources are to be freed with bpf_cleanup() call. All the
|
|
|
|
|
* transient state is passed around in struct bpf_link_primer.
|
|
|
|
|
* This is preferred way to create and initialize bpf_link, especially when
|
2022-02-20 18:40:55 +00:00
|
|
|
* there are complicated and expensive operations in between creating bpf_link
|
2020-04-29 00:16:06 +00:00
|
|
|
* itself and attaching it to BPF hook. By using bpf_link_prime() and
|
|
|
|
|
* bpf_link_settle() kernel code using bpf_link doesn't have to perform
|
|
|
|
|
* expensive (and potentially failing) roll back operations in a rare case
|
|
|
|
|
* that file, FD, or ID can't be allocated.
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
*/
|
2020-04-29 00:16:06 +00:00
|
|
|
int bpf_link_prime(struct bpf_link *link, struct bpf_link_primer *primer)
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
{
|
|
|
|
|
struct file *file;
|
2020-04-29 00:16:06 +00:00
|
|
|
int fd, id;
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
|
|
|
|
|
fd = get_unused_fd_flags(O_CLOEXEC);
|
|
|
|
|
if (fd < 0)
|
2020-04-29 00:16:06 +00:00
|
|
|
return fd;
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
id = bpf_link_alloc_id(link);
|
|
|
|
|
if (id < 0) {
|
|
|
|
|
put_unused_fd(fd);
|
|
|
|
|
return id;
|
|
|
|
|
}
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
|
|
|
|
|
file = anon_inode_getfile("bpf_link", &bpf_link_fops, link, O_CLOEXEC);
|
|
|
|
|
if (IS_ERR(file)) {
|
2020-05-01 18:56:22 +00:00
|
|
|
bpf_link_free_id(id);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
put_unused_fd(fd);
|
2020-05-01 18:56:22 +00:00
|
|
|
return PTR_ERR(file);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
primer->link = link;
|
|
|
|
|
primer->file = file;
|
|
|
|
|
primer->fd = fd;
|
|
|
|
|
primer->id = id;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int bpf_link_settle(struct bpf_link_primer *primer)
|
|
|
|
|
{
|
|
|
|
|
/* make bpf_link fetchable by ID */
|
|
|
|
|
spin_lock_bh(&link_idr_lock);
|
|
|
|
|
primer->link->id = primer->id;
|
|
|
|
|
spin_unlock_bh(&link_idr_lock);
|
|
|
|
|
/* make bpf_link fetchable by FD */
|
|
|
|
|
fd_install(primer->fd, primer->file);
|
|
|
|
|
/* pass through installed FD */
|
|
|
|
|
return primer->fd;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int bpf_link_new_fd(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
return anon_inode_getfd("bpf-link", &bpf_link_fops, link, O_CLOEXEC);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
struct bpf_link *bpf_link_get_from_fd(u32 ufd)
|
|
|
|
|
{
|
|
|
|
|
struct fd f = fdget(ufd);
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
|
|
|
|
|
if (!f.file)
|
|
|
|
|
return ERR_PTR(-EBADF);
|
|
|
|
|
if (f.file->f_op != &bpf_link_fops) {
|
|
|
|
|
fdput(f);
|
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
link = f.file->private_data;
|
|
|
|
|
bpf_link_inc(link);
|
|
|
|
|
fdput(f);
|
|
|
|
|
|
|
|
|
|
return link;
|
|
|
|
|
}
|
2022-02-09 23:20:01 +00:00
|
|
|
EXPORT_SYMBOL(bpf_link_get_from_fd);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
|
|
|
|
static void bpf_tracing_link_release(struct bpf_link *link)
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
{
|
2020-09-29 12:45:50 +00:00
|
|
|
struct bpf_tracing_link *tr_link =
|
2022-05-10 20:59:19 +00:00
|
|
|
container_of(link, struct bpf_tracing_link, link.link);
|
2020-09-29 12:45:50 +00:00
|
|
|
|
2022-05-10 20:59:19 +00:00
|
|
|
WARN_ON_ONCE(bpf_trampoline_unlink_prog(&tr_link->link,
|
2020-09-29 12:45:50 +00:00
|
|
|
tr_link->trampoline));
|
|
|
|
|
|
|
|
|
|
bpf_trampoline_put(tr_link->trampoline);
|
|
|
|
|
|
|
|
|
|
/* tgt_prog is NULL if target is a kernel function */
|
|
|
|
|
if (tr_link->tgt_prog)
|
|
|
|
|
bpf_prog_put(tr_link->tgt_prog);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_tracing_link_dealloc(struct bpf_link *link)
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
{
|
|
|
|
|
struct bpf_tracing_link *tr_link =
|
2022-05-10 20:59:19 +00:00
|
|
|
container_of(link, struct bpf_tracing_link, link.link);
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
|
|
|
|
kfree(tr_link);
|
|
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:08 +00:00
|
|
|
static void bpf_tracing_link_show_fdinfo(const struct bpf_link *link,
|
|
|
|
|
struct seq_file *seq)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_tracing_link *tr_link =
|
2022-05-10 20:59:19 +00:00
|
|
|
container_of(link, struct bpf_tracing_link, link.link);
|
2023-05-17 10:31:25 +00:00
|
|
|
u32 target_btf_id, target_obj_id;
|
2020-04-29 00:16:08 +00:00
|
|
|
|
2023-05-17 10:31:25 +00:00
|
|
|
bpf_trampoline_unpack_key(tr_link->trampoline->key,
|
|
|
|
|
&target_obj_id, &target_btf_id);
|
2020-04-29 00:16:08 +00:00
|
|
|
seq_printf(seq,
|
2023-05-17 10:31:25 +00:00
|
|
|
"attach_type:\t%d\n"
|
|
|
|
|
"target_obj_id:\t%u\n"
|
|
|
|
|
"target_btf_id:\t%u\n",
|
|
|
|
|
tr_link->attach_type,
|
|
|
|
|
target_obj_id,
|
|
|
|
|
target_btf_id);
|
2020-04-29 00:16:08 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_tracing_link_fill_link_info(const struct bpf_link *link,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_tracing_link *tr_link =
|
2022-05-10 20:59:19 +00:00
|
|
|
container_of(link, struct bpf_tracing_link, link.link);
|
2020-04-29 00:16:08 +00:00
|
|
|
|
|
|
|
|
info->tracing.attach_type = tr_link->attach_type;
|
2021-04-13 09:16:06 +00:00
|
|
|
bpf_trampoline_unpack_key(tr_link->trampoline->key,
|
|
|
|
|
&info->tracing.target_obj_id,
|
|
|
|
|
&info->tracing.target_btf_id);
|
2020-04-29 00:16:08 +00:00
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
static const struct bpf_link_ops bpf_tracing_link_lops = {
|
|
|
|
|
.release = bpf_tracing_link_release,
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
.dealloc = bpf_tracing_link_dealloc,
|
2020-04-29 00:16:08 +00:00
|
|
|
.show_fdinfo = bpf_tracing_link_show_fdinfo,
|
|
|
|
|
.fill_link_info = bpf_tracing_link_fill_link_info,
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
};
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
static int bpf_tracing_prog_attach(struct bpf_prog *prog,
|
|
|
|
|
int tgt_prog_fd,
|
2022-05-10 20:59:21 +00:00
|
|
|
u32 btf_id,
|
|
|
|
|
u64 bpf_cookie)
|
2019-11-14 18:57:04 +00:00
|
|
|
{
|
2020-04-29 00:16:06 +00:00
|
|
|
struct bpf_link_primer link_primer;
|
2020-09-29 12:45:50 +00:00
|
|
|
struct bpf_prog *tgt_prog = NULL;
|
2020-09-29 12:45:51 +00:00
|
|
|
struct bpf_trampoline *tr = NULL;
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
struct bpf_tracing_link *link;
|
2020-09-29 12:45:51 +00:00
|
|
|
u64 key = 0;
|
2020-04-29 00:16:06 +00:00
|
|
|
int err;
|
2019-11-14 18:57:04 +00:00
|
|
|
|
2020-03-29 00:43:52 +00:00
|
|
|
switch (prog->type) {
|
|
|
|
|
case BPF_PROG_TYPE_TRACING:
|
|
|
|
|
if (prog->expected_attach_type != BPF_TRACE_FENTRY &&
|
|
|
|
|
prog->expected_attach_type != BPF_TRACE_FEXIT &&
|
|
|
|
|
prog->expected_attach_type != BPF_MODIFY_RETURN) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case BPF_PROG_TYPE_EXT:
|
|
|
|
|
if (prog->expected_attach_type != 0) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case BPF_PROG_TYPE_LSM:
|
|
|
|
|
if (prog->expected_attach_type != BPF_LSM_MAC) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
2019-11-14 18:57:04 +00:00
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
if (!!tgt_prog_fd != !!btf_id) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (tgt_prog_fd) {
|
2024-01-03 19:05:44 +00:00
|
|
|
/*
|
|
|
|
|
* For now we only allow new targets for BPF_PROG_TYPE_EXT. If this
|
|
|
|
|
* part would be changed to implement the same for
|
|
|
|
|
* BPF_PROG_TYPE_TRACING, do not forget to update the way how
|
|
|
|
|
* attach_tracing_prog flag is set.
|
|
|
|
|
*/
|
2020-09-29 12:45:51 +00:00
|
|
|
if (prog->type != BPF_PROG_TYPE_EXT) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
tgt_prog = bpf_prog_get(tgt_prog_fd);
|
|
|
|
|
if (IS_ERR(tgt_prog)) {
|
|
|
|
|
err = PTR_ERR(tgt_prog);
|
|
|
|
|
tgt_prog = NULL;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier
Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.
This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.
Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.
Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org
2020-12-03 20:46:29 +00:00
|
|
|
key = bpf_trampoline_compute_key(tgt_prog, NULL, btf_id);
|
2020-09-29 12:45:51 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
link = kzalloc(sizeof(*link), GFP_USER);
|
|
|
|
|
if (!link) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_put_prog;
|
|
|
|
|
}
|
2022-05-10 20:59:19 +00:00
|
|
|
bpf_link_init(&link->link.link, BPF_LINK_TYPE_TRACING,
|
2020-04-29 00:16:08 +00:00
|
|
|
&bpf_tracing_link_lops, prog);
|
|
|
|
|
link->attach_type = prog->expected_attach_type;
|
2022-05-10 20:59:21 +00:00
|
|
|
link->link.cookie = bpf_cookie;
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
|
2020-09-29 12:45:50 +00:00
|
|
|
mutex_lock(&prog->aux->dst_mutex);
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
/* There are a few possible cases here:
|
|
|
|
|
*
|
|
|
|
|
* - if prog->aux->dst_trampoline is set, the program was just loaded
|
|
|
|
|
* and not yet attached to anything, so we can use the values stored
|
|
|
|
|
* in prog->aux
|
|
|
|
|
*
|
|
|
|
|
* - if prog->aux->dst_trampoline is NULL, the program has already been
|
|
|
|
|
* attached to a target and its initial target was cleared (below)
|
|
|
|
|
*
|
|
|
|
|
* - if tgt_prog != NULL, the caller specified tgt_prog_fd +
|
|
|
|
|
* target_btf_id using the link_create API.
|
|
|
|
|
*
|
|
|
|
|
* - if tgt_prog == NULL when this function was called using the old
|
2021-04-14 19:51:41 +00:00
|
|
|
* raw_tracepoint_open API, and we need a target from prog->aux
|
|
|
|
|
*
|
|
|
|
|
* - if prog->aux->dst_trampoline and tgt_prog is NULL, the program
|
|
|
|
|
* was detached and is going for re-attachment.
|
2024-01-03 19:05:46 +00:00
|
|
|
*
|
|
|
|
|
* - if prog->aux->dst_trampoline is NULL and tgt_prog and prog->aux->attach_btf
|
|
|
|
|
* are NULL, then program was already attached and user did not provide
|
|
|
|
|
* tgt_prog_fd so we have no way to find out or create trampoline
|
2020-09-29 12:45:51 +00:00
|
|
|
*/
|
|
|
|
|
if (!prog->aux->dst_trampoline && !tgt_prog) {
|
2021-04-14 19:51:41 +00:00
|
|
|
/*
|
|
|
|
|
* Allow re-attach for TRACING and LSM programs. If it's
|
|
|
|
|
* currently linked, bpf_trampoline_link_prog will fail.
|
|
|
|
|
* EXT programs need to specify tgt_prog_fd, so they
|
|
|
|
|
* re-attach in separate code path.
|
|
|
|
|
*/
|
|
|
|
|
if (prog->type != BPF_PROG_TYPE_TRACING &&
|
|
|
|
|
prog->type != BPF_PROG_TYPE_LSM) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
2024-01-03 19:05:46 +00:00
|
|
|
/* We can allow re-attach only if we have valid attach_btf. */
|
|
|
|
|
if (!prog->aux->attach_btf) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
2021-04-14 19:51:41 +00:00
|
|
|
btf_id = prog->aux->attach_btf_id;
|
|
|
|
|
key = bpf_trampoline_compute_key(NULL, prog->aux->attach_btf, btf_id);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
2020-09-29 12:45:51 +00:00
|
|
|
|
|
|
|
|
if (!prog->aux->dst_trampoline ||
|
|
|
|
|
(key && key != prog->aux->dst_trampoline->key)) {
|
|
|
|
|
/* If there is no saved target, or the specified target is
|
|
|
|
|
* different from the destination specified at load time, we
|
|
|
|
|
* need a new trampoline and a check for compatibility
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_attach_target_info tgt_info = {};
|
|
|
|
|
|
|
|
|
|
err = bpf_check_attach_target(NULL, prog, tgt_prog, btf_id,
|
|
|
|
|
&tgt_info);
|
|
|
|
|
if (err)
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
2023-03-10 07:40:59 +00:00
|
|
|
if (tgt_info.tgt_mod) {
|
|
|
|
|
module_put(prog->aux->mod);
|
|
|
|
|
prog->aux->mod = tgt_info.tgt_mod;
|
|
|
|
|
}
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
tr = bpf_trampoline_get(key, &tgt_info);
|
|
|
|
|
if (!tr) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
/* The caller didn't specify a target, or the target was the
|
|
|
|
|
* same as the destination supplied during program load. This
|
|
|
|
|
* means we can reuse the trampoline and reference from program
|
|
|
|
|
* load time, and there is no need to allocate a new one. This
|
|
|
|
|
* can only happen once for any program, as the saved values in
|
|
|
|
|
* prog->aux are cleared below.
|
|
|
|
|
*/
|
|
|
|
|
tr = prog->aux->dst_trampoline;
|
|
|
|
|
tgt_prog = prog->aux->dst_prog;
|
|
|
|
|
}
|
2020-09-29 12:45:50 +00:00
|
|
|
|
2022-05-10 20:59:19 +00:00
|
|
|
err = bpf_link_prime(&link->link.link, &link_primer);
|
2020-09-29 12:45:50 +00:00
|
|
|
if (err)
|
|
|
|
|
goto out_unlock;
|
2019-11-14 18:57:04 +00:00
|
|
|
|
2022-05-10 20:59:19 +00:00
|
|
|
err = bpf_trampoline_link_prog(&link->link, tr);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
if (err) {
|
2020-04-29 00:16:06 +00:00
|
|
|
bpf_link_cleanup(&link_primer);
|
2020-09-29 12:45:50 +00:00
|
|
|
link = NULL;
|
|
|
|
|
goto out_unlock;
|
2019-11-14 18:57:04 +00:00
|
|
|
}
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
|
2020-09-29 12:45:50 +00:00
|
|
|
link->tgt_prog = tgt_prog;
|
|
|
|
|
link->trampoline = tr;
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
/* Always clear the trampoline and target prog from prog->aux to make
|
|
|
|
|
* sure the original attach destination is not kept alive after a
|
|
|
|
|
* program is (re-)attached to another target.
|
|
|
|
|
*/
|
|
|
|
|
if (prog->aux->dst_prog &&
|
|
|
|
|
(tgt_prog_fd || tr != prog->aux->dst_trampoline))
|
|
|
|
|
/* got extra prog ref from syscall, or attaching to different prog */
|
|
|
|
|
bpf_prog_put(prog->aux->dst_prog);
|
|
|
|
|
if (prog->aux->dst_trampoline && tr != prog->aux->dst_trampoline)
|
|
|
|
|
/* we allocated a new trampoline, so free the old one */
|
|
|
|
|
bpf_trampoline_put(prog->aux->dst_trampoline);
|
|
|
|
|
|
2020-09-29 12:45:50 +00:00
|
|
|
prog->aux->dst_prog = NULL;
|
|
|
|
|
prog->aux->dst_trampoline = NULL;
|
|
|
|
|
mutex_unlock(&prog->aux->dst_mutex);
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
return bpf_link_settle(&link_primer);
|
2020-09-29 12:45:50 +00:00
|
|
|
out_unlock:
|
2020-09-29 12:45:51 +00:00
|
|
|
if (tr && tr != prog->aux->dst_trampoline)
|
|
|
|
|
bpf_trampoline_put(tr);
|
2020-09-29 12:45:50 +00:00
|
|
|
mutex_unlock(&prog->aux->dst_mutex);
|
|
|
|
|
kfree(link);
|
2019-11-14 18:57:04 +00:00
|
|
|
out_put_prog:
|
2020-09-29 12:45:51 +00:00
|
|
|
if (tgt_prog_fd && tgt_prog)
|
|
|
|
|
bpf_prog_put(tgt_prog);
|
2019-11-14 18:57:04 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
struct bpf_raw_tp_link {
|
|
|
|
|
struct bpf_link link;
|
2018-03-28 19:05:37 +00:00
|
|
|
struct bpf_raw_event_map *btp;
|
|
|
|
|
};
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
static void bpf_raw_tp_link_release(struct bpf_link *link)
|
2018-03-28 19:05:37 +00:00
|
|
|
{
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
struct bpf_raw_tp_link *raw_tp =
|
|
|
|
|
container_of(link, struct bpf_raw_tp_link, link);
|
2018-03-28 19:05:37 +00:00
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
bpf_probe_unregister(raw_tp->btp, raw_tp->link.prog);
|
2018-12-13 00:42:37 +00:00
|
|
|
bpf_put_raw_tracepoint(raw_tp->btp);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_raw_tp_link_dealloc(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_raw_tp_link *raw_tp =
|
|
|
|
|
container_of(link, struct bpf_raw_tp_link, link);
|
|
|
|
|
|
2018-03-28 19:05:37 +00:00
|
|
|
kfree(raw_tp);
|
|
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:08 +00:00
|
|
|
static void bpf_raw_tp_link_show_fdinfo(const struct bpf_link *link,
|
|
|
|
|
struct seq_file *seq)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_raw_tp_link *raw_tp_link =
|
|
|
|
|
container_of(link, struct bpf_raw_tp_link, link);
|
|
|
|
|
|
|
|
|
|
seq_printf(seq,
|
|
|
|
|
"tp_name:\t%s\n",
|
|
|
|
|
raw_tp_link->btp->tp->name);
|
|
|
|
|
}
|
|
|
|
|
|
2023-07-09 02:56:27 +00:00
|
|
|
static int bpf_copy_to_user(char __user *ubuf, const char *buf, u32 ulen,
|
|
|
|
|
u32 len)
|
|
|
|
|
{
|
|
|
|
|
if (ulen >= len + 1) {
|
|
|
|
|
if (copy_to_user(ubuf, buf, len + 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
} else {
|
|
|
|
|
char zero = '\0';
|
|
|
|
|
|
|
|
|
|
if (copy_to_user(ubuf, buf, ulen - 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
if (put_user(zero, ubuf + ulen - 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
return -ENOSPC;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:08 +00:00
|
|
|
static int bpf_raw_tp_link_fill_link_info(const struct bpf_link *link,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_raw_tp_link *raw_tp_link =
|
|
|
|
|
container_of(link, struct bpf_raw_tp_link, link);
|
|
|
|
|
char __user *ubuf = u64_to_user_ptr(info->raw_tracepoint.tp_name);
|
|
|
|
|
const char *tp_name = raw_tp_link->btp->tp->name;
|
|
|
|
|
u32 ulen = info->raw_tracepoint.tp_name_len;
|
|
|
|
|
size_t tp_len = strlen(tp_name);
|
|
|
|
|
|
2020-08-21 19:10:54 +00:00
|
|
|
if (!ulen ^ !ubuf)
|
2020-04-29 00:16:08 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
info->raw_tracepoint.tp_name_len = tp_len + 1;
|
|
|
|
|
|
|
|
|
|
if (!ubuf)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2023-07-09 02:56:27 +00:00
|
|
|
return bpf_copy_to_user(ubuf, tp_name, ulen, tp_len);
|
2020-04-29 00:16:08 +00:00
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
static const struct bpf_link_ops bpf_raw_tp_link_lops = {
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
.release = bpf_raw_tp_link_release,
|
bpf: support deferring bpf_link dealloc to after RCU grace period
BPF link for some program types is passed as a "context" which can be
used by those BPF programs to look up additional information. E.g., for
multi-kprobes and multi-uprobes, link is used to fetch BPF cookie values.
Because of this runtime dependency, when bpf_link refcnt drops to zero
there could still be active BPF programs running accessing link data.
This patch adds generic support to defer bpf_link dealloc callback to
after RCU GP, if requested. This is done by exposing two different
deallocation callbacks, one synchronous and one deferred. If deferred
one is provided, bpf_link_free() will schedule dealloc_deferred()
callback to happen after RCU GP.
BPF is using two flavors of RCU: "classic" non-sleepable one and RCU
tasks trace one. The latter is used when sleepable BPF programs are
used. bpf_link_free() accommodates that by checking underlying BPF
program's sleepable flag, and goes either through normal RCU GP only for
non-sleepable, or through RCU tasks trace GP *and* then normal RCU GP
(taking into account rcu_trace_implies_rcu_gp() optimization), if BPF
program is sleepable.
We use this for multi-kprobe and multi-uprobe links, which dereference
link during program run. We also preventively switch raw_tp link to use
deferred dealloc callback, as upcoming changes in bpf-next tree expose
raw_tp link data (specifically, cookie value) to BPF program at runtime
as well.
Fixes: 0dcac2725406 ("bpf: Add multi kprobe link")
Fixes: 89ae89f53d20 ("bpf: Add multi uprobe link")
Reported-by: syzbot+981935d9485a560bfbcb@syzkaller.appspotmail.com
Reported-by: syzbot+2cb5a6c573e98db598cc@syzkaller.appspotmail.com
Reported-by: syzbot+62d8b26793e8a2bd0516@syzkaller.appspotmail.com
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20240328052426.3042617-2-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-28 05:24:26 +00:00
|
|
|
.dealloc_deferred = bpf_raw_tp_link_dealloc,
|
2020-04-29 00:16:08 +00:00
|
|
|
.show_fdinfo = bpf_raw_tp_link_show_fdinfo,
|
|
|
|
|
.fill_link_info = bpf_raw_tp_link_fill_link_info,
|
2018-03-28 19:05:37 +00:00
|
|
|
};
|
|
|
|
|
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
#ifdef CONFIG_PERF_EVENTS
|
|
|
|
|
struct bpf_perf_link {
|
|
|
|
|
struct bpf_link link;
|
|
|
|
|
struct file *perf_file;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static void bpf_perf_link_release(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_perf_link *perf_link = container_of(link, struct bpf_perf_link, link);
|
|
|
|
|
struct perf_event *event = perf_link->perf_file->private_data;
|
|
|
|
|
|
|
|
|
|
perf_event_free_bpf_prog(event);
|
|
|
|
|
fput(perf_link->perf_file);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_perf_link_dealloc(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_perf_link *perf_link = container_of(link, struct bpf_perf_link, link);
|
|
|
|
|
|
|
|
|
|
kfree(perf_link);
|
|
|
|
|
}
|
|
|
|
|
|
2023-07-09 02:56:28 +00:00
|
|
|
static int bpf_perf_link_fill_common(const struct perf_event *event,
|
|
|
|
|
char __user *uname, u32 ulen,
|
|
|
|
|
u64 *probe_offset, u64 *probe_addr,
|
2023-09-20 21:31:39 +00:00
|
|
|
u32 *fd_type, unsigned long *missed)
|
2023-07-09 02:56:28 +00:00
|
|
|
{
|
|
|
|
|
const char *buf;
|
|
|
|
|
u32 prog_id;
|
|
|
|
|
size_t len;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (!ulen ^ !uname)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
err = bpf_get_perf_event_info(event, &prog_id, fd_type, &buf,
|
2023-09-20 21:31:39 +00:00
|
|
|
probe_offset, probe_addr, missed);
|
2023-07-09 02:56:28 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
2023-08-13 14:18:59 +00:00
|
|
|
if (!uname)
|
|
|
|
|
return 0;
|
2023-07-09 02:56:28 +00:00
|
|
|
if (buf) {
|
|
|
|
|
len = strlen(buf);
|
|
|
|
|
err = bpf_copy_to_user(uname, buf, ulen, len);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
} else {
|
|
|
|
|
char zero = '\0';
|
|
|
|
|
|
|
|
|
|
if (put_user(zero, uname))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_KPROBE_EVENTS
|
|
|
|
|
static int bpf_perf_link_fill_kprobe(const struct perf_event *event,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
2023-09-20 21:31:39 +00:00
|
|
|
unsigned long missed;
|
2023-07-09 02:56:28 +00:00
|
|
|
char __user *uname;
|
|
|
|
|
u64 addr, offset;
|
|
|
|
|
u32 ulen, type;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
uname = u64_to_user_ptr(info->perf_event.kprobe.func_name);
|
|
|
|
|
ulen = info->perf_event.kprobe.name_len;
|
|
|
|
|
err = bpf_perf_link_fill_common(event, uname, ulen, &offset, &addr,
|
2023-09-20 21:31:39 +00:00
|
|
|
&type, &missed);
|
2023-07-09 02:56:28 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
if (type == BPF_FD_TYPE_KRETPROBE)
|
|
|
|
|
info->perf_event.type = BPF_PERF_EVENT_KRETPROBE;
|
|
|
|
|
else
|
|
|
|
|
info->perf_event.type = BPF_PERF_EVENT_KPROBE;
|
|
|
|
|
|
|
|
|
|
info->perf_event.kprobe.offset = offset;
|
2023-09-20 21:31:39 +00:00
|
|
|
info->perf_event.kprobe.missed = missed;
|
2023-07-09 02:56:28 +00:00
|
|
|
if (!kallsyms_show_value(current_cred()))
|
|
|
|
|
addr = 0;
|
|
|
|
|
info->perf_event.kprobe.addr = addr;
|
2024-01-19 11:04:58 +00:00
|
|
|
info->perf_event.kprobe.cookie = event->bpf_cookie;
|
2023-07-09 02:56:28 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_UPROBE_EVENTS
|
|
|
|
|
static int bpf_perf_link_fill_uprobe(const struct perf_event *event,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
char __user *uname;
|
|
|
|
|
u64 addr, offset;
|
|
|
|
|
u32 ulen, type;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
uname = u64_to_user_ptr(info->perf_event.uprobe.file_name);
|
|
|
|
|
ulen = info->perf_event.uprobe.name_len;
|
|
|
|
|
err = bpf_perf_link_fill_common(event, uname, ulen, &offset, &addr,
|
2023-09-20 21:31:39 +00:00
|
|
|
&type, NULL);
|
2023-07-09 02:56:28 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
|
|
|
|
|
if (type == BPF_FD_TYPE_URETPROBE)
|
|
|
|
|
info->perf_event.type = BPF_PERF_EVENT_URETPROBE;
|
|
|
|
|
else
|
|
|
|
|
info->perf_event.type = BPF_PERF_EVENT_UPROBE;
|
|
|
|
|
info->perf_event.uprobe.offset = offset;
|
2024-01-19 11:04:58 +00:00
|
|
|
info->perf_event.uprobe.cookie = event->bpf_cookie;
|
2023-07-09 02:56:28 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
static int bpf_perf_link_fill_probe(const struct perf_event *event,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
#ifdef CONFIG_KPROBE_EVENTS
|
|
|
|
|
if (event->tp_event->flags & TRACE_EVENT_FL_KPROBE)
|
|
|
|
|
return bpf_perf_link_fill_kprobe(event, info);
|
|
|
|
|
#endif
|
|
|
|
|
#ifdef CONFIG_UPROBE_EVENTS
|
|
|
|
|
if (event->tp_event->flags & TRACE_EVENT_FL_UPROBE)
|
|
|
|
|
return bpf_perf_link_fill_uprobe(event, info);
|
|
|
|
|
#endif
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_perf_link_fill_tracepoint(const struct perf_event *event,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
char __user *uname;
|
|
|
|
|
u32 ulen;
|
|
|
|
|
|
|
|
|
|
uname = u64_to_user_ptr(info->perf_event.tracepoint.tp_name);
|
|
|
|
|
ulen = info->perf_event.tracepoint.name_len;
|
|
|
|
|
info->perf_event.type = BPF_PERF_EVENT_TRACEPOINT;
|
2024-01-19 11:04:58 +00:00
|
|
|
info->perf_event.tracepoint.cookie = event->bpf_cookie;
|
2023-09-20 21:31:39 +00:00
|
|
|
return bpf_perf_link_fill_common(event, uname, ulen, NULL, NULL, NULL, NULL);
|
2023-07-09 02:56:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_perf_link_fill_perf_event(const struct perf_event *event,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
info->perf_event.event.type = event->attr.type;
|
|
|
|
|
info->perf_event.event.config = event->attr.config;
|
2024-01-19 11:04:58 +00:00
|
|
|
info->perf_event.event.cookie = event->bpf_cookie;
|
2023-07-09 02:56:28 +00:00
|
|
|
info->perf_event.type = BPF_PERF_EVENT_EVENT;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_perf_link_fill_link_info(const struct bpf_link *link,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_perf_link *perf_link;
|
|
|
|
|
const struct perf_event *event;
|
|
|
|
|
|
|
|
|
|
perf_link = container_of(link, struct bpf_perf_link, link);
|
|
|
|
|
event = perf_get_event(perf_link->perf_file);
|
|
|
|
|
if (IS_ERR(event))
|
|
|
|
|
return PTR_ERR(event);
|
|
|
|
|
|
|
|
|
|
switch (event->prog->type) {
|
|
|
|
|
case BPF_PROG_TYPE_PERF_EVENT:
|
|
|
|
|
return bpf_perf_link_fill_perf_event(event, info);
|
|
|
|
|
case BPF_PROG_TYPE_TRACEPOINT:
|
|
|
|
|
return bpf_perf_link_fill_tracepoint(event, info);
|
|
|
|
|
case BPF_PROG_TYPE_KPROBE:
|
|
|
|
|
return bpf_perf_link_fill_probe(event, info);
|
|
|
|
|
default:
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
static const struct bpf_link_ops bpf_perf_link_lops = {
|
|
|
|
|
.release = bpf_perf_link_release,
|
|
|
|
|
.dealloc = bpf_perf_link_dealloc,
|
2023-07-09 02:56:28 +00:00
|
|
|
.fill_link_info = bpf_perf_link_fill_link_info,
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int bpf_perf_link_attach(const union bpf_attr *attr, struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link_primer link_primer;
|
|
|
|
|
struct bpf_perf_link *link;
|
|
|
|
|
struct perf_event *event;
|
|
|
|
|
struct file *perf_file;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (attr->link_create.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
perf_file = perf_event_get(attr->link_create.target_fd);
|
|
|
|
|
if (IS_ERR(perf_file))
|
|
|
|
|
return PTR_ERR(perf_file);
|
|
|
|
|
|
|
|
|
|
link = kzalloc(sizeof(*link), GFP_USER);
|
|
|
|
|
if (!link) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_put_file;
|
|
|
|
|
}
|
|
|
|
|
bpf_link_init(&link->link, BPF_LINK_TYPE_PERF_EVENT, &bpf_perf_link_lops, prog);
|
|
|
|
|
link->perf_file = perf_file;
|
|
|
|
|
|
|
|
|
|
err = bpf_link_prime(&link->link, &link_primer);
|
|
|
|
|
if (err) {
|
|
|
|
|
kfree(link);
|
|
|
|
|
goto out_put_file;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
event = perf_file->private_data;
|
bpf: Allow to specify user-provided bpf_cookie for BPF perf links
Add ability for users to specify custom u64 value (bpf_cookie) when creating
BPF link for perf_event-backed BPF programs (kprobe/uprobe, perf_event,
tracepoints).
This is useful for cases when the same BPF program is used for attaching and
processing invocation of different tracepoints/kprobes/uprobes in a generic
fashion, but such that each invocation is distinguished from each other (e.g.,
BPF program can look up additional information associated with a specific
kernel function without having to rely on function IP lookups). This enables
new use cases to be implemented simply and efficiently that previously were
possible only through code generation (and thus multiple instances of almost
identical BPF program) or compilation at runtime (BCC-style) on target hosts
(even more expensive resource-wise). For uprobes it is not even possible in
some cases to know function IP before hand (e.g., when attaching to shared
library without PID filtering, in which case base load address is not known
for a library).
This is done by storing u64 bpf_cookie in struct bpf_prog_array_item,
corresponding to each attached and run BPF program. Given cgroup BPF programs
already use two 8-byte pointers for their needs and cgroup BPF programs don't
have (yet?) support for bpf_cookie, reuse that space through union of
cgroup_storage and new bpf_cookie field.
Make it available to kprobe/tracepoint BPF programs through bpf_trace_run_ctx.
This is set by BPF_PROG_RUN_ARRAY, used by kprobe/uprobe/tracepoint BPF
program execution code, which luckily is now also split from
BPF_PROG_RUN_ARRAY_CG. This run context will be utilized by a new BPF helper
giving access to this user-provided cookie value from inside a BPF program.
Generic perf_event BPF programs will access this value from perf_event itself
through passed in BPF program context.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-6-andrii@kernel.org
2021-08-15 07:05:58 +00:00
|
|
|
err = perf_event_set_bpf_prog(event, prog, attr->link_create.perf_event.bpf_cookie);
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
if (err) {
|
|
|
|
|
bpf_link_cleanup(&link_primer);
|
|
|
|
|
goto out_put_file;
|
|
|
|
|
}
|
|
|
|
|
/* perf_event_set_bpf_prog() doesn't take its own refcnt on prog */
|
|
|
|
|
bpf_prog_inc(prog);
|
|
|
|
|
|
|
|
|
|
return bpf_link_settle(&link_primer);
|
|
|
|
|
|
|
|
|
|
out_put_file:
|
|
|
|
|
fput(perf_file);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
2022-03-16 12:24:09 +00:00
|
|
|
#else
|
|
|
|
|
static int bpf_perf_link_attach(const union bpf_attr *attr, struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
}
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
#endif /* CONFIG_PERF_EVENTS */
|
|
|
|
|
|
2022-04-21 03:39:43 +00:00
|
|
|
static int bpf_raw_tp_link_attach(struct bpf_prog *prog,
|
|
|
|
|
const char __user *user_tp_name)
|
2018-03-28 19:05:37 +00:00
|
|
|
{
|
2020-04-29 00:16:06 +00:00
|
|
|
struct bpf_link_primer link_primer;
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
struct bpf_raw_tp_link *link;
|
2018-03-28 19:05:37 +00:00
|
|
|
struct bpf_raw_event_map *btp;
|
2019-10-16 03:25:01 +00:00
|
|
|
const char *tp_name;
|
|
|
|
|
char buf[128];
|
2020-04-29 00:16:06 +00:00
|
|
|
int err;
|
2018-03-28 19:05:37 +00:00
|
|
|
|
2020-03-29 00:43:52 +00:00
|
|
|
switch (prog->type) {
|
|
|
|
|
case BPF_PROG_TYPE_TRACING:
|
|
|
|
|
case BPF_PROG_TYPE_EXT:
|
|
|
|
|
case BPF_PROG_TYPE_LSM:
|
2022-04-21 03:39:43 +00:00
|
|
|
if (user_tp_name)
|
2019-11-14 18:57:04 +00:00
|
|
|
/* The attach point for this category of programs
|
|
|
|
|
* should be specified via btf_id during program load.
|
2019-10-16 03:25:01 +00:00
|
|
|
*/
|
2022-04-21 03:39:43 +00:00
|
|
|
return -EINVAL;
|
2020-03-29 00:43:52 +00:00
|
|
|
if (prog->type == BPF_PROG_TYPE_TRACING &&
|
|
|
|
|
prog->expected_attach_type == BPF_TRACE_RAW_TP) {
|
2019-11-14 18:57:04 +00:00
|
|
|
tp_name = prog->aux->attach_func_name;
|
2020-03-29 00:43:52 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2022-05-10 20:59:21 +00:00
|
|
|
return bpf_tracing_prog_attach(prog, 0, 0, 0);
|
2020-03-29 00:43:52 +00:00
|
|
|
case BPF_PROG_TYPE_RAW_TRACEPOINT:
|
|
|
|
|
case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE:
|
2022-04-21 03:39:43 +00:00
|
|
|
if (strncpy_from_user(buf, user_tp_name, sizeof(buf) - 1) < 0)
|
|
|
|
|
return -EFAULT;
|
2019-10-16 03:25:01 +00:00
|
|
|
buf[sizeof(buf) - 1] = 0;
|
|
|
|
|
tp_name = buf;
|
2020-03-29 00:43:52 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
2022-04-21 03:39:43 +00:00
|
|
|
return -EINVAL;
|
2019-10-16 03:25:01 +00:00
|
|
|
}
|
2018-03-28 19:05:37 +00:00
|
|
|
|
2018-12-13 00:42:37 +00:00
|
|
|
btp = bpf_get_raw_tracepoint(tp_name);
|
2022-04-21 03:39:43 +00:00
|
|
|
if (!btp)
|
|
|
|
|
return -ENOENT;
|
2018-03-28 19:05:37 +00:00
|
|
|
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
link = kzalloc(sizeof(*link), GFP_USER);
|
|
|
|
|
if (!link) {
|
2018-12-13 00:42:37 +00:00
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_put_btp;
|
|
|
|
|
}
|
2020-04-29 00:16:08 +00:00
|
|
|
bpf_link_init(&link->link, BPF_LINK_TYPE_RAW_TRACEPOINT,
|
|
|
|
|
&bpf_raw_tp_link_lops, prog);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
link->btp = btp;
|
2018-03-28 19:05:37 +00:00
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
err = bpf_link_prime(&link->link, &link_primer);
|
|
|
|
|
if (err) {
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
kfree(link);
|
|
|
|
|
goto out_put_btp;
|
|
|
|
|
}
|
2018-03-28 19:05:37 +00:00
|
|
|
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
err = bpf_probe_register(link->btp, prog);
|
|
|
|
|
if (err) {
|
2020-04-29 00:16:06 +00:00
|
|
|
bpf_link_cleanup(&link_primer);
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
goto out_put_btp;
|
2018-03-28 19:05:37 +00:00
|
|
|
}
|
bpf: Add bpf_link_new_file that doesn't install FD
Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.
After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.
Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
2020-03-09 23:10:51 +00:00
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
return bpf_link_settle(&link_primer);
|
2018-03-28 19:05:37 +00:00
|
|
|
|
2018-12-13 00:42:37 +00:00
|
|
|
out_put_btp:
|
|
|
|
|
bpf_put_raw_tracepoint(btp);
|
2018-03-28 19:05:37 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2022-04-21 03:39:43 +00:00
|
|
|
#define BPF_RAW_TRACEPOINT_OPEN_LAST_FIELD raw_tracepoint.prog_fd
|
|
|
|
|
|
|
|
|
|
static int bpf_raw_tracepoint_open(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_RAW_TRACEPOINT_OPEN))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
prog = bpf_prog_get(attr->raw_tracepoint.prog_fd);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
fd = bpf_raw_tp_link_attach(prog, u64_to_user_ptr(attr->raw_tracepoint.name));
|
|
|
|
|
if (fd < 0)
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return fd;
|
|
|
|
|
}
|
|
|
|
|
|
2020-03-25 06:57:42 +00:00
|
|
|
static enum bpf_prog_type
|
|
|
|
|
attach_type_to_prog_type(enum bpf_attach_type attach_type)
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
{
|
2020-03-25 06:57:42 +00:00
|
|
|
switch (attach_type) {
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
case BPF_CGROUP_INET_INGRESS:
|
|
|
|
|
case BPF_CGROUP_INET_EGRESS:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_SKB;
|
2016-12-01 16:48:04 +00:00
|
|
|
case BPF_CGROUP_INET_SOCK_CREATE:
|
2020-07-06 23:01:25 +00:00
|
|
|
case BPF_CGROUP_INET_SOCK_RELEASE:
|
2018-03-30 22:08:07 +00:00
|
|
|
case BPF_CGROUP_INET4_POST_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_POST_BIND:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_SOCK;
|
2018-03-30 22:08:02 +00:00
|
|
|
case BPF_CGROUP_INET4_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_BIND:
|
2018-03-30 22:08:05 +00:00
|
|
|
case BPF_CGROUP_INET4_CONNECT:
|
|
|
|
|
case BPF_CGROUP_INET6_CONNECT:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_CONNECT:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETPEERNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETPEERNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETPEERNAME:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETSOCKNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETSOCKNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETSOCKNAME:
|
2018-05-25 15:55:23 +00:00
|
|
|
case BPF_CGROUP_UDP4_SENDMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_SENDMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_SENDMSG:
|
bpf: fix unconnected udp hooks
Intention of cgroup bind/connect/sendmsg BPF hooks is to act transparently
to applications as also stated in original motivation in 7828f20e3779 ("Merge
branch 'bpf-cgroup-bind-connect'"). When recently integrating the latter
two hooks into Cilium to enable host based load-balancing with Kubernetes,
I ran into the issue that pods couldn't start up as DNS got broken. Kubernetes
typically sets up DNS as a service and is thus subject to load-balancing.
Upon further debugging, it turns out that the cgroupv2 sendmsg BPF hooks API
is currently insufficient and thus not usable as-is for standard applications
shipped with most distros. To break down the issue we ran into with a simple
example:
# cat /etc/resolv.conf
nameserver 147.75.207.207
nameserver 147.75.207.208
For the purpose of a simple test, we set up above IPs as service IPs and
transparently redirect traffic to a different DNS backend server for that
node:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
The attached BPF program is basically selecting one of the backends if the
service IP/port matches on the cgroup hook. DNS breaks here, because the
hooks are not transparent enough to applications which have built-in msg_name
address checks:
# nslookup 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
;; connection timed out; no servers could be reached
# dig 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; connection timed out; no servers could be reached
For comparison, if none of the service IPs is used, and we tell nslookup
to use 8.8.8.8 directly it works just fine, of course:
# nslookup 1.1.1.1 8.8.8.8
1.1.1.1.in-addr.arpa name = one.one.one.one.
In order to fix this and thus act more transparent to the application,
this needs reverse translation on recvmsg() side. A minimal fix for this
API is to add similar recvmsg() hooks behind the BPF cgroups static key
such that the program can track state and replace the current sockaddr_in{,6}
with the original service IP. From BPF side, this basically tracks the
service tuple plus socket cookie in an LRU map where the reverse NAT can
then be retrieved via map value as one example. Side-note: the BPF cgroups
static key should be converted to a per-hook static key in future.
Same example after this fix:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
Lookups work fine now:
# nslookup 1.1.1.1
1.1.1.1.in-addr.arpa name = one.one.one.one.
Authoritative answers can be found from:
# dig 1.1.1.1
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 51550
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 512
;; QUESTION SECTION:
;1.1.1.1. IN A
;; AUTHORITY SECTION:
. 23426 IN SOA a.root-servers.net. nstld.verisign-grs.com. 2019052001 1800 900 604800 86400
;; Query time: 17 msec
;; SERVER: 147.75.207.207#53(147.75.207.207)
;; WHEN: Tue May 21 12:59:38 UTC 2019
;; MSG SIZE rcvd: 111
And from an actual packet level it shows that we're using the back end
server when talking via 147.75.207.20{7,8} front end:
# tcpdump -i any udp
[...]
12:59:52.698732 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.698735 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
[...]
In order to be flexible and to have same semantics as in sendmsg BPF
programs, we only allow return codes in [1,1] range. In the sendmsg case
the program is called if msg->msg_name is present which can be the case
in both, connected and unconnected UDP.
The former only relies on the sockaddr_in{,6} passed via connect(2) if
passed msg->msg_name was NULL. Therefore, on recvmsg side, we act in similar
way to call into the BPF program whenever a non-NULL msg->msg_name was
passed independent of sk->sk_state being TCP_ESTABLISHED or not. Note
that for TCP case, the msg->msg_name is ignored in the regular recvmsg
path and therefore not relevant.
For the case of ip{,v6}_recv_error() paths, picked up via MSG_ERRQUEUE,
the hook is not called. This is intentional as it aligns with the same
semantics as in case of TCP cgroup BPF hooks right now. This might be
better addressed in future through a different bpf_attach_type such
that this case can be distinguished from the regular recvmsg paths,
for example.
Fixes: 1cedee13d25a ("bpf: Hooks for sys_sendmsg")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrey Ignatov <rdna@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Martynas Pumputis <m@lambda.lt>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-06 23:48:57 +00:00
|
|
|
case BPF_CGROUP_UDP4_RECVMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_RECVMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_RECVMSG:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_SOCK_ADDR;
|
bpf: BPF support for sock_ops
Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding
struct that allows BPF programs of this type to access some of the
socket's fields (such as IP addresses, ports, etc.). It uses the
existing bpf cgroups infrastructure so the programs can be attached per
cgroup with full inheritance support. The program will be called at
appropriate times to set relevant connections parameters such as buffer
sizes, SYN and SYN-ACK RTOs, etc., based on connection information such
as IP addresses, port numbers, etc.
Alghough there are already 3 mechanisms to set parameters (sysctls,
route metrics and setsockopts), this new mechanism provides some
distinct advantages. Unlike sysctls, it can set parameters per
connection. In contrast to route metrics, it can also use port numbers
and information provided by a user level program. In addition, it could
set parameters probabilistically for evaluation purposes (i.e. do
something different on 10% of the flows and compare results with the
other 90% of the flows). Also, in cases where IPv6 addresses contain
geographic information, the rules to make changes based on the distance
(or RTT) between the hosts are much easier than route metric rules and
can be global. Finally, unlike setsockopt, it oes not require
application changes and it can be updated easily at any time.
Although the bpf cgroup framework already contains a sock related
program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type
(BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called
only once during the connections's lifetime. In contrast, the new
program type will be called multiple times from different places in the
network stack code. For example, before sending SYN and SYN-ACKs to set
an appropriate timeout, when the connection is established to set
congestion control, etc. As a result it has "op" field to specify the
type of operation requested.
The purpose of this new program type is to simplify setting connection
parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is
easy to use facebook's internal IPv6 addresses to determine if both hosts
of a connection are in the same datacenter. Therefore, it is easy to
write a BPF program to choose a small SYN RTO value when both hosts are
in the same datacenter.
This patch only contains the framework to support the new BPF program
type, following patches add the functionality to set various connection
parameters.
This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS
and a new bpf syscall command to load a new program of this type:
BPF_PROG_LOAD_SOCKET_OPS.
Two new corresponding structs (one for the kernel one for the user/BPF
program):
/* kernel version */
struct bpf_sock_ops_kern {
struct sock *sk;
__u32 op;
union {
__u32 reply;
__u32 replylong[4];
};
};
/* user version
* Some fields are in network byte order reflecting the sock struct
* Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to
* convert them to host byte order.
*/
struct bpf_sock_ops {
__u32 op;
union {
__u32 reply;
__u32 replylong[4];
};
__u32 family;
__u32 remote_ip4; /* In network byte order */
__u32 local_ip4; /* In network byte order */
__u32 remote_ip6[4]; /* In network byte order */
__u32 local_ip6[4]; /* In network byte order */
__u32 remote_port; /* In network byte order */
__u32 local_port; /* In host byte horder */
};
Currently there are two types of ops. The first type expects the BPF
program to return a value which is then used by the caller (or a
negative value to indicate the operation is not supported). The second
type expects state changes to be done by the BPF program, for example
through a setsockopt BPF helper function, and they ignore the return
value.
The reply fields of the bpf_sockt_ops struct are there in case a bpf
program needs to return a value larger than an integer.
Signed-off-by: Lawrence Brakmo <brakmo@fb.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 03:02:40 +00:00
|
|
|
case BPF_CGROUP_SOCK_OPS:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_SOCK_OPS;
|
2017-11-05 13:15:32 +00:00
|
|
|
case BPF_CGROUP_DEVICE:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_DEVICE;
|
bpf: create tcp_bpf_ulp allowing BPF to monitor socket TX/RX data
This implements a BPF ULP layer to allow policy enforcement and
monitoring at the socket layer. In order to support this a new
program type BPF_PROG_TYPE_SK_MSG is used to run the policy at
the sendmsg/sendpage hook. To attach the policy to sockets a
sockmap is used with a new program attach type BPF_SK_MSG_VERDICT.
Similar to previous sockmap usages when a sock is added to a
sockmap, via a map update, if the map contains a BPF_SK_MSG_VERDICT
program type attached then the BPF ULP layer is created on the
socket and the attached BPF_PROG_TYPE_SK_MSG program is run for
every msg in sendmsg case and page/offset in sendpage case.
BPF_PROG_TYPE_SK_MSG Semantics/API:
BPF_PROG_TYPE_SK_MSG supports only two return codes SK_PASS and
SK_DROP. Returning SK_DROP free's the copied data in the sendmsg
case and in the sendpage case leaves the data untouched. Both cases
return -EACESS to the user. Returning SK_PASS will allow the msg to
be sent.
In the sendmsg case data is copied into kernel space buffers before
running the BPF program. The kernel space buffers are stored in a
scatterlist object where each element is a kernel memory buffer.
Some effort is made to coalesce data from the sendmsg call here.
For example a sendmsg call with many one byte iov entries will
likely be pushed into a single entry. The BPF program is run with
data pointers (start/end) pointing to the first sg element.
In the sendpage case data is not copied. We opt not to copy the
data by default here, because the BPF infrastructure does not
know what bytes will be needed nor when they will be needed. So
copying all bytes may be wasteful. Because of this the initial
start/end data pointers are (0,0). Meaning no data can be read or
written. This avoids reading data that may be modified by the
user. A new helper is added later in this series if reading and
writing the data is needed. The helper call will do a copy by
default so that the page is exclusively owned by the BPF call.
The verdict from the BPF_PROG_TYPE_SK_MSG applies to the entire msg
in the sendmsg() case and the entire page/offset in the sendpage case.
This avoids ambiguity on how to handle mixed return codes in the
sendmsg case. Again a helper is added later in the series if
a verdict needs to apply to multiple system calls and/or only
a subpart of the currently being processed message.
The helper msg_redirect_map() can be used to select the socket to
send the data on. This is used similar to existing redirect use
cases. This allows policy to redirect msgs.
Pseudo code simple example:
The basic logic to attach a program to a socket is as follows,
// load the programs
bpf_prog_load(SOCKMAP_TCP_MSG_PROG, BPF_PROG_TYPE_SK_MSG,
&obj, &msg_prog);
// lookup the sockmap
bpf_map_msg = bpf_object__find_map_by_name(obj, "my_sock_map");
// get fd for sockmap
map_fd_msg = bpf_map__fd(bpf_map_msg);
// attach program to sockmap
bpf_prog_attach(msg_prog, map_fd_msg, BPF_SK_MSG_VERDICT, 0);
Adding sockets to the map is done in the normal way,
// Add a socket 'fd' to sockmap at location 'i'
bpf_map_update_elem(map_fd_msg, &i, fd, BPF_ANY);
After the above any socket attached to "my_sock_map", in this case
'fd', will run the BPF msg verdict program (msg_prog) on every
sendmsg and sendpage system call.
For a complete example see BPF selftests or sockmap samples.
Implementation notes:
It seemed the simplest, to me at least, to use a refcnt to ensure
psock is not lost across the sendmsg copy into the sg, the bpf program
running on the data in sg_data, and the final pass to the TCP stack.
Some performance testing may show a better method to do this and avoid
the refcnt cost, but for now use the simpler method.
Another item that will come after basic support is in place is
supporting MSG_MORE flag. At the moment we call sendpages even if
the MSG_MORE flag is set. An enhancement would be to collect the
pages into a larger scatterlist and pass down the stack. Notice that
bpf_tcp_sendmsg() could support this with some additional state saved
across sendmsg calls. I built the code to support this without having
to do refactoring work. Other features TBD include ZEROCOPY and the
TCP_RECV_QUEUE/TCP_NO_QUEUE support. This will follow initial series
shortly.
Future work could improve size limits on the scatterlist rings used
here. Currently, we use MAX_SKB_FRAGS simply because this was being
used already in the TLS case. Future work could extend the kernel sk
APIs to tune this depending on workload. This is a trade-off
between memory usage and throughput performance.
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-18 19:57:10 +00:00
|
|
|
case BPF_SK_MSG_VERDICT:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_SK_MSG;
|
2017-08-28 14:10:04 +00:00
|
|
|
case BPF_SK_SKB_STREAM_PARSER:
|
|
|
|
|
case BPF_SK_SKB_STREAM_VERDICT:
|
2021-03-31 02:32:30 +00:00
|
|
|
case BPF_SK_SKB_VERDICT:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_SK_SKB;
|
2018-05-27 11:24:09 +00:00
|
|
|
case BPF_LIRC_MODE2:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_LIRC_MODE2;
|
2018-09-14 14:46:18 +00:00
|
|
|
case BPF_FLOW_DISSECTOR:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_FLOW_DISSECTOR;
|
2019-02-27 20:59:24 +00:00
|
|
|
case BPF_CGROUP_SYSCTL:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_SYSCTL;
|
bpf: implement getsockopt and setsockopt hooks
Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and
BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks.
BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before
passing them down to the kernel or bypass kernel completely.
BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that
kernel returns.
Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure.
The buffer memory is pre-allocated (because I don't think there is
a precedent for working with __user memory from bpf). This might be
slow to do for each {s,g}etsockopt call, that's why I've added
__cgroup_bpf_prog_array_is_empty that exits early if there is nothing
attached to a cgroup. Note, however, that there is a race between
__cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup
program layout might have changed; this should not be a problem
because in general there is a race between multiple calls to
{s,g}etsocktop and user adding/removing bpf progs from a cgroup.
The return code of the BPF program is handled as follows:
* 0: EPERM
* 1: success, continue with next BPF program in the cgroup chain
v9:
* allow overwriting setsockopt arguments (Alexei Starovoitov):
* use set_fs (same as kernel_setsockopt)
* buffer is always kzalloc'd (no small on-stack buffer)
v8:
* use s32 for optlen (Andrii Nakryiko)
v7:
* return only 0 or 1 (Alexei Starovoitov)
* always run all progs (Alexei Starovoitov)
* use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov)
(decided to use optval=-1 instead, optval=0 might be a valid input)
* call getsockopt hook after kernel handlers (Alexei Starovoitov)
v6:
* rework cgroup chaining; stop as soon as bpf program returns
0 or 2; see patch with the documentation for the details
* drop Andrii's and Martin's Acked-by (not sure they are comfortable
with the new state of things)
v5:
* skip copy_to_user() and put_user() when ret == 0 (Martin Lau)
v4:
* don't export bpf_sk_fullsock helper (Martin Lau)
* size != sizeof(__u64) for uapi pointers (Martin Lau)
* offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau)
v3:
* typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko)
* reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii
Nakryiko)
* use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau)
* use BPF_FIELD_SIZEOF() for consistency (Martin Lau)
* new CG_SOCKOPT_ACCESS macro to wrap repeated parts
v2:
* moved bpf_sockopt_kern fields around to remove a hole (Martin Lau)
* aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau)
* bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau)
* added [0,2] return code check to verifier (Martin Lau)
* dropped unused buf[64] from the stack (Martin Lau)
* use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau)
* dropped bpf_target_off from ctx rewrites (Martin Lau)
* use return code for kernel bypass (Martin Lau & Andrii Nakryiko)
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Martin Lau <kafai@fb.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-27 20:38:47 +00:00
|
|
|
case BPF_CGROUP_GETSOCKOPT:
|
|
|
|
|
case BPF_CGROUP_SETSOCKOPT:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_CGROUP_SOCKOPT;
|
2020-05-09 17:59:01 +00:00
|
|
|
case BPF_TRACE_ITER:
|
2022-04-21 03:39:43 +00:00
|
|
|
case BPF_TRACE_RAW_TP:
|
|
|
|
|
case BPF_TRACE_FENTRY:
|
|
|
|
|
case BPF_TRACE_FEXIT:
|
|
|
|
|
case BPF_MODIFY_RETURN:
|
2020-05-09 17:59:01 +00:00
|
|
|
return BPF_PROG_TYPE_TRACING;
|
2022-04-21 03:39:43 +00:00
|
|
|
case BPF_LSM_MAC:
|
|
|
|
|
return BPF_PROG_TYPE_LSM;
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 10:35:23 +00:00
|
|
|
case BPF_SK_LOOKUP:
|
|
|
|
|
return BPF_PROG_TYPE_SK_LOOKUP;
|
2020-07-22 06:45:57 +00:00
|
|
|
case BPF_XDP:
|
|
|
|
|
return BPF_PROG_TYPE_XDP;
|
2022-06-28 17:43:06 +00:00
|
|
|
case BPF_LSM_CGROUP:
|
|
|
|
|
return BPF_PROG_TYPE_LSM;
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
case BPF_TCX_INGRESS:
|
|
|
|
|
case BPF_TCX_EGRESS:
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
case BPF_NETKIT_PRIMARY:
|
|
|
|
|
case BPF_NETKIT_PEER:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
return BPF_PROG_TYPE_SCHED_CLS;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
default:
|
2020-03-25 06:57:42 +00:00
|
|
|
return BPF_PROG_TYPE_UNSPEC;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
}
|
2020-03-25 06:57:42 +00:00
|
|
|
}
|
|
|
|
|
|
2023-08-09 08:34:14 +00:00
|
|
|
static int bpf_prog_attach_check_attach_type(const struct bpf_prog *prog,
|
|
|
|
|
enum bpf_attach_type attach_type)
|
|
|
|
|
{
|
|
|
|
|
enum bpf_prog_type ptype;
|
|
|
|
|
|
|
|
|
|
switch (prog->type) {
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
case BPF_PROG_TYPE_SK_LOOKUP:
|
|
|
|
|
return attach_type == prog->expected_attach_type ? 0 : -EINVAL;
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
2024-01-24 02:21:03 +00:00
|
|
|
if (!bpf_token_capable(prog->aux->token, CAP_NET_ADMIN))
|
2023-08-09 08:34:14 +00:00
|
|
|
/* cg-skb progs can be loaded by unpriv user.
|
|
|
|
|
* check permissions at attach time.
|
|
|
|
|
*/
|
|
|
|
|
return -EPERM;
|
|
|
|
|
return prog->enforce_expected_attach_type &&
|
|
|
|
|
prog->expected_attach_type != attach_type ?
|
|
|
|
|
-EINVAL : 0;
|
|
|
|
|
case BPF_PROG_TYPE_EXT:
|
|
|
|
|
return 0;
|
|
|
|
|
case BPF_PROG_TYPE_NETFILTER:
|
|
|
|
|
if (attach_type != BPF_NETFILTER)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
return 0;
|
|
|
|
|
case BPF_PROG_TYPE_PERF_EVENT:
|
|
|
|
|
case BPF_PROG_TYPE_TRACEPOINT:
|
|
|
|
|
if (attach_type != BPF_PERF_EVENT)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
return 0;
|
|
|
|
|
case BPF_PROG_TYPE_KPROBE:
|
|
|
|
|
if (prog->expected_attach_type == BPF_TRACE_KPROBE_MULTI &&
|
|
|
|
|
attach_type != BPF_TRACE_KPROBE_MULTI)
|
|
|
|
|
return -EINVAL;
|
2023-08-09 08:34:15 +00:00
|
|
|
if (prog->expected_attach_type == BPF_TRACE_UPROBE_MULTI &&
|
|
|
|
|
attach_type != BPF_TRACE_UPROBE_MULTI)
|
|
|
|
|
return -EINVAL;
|
2023-08-09 08:34:14 +00:00
|
|
|
if (attach_type != BPF_PERF_EVENT &&
|
2023-08-09 08:34:15 +00:00
|
|
|
attach_type != BPF_TRACE_KPROBE_MULTI &&
|
|
|
|
|
attach_type != BPF_TRACE_UPROBE_MULTI)
|
2023-08-09 08:34:14 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
return 0;
|
|
|
|
|
case BPF_PROG_TYPE_SCHED_CLS:
|
|
|
|
|
if (attach_type != BPF_TCX_INGRESS &&
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
attach_type != BPF_TCX_EGRESS &&
|
|
|
|
|
attach_type != BPF_NETKIT_PRIMARY &&
|
|
|
|
|
attach_type != BPF_NETKIT_PEER)
|
2023-08-09 08:34:14 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
return 0;
|
|
|
|
|
default:
|
|
|
|
|
ptype = attach_type_to_prog_type(attach_type);
|
|
|
|
|
if (ptype == BPF_PROG_TYPE_UNSPEC || ptype != prog->type)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
#define BPF_PROG_ATTACH_LAST_FIELD expected_revision
|
|
|
|
|
|
|
|
|
|
#define BPF_F_ATTACH_MASK_BASE \
|
|
|
|
|
(BPF_F_ALLOW_OVERRIDE | \
|
|
|
|
|
BPF_F_ALLOW_MULTI | \
|
|
|
|
|
BPF_F_REPLACE)
|
2020-03-25 06:57:42 +00:00
|
|
|
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
#define BPF_F_ATTACH_MASK_MPROG \
|
|
|
|
|
(BPF_F_REPLACE | \
|
|
|
|
|
BPF_F_BEFORE | \
|
|
|
|
|
BPF_F_AFTER | \
|
|
|
|
|
BPF_F_ID | \
|
|
|
|
|
BPF_F_LINK)
|
2020-03-25 06:57:42 +00:00
|
|
|
|
|
|
|
|
static int bpf_prog_attach(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
enum bpf_prog_type ptype;
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_ATTACH))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
ptype = attach_type_to_prog_type(attr->attach_type);
|
|
|
|
|
if (ptype == BPF_PROG_TYPE_UNSPEC)
|
|
|
|
|
return -EINVAL;
|
2023-10-06 22:06:51 +00:00
|
|
|
if (bpf_mprog_supported(ptype)) {
|
|
|
|
|
if (attr->attach_flags & ~BPF_F_ATTACH_MASK_MPROG)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
} else {
|
|
|
|
|
if (attr->attach_flags & ~BPF_F_ATTACH_MASK_BASE)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (attr->relative_fd ||
|
|
|
|
|
attr->expected_revision)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
|
2016-12-01 16:48:03 +00:00
|
|
|
prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
2018-03-30 22:08:00 +00:00
|
|
|
if (bpf_prog_attach_check_attach_type(prog, attr->attach_type)) {
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
2018-06-18 23:04:24 +00:00
|
|
|
switch (ptype) {
|
|
|
|
|
case BPF_PROG_TYPE_SK_SKB:
|
|
|
|
|
case BPF_PROG_TYPE_SK_MSG:
|
bpf, sockmap: convert to generic sk_msg interface
Add a generic sk_msg layer, and convert current sockmap and later
kTLS over to make use of it. While sk_buff handles network packet
representation from netdevice up to socket, sk_msg handles data
representation from application to socket layer.
This means that sk_msg framework spans across ULP users in the
kernel, and enables features such as introspection or filtering
of data with the help of BPF programs that operate on this data
structure.
Latter becomes in particular useful for kTLS where data encryption
is deferred into the kernel, and as such enabling the kernel to
perform L7 introspection and policy based on BPF for TLS connections
where the record is being encrypted after BPF has run and came to
a verdict. In order to get there, first step is to transform open
coding of scatter-gather list handling into a common core framework
that subsystems can use.
The code itself has been split and refactored into three bigger
pieces: i) the generic sk_msg API which deals with managing the
scatter gather ring, providing helpers for walking and mangling,
transferring application data from user space into it, and preparing
it for BPF pre/post-processing, ii) the plain sock map itself
where sockets can be attached to or detached from; these bits
are independent of i) which can now be used also without sock
map, and iii) the integration with plain TCP as one protocol
to be used for processing L7 application data (later this could
e.g. also be extended to other protocols like UDP). The semantics
are the same with the old sock map code and therefore no change
of user facing behavior or APIs. While pursuing this work it
also helped finding a number of bugs in the old sockmap code
that we've fixed already in earlier commits. The test_sockmap
kselftest suite passes through fine as well.
Joint work with John.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 00:45:58 +00:00
|
|
|
ret = sock_map_get_from_fd(attr, prog);
|
2018-06-18 23:04:24 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_PROG_TYPE_LIRC_MODE2:
|
|
|
|
|
ret = lirc_prog_attach(attr, prog);
|
|
|
|
|
break;
|
2018-09-14 14:46:18 +00:00
|
|
|
case BPF_PROG_TYPE_FLOW_DISSECTOR:
|
2020-05-31 08:28:36 +00:00
|
|
|
ret = netns_bpf_prog_attach(attr, prog);
|
2018-09-14 14:46:18 +00:00
|
|
|
break;
|
2020-03-25 06:57:42 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_DEVICE:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SYSCTL:
|
|
|
|
|
case BPF_PROG_TYPE_SOCK_OPS:
|
2022-06-28 17:43:06 +00:00
|
|
|
case BPF_PROG_TYPE_LSM:
|
|
|
|
|
if (ptype == BPF_PROG_TYPE_LSM &&
|
|
|
|
|
prog->expected_attach_type != BPF_LSM_CGROUP)
|
2022-12-13 17:57:14 +00:00
|
|
|
ret = -EINVAL;
|
|
|
|
|
else
|
|
|
|
|
ret = cgroup_bpf_prog_attach(attr, ptype, prog);
|
2020-03-25 06:57:42 +00:00
|
|
|
break;
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
case BPF_PROG_TYPE_SCHED_CLS:
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
if (attr->attach_type == BPF_TCX_INGRESS ||
|
|
|
|
|
attr->attach_type == BPF_TCX_EGRESS)
|
|
|
|
|
ret = tcx_prog_attach(attr, prog);
|
|
|
|
|
else
|
|
|
|
|
ret = netkit_prog_attach(attr, prog);
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
break;
|
2020-03-25 06:57:42 +00:00
|
|
|
default:
|
|
|
|
|
ret = -EINVAL;
|
2016-12-01 16:48:03 +00:00
|
|
|
}
|
|
|
|
|
|
2017-02-11 04:28:24 +00:00
|
|
|
if (ret)
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return ret;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
}
|
|
|
|
|
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
#define BPF_PROG_DETACH_LAST_FIELD expected_revision
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
|
|
|
|
|
static int bpf_prog_detach(const union bpf_attr *attr)
|
|
|
|
|
{
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
struct bpf_prog *prog = NULL;
|
2017-10-03 05:50:21 +00:00
|
|
|
enum bpf_prog_type ptype;
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
int ret;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_DETACH))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-03-25 06:57:42 +00:00
|
|
|
ptype = attach_type_to_prog_type(attr->attach_type);
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
if (bpf_mprog_supported(ptype)) {
|
|
|
|
|
if (ptype == BPF_PROG_TYPE_UNSPEC)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (attr->attach_flags & ~BPF_F_ATTACH_MASK_MPROG)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (attr->attach_bpf_fd) {
|
|
|
|
|
prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
}
|
2023-10-06 22:06:51 +00:00
|
|
|
} else if (attr->attach_flags ||
|
|
|
|
|
attr->relative_fd ||
|
|
|
|
|
attr->expected_revision) {
|
|
|
|
|
return -EINVAL;
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
}
|
2020-03-25 06:57:42 +00:00
|
|
|
|
|
|
|
|
switch (ptype) {
|
|
|
|
|
case BPF_PROG_TYPE_SK_MSG:
|
|
|
|
|
case BPF_PROG_TYPE_SK_SKB:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
ret = sock_map_prog_detach(attr, ptype);
|
|
|
|
|
break;
|
2020-03-25 06:57:42 +00:00
|
|
|
case BPF_PROG_TYPE_LIRC_MODE2:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
ret = lirc_prog_detach(attr);
|
|
|
|
|
break;
|
2020-03-25 06:57:42 +00:00
|
|
|
case BPF_PROG_TYPE_FLOW_DISSECTOR:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
ret = netns_bpf_prog_detach(attr, ptype);
|
|
|
|
|
break;
|
2020-03-25 06:57:42 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_DEVICE:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SYSCTL:
|
|
|
|
|
case BPF_PROG_TYPE_SOCK_OPS:
|
2022-06-28 17:43:06 +00:00
|
|
|
case BPF_PROG_TYPE_LSM:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
ret = cgroup_bpf_prog_detach(attr, ptype);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_PROG_TYPE_SCHED_CLS:
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
if (attr->attach_type == BPF_TCX_INGRESS ||
|
|
|
|
|
attr->attach_type == BPF_TCX_EGRESS)
|
|
|
|
|
ret = tcx_prog_detach(attr, prog);
|
|
|
|
|
else
|
|
|
|
|
ret = netkit_prog_detach(attr, prog);
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
break;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
default:
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
ret = -EINVAL;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
}
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
|
|
|
|
|
if (prog)
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return ret;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
}
|
bpf: BPF support for sock_ops
Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding
struct that allows BPF programs of this type to access some of the
socket's fields (such as IP addresses, ports, etc.). It uses the
existing bpf cgroups infrastructure so the programs can be attached per
cgroup with full inheritance support. The program will be called at
appropriate times to set relevant connections parameters such as buffer
sizes, SYN and SYN-ACK RTOs, etc., based on connection information such
as IP addresses, port numbers, etc.
Alghough there are already 3 mechanisms to set parameters (sysctls,
route metrics and setsockopts), this new mechanism provides some
distinct advantages. Unlike sysctls, it can set parameters per
connection. In contrast to route metrics, it can also use port numbers
and information provided by a user level program. In addition, it could
set parameters probabilistically for evaluation purposes (i.e. do
something different on 10% of the flows and compare results with the
other 90% of the flows). Also, in cases where IPv6 addresses contain
geographic information, the rules to make changes based on the distance
(or RTT) between the hosts are much easier than route metric rules and
can be global. Finally, unlike setsockopt, it oes not require
application changes and it can be updated easily at any time.
Although the bpf cgroup framework already contains a sock related
program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type
(BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called
only once during the connections's lifetime. In contrast, the new
program type will be called multiple times from different places in the
network stack code. For example, before sending SYN and SYN-ACKs to set
an appropriate timeout, when the connection is established to set
congestion control, etc. As a result it has "op" field to specify the
type of operation requested.
The purpose of this new program type is to simplify setting connection
parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is
easy to use facebook's internal IPv6 addresses to determine if both hosts
of a connection are in the same datacenter. Therefore, it is easy to
write a BPF program to choose a small SYN RTO value when both hosts are
in the same datacenter.
This patch only contains the framework to support the new BPF program
type, following patches add the functionality to set various connection
parameters.
This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS
and a new bpf syscall command to load a new program of this type:
BPF_PROG_LOAD_SOCKET_OPS.
Two new corresponding structs (one for the kernel one for the user/BPF
program):
/* kernel version */
struct bpf_sock_ops_kern {
struct sock *sk;
__u32 op;
union {
__u32 reply;
__u32 replylong[4];
};
};
/* user version
* Some fields are in network byte order reflecting the sock struct
* Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to
* convert them to host byte order.
*/
struct bpf_sock_ops {
__u32 op;
union {
__u32 reply;
__u32 replylong[4];
};
__u32 family;
__u32 remote_ip4; /* In network byte order */
__u32 local_ip4; /* In network byte order */
__u32 remote_ip6[4]; /* In network byte order */
__u32 local_ip6[4]; /* In network byte order */
__u32 remote_port; /* In network byte order */
__u32 local_port; /* In host byte horder */
};
Currently there are two types of ops. The first type expects the BPF
program to return a value which is then used by the caller (or a
negative value to indicate the operation is not supported). The second
type expects state changes to be done by the BPF program, for example
through a setsockopt BPF helper function, and they ignore the return
value.
The reply fields of the bpf_sockt_ops struct are there in case a bpf
program needs to return a value larger than an integer.
Signed-off-by: Lawrence Brakmo <brakmo@fb.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 03:02:40 +00:00
|
|
|
|
2023-10-06 22:06:49 +00:00
|
|
|
#define BPF_PROG_QUERY_LAST_FIELD query.revision
|
2017-10-03 05:50:22 +00:00
|
|
|
|
|
|
|
|
static int bpf_prog_query(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
2024-01-24 02:20:58 +00:00
|
|
|
if (!bpf_net_capable())
|
2017-10-03 05:50:22 +00:00
|
|
|
return -EPERM;
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_QUERY))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (attr->query.query_flags & ~BPF_F_QUERY_EFFECTIVE)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
switch (attr->query.attach_type) {
|
|
|
|
|
case BPF_CGROUP_INET_INGRESS:
|
|
|
|
|
case BPF_CGROUP_INET_EGRESS:
|
|
|
|
|
case BPF_CGROUP_INET_SOCK_CREATE:
|
2020-07-06 23:01:25 +00:00
|
|
|
case BPF_CGROUP_INET_SOCK_RELEASE:
|
2018-03-30 22:08:02 +00:00
|
|
|
case BPF_CGROUP_INET4_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_BIND:
|
2018-03-30 22:08:07 +00:00
|
|
|
case BPF_CGROUP_INET4_POST_BIND:
|
|
|
|
|
case BPF_CGROUP_INET6_POST_BIND:
|
2018-03-30 22:08:05 +00:00
|
|
|
case BPF_CGROUP_INET4_CONNECT:
|
|
|
|
|
case BPF_CGROUP_INET6_CONNECT:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_CONNECT:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETPEERNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETPEERNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETPEERNAME:
|
bpf: Add get{peer, sock}name attach types for sock_addr
As stated in 983695fa6765 ("bpf: fix unconnected udp hooks"), the objective
for the existing cgroup connect/sendmsg/recvmsg/bind BPF hooks is to be
transparent to applications. In Cilium we make use of these hooks [0] in
order to enable E-W load balancing for existing Kubernetes service types
for all Cilium managed nodes in the cluster. Those backends can be local
or remote. The main advantage of this approach is that it operates as close
as possible to the socket, and therefore allows to avoid packet-based NAT
given in connect/sendmsg/recvmsg hooks we only need to xlate sock addresses.
This also allows to expose NodePort services on loopback addresses in the
host namespace, for example. As another advantage, this also efficiently
blocks bind requests for applications in the host namespace for exposed
ports. However, one missing item is that we also need to perform reverse
xlation for inet{,6}_getname() hooks such that we can return the service
IP/port tuple back to the application instead of the remote peer address.
The vast majority of applications does not bother about getpeername(), but
in a few occasions we've seen breakage when validating the peer's address
since it returns unexpectedly the backend tuple instead of the service one.
Therefore, this trivial patch allows to customise and adds a getpeername()
as well as getsockname() BPF cgroup hook for both IPv4 and IPv6 in order
to address this situation.
Simple example:
# ./cilium/cilium service list
ID Frontend Service Type Backend
1 1.2.3.4:80 ClusterIP 1 => 10.0.0.10:80
Before; curl's verbose output example, no getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (10.0.0.10) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
After; with getpeername() reverse xlation:
# curl --verbose 1.2.3.4
* Rebuilt URL to: 1.2.3.4/
* Trying 1.2.3.4...
* TCP_NODELAY set
* Connected to 1.2.3.4 (1.2.3.4) port 80 (#0)
> GET / HTTP/1.1
> Host: 1.2.3.4
> User-Agent: curl/7.58.0
> Accept: */*
[...]
Originally, I had both under a BPF_CGROUP_INET{4,6}_GETNAME type and exposed
peer to the context similar as in inet{,6}_getname() fashion, but API-wise
this is suboptimal as it always enforces programs having to test for ctx->peer
which can easily be missed, hence BPF_CGROUP_INET{4,6}_GET{PEER,SOCK}NAME split.
Similarly, the checked return code is on tnum_range(1, 1), but if a use case
comes up in future, it can easily be changed to return an error code instead.
Helper and ctx member access is the same as with connect/sendmsg/etc hooks.
[0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.c
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrey Ignatov <rdna@fb.com>
Link: https://lore.kernel.org/bpf/61a479d759b2482ae3efb45546490bacd796a220.1589841594.git.daniel@iogearbox.net
2020-05-18 22:45:45 +00:00
|
|
|
case BPF_CGROUP_INET4_GETSOCKNAME:
|
|
|
|
|
case BPF_CGROUP_INET6_GETSOCKNAME:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_GETSOCKNAME:
|
2018-05-25 15:55:23 +00:00
|
|
|
case BPF_CGROUP_UDP4_SENDMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_SENDMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_SENDMSG:
|
bpf: fix unconnected udp hooks
Intention of cgroup bind/connect/sendmsg BPF hooks is to act transparently
to applications as also stated in original motivation in 7828f20e3779 ("Merge
branch 'bpf-cgroup-bind-connect'"). When recently integrating the latter
two hooks into Cilium to enable host based load-balancing with Kubernetes,
I ran into the issue that pods couldn't start up as DNS got broken. Kubernetes
typically sets up DNS as a service and is thus subject to load-balancing.
Upon further debugging, it turns out that the cgroupv2 sendmsg BPF hooks API
is currently insufficient and thus not usable as-is for standard applications
shipped with most distros. To break down the issue we ran into with a simple
example:
# cat /etc/resolv.conf
nameserver 147.75.207.207
nameserver 147.75.207.208
For the purpose of a simple test, we set up above IPs as service IPs and
transparently redirect traffic to a different DNS backend server for that
node:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
The attached BPF program is basically selecting one of the backends if the
service IP/port matches on the cgroup hook. DNS breaks here, because the
hooks are not transparent enough to applications which have built-in msg_name
address checks:
# nslookup 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
;; connection timed out; no servers could be reached
# dig 1.1.1.1
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53
;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53
[...]
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; connection timed out; no servers could be reached
For comparison, if none of the service IPs is used, and we tell nslookup
to use 8.8.8.8 directly it works just fine, of course:
# nslookup 1.1.1.1 8.8.8.8
1.1.1.1.in-addr.arpa name = one.one.one.one.
In order to fix this and thus act more transparent to the application,
this needs reverse translation on recvmsg() side. A minimal fix for this
API is to add similar recvmsg() hooks behind the BPF cgroups static key
such that the program can track state and replace the current sockaddr_in{,6}
with the original service IP. From BPF side, this basically tracks the
service tuple plus socket cookie in an LRU map where the reverse NAT can
then be retrieved via map value as one example. Side-note: the BPF cgroups
static key should be converted to a per-hook static key in future.
Same example after this fix:
# cilium service list
ID Frontend Backend
1 147.75.207.207:53 1 => 8.8.8.8:53
2 147.75.207.208:53 1 => 8.8.8.8:53
Lookups work fine now:
# nslookup 1.1.1.1
1.1.1.1.in-addr.arpa name = one.one.one.one.
Authoritative answers can be found from:
# dig 1.1.1.1
; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 51550
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 512
;; QUESTION SECTION:
;1.1.1.1. IN A
;; AUTHORITY SECTION:
. 23426 IN SOA a.root-servers.net. nstld.verisign-grs.com. 2019052001 1800 900 604800 86400
;; Query time: 17 msec
;; SERVER: 147.75.207.207#53(147.75.207.207)
;; WHEN: Tue May 21 12:59:38 UTC 2019
;; MSG SIZE rcvd: 111
And from an actual packet level it shows that we're using the back end
server when talking via 147.75.207.20{7,8} front end:
# tcpdump -i any udp
[...]
12:59:52.698732 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.698735 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67)
[...]
In order to be flexible and to have same semantics as in sendmsg BPF
programs, we only allow return codes in [1,1] range. In the sendmsg case
the program is called if msg->msg_name is present which can be the case
in both, connected and unconnected UDP.
The former only relies on the sockaddr_in{,6} passed via connect(2) if
passed msg->msg_name was NULL. Therefore, on recvmsg side, we act in similar
way to call into the BPF program whenever a non-NULL msg->msg_name was
passed independent of sk->sk_state being TCP_ESTABLISHED or not. Note
that for TCP case, the msg->msg_name is ignored in the regular recvmsg
path and therefore not relevant.
For the case of ip{,v6}_recv_error() paths, picked up via MSG_ERRQUEUE,
the hook is not called. This is intentional as it aligns with the same
semantics as in case of TCP cgroup BPF hooks right now. This might be
better addressed in future through a different bpf_attach_type such
that this case can be distinguished from the regular recvmsg paths,
for example.
Fixes: 1cedee13d25a ("bpf: Hooks for sys_sendmsg")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrey Ignatov <rdna@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Martynas Pumputis <m@lambda.lt>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-06 23:48:57 +00:00
|
|
|
case BPF_CGROUP_UDP4_RECVMSG:
|
|
|
|
|
case BPF_CGROUP_UDP6_RECVMSG:
|
2023-10-11 18:51:06 +00:00
|
|
|
case BPF_CGROUP_UNIX_RECVMSG:
|
2017-10-03 05:50:22 +00:00
|
|
|
case BPF_CGROUP_SOCK_OPS:
|
2017-11-05 13:15:32 +00:00
|
|
|
case BPF_CGROUP_DEVICE:
|
2019-02-27 20:59:24 +00:00
|
|
|
case BPF_CGROUP_SYSCTL:
|
bpf: implement getsockopt and setsockopt hooks
Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and
BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks.
BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before
passing them down to the kernel or bypass kernel completely.
BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that
kernel returns.
Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure.
The buffer memory is pre-allocated (because I don't think there is
a precedent for working with __user memory from bpf). This might be
slow to do for each {s,g}etsockopt call, that's why I've added
__cgroup_bpf_prog_array_is_empty that exits early if there is nothing
attached to a cgroup. Note, however, that there is a race between
__cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup
program layout might have changed; this should not be a problem
because in general there is a race between multiple calls to
{s,g}etsocktop and user adding/removing bpf progs from a cgroup.
The return code of the BPF program is handled as follows:
* 0: EPERM
* 1: success, continue with next BPF program in the cgroup chain
v9:
* allow overwriting setsockopt arguments (Alexei Starovoitov):
* use set_fs (same as kernel_setsockopt)
* buffer is always kzalloc'd (no small on-stack buffer)
v8:
* use s32 for optlen (Andrii Nakryiko)
v7:
* return only 0 or 1 (Alexei Starovoitov)
* always run all progs (Alexei Starovoitov)
* use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov)
(decided to use optval=-1 instead, optval=0 might be a valid input)
* call getsockopt hook after kernel handlers (Alexei Starovoitov)
v6:
* rework cgroup chaining; stop as soon as bpf program returns
0 or 2; see patch with the documentation for the details
* drop Andrii's and Martin's Acked-by (not sure they are comfortable
with the new state of things)
v5:
* skip copy_to_user() and put_user() when ret == 0 (Martin Lau)
v4:
* don't export bpf_sk_fullsock helper (Martin Lau)
* size != sizeof(__u64) for uapi pointers (Martin Lau)
* offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau)
v3:
* typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko)
* reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii
Nakryiko)
* use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau)
* use BPF_FIELD_SIZEOF() for consistency (Martin Lau)
* new CG_SOCKOPT_ACCESS macro to wrap repeated parts
v2:
* moved bpf_sockopt_kern fields around to remove a hole (Martin Lau)
* aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau)
* bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau)
* added [0,2] return code check to verifier (Martin Lau)
* dropped unused buf[64] from the stack (Martin Lau)
* use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau)
* dropped bpf_target_off from ctx rewrites (Martin Lau)
* use return code for kernel bypass (Martin Lau & Andrii Nakryiko)
Cc: Andrii Nakryiko <andriin@fb.com>
Cc: Martin Lau <kafai@fb.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-27 20:38:47 +00:00
|
|
|
case BPF_CGROUP_GETSOCKOPT:
|
|
|
|
|
case BPF_CGROUP_SETSOCKOPT:
|
2022-06-28 17:43:08 +00:00
|
|
|
case BPF_LSM_CGROUP:
|
2020-03-25 06:57:42 +00:00
|
|
|
return cgroup_bpf_prog_query(attr, uattr);
|
2018-05-27 11:24:09 +00:00
|
|
|
case BPF_LIRC_MODE2:
|
|
|
|
|
return lirc_prog_query(attr, uattr);
|
2019-04-25 21:37:23 +00:00
|
|
|
case BPF_FLOW_DISSECTOR:
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 10:35:23 +00:00
|
|
|
case BPF_SK_LOOKUP:
|
2020-05-31 08:28:36 +00:00
|
|
|
return netns_bpf_prog_query(attr, uattr);
|
2022-01-19 01:40:04 +00:00
|
|
|
case BPF_SK_SKB_STREAM_PARSER:
|
|
|
|
|
case BPF_SK_SKB_STREAM_VERDICT:
|
|
|
|
|
case BPF_SK_MSG_VERDICT:
|
|
|
|
|
case BPF_SK_SKB_VERDICT:
|
|
|
|
|
return sock_map_bpf_prog_query(attr, uattr);
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
case BPF_TCX_INGRESS:
|
|
|
|
|
case BPF_TCX_EGRESS:
|
|
|
|
|
return tcx_prog_query(attr, uattr);
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
case BPF_NETKIT_PRIMARY:
|
|
|
|
|
case BPF_NETKIT_PEER:
|
|
|
|
|
return netkit_prog_query(attr, uattr);
|
2017-10-03 05:50:22 +00:00
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
}
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
|
bpf: Add "live packet" mode for XDP in BPF_PROG_RUN
This adds support for running XDP programs through BPF_PROG_RUN in a mode
that enables live packet processing of the resulting frames. Previous uses
of BPF_PROG_RUN for XDP returned the XDP program return code and the
modified packet data to userspace, which is useful for unit testing of XDP
programs.
The existing BPF_PROG_RUN for XDP allows userspace to set the ingress
ifindex and RXQ number as part of the context object being passed to the
kernel. This patch reuses that code, but adds a new mode with different
semantics, which can be selected with the new BPF_F_TEST_XDP_LIVE_FRAMES
flag.
When running BPF_PROG_RUN in this mode, the XDP program return codes will
be honoured: returning XDP_PASS will result in the frame being injected
into the networking stack as if it came from the selected networking
interface, while returning XDP_TX and XDP_REDIRECT will result in the frame
being transmitted out that interface. XDP_TX is translated into an
XDP_REDIRECT operation to the same interface, since the real XDP_TX action
is only possible from within the network drivers themselves, not from the
process context where BPF_PROG_RUN is executed.
Internally, this new mode of operation creates a page pool instance while
setting up the test run, and feeds pages from that into the XDP program.
The setup cost of this is amortised over the number of repetitions
specified by userspace.
To support the performance testing use case, we further optimise the setup
step so that all pages in the pool are pre-initialised with the packet
data, and pre-computed context and xdp_frame objects stored at the start of
each page. This makes it possible to entirely avoid touching the page
content on each XDP program invocation, and enables sending up to 9
Mpps/core on my test box.
Because the data pages are recycled by the page pool, and the test runner
doesn't re-initialise them for each run, subsequent invocations of the XDP
program will see the packet data in the state it was after the last time it
ran on that particular page. This means that an XDP program that modifies
the packet before redirecting it has to be careful about which assumptions
it makes about the packet content, but that is only an issue for the most
naively written programs.
Enabling the new flag is only allowed when not setting ctx_out and data_out
in the test specification, since using it means frames will be redirected
somewhere else, so they can't be returned.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20220309105346.100053-2-toke@redhat.com
2022-03-09 10:53:42 +00:00
|
|
|
#define BPF_PROG_TEST_RUN_LAST_FIELD test.batch_size
|
2017-03-31 04:45:38 +00:00
|
|
|
|
|
|
|
|
static int bpf_prog_test_run(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
int ret = -ENOTSUPP;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_TEST_RUN))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2019-04-09 18:49:09 +00:00
|
|
|
if ((attr->test.ctx_size_in && !attr->test.ctx_in) ||
|
|
|
|
|
(!attr->test.ctx_size_in && attr->test.ctx_in))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if ((attr->test.ctx_size_out && !attr->test.ctx_out) ||
|
|
|
|
|
(!attr->test.ctx_size_out && attr->test.ctx_out))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2017-03-31 04:45:38 +00:00
|
|
|
prog = bpf_prog_get(attr->test.prog_fd);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
if (prog->aux->ops->test_run)
|
|
|
|
|
ret = prog->aux->ops->test_run(prog, attr, uattr);
|
|
|
|
|
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:48 +00:00
|
|
|
#define BPF_OBJ_GET_NEXT_ID_LAST_FIELD next_id
|
|
|
|
|
|
|
|
|
|
static int bpf_obj_get_next_id(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr,
|
|
|
|
|
struct idr *idr,
|
|
|
|
|
spinlock_t *lock)
|
|
|
|
|
{
|
|
|
|
|
u32 next_id = attr->start_id;
|
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_OBJ_GET_NEXT_ID) || next_id >= INT_MAX)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
next_id++;
|
|
|
|
|
spin_lock_bh(lock);
|
|
|
|
|
if (!idr_get_next(idr, &next_id))
|
|
|
|
|
err = -ENOENT;
|
|
|
|
|
spin_unlock_bh(lock);
|
|
|
|
|
|
|
|
|
|
if (!err)
|
|
|
|
|
err = put_user(next_id, &uattr->next_id);
|
|
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-09 17:59:09 +00:00
|
|
|
struct bpf_map *bpf_map_get_curr_or_next(u32 *id)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
|
|
|
|
|
spin_lock_bh(&map_idr_lock);
|
|
|
|
|
again:
|
|
|
|
|
map = idr_get_next(&map_idr, id);
|
|
|
|
|
if (map) {
|
|
|
|
|
map = __bpf_map_inc_not_zero(map, false);
|
|
|
|
|
if (IS_ERR(map)) {
|
|
|
|
|
(*id)++;
|
|
|
|
|
goto again;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_bh(&map_idr_lock);
|
|
|
|
|
|
|
|
|
|
return map;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-02 01:10:18 +00:00
|
|
|
struct bpf_prog *bpf_prog_get_curr_or_next(u32 *id)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
|
|
|
|
|
spin_lock_bh(&prog_idr_lock);
|
|
|
|
|
again:
|
|
|
|
|
prog = idr_get_next(&prog_idr, id);
|
|
|
|
|
if (prog) {
|
|
|
|
|
prog = bpf_prog_inc_not_zero(prog);
|
|
|
|
|
if (IS_ERR(prog)) {
|
|
|
|
|
(*id)++;
|
|
|
|
|
goto again;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_bh(&prog_idr_lock);
|
|
|
|
|
|
|
|
|
|
return prog;
|
|
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:49 +00:00
|
|
|
#define BPF_PROG_GET_FD_BY_ID_LAST_FIELD prog_id
|
|
|
|
|
|
2019-12-13 17:51:09 +00:00
|
|
|
struct bpf_prog *bpf_prog_by_id(u32 id)
|
2017-06-05 19:15:49 +00:00
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
|
2019-12-13 17:51:09 +00:00
|
|
|
if (!id)
|
|
|
|
|
return ERR_PTR(-ENOENT);
|
2017-06-05 19:15:49 +00:00
|
|
|
|
|
|
|
|
spin_lock_bh(&prog_idr_lock);
|
|
|
|
|
prog = idr_find(&prog_idr, id);
|
|
|
|
|
if (prog)
|
|
|
|
|
prog = bpf_prog_inc_not_zero(prog);
|
|
|
|
|
else
|
|
|
|
|
prog = ERR_PTR(-ENOENT);
|
|
|
|
|
spin_unlock_bh(&prog_idr_lock);
|
2019-12-13 17:51:09 +00:00
|
|
|
return prog;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_prog_get_fd_by_id(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
u32 id = attr->prog_id;
|
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_GET_FD_BY_ID))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
2017-06-05 19:15:49 +00:00
|
|
|
|
2019-12-13 17:51:09 +00:00
|
|
|
prog = bpf_prog_by_id(id);
|
2017-06-05 19:15:49 +00:00
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
fd = bpf_prog_new_fd(prog);
|
|
|
|
|
if (fd < 0)
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
|
|
|
|
|
return fd;
|
|
|
|
|
}
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
#define BPF_MAP_GET_FD_BY_ID_LAST_FIELD open_flags
|
2017-06-05 19:15:50 +00:00
|
|
|
|
|
|
|
|
static int bpf_map_get_fd_by_id(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
u32 id = attr->map_id;
|
2017-10-18 20:00:22 +00:00
|
|
|
int f_flags;
|
2017-06-05 19:15:50 +00:00
|
|
|
int fd;
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
if (CHECK_ATTR(BPF_MAP_GET_FD_BY_ID) ||
|
|
|
|
|
attr->open_flags & ~BPF_OBJ_FLAG_MASK)
|
2017-06-05 19:15:50 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
f_flags = bpf_get_file_flag(attr->open_flags);
|
|
|
|
|
if (f_flags < 0)
|
|
|
|
|
return f_flags;
|
|
|
|
|
|
2017-06-05 19:15:50 +00:00
|
|
|
spin_lock_bh(&map_idr_lock);
|
|
|
|
|
map = idr_find(&map_idr, id);
|
|
|
|
|
if (map)
|
2019-08-14 17:37:48 +00:00
|
|
|
map = __bpf_map_inc_not_zero(map, true);
|
2017-06-05 19:15:50 +00:00
|
|
|
else
|
|
|
|
|
map = ERR_PTR(-ENOENT);
|
|
|
|
|
spin_unlock_bh(&map_idr_lock);
|
|
|
|
|
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
|
|
|
|
|
2017-10-18 20:00:22 +00:00
|
|
|
fd = bpf_map_new_fd(map, f_flags);
|
2017-06-05 19:15:50 +00:00
|
|
|
if (fd < 0)
|
2019-02-26 14:15:37 +00:00
|
|
|
bpf_map_put_with_uref(map);
|
2017-06-05 19:15:50 +00:00
|
|
|
|
|
|
|
|
return fd;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
static const struct bpf_map *bpf_map_from_imm(const struct bpf_prog *prog,
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
unsigned long addr, u32 *off,
|
|
|
|
|
u32 *type)
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
{
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
const struct bpf_map *map;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
int i;
|
|
|
|
|
|
2020-09-15 23:45:39 +00:00
|
|
|
mutex_lock(&prog->aux->used_maps_mutex);
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
for (i = 0, *off = 0; i < prog->aux->used_map_cnt; i++) {
|
|
|
|
|
map = prog->aux->used_maps[i];
|
|
|
|
|
if (map == (void *)addr) {
|
|
|
|
|
*type = BPF_PSEUDO_MAP_FD;
|
2020-09-15 23:45:39 +00:00
|
|
|
goto out;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
}
|
|
|
|
|
if (!map->ops->map_direct_value_meta)
|
|
|
|
|
continue;
|
|
|
|
|
if (!map->ops->map_direct_value_meta(map, addr, off)) {
|
|
|
|
|
*type = BPF_PSEUDO_MAP_VALUE;
|
2020-09-15 23:45:39 +00:00
|
|
|
goto out;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
}
|
|
|
|
|
}
|
2020-09-15 23:45:39 +00:00
|
|
|
map = NULL;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
|
2020-09-15 23:45:39 +00:00
|
|
|
out:
|
|
|
|
|
mutex_unlock(&prog->aux->used_maps_mutex);
|
|
|
|
|
return map;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
}
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
static struct bpf_insn *bpf_insn_prepare_dump(const struct bpf_prog *prog,
|
|
|
|
|
const struct cred *f_cred)
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
{
|
|
|
|
|
const struct bpf_map *map;
|
|
|
|
|
struct bpf_insn *insns;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
u32 off, type;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
u64 imm;
|
2020-06-13 00:21:15 +00:00
|
|
|
u8 code;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
insns = kmemdup(prog->insnsi, bpf_prog_insn_size(prog),
|
|
|
|
|
GFP_USER);
|
|
|
|
|
if (!insns)
|
|
|
|
|
return insns;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < prog->len; i++) {
|
2020-06-13 00:21:15 +00:00
|
|
|
code = insns[i].code;
|
|
|
|
|
|
|
|
|
|
if (code == (BPF_JMP | BPF_TAIL_CALL)) {
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
insns[i].code = BPF_JMP | BPF_CALL;
|
|
|
|
|
insns[i].imm = BPF_FUNC_tail_call;
|
|
|
|
|
/* fall-through */
|
|
|
|
|
}
|
2020-06-13 00:21:15 +00:00
|
|
|
if (code == (BPF_JMP | BPF_CALL) ||
|
|
|
|
|
code == (BPF_JMP | BPF_CALL_ARGS)) {
|
|
|
|
|
if (code == (BPF_JMP | BPF_CALL_ARGS))
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
insns[i].code = BPF_JMP | BPF_CALL;
|
2020-07-02 22:45:23 +00:00
|
|
|
if (!bpf_dump_raw_ok(f_cred))
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
insns[i].imm = 0;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
2020-06-13 00:21:15 +00:00
|
|
|
if (BPF_CLASS(code) == BPF_LDX && BPF_MODE(code) == BPF_PROBE_MEM) {
|
|
|
|
|
insns[i].code = BPF_LDX | BPF_SIZE(code) | BPF_MEM;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
|
2024-03-08 01:08:03 +00:00
|
|
|
if ((BPF_CLASS(code) == BPF_LDX || BPF_CLASS(code) == BPF_STX ||
|
|
|
|
|
BPF_CLASS(code) == BPF_ST) && BPF_MODE(code) == BPF_PROBE_MEM32) {
|
|
|
|
|
insns[i].code = BPF_CLASS(code) | BPF_SIZE(code) | BPF_MEM;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-13 00:21:15 +00:00
|
|
|
if (code != (BPF_LD | BPF_IMM | BPF_DW))
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
imm = ((u64)insns[i + 1].imm << 32) | (u32)insns[i].imm;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
map = bpf_map_from_imm(prog, imm, &off, &type);
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
if (map) {
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
insns[i].src_reg = type;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
insns[i].imm = map->id;
|
bpf: implement lookup-free direct value access for maps
This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!
The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.
The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.
This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.
The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:03 +00:00
|
|
|
insns[i + 1].imm = off;
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return insns;
|
|
|
|
|
}
|
|
|
|
|
|
2018-12-08 00:42:25 +00:00
|
|
|
static int set_info_rec_size(struct bpf_prog_info *info)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* Ensure info.*_rec_size is the same as kernel expected size
|
|
|
|
|
*
|
|
|
|
|
* or
|
|
|
|
|
*
|
|
|
|
|
* Only allow zero *_rec_size if both _rec_size and _cnt are
|
|
|
|
|
* zero. In this case, the kernel will set the expected
|
|
|
|
|
* _rec_size back to the info.
|
|
|
|
|
*/
|
|
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
if ((info->nr_func_info || info->func_info_rec_size) &&
|
2018-12-08 00:42:25 +00:00
|
|
|
info->func_info_rec_size != sizeof(struct bpf_func_info))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
if ((info->nr_line_info || info->line_info_rec_size) &&
|
2018-12-08 00:42:25 +00:00
|
|
|
info->line_info_rec_size != sizeof(struct bpf_line_info))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
if ((info->nr_jited_line_info || info->jited_line_info_rec_size) &&
|
2018-12-08 00:42:25 +00:00
|
|
|
info->jited_line_info_rec_size != sizeof(__u64))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
info->func_info_rec_size = sizeof(struct bpf_func_info);
|
|
|
|
|
info->line_info_rec_size = sizeof(struct bpf_line_info);
|
|
|
|
|
info->jited_line_info_rec_size = sizeof(__u64);
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
static int bpf_prog_get_info_by_fd(struct file *file,
|
|
|
|
|
struct bpf_prog *prog,
|
2017-06-05 19:15:52 +00:00
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog_info __user *uinfo = u64_to_user_ptr(attr->info.info);
|
2022-08-04 20:11:39 +00:00
|
|
|
struct btf *attach_btf = bpf_prog_get_target_btf(prog);
|
2020-03-20 16:22:58 +00:00
|
|
|
struct bpf_prog_info info;
|
2017-06-05 19:15:52 +00:00
|
|
|
u32 info_len = attr->info.info_len;
|
2021-10-26 21:41:33 +00:00
|
|
|
struct bpf_prog_kstats stats;
|
2017-06-05 19:15:52 +00:00
|
|
|
char __user *uinsns;
|
|
|
|
|
u32 ulen;
|
|
|
|
|
int err;
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_check_uarg_tail_zero(USER_BPFPTR(uinfo), sizeof(info), info_len);
|
2017-06-05 19:15:52 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
info_len = min_t(u32, sizeof(info), info_len);
|
|
|
|
|
|
2020-03-20 16:22:58 +00:00
|
|
|
memset(&info, 0, sizeof(info));
|
2017-06-05 19:15:52 +00:00
|
|
|
if (copy_from_user(&info, uinfo, info_len))
|
2017-07-27 19:02:46 +00:00
|
|
|
return -EFAULT;
|
2017-06-05 19:15:52 +00:00
|
|
|
|
|
|
|
|
info.type = prog->type;
|
|
|
|
|
info.id = prog->aux->id;
|
2017-09-27 21:37:52 +00:00
|
|
|
info.load_time = prog->aux->load_time;
|
|
|
|
|
info.created_by_uid = from_kuid_munged(current_user_ns(),
|
|
|
|
|
prog->aux->user->uid);
|
2018-04-25 17:41:06 +00:00
|
|
|
info.gpl_compatible = prog->gpl_compatible;
|
2017-06-05 19:15:52 +00:00
|
|
|
|
|
|
|
|
memcpy(info.tag, prog->tag, sizeof(prog->tag));
|
2017-09-27 21:37:52 +00:00
|
|
|
memcpy(info.name, prog->aux->name, sizeof(prog->aux->name));
|
|
|
|
|
|
2020-09-15 23:45:39 +00:00
|
|
|
mutex_lock(&prog->aux->used_maps_mutex);
|
2017-09-27 21:37:52 +00:00
|
|
|
ulen = info.nr_map_ids;
|
|
|
|
|
info.nr_map_ids = prog->aux->used_map_cnt;
|
|
|
|
|
ulen = min_t(u32, info.nr_map_ids, ulen);
|
|
|
|
|
if (ulen) {
|
2017-09-29 17:52:17 +00:00
|
|
|
u32 __user *user_map_ids = u64_to_user_ptr(info.map_ids);
|
2017-09-27 21:37:52 +00:00
|
|
|
u32 i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < ulen; i++)
|
|
|
|
|
if (put_user(prog->aux->used_maps[i]->id,
|
2020-09-15 23:45:39 +00:00
|
|
|
&user_map_ids[i])) {
|
|
|
|
|
mutex_unlock(&prog->aux->used_maps_mutex);
|
2017-09-27 21:37:52 +00:00
|
|
|
return -EFAULT;
|
2020-09-15 23:45:39 +00:00
|
|
|
}
|
2017-09-27 21:37:52 +00:00
|
|
|
}
|
2020-09-15 23:45:39 +00:00
|
|
|
mutex_unlock(&prog->aux->used_maps_mutex);
|
2017-06-05 19:15:52 +00:00
|
|
|
|
2018-12-08 00:42:25 +00:00
|
|
|
err = set_info_rec_size(&info);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
2018-12-06 01:35:43 +00:00
|
|
|
|
2019-02-25 22:28:40 +00:00
|
|
|
bpf_prog_get_stats(prog, &stats);
|
|
|
|
|
info.run_time_ns = stats.nsecs;
|
|
|
|
|
info.run_cnt = stats.cnt;
|
2021-02-10 03:36:31 +00:00
|
|
|
info.recursion_misses = stats.misses;
|
2019-02-25 22:28:40 +00:00
|
|
|
|
2021-10-20 07:48:17 +00:00
|
|
|
info.verified_insns = prog->aux->verified_insns;
|
|
|
|
|
|
2020-05-13 23:03:54 +00:00
|
|
|
if (!bpf_capable()) {
|
2017-06-05 19:15:52 +00:00
|
|
|
info.jited_prog_len = 0;
|
|
|
|
|
info.xlated_prog_len = 0;
|
2018-05-24 06:56:48 +00:00
|
|
|
info.nr_jited_ksyms = 0;
|
2018-11-02 10:35:46 +00:00
|
|
|
info.nr_jited_func_lens = 0;
|
2018-12-10 22:14:08 +00:00
|
|
|
info.nr_func_info = 0;
|
|
|
|
|
info.nr_line_info = 0;
|
|
|
|
|
info.nr_jited_line_info = 0;
|
2017-06-05 19:15:52 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ulen = info.xlated_prog_len;
|
2017-07-28 15:05:25 +00:00
|
|
|
info.xlated_prog_len = bpf_prog_insn_size(prog);
|
2017-06-05 19:15:52 +00:00
|
|
|
if (info.xlated_prog_len && ulen) {
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
struct bpf_insn *insns_sanitized;
|
|
|
|
|
bool fault;
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
if (prog->blinded && !bpf_dump_raw_ok(file->f_cred)) {
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
info.xlated_prog_insns = 0;
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
2020-07-02 22:45:23 +00:00
|
|
|
insns_sanitized = bpf_insn_prepare_dump(prog, file->f_cred);
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
if (!insns_sanitized)
|
|
|
|
|
return -ENOMEM;
|
2017-06-05 19:15:52 +00:00
|
|
|
uinsns = u64_to_user_ptr(info.xlated_prog_insns);
|
|
|
|
|
ulen = min_t(u32, info.xlated_prog_len, ulen);
|
bpf: allow for correlation of maps and helpers in dump
Currently a dump of an xlated prog (post verifier stage) doesn't
correlate used helpers as well as maps. The prog info lists
involved map ids, however there's no correlation of where in the
program they are used as of today. Likewise, bpftool does not
correlate helper calls with the target functions.
The latter can be done w/o any kernel changes through kallsyms,
and also has the advantage that this works with inlined helpers
and BPF calls.
Example, via interpreter:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 1 tag c74773051b364165 <-- prog id:1
* Output before patch (calls/maps remain unclear):
# bpftool prog dump xlated id 1 <-- dump prog id:1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = 0xffff95c47a8d4800
6: (85) call unknown#73040
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call unknown#73040
12: (15) if r0 == 0x0 goto pc+23
[...]
* Output after patch:
# bpftool prog dump xlated id 1
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call bpf_map_lookup_elem#73424 <-- helper call
7: (15) if r0 == 0x0 goto pc+18
8: (bf) r2 = r10
9: (07) r2 += -4
10: (bf) r1 = r0
11: (85) call bpf_map_lookup_elem#73424
12: (15) if r0 == 0x0 goto pc+23
[...]
# bpftool map show id 2 <-- show/dump/etc map id:2
2: hash_of_maps flags 0x0
key 4B value 4B max_entries 3 memlock 4096B
Example, JITed, same prog:
# tc filter show dev foo ingress
filter protocol all pref 49152 bpf chain 0
filter protocol all pref 49152 bpf chain 0 handle 0x1 foo.o:[ingress] \
direct-action not_in_hw id 3 tag c74773051b364165 jited
# bpftool prog show id 3
3: sched_cls tag c74773051b364165
loaded_at Dec 19/13:48 uid 0
xlated 384B jited 257B memlock 4096B map_ids 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2] <-- map id:2
6: (85) call __htab_map_lookup_elem#77408 <-+ inlined rewrite
7: (15) if r0 == 0x0 goto pc+2 |
8: (07) r0 += 56 |
9: (79) r0 = *(u64 *)(r0 +0) <-+
10: (15) if r0 == 0x0 goto pc+24
11: (bf) r2 = r10
12: (07) r2 += -4
[...]
Example, same prog, but kallsyms disabled (in that case we are
also not allowed to pass any relative offsets, etc, so prog
becomes pointer sanitized on dump):
# sysctl kernel.kptr_restrict=2
kernel.kptr_restrict = 2
# bpftool prog dump xlated id 3
0: (b7) r1 = 2
1: (63) *(u32 *)(r10 -4) = r1
2: (bf) r2 = r10
3: (07) r2 += -4
4: (18) r1 = map[id:2]
6: (85) call bpf_unspec#0
7: (15) if r0 == 0x0 goto pc+2
[...]
Example, BPF calls via interpreter:
# bpftool prog dump xlated id 1
0: (85) call pc+2#__bpf_prog_run_args32
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
Example, BPF calls via JIT:
# sysctl net.core.bpf_jit_enable=1
net.core.bpf_jit_enable = 1
# sysctl net.core.bpf_jit_kallsyms=1
net.core.bpf_jit_kallsyms = 1
# bpftool prog dump xlated id 1
0: (85) call pc+2#bpf_prog_3b185187f1855c4c_F
1: (b7) r0 = 1
2: (95) exit
3: (b7) r0 = 2
4: (95) exit
And finally, an example for tail calls that is now working
as well wrt correlation:
# bpftool prog dump xlated id 2
[...]
10: (b7) r2 = 8
11: (85) call bpf_trace_printk#-41312
12: (bf) r1 = r6
13: (18) r2 = map[id:1]
15: (b7) r3 = 0
16: (85) call bpf_tail_call#12
17: (b7) r1 = 42
18: (6b) *(u16 *)(r6 +46) = r1
19: (b7) r0 = 0
20: (95) exit
# bpftool map show id 1
1: prog_array flags 0x0
key 4B value 4B max_entries 1 memlock 4096B
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2017-12-20 12:42:57 +00:00
|
|
|
fault = copy_to_user(uinsns, insns_sanitized, ulen);
|
|
|
|
|
kfree(insns_sanitized);
|
|
|
|
|
if (fault)
|
2017-06-05 19:15:52 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_prog_is_offloaded(prog->aux)) {
|
2017-12-28 02:39:09 +00:00
|
|
|
err = bpf_prog_offload_info_fill(&info, prog);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
2018-01-17 00:05:19 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* NOTE: the following code is supposed to be skipped for offload.
|
|
|
|
|
* bpf_prog_offload_info_fill() is the place to fill similar fields
|
|
|
|
|
* for offload.
|
|
|
|
|
*/
|
|
|
|
|
ulen = info.jited_prog_len;
|
bpf: fix multi-function JITed dump obtained via syscall
Currently, for multi-function programs, we cannot get the JITed
instructions using the bpf system call's BPF_OBJ_GET_INFO_BY_FD
command. Because of this, userspace tools such as bpftool fail
to identify a multi-function program as being JITed or not.
With the JIT enabled and the test program running, this can be
verified as follows:
# cat /proc/sys/net/core/bpf_jit_enable
1
Before applying this patch:
# bpftool prog list
1: kprobe name foo tag b811aab41a39ad3d gpl
loaded_at 2018-05-16T11:43:38+0530 uid 0
xlated 216B not jited memlock 65536B
...
# bpftool prog dump jited id 1
no instructions returned
After applying this patch:
# bpftool prog list
1: kprobe name foo tag b811aab41a39ad3d gpl
loaded_at 2018-05-16T12:13:01+0530 uid 0
xlated 216B jited 308B memlock 65536B
...
# bpftool prog dump jited id 1
0: nop
4: nop
8: mflr r0
c: std r0,16(r1)
10: stdu r1,-112(r1)
14: std r31,104(r1)
18: addi r31,r1,48
1c: li r3,10
...
Signed-off-by: Sandipan Das <sandipan@linux.vnet.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-24 06:56:51 +00:00
|
|
|
if (prog->aux->func_cnt) {
|
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
|
|
info.jited_prog_len = 0;
|
|
|
|
|
for (i = 0; i < prog->aux->func_cnt; i++)
|
|
|
|
|
info.jited_prog_len += prog->aux->func[i]->jited_len;
|
|
|
|
|
} else {
|
|
|
|
|
info.jited_prog_len = prog->jited_len;
|
|
|
|
|
}
|
|
|
|
|
|
2018-01-17 00:05:19 +00:00
|
|
|
if (info.jited_prog_len && ulen) {
|
2020-07-02 22:45:23 +00:00
|
|
|
if (bpf_dump_raw_ok(file->f_cred)) {
|
2018-01-17 00:05:19 +00:00
|
|
|
uinsns = u64_to_user_ptr(info.jited_prog_insns);
|
|
|
|
|
ulen = min_t(u32, info.jited_prog_len, ulen);
|
bpf: fix multi-function JITed dump obtained via syscall
Currently, for multi-function programs, we cannot get the JITed
instructions using the bpf system call's BPF_OBJ_GET_INFO_BY_FD
command. Because of this, userspace tools such as bpftool fail
to identify a multi-function program as being JITed or not.
With the JIT enabled and the test program running, this can be
verified as follows:
# cat /proc/sys/net/core/bpf_jit_enable
1
Before applying this patch:
# bpftool prog list
1: kprobe name foo tag b811aab41a39ad3d gpl
loaded_at 2018-05-16T11:43:38+0530 uid 0
xlated 216B not jited memlock 65536B
...
# bpftool prog dump jited id 1
no instructions returned
After applying this patch:
# bpftool prog list
1: kprobe name foo tag b811aab41a39ad3d gpl
loaded_at 2018-05-16T12:13:01+0530 uid 0
xlated 216B jited 308B memlock 65536B
...
# bpftool prog dump jited id 1
0: nop
4: nop
8: mflr r0
c: std r0,16(r1)
10: stdu r1,-112(r1)
14: std r31,104(r1)
18: addi r31,r1,48
1c: li r3,10
...
Signed-off-by: Sandipan Das <sandipan@linux.vnet.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-24 06:56:51 +00:00
|
|
|
|
|
|
|
|
/* for multi-function programs, copy the JITed
|
|
|
|
|
* instructions for all the functions
|
|
|
|
|
*/
|
|
|
|
|
if (prog->aux->func_cnt) {
|
|
|
|
|
u32 len, free, i;
|
|
|
|
|
u8 *img;
|
|
|
|
|
|
|
|
|
|
free = ulen;
|
|
|
|
|
for (i = 0; i < prog->aux->func_cnt; i++) {
|
|
|
|
|
len = prog->aux->func[i]->jited_len;
|
|
|
|
|
len = min_t(u32, len, free);
|
|
|
|
|
img = (u8 *) prog->aux->func[i]->bpf_func;
|
|
|
|
|
if (copy_to_user(uinsns, img, len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
uinsns += len;
|
|
|
|
|
free -= len;
|
|
|
|
|
if (!free)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
if (copy_to_user(uinsns, prog->bpf_func, ulen))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
2018-01-17 00:05:19 +00:00
|
|
|
} else {
|
|
|
|
|
info.jited_prog_insns = 0;
|
|
|
|
|
}
|
2017-12-28 02:39:09 +00:00
|
|
|
}
|
|
|
|
|
|
2018-05-24 06:56:48 +00:00
|
|
|
ulen = info.nr_jited_ksyms;
|
2018-11-02 17:16:17 +00:00
|
|
|
info.nr_jited_ksyms = prog->aux->func_cnt ? : 1;
|
2018-12-10 19:17:50 +00:00
|
|
|
if (ulen) {
|
2020-07-02 22:45:23 +00:00
|
|
|
if (bpf_dump_raw_ok(file->f_cred)) {
|
2018-11-02 17:16:17 +00:00
|
|
|
unsigned long ksym_addr;
|
2018-05-24 06:56:48 +00:00
|
|
|
u64 __user *user_ksyms;
|
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
|
|
/* copy the address of the kernel symbol
|
|
|
|
|
* corresponding to each function
|
|
|
|
|
*/
|
|
|
|
|
ulen = min_t(u32, info.nr_jited_ksyms, ulen);
|
|
|
|
|
user_ksyms = u64_to_user_ptr(info.jited_ksyms);
|
2018-11-02 17:16:17 +00:00
|
|
|
if (prog->aux->func_cnt) {
|
|
|
|
|
for (i = 0; i < ulen; i++) {
|
|
|
|
|
ksym_addr = (unsigned long)
|
|
|
|
|
prog->aux->func[i]->bpf_func;
|
|
|
|
|
if (put_user((u64) ksym_addr,
|
|
|
|
|
&user_ksyms[i]))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
ksym_addr = (unsigned long) prog->bpf_func;
|
|
|
|
|
if (put_user((u64) ksym_addr, &user_ksyms[0]))
|
2018-05-24 06:56:48 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
info.jited_ksyms = 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-05-24 06:56:52 +00:00
|
|
|
ulen = info.nr_jited_func_lens;
|
2018-11-02 17:16:17 +00:00
|
|
|
info.nr_jited_func_lens = prog->aux->func_cnt ? : 1;
|
2018-12-10 19:17:50 +00:00
|
|
|
if (ulen) {
|
2020-07-02 22:45:23 +00:00
|
|
|
if (bpf_dump_raw_ok(file->f_cred)) {
|
2018-05-24 06:56:52 +00:00
|
|
|
u32 __user *user_lens;
|
|
|
|
|
u32 func_len, i;
|
|
|
|
|
|
|
|
|
|
/* copy the JITed image lengths for each function */
|
|
|
|
|
ulen = min_t(u32, info.nr_jited_func_lens, ulen);
|
|
|
|
|
user_lens = u64_to_user_ptr(info.jited_func_lens);
|
2018-11-02 17:16:17 +00:00
|
|
|
if (prog->aux->func_cnt) {
|
|
|
|
|
for (i = 0; i < ulen; i++) {
|
|
|
|
|
func_len =
|
|
|
|
|
prog->aux->func[i]->jited_len;
|
|
|
|
|
if (put_user(func_len, &user_lens[i]))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
func_len = prog->jited_len;
|
|
|
|
|
if (put_user(func_len, &user_lens[0]))
|
2018-05-24 06:56:52 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
info.jited_func_lens = 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-12-06 01:35:43 +00:00
|
|
|
if (prog->aux->btf)
|
bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier
Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.
This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.
Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.
Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org
2020-12-03 20:46:29 +00:00
|
|
|
info.btf_id = btf_obj_id(prog->aux->btf);
|
2022-06-28 17:43:08 +00:00
|
|
|
info.attach_btf_id = prog->aux->attach_btf_id;
|
2022-08-04 20:11:39 +00:00
|
|
|
if (attach_btf)
|
|
|
|
|
info.attach_btf_obj_id = btf_obj_id(attach_btf);
|
bpf: Introduce bpf_func_info
This patch added interface to load a program with the following
additional information:
. prog_btf_fd
. func_info, func_info_rec_size and func_info_cnt
where func_info will provide function range and type_id
corresponding to each function.
The func_info_rec_size is introduced in the UAPI to specify
struct bpf_func_info size passed from user space. This
intends to make bpf_func_info structure growable in the future.
If the kernel gets a different bpf_func_info size from userspace,
it will try to handle user request with part of bpf_func_info
it can understand. In this patch, kernel can understand
struct bpf_func_info {
__u32 insn_offset;
__u32 type_id;
};
If user passed a bpf func_info record size of 16 bytes, the
kernel can still handle part of records with the above definition.
If verifier agrees with function range provided by the user,
the bpf_prog ksym for each function will use the func name
provided in the type_id, which is supposed to provide better
encoding as it is not limited by 16 bytes program name
limitation and this is better for bpf program which contains
multiple subprograms.
The bpf_prog_info interface is also extended to
return btf_id, func_info, func_info_rec_size and func_info_cnt
to userspace, so userspace can print out the function prototype
for each xlated function. The insn_offset in the returned
func_info corresponds to the insn offset for xlated functions.
With other jit related fields in bpf_prog_info, userspace can also
print out function prototypes for each jited function.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-11-19 23:29:11 +00:00
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
ulen = info.nr_func_info;
|
|
|
|
|
info.nr_func_info = prog->aux->func_info_cnt;
|
|
|
|
|
if (info.nr_func_info && ulen) {
|
2018-12-12 18:18:21 +00:00
|
|
|
char __user *user_finfo;
|
2018-12-06 01:35:43 +00:00
|
|
|
|
2018-12-12 18:18:21 +00:00
|
|
|
user_finfo = u64_to_user_ptr(info.func_info);
|
|
|
|
|
ulen = min_t(u32, info.nr_func_info, ulen);
|
|
|
|
|
if (copy_to_user(user_finfo, prog->aux->func_info,
|
|
|
|
|
info.func_info_rec_size * ulen))
|
|
|
|
|
return -EFAULT;
|
bpf: Introduce bpf_func_info
This patch added interface to load a program with the following
additional information:
. prog_btf_fd
. func_info, func_info_rec_size and func_info_cnt
where func_info will provide function range and type_id
corresponding to each function.
The func_info_rec_size is introduced in the UAPI to specify
struct bpf_func_info size passed from user space. This
intends to make bpf_func_info structure growable in the future.
If the kernel gets a different bpf_func_info size from userspace,
it will try to handle user request with part of bpf_func_info
it can understand. In this patch, kernel can understand
struct bpf_func_info {
__u32 insn_offset;
__u32 type_id;
};
If user passed a bpf func_info record size of 16 bytes, the
kernel can still handle part of records with the above definition.
If verifier agrees with function range provided by the user,
the bpf_prog ksym for each function will use the func name
provided in the type_id, which is supposed to provide better
encoding as it is not limited by 16 bytes program name
limitation and this is better for bpf program which contains
multiple subprograms.
The bpf_prog_info interface is also extended to
return btf_id, func_info, func_info_rec_size and func_info_cnt
to userspace, so userspace can print out the function prototype
for each xlated function. The insn_offset in the returned
func_info corresponds to the insn offset for xlated functions.
With other jit related fields in bpf_prog_info, userspace can also
print out function prototypes for each jited function.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-11-19 23:29:11 +00:00
|
|
|
}
|
|
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
ulen = info.nr_line_info;
|
|
|
|
|
info.nr_line_info = prog->aux->nr_linfo;
|
|
|
|
|
if (info.nr_line_info && ulen) {
|
2018-12-12 18:18:21 +00:00
|
|
|
__u8 __user *user_linfo;
|
2018-12-08 00:42:25 +00:00
|
|
|
|
2018-12-12 18:18:21 +00:00
|
|
|
user_linfo = u64_to_user_ptr(info.line_info);
|
|
|
|
|
ulen = min_t(u32, info.nr_line_info, ulen);
|
|
|
|
|
if (copy_to_user(user_linfo, prog->aux->linfo,
|
|
|
|
|
info.line_info_rec_size * ulen))
|
|
|
|
|
return -EFAULT;
|
2018-12-08 00:42:25 +00:00
|
|
|
}
|
|
|
|
|
|
2018-12-10 22:14:08 +00:00
|
|
|
ulen = info.nr_jited_line_info;
|
2018-12-08 00:42:25 +00:00
|
|
|
if (prog->aux->jited_linfo)
|
2018-12-10 22:14:08 +00:00
|
|
|
info.nr_jited_line_info = prog->aux->nr_linfo;
|
2018-12-08 00:42:25 +00:00
|
|
|
else
|
2018-12-10 22:14:08 +00:00
|
|
|
info.nr_jited_line_info = 0;
|
|
|
|
|
if (info.nr_jited_line_info && ulen) {
|
2020-07-02 22:45:23 +00:00
|
|
|
if (bpf_dump_raw_ok(file->f_cred)) {
|
2022-05-30 09:28:10 +00:00
|
|
|
unsigned long line_addr;
|
2018-12-08 00:42:25 +00:00
|
|
|
__u64 __user *user_linfo;
|
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
|
|
user_linfo = u64_to_user_ptr(info.jited_line_info);
|
2018-12-10 22:14:08 +00:00
|
|
|
ulen = min_t(u32, info.nr_jited_line_info, ulen);
|
2018-12-08 00:42:25 +00:00
|
|
|
for (i = 0; i < ulen; i++) {
|
2022-05-30 09:28:10 +00:00
|
|
|
line_addr = (unsigned long)prog->aux->jited_linfo[i];
|
|
|
|
|
if (put_user((__u64)line_addr, &user_linfo[i]))
|
2018-12-08 00:42:25 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
info.jited_line_info = 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2018-12-12 17:37:46 +00:00
|
|
|
ulen = info.nr_prog_tags;
|
|
|
|
|
info.nr_prog_tags = prog->aux->func_cnt ? : 1;
|
|
|
|
|
if (ulen) {
|
|
|
|
|
__u8 __user (*user_prog_tags)[BPF_TAG_SIZE];
|
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
|
|
user_prog_tags = u64_to_user_ptr(info.prog_tags);
|
|
|
|
|
ulen = min_t(u32, info.nr_prog_tags, ulen);
|
|
|
|
|
if (prog->aux->func_cnt) {
|
|
|
|
|
for (i = 0; i < ulen; i++) {
|
|
|
|
|
if (copy_to_user(user_prog_tags[i],
|
|
|
|
|
prog->aux->func[i]->tag,
|
|
|
|
|
BPF_TAG_SIZE))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
if (copy_to_user(user_prog_tags[0],
|
|
|
|
|
prog->tag, BPF_TAG_SIZE))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:52 +00:00
|
|
|
done:
|
|
|
|
|
if (copy_to_user(uinfo, &info, info_len) ||
|
|
|
|
|
put_user(info_len, &uattr->info.info_len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
static int bpf_map_get_info_by_fd(struct file *file,
|
|
|
|
|
struct bpf_map *map,
|
2017-06-05 19:15:52 +00:00
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map_info __user *uinfo = u64_to_user_ptr(attr->info.info);
|
2020-03-20 16:22:58 +00:00
|
|
|
struct bpf_map_info info;
|
2017-06-05 19:15:52 +00:00
|
|
|
u32 info_len = attr->info.info_len;
|
|
|
|
|
int err;
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_check_uarg_tail_zero(USER_BPFPTR(uinfo), sizeof(info), info_len);
|
2017-06-05 19:15:52 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
info_len = min_t(u32, sizeof(info), info_len);
|
|
|
|
|
|
2020-03-20 16:22:58 +00:00
|
|
|
memset(&info, 0, sizeof(info));
|
2017-06-05 19:15:52 +00:00
|
|
|
info.type = map->map_type;
|
|
|
|
|
info.id = map->id;
|
|
|
|
|
info.key_size = map->key_size;
|
|
|
|
|
info.value_size = map->value_size;
|
|
|
|
|
info.max_entries = map->max_entries;
|
|
|
|
|
info.map_flags = map->map_flags;
|
bpf: Add bloom filter map implementation
This patch adds the kernel-side changes for the implementation of
a bpf bloom filter map.
The bloom filter map supports peek (determining whether an element
is present in the map) and push (adding an element to the map)
operations.These operations are exposed to userspace applications
through the already existing syscalls in the following way:
BPF_MAP_LOOKUP_ELEM -> peek
BPF_MAP_UPDATE_ELEM -> push
The bloom filter map does not have keys, only values. In light of
this, the bloom filter map's API matches that of queue stack maps:
user applications use BPF_MAP_LOOKUP_ELEM/BPF_MAP_UPDATE_ELEM
which correspond internally to bpf_map_peek_elem/bpf_map_push_elem,
and bpf programs must use the bpf_map_peek_elem and bpf_map_push_elem
APIs to query or add an element to the bloom filter map. When the
bloom filter map is created, it must be created with a key_size of 0.
For updates, the user will pass in the element to add to the map
as the value, with a NULL key. For lookups, the user will pass in the
element to query in the map as the value, with a NULL key. In the
verifier layer, this requires us to modify the argument type of
a bloom filter's BPF_FUNC_map_peek_elem call to ARG_PTR_TO_MAP_VALUE;
as well, in the syscall layer, we need to copy over the user value
so that in bpf_map_peek_elem, we know which specific value to query.
A few things to please take note of:
* If there are any concurrent lookups + updates, the user is
responsible for synchronizing this to ensure no false negative lookups
occur.
* The number of hashes to use for the bloom filter is configurable from
userspace. If no number is specified, the default used will be 5 hash
functions. The benchmarks later in this patchset can help compare the
performance of using different number of hashes on different entry
sizes. In general, using more hashes decreases both the false positive
rate and the speed of a lookup.
* Deleting an element in the bloom filter map is not supported.
* The bloom filter map may be used as an inner map.
* The "max_entries" size that is specified at map creation time is used
to approximate a reasonable bitmap size for the bloom filter, and is not
otherwise strictly enforced. If the user wishes to insert more entries
into the bloom filter than "max_entries", they may do so but they should
be aware that this may lead to a higher false positive rate.
Signed-off-by: Joanne Koong <joannekoong@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211027234504.30744-2-joannekoong@fb.com
2021-10-27 23:45:00 +00:00
|
|
|
info.map_extra = map->map_extra;
|
2017-09-27 21:37:53 +00:00
|
|
|
memcpy(info.name, map->name, sizeof(map->name));
|
2017-06-05 19:15:52 +00:00
|
|
|
|
2018-05-04 21:49:51 +00:00
|
|
|
if (map->btf) {
|
bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier
Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.
This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.
Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.
Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org
2020-12-03 20:46:29 +00:00
|
|
|
info.btf_id = btf_obj_id(map->btf);
|
2018-05-22 21:57:21 +00:00
|
|
|
info.btf_key_type_id = map->btf_key_type_id;
|
|
|
|
|
info.btf_value_type_id = map->btf_value_type_id;
|
2018-05-04 21:49:51 +00:00
|
|
|
}
|
bpf: Introduce BPF_MAP_TYPE_STRUCT_OPS
The patch introduces BPF_MAP_TYPE_STRUCT_OPS. The map value
is a kernel struct with its func ptr implemented in bpf prog.
This new map is the interface to register/unregister/introspect
a bpf implemented kernel struct.
The kernel struct is actually embedded inside another new struct
(or called the "value" struct in the code). For example,
"struct tcp_congestion_ops" is embbeded in:
struct bpf_struct_ops_tcp_congestion_ops {
refcount_t refcnt;
enum bpf_struct_ops_state state;
struct tcp_congestion_ops data; /* <-- kernel subsystem struct here */
}
The map value is "struct bpf_struct_ops_tcp_congestion_ops".
The "bpftool map dump" will then be able to show the
state ("inuse"/"tobefree") and the number of subsystem's refcnt (e.g.
number of tcp_sock in the tcp_congestion_ops case). This "value" struct
is created automatically by a macro. Having a separate "value" struct
will also make extending "struct bpf_struct_ops_XYZ" easier (e.g. adding
"void (*init)(void)" to "struct bpf_struct_ops_XYZ" to do some
initialization works before registering the struct_ops to the kernel
subsystem). The libbpf will take care of finding and populating the
"struct bpf_struct_ops_XYZ" from "struct XYZ".
Register a struct_ops to a kernel subsystem:
1. Load all needed BPF_PROG_TYPE_STRUCT_OPS prog(s)
2. Create a BPF_MAP_TYPE_STRUCT_OPS with attr->btf_vmlinux_value_type_id
set to the btf id "struct bpf_struct_ops_tcp_congestion_ops" of the
running kernel.
Instead of reusing the attr->btf_value_type_id,
btf_vmlinux_value_type_id s added such that attr->btf_fd can still be
used as the "user" btf which could store other useful sysadmin/debug
info that may be introduced in the furture,
e.g. creation-date/compiler-details/map-creator...etc.
3. Create a "struct bpf_struct_ops_tcp_congestion_ops" object as described
in the running kernel btf. Populate the value of this object.
The function ptr should be populated with the prog fds.
4. Call BPF_MAP_UPDATE with the object created in (3) as
the map value. The key is always "0".
During BPF_MAP_UPDATE, the code that saves the kernel-func-ptr's
args as an array of u64 is generated. BPF_MAP_UPDATE also allows
the specific struct_ops to do some final checks in "st_ops->init_member()"
(e.g. ensure all mandatory func ptrs are implemented).
If everything looks good, it will register this kernel struct
to the kernel subsystem. The map will not allow further update
from this point.
Unregister a struct_ops from the kernel subsystem:
BPF_MAP_DELETE with key "0".
Introspect a struct_ops:
BPF_MAP_LOOKUP_ELEM with key "0". The map value returned will
have the prog _id_ populated as the func ptr.
The map value state (enum bpf_struct_ops_state) will transit from:
INIT (map created) =>
INUSE (map updated, i.e. reg) =>
TOBEFREE (map value deleted, i.e. unreg)
The kernel subsystem needs to call bpf_struct_ops_get() and
bpf_struct_ops_put() to manage the "refcnt" in the
"struct bpf_struct_ops_XYZ". This patch uses a separate refcnt
for the purose of tracking the subsystem usage. Another approach
is to reuse the map->refcnt and then "show" (i.e. during map_lookup)
the subsystem's usage by doing map->refcnt - map->usercnt to filter out
the map-fd/pinned-map usage. However, that will also tie down the
future semantics of map->refcnt and map->usercnt.
The very first subsystem's refcnt (during reg()) holds one
count to map->refcnt. When the very last subsystem's refcnt
is gone, it will also release the map->refcnt. All bpf_prog will be
freed when the map->refcnt reaches 0 (i.e. during map_free()).
Here is how the bpftool map command will look like:
[root@arch-fb-vm1 bpf]# bpftool map show
6: struct_ops name dctcp flags 0x0
key 4B value 256B max_entries 1 memlock 4096B
btf_id 6
[root@arch-fb-vm1 bpf]# bpftool map dump id 6
[{
"value": {
"refcnt": {
"refs": {
"counter": 1
}
},
"state": 1,
"data": {
"list": {
"next": 0,
"prev": 0
},
"key": 0,
"flags": 2,
"init": 24,
"release": 0,
"ssthresh": 25,
"cong_avoid": 30,
"set_state": 27,
"cwnd_event": 28,
"in_ack_event": 26,
"undo_cwnd": 29,
"pkts_acked": 0,
"min_tso_segs": 0,
"sndbuf_expand": 0,
"cong_control": 0,
"get_info": 0,
"name": [98,112,102,95,100,99,116,99,112,0,0,0,0,0,0,0
],
"owner": 0
}
}
}
]
Misc Notes:
* bpf_struct_ops_map_sys_lookup_elem() is added for syscall lookup.
It does an inplace update on "*value" instead returning a pointer
to syscall.c. Otherwise, it needs a separate copy of "zero" value
for the BPF_STRUCT_OPS_STATE_INIT to avoid races.
* The bpf_struct_ops_map_delete_elem() is also called without
preempt_disable() from map_delete_elem(). It is because
the "->unreg()" may requires sleepable context, e.g.
the "tcp_unregister_congestion_control()".
* "const" is added to some of the existing "struct btf_func_model *"
function arg to avoid a compiler warning caused by this patch.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003505.3855919-1-kafai@fb.com
2020-01-09 00:35:05 +00:00
|
|
|
info.btf_vmlinux_value_type_id = map->btf_vmlinux_value_type_id;
|
2024-01-19 22:49:57 +00:00
|
|
|
if (map->map_type == BPF_MAP_TYPE_STRUCT_OPS)
|
|
|
|
|
bpf_map_struct_ops_info_fill(&info, map);
|
2018-05-04 21:49:51 +00:00
|
|
|
|
2023-01-19 22:15:21 +00:00
|
|
|
if (bpf_map_is_offloaded(map)) {
|
2018-01-18 03:13:28 +00:00
|
|
|
err = bpf_map_offload_info_fill(&info, map);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2017-06-05 19:15:52 +00:00
|
|
|
if (copy_to_user(uinfo, &info, info_len) ||
|
|
|
|
|
put_user(info_len, &uattr->info.info_len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
static int bpf_btf_get_info_by_fd(struct file *file,
|
|
|
|
|
struct btf *btf,
|
2018-05-04 21:49:52 +00:00
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_btf_info __user *uinfo = u64_to_user_ptr(attr->info.info);
|
|
|
|
|
u32 info_len = attr->info.info_len;
|
|
|
|
|
int err;
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_check_uarg_tail_zero(USER_BPFPTR(uinfo), sizeof(*uinfo), info_len);
|
2018-05-04 21:49:52 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
|
|
|
|
|
return btf_get_info_by_fd(btf, attr, uattr);
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-02 22:45:23 +00:00
|
|
|
static int bpf_link_get_info_by_fd(struct file *file,
|
|
|
|
|
struct bpf_link *link,
|
2020-04-29 00:16:08 +00:00
|
|
|
const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link_info __user *uinfo = u64_to_user_ptr(attr->info.info);
|
|
|
|
|
struct bpf_link_info info;
|
|
|
|
|
u32 info_len = attr->info.info_len;
|
|
|
|
|
int err;
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_check_uarg_tail_zero(USER_BPFPTR(uinfo), sizeof(info), info_len);
|
2020-04-29 00:16:08 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
info_len = min_t(u32, sizeof(info), info_len);
|
|
|
|
|
|
|
|
|
|
memset(&info, 0, sizeof(info));
|
|
|
|
|
if (copy_from_user(&info, uinfo, info_len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
|
|
info.type = link->type;
|
|
|
|
|
info.id = link->id;
|
bpf: Create links for BPF struct_ops maps.
Make bpf_link support struct_ops. Previously, struct_ops were always
used alone without any associated links. Upon updating its value, a
struct_ops would be activated automatically. Yet other BPF program
types required to make a bpf_link with their instances before they
could become active. Now, however, you can create an inactive
struct_ops, and create a link to activate it later.
With bpf_links, struct_ops has a behavior similar to other BPF program
types. You can pin/unpin them from their links and the struct_ops will
be deactivated when its link is removed while previously need someone
to delete the value for it to be deactivated.
bpf_links are responsible for registering their associated
struct_ops. You can only use a struct_ops that has the BPF_F_LINK flag
set to create a bpf_link, while a structs without this flag behaves in
the same manner as before and is registered upon updating its value.
The BPF_LINK_TYPE_STRUCT_OPS serves a dual purpose. Not only is it
used to craft the links for BPF struct_ops programs, but also to
create links for BPF struct_ops them-self. Since the links of BPF
struct_ops programs are only used to create trampolines internally,
they are never seen in other contexts. Thus, they can be reused for
struct_ops themself.
To maintain a reference to the map supporting this link, we add
bpf_struct_ops_link as an additional type. The pointer of the map is
RCU and won't be necessary until later in the patchset.
Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Link: https://lore.kernel.org/r/20230323032405.3735486-4-kuifeng@meta.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-03-23 03:24:00 +00:00
|
|
|
if (link->prog)
|
|
|
|
|
info.prog_id = link->prog->aux->id;
|
2020-04-29 00:16:08 +00:00
|
|
|
|
|
|
|
|
if (link->ops->fill_link_info) {
|
|
|
|
|
err = link->ops->fill_link_info(link, &info);
|
|
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (copy_to_user(uinfo, &info, info_len) ||
|
|
|
|
|
put_user(info_len, &uattr->info.info_len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2017-06-05 19:15:52 +00:00
|
|
|
#define BPF_OBJ_GET_INFO_BY_FD_LAST_FIELD info.info
|
|
|
|
|
|
|
|
|
|
static int bpf_obj_get_info_by_fd(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
int ufd = attr->info.bpf_fd;
|
|
|
|
|
struct fd f;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_OBJ_GET_INFO_BY_FD))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
f = fdget(ufd);
|
|
|
|
|
if (!f.file)
|
|
|
|
|
return -EBADFD;
|
|
|
|
|
|
|
|
|
|
if (f.file->f_op == &bpf_prog_fops)
|
2020-07-02 22:45:23 +00:00
|
|
|
err = bpf_prog_get_info_by_fd(f.file, f.file->private_data, attr,
|
2017-06-05 19:15:52 +00:00
|
|
|
uattr);
|
|
|
|
|
else if (f.file->f_op == &bpf_map_fops)
|
2020-07-02 22:45:23 +00:00
|
|
|
err = bpf_map_get_info_by_fd(f.file, f.file->private_data, attr,
|
2017-06-05 19:15:52 +00:00
|
|
|
uattr);
|
2018-04-18 22:56:02 +00:00
|
|
|
else if (f.file->f_op == &btf_fops)
|
2020-07-02 22:45:23 +00:00
|
|
|
err = bpf_btf_get_info_by_fd(f.file, f.file->private_data, attr, uattr);
|
2020-04-29 00:16:08 +00:00
|
|
|
else if (f.file->f_op == &bpf_link_fops)
|
2020-07-02 22:45:23 +00:00
|
|
|
err = bpf_link_get_info_by_fd(f.file, f.file->private_data,
|
2020-04-29 00:16:08 +00:00
|
|
|
attr, uattr);
|
2017-06-05 19:15:52 +00:00
|
|
|
else
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
|
|
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2024-01-24 02:21:02 +00:00
|
|
|
#define BPF_BTF_LOAD_LAST_FIELD btf_token_fd
|
2018-04-18 22:56:01 +00:00
|
|
|
|
2023-04-06 23:41:58 +00:00
|
|
|
static int bpf_btf_load(const union bpf_attr *attr, bpfptr_t uattr, __u32 uattr_size)
|
2018-04-18 22:56:01 +00:00
|
|
|
{
|
2024-01-24 02:21:02 +00:00
|
|
|
struct bpf_token *token = NULL;
|
|
|
|
|
|
2018-04-18 22:56:01 +00:00
|
|
|
if (CHECK_ATTR(BPF_BTF_LOAD))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2024-01-24 02:21:02 +00:00
|
|
|
if (attr->btf_flags & ~BPF_F_TOKEN_FD)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (attr->btf_flags & BPF_F_TOKEN_FD) {
|
|
|
|
|
token = bpf_token_get_from_fd(attr->btf_token_fd);
|
|
|
|
|
if (IS_ERR(token))
|
|
|
|
|
return PTR_ERR(token);
|
|
|
|
|
if (!bpf_token_allow_cmd(token, BPF_BTF_LOAD)) {
|
|
|
|
|
bpf_token_put(token);
|
|
|
|
|
token = NULL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!bpf_token_capable(token, CAP_BPF)) {
|
|
|
|
|
bpf_token_put(token);
|
2018-04-18 22:56:01 +00:00
|
|
|
return -EPERM;
|
2024-01-24 02:21:02 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bpf_token_put(token);
|
2018-04-18 22:56:01 +00:00
|
|
|
|
2023-04-06 23:41:58 +00:00
|
|
|
return btf_new_fd(attr, uattr, uattr_size);
|
2018-04-18 22:56:01 +00:00
|
|
|
}
|
|
|
|
|
|
2018-05-04 21:49:51 +00:00
|
|
|
#define BPF_BTF_GET_FD_BY_ID_LAST_FIELD btf_id
|
|
|
|
|
|
|
|
|
|
static int bpf_btf_get_fd_by_id(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
if (CHECK_ATTR(BPF_BTF_GET_FD_BY_ID))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
return btf_get_fd_by_id(attr->btf_id);
|
|
|
|
|
}
|
|
|
|
|
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
static int bpf_task_fd_query_copy(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr,
|
|
|
|
|
u32 prog_id, u32 fd_type,
|
|
|
|
|
const char *buf, u64 probe_offset,
|
|
|
|
|
u64 probe_addr)
|
|
|
|
|
{
|
|
|
|
|
char __user *ubuf = u64_to_user_ptr(attr->task_fd_query.buf);
|
|
|
|
|
u32 len = buf ? strlen(buf) : 0, input_len;
|
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
|
|
if (put_user(len, &uattr->task_fd_query.buf_len))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
input_len = attr->task_fd_query.buf_len;
|
|
|
|
|
if (input_len && ubuf) {
|
|
|
|
|
if (!len) {
|
|
|
|
|
/* nothing to copy, just make ubuf NULL terminated */
|
|
|
|
|
char zero = '\0';
|
|
|
|
|
|
|
|
|
|
if (put_user(zero, ubuf))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
} else if (input_len >= len + 1) {
|
|
|
|
|
/* ubuf can hold the string with NULL terminator */
|
|
|
|
|
if (copy_to_user(ubuf, buf, len + 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
} else {
|
|
|
|
|
/* ubuf cannot hold the string with NULL terminator,
|
|
|
|
|
* do a partial copy with NULL terminator.
|
|
|
|
|
*/
|
|
|
|
|
char zero = '\0';
|
|
|
|
|
|
|
|
|
|
err = -ENOSPC;
|
|
|
|
|
if (copy_to_user(ubuf, buf, input_len - 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
if (put_user(zero, ubuf + input_len - 1))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (put_user(prog_id, &uattr->task_fd_query.prog_id) ||
|
|
|
|
|
put_user(fd_type, &uattr->task_fd_query.fd_type) ||
|
|
|
|
|
put_user(probe_offset, &uattr->task_fd_query.probe_offset) ||
|
|
|
|
|
put_user(probe_addr, &uattr->task_fd_query.probe_addr))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define BPF_TASK_FD_QUERY_LAST_FIELD task_fd_query.probe_addr
|
|
|
|
|
|
|
|
|
|
static int bpf_task_fd_query(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
pid_t pid = attr->task_fd_query.pid;
|
|
|
|
|
u32 fd = attr->task_fd_query.fd;
|
|
|
|
|
const struct perf_event *event;
|
|
|
|
|
struct task_struct *task;
|
|
|
|
|
struct file *file;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_TASK_FD_QUERY))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
if (attr->task_fd_query.flags != 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2022-09-12 13:38:55 +00:00
|
|
|
rcu_read_lock();
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
task = get_pid_task(find_vpid(pid), PIDTYPE_PID);
|
2022-09-12 13:38:55 +00:00
|
|
|
rcu_read_unlock();
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
if (!task)
|
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
|
|
err = 0;
|
2020-11-20 23:14:22 +00:00
|
|
|
file = fget_task(task, fd);
|
|
|
|
|
put_task_struct(task);
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
if (!file)
|
2020-11-20 23:14:22 +00:00
|
|
|
return -EBADF;
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
if (file->f_op == &bpf_link_fops) {
|
|
|
|
|
struct bpf_link *link = file->private_data;
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
|
2020-04-29 00:16:06 +00:00
|
|
|
if (link->ops == &bpf_raw_tp_link_lops) {
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
struct bpf_raw_tp_link *raw_tp =
|
|
|
|
|
container_of(link, struct bpf_raw_tp_link, link);
|
|
|
|
|
struct bpf_raw_event_map *btp = raw_tp->btp;
|
|
|
|
|
|
|
|
|
|
err = bpf_task_fd_query_copy(attr, uattr,
|
|
|
|
|
raw_tp->link.prog->aux->id,
|
|
|
|
|
BPF_FD_TYPE_RAW_TRACEPOINT,
|
|
|
|
|
btp->tp->name, 0, 0);
|
|
|
|
|
goto put_file;
|
|
|
|
|
}
|
|
|
|
|
goto out_not_supp;
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
event = perf_get_event(file);
|
|
|
|
|
if (!IS_ERR(event)) {
|
|
|
|
|
u64 probe_offset, probe_addr;
|
|
|
|
|
u32 prog_id, fd_type;
|
|
|
|
|
const char *buf;
|
|
|
|
|
|
|
|
|
|
err = bpf_get_perf_event_info(event, &prog_id, &fd_type,
|
|
|
|
|
&buf, &probe_offset,
|
2023-09-20 21:31:39 +00:00
|
|
|
&probe_addr, NULL);
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
if (!err)
|
|
|
|
|
err = bpf_task_fd_query_copy(attr, uattr, prog_id,
|
|
|
|
|
fd_type, buf,
|
|
|
|
|
probe_offset,
|
|
|
|
|
probe_addr);
|
|
|
|
|
goto put_file;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce pinnable bpf_link abstraction
Introduce bpf_link abstraction, representing an attachment of BPF program to
a BPF hook point (e.g., tracepoint, perf event, etc). bpf_link encapsulates
ownership of attached BPF program, reference counting of a link itself, when
reference from multiple anonymous inodes, as well as ensures that release
callback will be called from a process context, so that users can safely take
mutex locks and sleep.
Additionally, with a new abstraction it's now possible to generalize pinning
of a link object in BPF FS, allowing to explicitly prevent BPF program
detachment on process exit by pinning it in a BPF FS and let it open from
independent other process to keep working with it.
Convert two existing bpf_link-like objects (raw tracepoint and tracing BPF
program attachments) into utilizing bpf_link framework, making them pinnable
in BPF FS. More FD-based bpf_links will be added in follow up patches.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200303043159.323675-2-andriin@fb.com
2020-03-03 04:31:57 +00:00
|
|
|
out_not_supp:
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
err = -ENOTSUPP;
|
|
|
|
|
put_file:
|
|
|
|
|
fput(file);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-01-15 18:43:01 +00:00
|
|
|
#define BPF_MAP_BATCH_LAST_FIELD batch.flags
|
|
|
|
|
|
2022-11-16 07:50:58 +00:00
|
|
|
#define BPF_DO_BATCH(fn, ...) \
|
2020-01-15 18:43:01 +00:00
|
|
|
do { \
|
|
|
|
|
if (!fn) { \
|
|
|
|
|
err = -ENOTSUPP; \
|
|
|
|
|
goto err_put; \
|
|
|
|
|
} \
|
2022-11-16 07:50:58 +00:00
|
|
|
err = fn(__VA_ARGS__); \
|
2020-01-15 18:43:01 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
|
|
|
|
static int bpf_map_do_batch(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr,
|
|
|
|
|
int cmd)
|
|
|
|
|
{
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bool has_read = cmd == BPF_MAP_LOOKUP_BATCH ||
|
|
|
|
|
cmd == BPF_MAP_LOOKUP_AND_DELETE_BATCH;
|
|
|
|
|
bool has_write = cmd != BPF_MAP_LOOKUP_BATCH;
|
2020-01-15 18:43:01 +00:00
|
|
|
struct bpf_map *map;
|
|
|
|
|
int err, ufd;
|
|
|
|
|
struct fd f;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_MAP_BATCH))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
ufd = attr->batch.map_fd;
|
|
|
|
|
f = fdget(ufd);
|
|
|
|
|
map = __bpf_map_get(f);
|
|
|
|
|
if (IS_ERR(map))
|
|
|
|
|
return PTR_ERR(map);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
if (has_write)
|
|
|
|
|
bpf_map_write_active_inc(map);
|
|
|
|
|
if (has_read && !(map_get_sys_perms(map, f) & FMODE_CAN_READ)) {
|
2020-01-15 18:43:01 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
if (has_write && !(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) {
|
2020-01-15 18:43:01 +00:00
|
|
|
err = -EPERM;
|
|
|
|
|
goto err_put;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (cmd == BPF_MAP_LOOKUP_BATCH)
|
2022-11-16 07:50:58 +00:00
|
|
|
BPF_DO_BATCH(map->ops->map_lookup_batch, map, attr, uattr);
|
2020-01-15 18:43:04 +00:00
|
|
|
else if (cmd == BPF_MAP_LOOKUP_AND_DELETE_BATCH)
|
2022-11-16 07:50:58 +00:00
|
|
|
BPF_DO_BATCH(map->ops->map_lookup_and_delete_batch, map, attr, uattr);
|
2020-01-15 18:43:02 +00:00
|
|
|
else if (cmd == BPF_MAP_UPDATE_BATCH)
|
2022-11-16 07:50:58 +00:00
|
|
|
BPF_DO_BATCH(map->ops->map_update_batch, map, f.file, attr, uattr);
|
2020-01-15 18:43:02 +00:00
|
|
|
else
|
2022-11-16 07:50:58 +00:00
|
|
|
BPF_DO_BATCH(map->ops->map_delete_batch, map, attr, uattr);
|
2020-01-15 18:43:01 +00:00
|
|
|
err_put:
|
2023-12-08 10:23:51 +00:00
|
|
|
if (has_write) {
|
|
|
|
|
maybe_wait_bpf_programs(map);
|
bpf: Fix toctou on read-only map's constant scalar tracking
Commit a23740ec43ba ("bpf: Track contents of read-only maps as scalars") is
checking whether maps are read-only both from BPF program side and user space
side, and then, given their content is constant, reading out their data via
map->ops->map_direct_value_addr() which is then subsequently used as known
scalar value for the register, that is, it is marked as __mark_reg_known()
with the read value at verification time. Before a23740ec43ba, the register
content was marked as an unknown scalar so the verifier could not make any
assumptions about the map content.
The current implementation however is prone to a TOCTOU race, meaning, the
value read as known scalar for the register is not guaranteed to be exactly
the same at a later point when the program is executed, and as such, the
prior made assumptions of the verifier with regards to the program will be
invalid which can cause issues such as OOB access, etc.
While the BPF_F_RDONLY_PROG map flag is always fixed and required to be
specified at map creation time, the map->frozen property is initially set to
false for the map given the map value needs to be populated, e.g. for global
data sections. Once complete, the loader "freezes" the map from user space
such that no subsequent updates/deletes are possible anymore. For the rest
of the lifetime of the map, this freeze one-time trigger cannot be undone
anymore after a successful BPF_MAP_FREEZE cmd return. Meaning, any new BPF_*
cmd calls which would update/delete map entries will be rejected with -EPERM
since map_get_sys_perms() removes the FMODE_CAN_WRITE permission. This also
means that pending update/delete map entries must still complete before this
guarantee is given. This corner case is not an issue for loaders since they
create and prepare such program private map in successive steps.
However, a malicious user is able to trigger this TOCTOU race in two different
ways: i) via userfaultfd, and ii) via batched updates. For i) userfaultfd is
used to expand the competition interval, so that map_update_elem() can modify
the contents of the map after map_freeze() and bpf_prog_load() were executed.
This works, because userfaultfd halts the parallel thread which triggered a
map_update_elem() at the time where we copy key/value from the user buffer and
this already passed the FMODE_CAN_WRITE capability test given at that time the
map was not "frozen". Then, the main thread performs the map_freeze() and
bpf_prog_load(), and once that had completed successfully, the other thread
is woken up to complete the pending map_update_elem() which then changes the
map content. For ii) the idea of the batched update is similar, meaning, when
there are a large number of updates to be processed, it can increase the
competition interval between the two. It is therefore possible in practice to
modify the contents of the map after executing map_freeze() and bpf_prog_load().
One way to fix both i) and ii) at the same time is to expand the use of the
map's map->writecnt. The latter was introduced in fc9702273e2e ("bpf: Add mmap()
support for BPF_MAP_TYPE_ARRAY") and further refined in 1f6cb19be2e2 ("bpf:
Prevent re-mmap()'ing BPF map as writable for initially r/o mapping") with
the rationale to make a writable mmap()'ing of a map mutually exclusive with
read-only freezing. The counter indicates writable mmap() mappings and then
prevents/fails the freeze operation. Its semantics can be expanded beyond
just mmap() by generally indicating ongoing write phases. This would essentially
span any parallel regular and batched flavor of update/delete operation and
then also have map_freeze() fail with -EBUSY. For the check_mem_access() in
the verifier we expand upon the bpf_map_is_rdonly() check ensuring that all
last pending writes have completed via bpf_map_write_active() test. Once the
map->frozen is set and bpf_map_write_active() indicates a map->writecnt of 0
only then we are really guaranteed to use the map's data as known constants.
For map->frozen being set and pending writes in process of still being completed
we fall back to marking that register as unknown scalar so we don't end up
making assumptions about it. With this, both TOCTOU reproducers from i) and
ii) are fixed.
Note that the map->writecnt has been converted into a atomic64 in the fix in
order to avoid a double freeze_mutex mutex_{un,}lock() pair when updating
map->writecnt in the various map update/delete BPF_* cmd flavors. Spanning
the freeze_mutex over entire map update/delete operations in syscall side
would not be possible due to then causing everything to be serialized.
Similarly, something like synchronize_rcu() after setting map->frozen to wait
for update/deletes to complete is not possible either since it would also
have to span the user copy which can sleep. On the libbpf side, this won't
break d66562fba1ce ("libbpf: Add BPF object skeleton support") as the
anonymous mmap()-ed "map initialization image" is remapped as a BPF map-backed
mmap()-ed memory where for .rodata it's non-writable.
Fixes: a23740ec43ba ("bpf: Track contents of read-only maps as scalars")
Reported-by: w1tcher.bupt@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2021-11-09 18:48:08 +00:00
|
|
|
bpf_map_write_active_dec(map);
|
2023-12-08 10:23:51 +00:00
|
|
|
}
|
2020-01-15 18:43:01 +00:00
|
|
|
fdput(f);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2023-08-09 08:34:17 +00:00
|
|
|
#define BPF_LINK_CREATE_LAST_FIELD link_create.uprobe_multi.pid
|
2021-05-14 00:36:05 +00:00
|
|
|
static int link_create(union bpf_attr *attr, bpfptr_t uattr)
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_LINK_CREATE))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
bpf: Create links for BPF struct_ops maps.
Make bpf_link support struct_ops. Previously, struct_ops were always
used alone without any associated links. Upon updating its value, a
struct_ops would be activated automatically. Yet other BPF program
types required to make a bpf_link with their instances before they
could become active. Now, however, you can create an inactive
struct_ops, and create a link to activate it later.
With bpf_links, struct_ops has a behavior similar to other BPF program
types. You can pin/unpin them from their links and the struct_ops will
be deactivated when its link is removed while previously need someone
to delete the value for it to be deactivated.
bpf_links are responsible for registering their associated
struct_ops. You can only use a struct_ops that has the BPF_F_LINK flag
set to create a bpf_link, while a structs without this flag behaves in
the same manner as before and is registered upon updating its value.
The BPF_LINK_TYPE_STRUCT_OPS serves a dual purpose. Not only is it
used to craft the links for BPF struct_ops programs, but also to
create links for BPF struct_ops them-self. Since the links of BPF
struct_ops programs are only used to create trampolines internally,
they are never seen in other contexts. Thus, they can be reused for
struct_ops themself.
To maintain a reference to the map supporting this link, we add
bpf_struct_ops_link as an additional type. The pointer of the map is
RCU and won't be necessary until later in the patchset.
Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Link: https://lore.kernel.org/r/20230323032405.3735486-4-kuifeng@meta.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-03-23 03:24:00 +00:00
|
|
|
if (attr->link_create.attach_type == BPF_STRUCT_OPS)
|
|
|
|
|
return bpf_struct_ops_link_create(attr);
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
prog = bpf_prog_get(attr->link_create.prog_fd);
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
ret = bpf_prog_attach_check_attach_type(prog,
|
|
|
|
|
attr->link_create.attach_type);
|
|
|
|
|
if (ret)
|
2020-09-29 12:45:51 +00:00
|
|
|
goto out;
|
|
|
|
|
|
2022-04-21 03:39:43 +00:00
|
|
|
switch (prog->type) {
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
case BPF_PROG_TYPE_CGROUP_SKB:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
|
|
|
|
|
case BPF_PROG_TYPE_SOCK_OPS:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_DEVICE:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SYSCTL:
|
|
|
|
|
case BPF_PROG_TYPE_CGROUP_SOCKOPT:
|
|
|
|
|
ret = cgroup_bpf_link_attach(attr, prog);
|
|
|
|
|
break;
|
2022-04-21 03:39:43 +00:00
|
|
|
case BPF_PROG_TYPE_EXT:
|
|
|
|
|
ret = bpf_tracing_prog_attach(prog,
|
|
|
|
|
attr->link_create.target_fd,
|
2022-05-10 20:59:21 +00:00
|
|
|
attr->link_create.target_btf_id,
|
|
|
|
|
attr->link_create.tracing.cookie);
|
2022-04-21 03:39:43 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_PROG_TYPE_LSM:
|
2020-05-09 17:59:01 +00:00
|
|
|
case BPF_PROG_TYPE_TRACING:
|
2022-04-21 03:39:43 +00:00
|
|
|
if (attr->link_create.attach_type != prog->expected_attach_type) {
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
if (prog->expected_attach_type == BPF_TRACE_RAW_TP)
|
|
|
|
|
ret = bpf_raw_tp_link_attach(prog, NULL);
|
|
|
|
|
else if (prog->expected_attach_type == BPF_TRACE_ITER)
|
|
|
|
|
ret = bpf_iter_link_attach(attr, uattr, prog);
|
2022-06-28 17:43:06 +00:00
|
|
|
else if (prog->expected_attach_type == BPF_LSM_CGROUP)
|
|
|
|
|
ret = cgroup_bpf_link_attach(attr, prog);
|
2022-04-21 03:39:43 +00:00
|
|
|
else
|
|
|
|
|
ret = bpf_tracing_prog_attach(prog,
|
|
|
|
|
attr->link_create.target_fd,
|
2022-05-10 20:59:21 +00:00
|
|
|
attr->link_create.target_btf_id,
|
|
|
|
|
attr->link_create.tracing.cookie);
|
2020-05-09 17:59:01 +00:00
|
|
|
break;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 08:28:38 +00:00
|
|
|
case BPF_PROG_TYPE_FLOW_DISSECTOR:
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 10:35:23 +00:00
|
|
|
case BPF_PROG_TYPE_SK_LOOKUP:
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 08:28:38 +00:00
|
|
|
ret = netns_bpf_link_create(attr, prog);
|
|
|
|
|
break;
|
2020-07-28 19:05:27 +00:00
|
|
|
#ifdef CONFIG_NET
|
2020-07-22 06:45:57 +00:00
|
|
|
case BPF_PROG_TYPE_XDP:
|
|
|
|
|
ret = bpf_xdp_link_attach(attr, prog);
|
2023-04-21 17:02:54 +00:00
|
|
|
break;
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
case BPF_PROG_TYPE_SCHED_CLS:
|
netkit, bpf: Add bpf programmable net device
This work adds a new, minimal BPF-programmable device called "netkit"
(former PoC code-name "meta") we recently presented at LSF/MM/BPF. The
core idea is that BPF programs are executed within the drivers xmit routine
and therefore e.g. in case of containers/Pods moving BPF processing closer
to the source.
One of the goals was that in case of Pod egress traffic, this allows to
move BPF programs from hostns tcx ingress into the device itself, providing
earlier drop or forward mechanisms, for example, if the BPF program
determines that the skb must be sent out of the node, then a redirect to
the physical device can take place directly without going through per-CPU
backlog queue. This helps to shift processing for such traffic from softirq
to process context, leading to better scheduling decisions/performance (see
measurements in the slides).
In this initial version, the netkit device ships as a pair, but we plan to
extend this further so it can also operate in single device mode. The pair
comes with a primary and a peer device. Only the primary device, typically
residing in hostns, can manage BPF programs for itself and its peer. The
peer device is designated for containers/Pods and cannot attach/detach
BPF programs. Upon the device creation, the user can set the default policy
to 'pass' or 'drop' for the case when no BPF program is attached.
Additionally, the device can be operated in L3 (default) or L2 mode. The
management of BPF programs is done via bpf_mprog, so that multi-attach is
supported right from the beginning with similar API and dependency controls
as tcx. For details on the latter see commit 053c8e1f235d ("bpf: Add generic
attach/detach/query API for multi-progs"). tc BPF compatibility is provided,
so that existing programs can be easily migrated.
Going forward, we plan to use netkit devices in Cilium as the main device
type for connecting Pods. They will be operated in L3 mode in order to
simplify a Pod's neighbor management and the peer will operate in default
drop mode, so that no traffic is leaving between the time when a Pod is
brought up by the CNI plugin and programs attached by the agent.
Additionally, the programs we attach via tcx on the physical devices are
using bpf_redirect_peer() for inbound traffic into netkit device, hence the
latter is also supporting the ndo_get_peer_dev callback. Similarly, we use
bpf_redirect_neigh() for the way out, pushing from netkit peer to phys device
directly. Also, BIG TCP is supported on netkit device. For the follow-up
work in single device mode, we plan to convert Cilium's cilium_host/_net
devices into a single one.
An extensive test suite for checking device operations and the BPF program
and link management API comes as BPF selftests in this series.
Co-developed-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Nikolay Aleksandrov <razor@blackwall.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://github.com/borkmann/iproute2/tree/pr/netkit
Link: http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf (24ff.)
Link: https://lore.kernel.org/r/20231024214904.29825-2-daniel@iogearbox.net
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
2023-10-24 21:48:58 +00:00
|
|
|
if (attr->link_create.attach_type == BPF_TCX_INGRESS ||
|
|
|
|
|
attr->link_create.attach_type == BPF_TCX_EGRESS)
|
|
|
|
|
ret = tcx_link_attach(attr, prog);
|
|
|
|
|
else
|
|
|
|
|
ret = netkit_link_attach(attr, prog);
|
bpf: Add fd-based tcx multi-prog infra with link support
This work refactors and adds a lightweight extension ("tcx") to the tc BPF
ingress and egress data path side for allowing BPF program management based
on fds via bpf() syscall through the newly added generic multi-prog API.
The main goal behind this work which we also presented at LPC [0] last year
and a recent update at LSF/MM/BPF this year [3] is to support long-awaited
BPF link functionality for tc BPF programs, which allows for a model of safe
ownership and program detachment.
Given the rise in tc BPF users in cloud native environments, this becomes
necessary to avoid hard to debug incidents either through stale leftover
programs or 3rd party applications accidentally stepping on each others toes.
As a recap, a BPF link represents the attachment of a BPF program to a BPF
hook point. The BPF link holds a single reference to keep BPF program alive.
Moreover, hook points do not reference a BPF link, only the application's
fd or pinning does. A BPF link holds meta-data specific to attachment and
implements operations for link creation, (atomic) BPF program update,
detachment and introspection. The motivation for BPF links for tc BPF programs
is multi-fold, for example:
- From Meta: "It's especially important for applications that are deployed
fleet-wide and that don't "control" hosts they are deployed to. If such
application crashes and no one notices and does anything about that, BPF
program will keep running draining resources or even just, say, dropping
packets. We at FB had outages due to such permanent BPF attachment
semantics. With fd-based BPF link we are getting a framework, which allows
safe, auto-detachable behavior by default, unless application explicitly
opts in by pinning the BPF link." [1]
- From Cilium-side the tc BPF programs we attach to host-facing veth devices
and phys devices build the core datapath for Kubernetes Pods, and they
implement forwarding, load-balancing, policy, EDT-management, etc, within
BPF. Currently there is no concept of 'safe' ownership, e.g. we've recently
experienced hard-to-debug issues in a user's staging environment where
another Kubernetes application using tc BPF attached to the same prio/handle
of cls_bpf, accidentally wiping all Cilium-based BPF programs from underneath
it. The goal is to establish a clear/safe ownership model via links which
cannot accidentally be overridden. [0,2]
BPF links for tc can co-exist with non-link attachments, and the semantics are
in line also with XDP links: BPF links cannot replace other BPF links, BPF
links cannot replace non-BPF links, non-BPF links cannot replace BPF links and
lastly only non-BPF links can replace non-BPF links. In case of Cilium, this
would solve mentioned issue of safe ownership model as 3rd party applications
would not be able to accidentally wipe Cilium programs, even if they are not
BPF link aware.
Earlier attempts [4] have tried to integrate BPF links into core tc machinery
to solve cls_bpf, which has been intrusive to the generic tc kernel API with
extensions only specific to cls_bpf and suboptimal/complex since cls_bpf could
be wiped from the qdisc also. Locking a tc BPF program in place this way, is
getting into layering hacks given the two object models are vastly different.
We instead implemented the tcx (tc 'express') layer which is an fd-based tc BPF
attach API, so that the BPF link implementation blends in naturally similar to
other link types which are fd-based and without the need for changing core tc
internal APIs. BPF programs for tc can then be successively migrated from classic
cls_bpf to the new tc BPF link without needing to change the program's source
code, just the BPF loader mechanics for attaching is sufficient.
For the current tc framework, there is no change in behavior with this change
and neither does this change touch on tc core kernel APIs. The gist of this
patch is that the ingress and egress hook have a lightweight, qdisc-less
extension for BPF to attach its tc BPF programs, in other words, a minimal
entry point for tc BPF. The name tcx has been suggested from discussion of
earlier revisions of this work as a good fit, and to more easily differ between
the classic cls_bpf attachment and the fd-based one.
For the ingress and egress tcx points, the device holds a cache-friendly array
with program pointers which is separated from control plane (slow-path) data.
Earlier versions of this work used priority to determine ordering and expression
of dependencies similar as with classic tc, but it was challenged that for
something more future-proof a better user experience is required. Hence this
resulted in the design and development of the generic attach/detach/query API
for multi-progs. See prior patch with its discussion on the API design. tcx is
the first user and later we plan to integrate also others, for example, one
candidate is multi-prog support for XDP which would benefit and have the same
'look and feel' from API perspective.
The goal with tcx is to have maximum compatibility to existing tc BPF programs,
so they don't need to be rewritten specifically. Compatibility to call into
classic tcf_classify() is also provided in order to allow successive migration
or both to cleanly co-exist where needed given its all one logical tc layer and
the tcx plus classic tc cls/act build one logical overall processing pipeline.
tcx supports the simplified return codes TCX_NEXT which is non-terminating (go
to next program) and terminating ones with TCX_PASS, TCX_DROP, TCX_REDIRECT.
The fd-based API is behind a static key, so that when unused the code is also
not entered. The struct tcx_entry's program array is currently static, but
could be made dynamic if necessary at a point in future. The a/b pair swap
design has been chosen so that for detachment there are no allocations which
otherwise could fail.
The work has been tested with tc-testing selftest suite which all passes, as
well as the tc BPF tests from the BPF CI, and also with Cilium's L4LB.
Thanks also to Nikolay Aleksandrov and Martin Lau for in-depth early reviews
of this work.
[0] https://lpc.events/event/16/contributions/1353/
[1] https://lore.kernel.org/bpf/CAEf4BzbokCJN33Nw_kg82sO=xppXnKWEncGTWCTB9vGCmLB6pw@mail.gmail.com
[2] https://colocatedeventseu2023.sched.com/event/1Jo6O/tales-from-an-ebpf-programs-murder-mystery-hemanth-malla-guillaume-fournier-datadog
[3] http://vger.kernel.org/bpfconf2023_material/tcx_meta_netdev_borkmann.pdf
[4] https://lore.kernel.org/bpf/20210604063116.234316-1-memxor@gmail.com
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230719140858.13224-3-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-07-19 14:08:52 +00:00
|
|
|
break;
|
2023-04-21 17:02:54 +00:00
|
|
|
case BPF_PROG_TYPE_NETFILTER:
|
|
|
|
|
ret = bpf_nf_link_attach(attr, prog);
|
2020-07-22 06:45:57 +00:00
|
|
|
break;
|
bpf: Implement minimal BPF perf link
Introduce a new type of BPF link - BPF perf link. This brings perf_event-based
BPF program attachments (perf_event, tracepoints, kprobes, and uprobes) into
the common BPF link infrastructure, allowing to list all active perf_event
based attachments, auto-detaching BPF program from perf_event when link's FD
is closed, get generic BPF link fdinfo/get_info functionality.
BPF_LINK_CREATE command expects perf_event's FD as target_fd. No extra flags
are currently supported.
Force-detaching and atomic BPF program updates are not yet implemented, but
with perf_event-based BPF links we now have common framework for this without
the need to extend ioctl()-based perf_event interface.
One interesting consideration is a new value for bpf_attach_type, which
BPF_LINK_CREATE command expects. Generally, it's either 1-to-1 mapping from
bpf_attach_type to bpf_prog_type, or many-to-1 mapping from a subset of
bpf_attach_types to one bpf_prog_type (e.g., see BPF_PROG_TYPE_SK_SKB or
BPF_PROG_TYPE_CGROUP_SOCK). In this case, though, we have three different
program types (KPROBE, TRACEPOINT, PERF_EVENT) using the same perf_event-based
mechanism, so it's many bpf_prog_types to one bpf_attach_type. I chose to
define a single BPF_PERF_EVENT attach type for all of them and adjust
link_create()'s logic for checking correspondence between attach type and
program type.
The alternative would be to define three new attach types (e.g., BPF_KPROBE,
BPF_TRACEPOINT, and BPF_PERF_EVENT), but that seemed like unnecessary overkill
and BPF_KPROBE will cause naming conflicts with BPF_KPROBE() macro, defined by
libbpf. I chose to not do this to avoid unnecessary proliferation of
bpf_attach_type enum values and not have to deal with naming conflicts.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20210815070609.987780-5-andrii@kernel.org
2021-08-15 07:05:57 +00:00
|
|
|
#endif
|
|
|
|
|
case BPF_PROG_TYPE_PERF_EVENT:
|
|
|
|
|
case BPF_PROG_TYPE_TRACEPOINT:
|
|
|
|
|
ret = bpf_perf_link_attach(attr, prog);
|
|
|
|
|
break;
|
2022-03-16 12:24:09 +00:00
|
|
|
case BPF_PROG_TYPE_KPROBE:
|
|
|
|
|
if (attr->link_create.attach_type == BPF_PERF_EVENT)
|
|
|
|
|
ret = bpf_perf_link_attach(attr, prog);
|
2023-08-09 08:34:15 +00:00
|
|
|
else if (attr->link_create.attach_type == BPF_TRACE_KPROBE_MULTI)
|
2022-03-16 12:24:09 +00:00
|
|
|
ret = bpf_kprobe_multi_link_attach(attr, prog);
|
2023-08-09 08:34:15 +00:00
|
|
|
else if (attr->link_create.attach_type == BPF_TRACE_UPROBE_MULTI)
|
|
|
|
|
ret = bpf_uprobe_multi_link_attach(attr, prog);
|
2022-03-16 12:24:09 +00:00
|
|
|
break;
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
default:
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
2020-09-29 12:45:51 +00:00
|
|
|
out:
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
if (ret < 0)
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2023-03-23 03:24:02 +00:00
|
|
|
static int link_update_map(struct bpf_link *link, union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_map *new_map, *old_map = NULL;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
new_map = bpf_map_get(attr->link_update.new_map_fd);
|
|
|
|
|
if (IS_ERR(new_map))
|
2023-03-24 18:42:41 +00:00
|
|
|
return PTR_ERR(new_map);
|
2023-03-23 03:24:02 +00:00
|
|
|
|
|
|
|
|
if (attr->link_update.flags & BPF_F_REPLACE) {
|
|
|
|
|
old_map = bpf_map_get(attr->link_update.old_map_fd);
|
|
|
|
|
if (IS_ERR(old_map)) {
|
2023-03-24 18:42:41 +00:00
|
|
|
ret = PTR_ERR(old_map);
|
2023-03-23 03:24:02 +00:00
|
|
|
goto out_put;
|
|
|
|
|
}
|
|
|
|
|
} else if (attr->link_update.old_map_fd) {
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
goto out_put;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ret = link->ops->update_map(link, new_map, old_map);
|
|
|
|
|
|
|
|
|
|
if (old_map)
|
|
|
|
|
bpf_map_put(old_map);
|
|
|
|
|
out_put:
|
|
|
|
|
bpf_map_put(new_map);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
#define BPF_LINK_UPDATE_LAST_FIELD link_update.old_prog_fd
|
|
|
|
|
|
|
|
|
|
static int link_update(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *old_prog = NULL, *new_prog;
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
u32 flags;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_LINK_UPDATE))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
flags = attr->link_update.flags;
|
|
|
|
|
if (flags & ~BPF_F_REPLACE)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
link = bpf_link_get_from_fd(attr->link_update.link_fd);
|
|
|
|
|
if (IS_ERR(link))
|
|
|
|
|
return PTR_ERR(link);
|
|
|
|
|
|
2023-03-23 03:24:02 +00:00
|
|
|
if (link->ops->update_map) {
|
|
|
|
|
ret = link_update_map(link, attr);
|
|
|
|
|
goto out_put_link;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
new_prog = bpf_prog_get(attr->link_update.new_prog_fd);
|
2020-04-24 05:20:44 +00:00
|
|
|
if (IS_ERR(new_prog)) {
|
|
|
|
|
ret = PTR_ERR(new_prog);
|
|
|
|
|
goto out_put_link;
|
|
|
|
|
}
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
|
|
|
|
|
if (flags & BPF_F_REPLACE) {
|
|
|
|
|
old_prog = bpf_prog_get(attr->link_update.old_prog_fd);
|
|
|
|
|
if (IS_ERR(old_prog)) {
|
|
|
|
|
ret = PTR_ERR(old_prog);
|
|
|
|
|
old_prog = NULL;
|
|
|
|
|
goto out_put_progs;
|
|
|
|
|
}
|
2020-04-24 05:20:44 +00:00
|
|
|
} else if (attr->link_update.old_prog_fd) {
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
goto out_put_progs;
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
}
|
|
|
|
|
|
2020-04-29 00:16:05 +00:00
|
|
|
if (link->ops->update_prog)
|
|
|
|
|
ret = link->ops->update_prog(link, new_prog, old_prog);
|
|
|
|
|
else
|
2020-05-25 12:29:28 +00:00
|
|
|
ret = -EINVAL;
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
|
|
|
|
|
out_put_progs:
|
|
|
|
|
if (old_prog)
|
|
|
|
|
bpf_prog_put(old_prog);
|
|
|
|
|
if (ret)
|
|
|
|
|
bpf_prog_put(new_prog);
|
2020-04-24 05:20:44 +00:00
|
|
|
out_put_link:
|
2023-06-14 08:34:30 +00:00
|
|
|
bpf_link_put_direct(link);
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-31 18:28:26 +00:00
|
|
|
#define BPF_LINK_DETACH_LAST_FIELD link_detach.link_fd
|
|
|
|
|
|
|
|
|
|
static int link_detach(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_LINK_DETACH))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
link = bpf_link_get_from_fd(attr->link_detach.link_fd);
|
|
|
|
|
if (IS_ERR(link))
|
|
|
|
|
return PTR_ERR(link);
|
|
|
|
|
|
|
|
|
|
if (link->ops->detach)
|
|
|
|
|
ret = link->ops->detach(link);
|
|
|
|
|
else
|
|
|
|
|
ret = -EOPNOTSUPP;
|
|
|
|
|
|
2023-06-14 08:34:30 +00:00
|
|
|
bpf_link_put_direct(link);
|
2020-07-31 18:28:26 +00:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2020-08-19 04:27:56 +00:00
|
|
|
static struct bpf_link *bpf_link_inc_not_zero(struct bpf_link *link)
|
2020-04-29 00:16:07 +00:00
|
|
|
{
|
2020-08-19 04:27:56 +00:00
|
|
|
return atomic64_fetch_add_unless(&link->refcnt, 1, 0) ? link : ERR_PTR(-ENOENT);
|
2020-04-29 00:16:07 +00:00
|
|
|
}
|
|
|
|
|
|
2020-08-19 04:27:56 +00:00
|
|
|
struct bpf_link *bpf_link_by_id(u32 id)
|
2020-04-29 00:16:07 +00:00
|
|
|
{
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
|
2020-08-19 04:27:56 +00:00
|
|
|
if (!id)
|
|
|
|
|
return ERR_PTR(-ENOENT);
|
2020-04-29 00:16:07 +00:00
|
|
|
|
|
|
|
|
spin_lock_bh(&link_idr_lock);
|
|
|
|
|
/* before link is "settled", ID is 0, pretend it doesn't exist yet */
|
2020-08-19 04:27:56 +00:00
|
|
|
link = idr_find(&link_idr, id);
|
2020-04-29 00:16:07 +00:00
|
|
|
if (link) {
|
|
|
|
|
if (link->id)
|
2020-08-19 04:27:56 +00:00
|
|
|
link = bpf_link_inc_not_zero(link);
|
2020-04-29 00:16:07 +00:00
|
|
|
else
|
2020-08-19 04:27:56 +00:00
|
|
|
link = ERR_PTR(-EAGAIN);
|
2020-04-29 00:16:07 +00:00
|
|
|
} else {
|
2020-08-19 04:27:56 +00:00
|
|
|
link = ERR_PTR(-ENOENT);
|
2020-04-29 00:16:07 +00:00
|
|
|
}
|
|
|
|
|
spin_unlock_bh(&link_idr_lock);
|
2020-08-19 04:27:56 +00:00
|
|
|
return link;
|
|
|
|
|
}
|
2020-04-29 00:16:07 +00:00
|
|
|
|
2022-05-10 15:52:30 +00:00
|
|
|
struct bpf_link *bpf_link_get_curr_or_next(u32 *id)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
|
|
|
|
|
spin_lock_bh(&link_idr_lock);
|
|
|
|
|
again:
|
|
|
|
|
link = idr_get_next(&link_idr, id);
|
|
|
|
|
if (link) {
|
|
|
|
|
link = bpf_link_inc_not_zero(link);
|
|
|
|
|
if (IS_ERR(link)) {
|
|
|
|
|
(*id)++;
|
|
|
|
|
goto again;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_bh(&link_idr_lock);
|
|
|
|
|
|
|
|
|
|
return link;
|
|
|
|
|
}
|
|
|
|
|
|
2020-08-19 04:27:56 +00:00
|
|
|
#define BPF_LINK_GET_FD_BY_ID_LAST_FIELD link_id
|
|
|
|
|
|
|
|
|
|
static int bpf_link_get_fd_by_id(const union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
u32 id = attr->link_id;
|
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_LINK_GET_FD_BY_ID))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
link = bpf_link_by_id(id);
|
|
|
|
|
if (IS_ERR(link))
|
|
|
|
|
return PTR_ERR(link);
|
2020-04-29 00:16:07 +00:00
|
|
|
|
|
|
|
|
fd = bpf_link_new_fd(link);
|
|
|
|
|
if (fd < 0)
|
2023-06-14 08:34:30 +00:00
|
|
|
bpf_link_put_direct(link);
|
2020-04-29 00:16:07 +00:00
|
|
|
|
|
|
|
|
return fd;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Sharing bpf runtime stats with BPF_ENABLE_STATS
Currently, sysctl kernel.bpf_stats_enabled controls BPF runtime stats.
Typical userspace tools use kernel.bpf_stats_enabled as follows:
1. Enable kernel.bpf_stats_enabled;
2. Check program run_time_ns;
3. Sleep for the monitoring period;
4. Check program run_time_ns again, calculate the difference;
5. Disable kernel.bpf_stats_enabled.
The problem with this approach is that only one userspace tool can toggle
this sysctl. If multiple tools toggle the sysctl at the same time, the
measurement may be inaccurate.
To fix this problem while keep backward compatibility, introduce a new
bpf command BPF_ENABLE_STATS. On success, this command enables stats and
returns a valid fd. BPF_ENABLE_STATS takes argument "type". Currently,
only one type, BPF_STATS_RUN_TIME, is supported. We can extend the
command to support other types of stats in the future.
With BPF_ENABLE_STATS, user space tool would have the following flow:
1. Get a fd with BPF_ENABLE_STATS, and make sure it is valid;
2. Check program run_time_ns;
3. Sleep for the monitoring period;
4. Check program run_time_ns again, calculate the difference;
5. Close the fd.
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200430071506.1408910-2-songliubraving@fb.com
2020-04-30 07:15:04 +00:00
|
|
|
DEFINE_MUTEX(bpf_stats_enabled_mutex);
|
|
|
|
|
|
|
|
|
|
static int bpf_stats_release(struct inode *inode, struct file *file)
|
|
|
|
|
{
|
|
|
|
|
mutex_lock(&bpf_stats_enabled_mutex);
|
|
|
|
|
static_key_slow_dec(&bpf_stats_enabled_key.key);
|
|
|
|
|
mutex_unlock(&bpf_stats_enabled_mutex);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static const struct file_operations bpf_stats_fops = {
|
|
|
|
|
.release = bpf_stats_release,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int bpf_enable_runtime_stats(void)
|
|
|
|
|
{
|
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
|
|
mutex_lock(&bpf_stats_enabled_mutex);
|
|
|
|
|
|
|
|
|
|
/* Set a very high limit to avoid overflow */
|
|
|
|
|
if (static_key_count(&bpf_stats_enabled_key.key) > INT_MAX / 2) {
|
|
|
|
|
mutex_unlock(&bpf_stats_enabled_mutex);
|
|
|
|
|
return -EBUSY;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
fd = anon_inode_getfd("bpf-stats", &bpf_stats_fops, NULL, O_CLOEXEC);
|
|
|
|
|
if (fd >= 0)
|
|
|
|
|
static_key_slow_inc(&bpf_stats_enabled_key.key);
|
|
|
|
|
|
|
|
|
|
mutex_unlock(&bpf_stats_enabled_mutex);
|
|
|
|
|
return fd;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define BPF_ENABLE_STATS_LAST_FIELD enable_stats.type
|
|
|
|
|
|
|
|
|
|
static int bpf_enable_stats(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_ENABLE_STATS))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
switch (attr->enable_stats.type) {
|
|
|
|
|
case BPF_STATS_RUN_TIME:
|
|
|
|
|
return bpf_enable_runtime_stats();
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-09 17:59:05 +00:00
|
|
|
#define BPF_ITER_CREATE_LAST_FIELD iter_create.flags
|
|
|
|
|
|
|
|
|
|
static int bpf_iter_create(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link *link;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_ITER_CREATE))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (attr->iter_create.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
link = bpf_link_get_from_fd(attr->iter_create.link_fd);
|
|
|
|
|
if (IS_ERR(link))
|
|
|
|
|
return PTR_ERR(link);
|
|
|
|
|
|
|
|
|
|
err = bpf_iter_new_fd(link);
|
2023-06-14 08:34:30 +00:00
|
|
|
bpf_link_put_direct(link);
|
2020-05-09 17:59:05 +00:00
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-09-15 23:45:40 +00:00
|
|
|
#define BPF_PROG_BIND_MAP_LAST_FIELD prog_bind_map.flags
|
|
|
|
|
|
|
|
|
|
static int bpf_prog_bind_map(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog *prog;
|
|
|
|
|
struct bpf_map *map;
|
|
|
|
|
struct bpf_map **used_maps_old, **used_maps_new;
|
|
|
|
|
int i, ret = 0;
|
|
|
|
|
|
|
|
|
|
if (CHECK_ATTR(BPF_PROG_BIND_MAP))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (attr->prog_bind_map.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
prog = bpf_prog_get(attr->prog_bind_map.prog_fd);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
map = bpf_map_get(attr->prog_bind_map.map_fd);
|
|
|
|
|
if (IS_ERR(map)) {
|
|
|
|
|
ret = PTR_ERR(map);
|
|
|
|
|
goto out_prog_put;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
mutex_lock(&prog->aux->used_maps_mutex);
|
|
|
|
|
|
|
|
|
|
used_maps_old = prog->aux->used_maps;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < prog->aux->used_map_cnt; i++)
|
2020-10-03 00:25:44 +00:00
|
|
|
if (used_maps_old[i] == map) {
|
|
|
|
|
bpf_map_put(map);
|
2020-09-15 23:45:40 +00:00
|
|
|
goto out_unlock;
|
2020-10-03 00:25:44 +00:00
|
|
|
}
|
2020-09-15 23:45:40 +00:00
|
|
|
|
|
|
|
|
used_maps_new = kmalloc_array(prog->aux->used_map_cnt + 1,
|
|
|
|
|
sizeof(used_maps_new[0]),
|
|
|
|
|
GFP_KERNEL);
|
|
|
|
|
if (!used_maps_new) {
|
|
|
|
|
ret = -ENOMEM;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
2023-12-04 14:04:23 +00:00
|
|
|
/* The bpf program will not access the bpf map, but for the sake of
|
|
|
|
|
* simplicity, increase sleepable_refcnt for sleepable program as well.
|
|
|
|
|
*/
|
2024-03-09 00:47:39 +00:00
|
|
|
if (prog->sleepable)
|
2023-12-04 14:04:23 +00:00
|
|
|
atomic64_inc(&map->sleepable_refcnt);
|
2020-09-15 23:45:40 +00:00
|
|
|
memcpy(used_maps_new, used_maps_old,
|
|
|
|
|
sizeof(used_maps_old[0]) * prog->aux->used_map_cnt);
|
|
|
|
|
used_maps_new[prog->aux->used_map_cnt] = map;
|
|
|
|
|
|
|
|
|
|
prog->aux->used_map_cnt++;
|
|
|
|
|
prog->aux->used_maps = used_maps_new;
|
|
|
|
|
|
|
|
|
|
kfree(used_maps_old);
|
|
|
|
|
|
|
|
|
|
out_unlock:
|
|
|
|
|
mutex_unlock(&prog->aux->used_maps_mutex);
|
|
|
|
|
|
|
|
|
|
if (ret)
|
|
|
|
|
bpf_map_put(map);
|
|
|
|
|
out_prog_put:
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Introduce BPF token object
Add new kind of BPF kernel object, BPF token. BPF token is meant to
allow delegating privileged BPF functionality, like loading a BPF
program or creating a BPF map, from privileged process to a *trusted*
unprivileged process, all while having a good amount of control over which
privileged operations could be performed using provided BPF token.
This is achieved through mounting BPF FS instance with extra delegation
mount options, which determine what operations are delegatable, and also
constraining it to the owning user namespace (as mentioned in the
previous patch).
BPF token itself is just a derivative from BPF FS and can be created
through a new bpf() syscall command, BPF_TOKEN_CREATE, which accepts BPF
FS FD, which can be attained through open() API by opening BPF FS mount
point. Currently, BPF token "inherits" delegated command, map types,
prog type, and attach type bit sets from BPF FS as is. In the future,
having an BPF token as a separate object with its own FD, we can allow
to further restrict BPF token's allowable set of things either at the
creation time or after the fact, allowing the process to guard itself
further from unintentionally trying to load undesired kind of BPF
programs. But for now we keep things simple and just copy bit sets as is.
When BPF token is created from BPF FS mount, we take reference to the
BPF super block's owning user namespace, and then use that namespace for
checking all the {CAP_BPF, CAP_PERFMON, CAP_NET_ADMIN, CAP_SYS_ADMIN}
capabilities that are normally only checked against init userns (using
capable()), but now we check them using ns_capable() instead (if BPF
token is provided). See bpf_token_capable() for details.
Such setup means that BPF token in itself is not sufficient to grant BPF
functionality. User namespaced process has to *also* have necessary
combination of capabilities inside that user namespace. So while
previously CAP_BPF was useless when granted within user namespace, now
it gains a meaning and allows container managers and sys admins to have
a flexible control over which processes can and need to use BPF
functionality within the user namespace (i.e., container in practice).
And BPF FS delegation mount options and derived BPF tokens serve as
a per-container "flag" to grant overall ability to use bpf() (plus further
restrict on which parts of bpf() syscalls are treated as namespaced).
Note also, BPF_TOKEN_CREATE command itself requires ns_capable(CAP_BPF)
within the BPF FS owning user namespace, rounding up the ns_capable()
story of BPF token. Also creating BPF token in init user namespace is
currently not supported, given BPF token doesn't have any effect in init
user namespace anyways.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-4-andrii@kernel.org
2024-01-24 02:21:00 +00:00
|
|
|
#define BPF_TOKEN_CREATE_LAST_FIELD token_create.bpffs_fd
|
|
|
|
|
|
|
|
|
|
static int token_create(union bpf_attr *attr)
|
|
|
|
|
{
|
|
|
|
|
if (CHECK_ATTR(BPF_TOKEN_CREATE))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
/* no flags are supported yet */
|
|
|
|
|
if (attr->token_create.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
return bpf_token_create(attr);
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
static int __sys_bpf(int cmd, bpfptr_t uattr, unsigned int size)
|
2014-09-26 07:16:57 +00:00
|
|
|
{
|
2020-03-20 09:48:13 +00:00
|
|
|
union bpf_attr attr;
|
2014-09-26 07:16:57 +00:00
|
|
|
int err;
|
|
|
|
|
|
2018-05-22 22:03:31 +00:00
|
|
|
err = bpf_check_uarg_tail_zero(uattr, sizeof(attr), size);
|
2017-06-05 19:15:52 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
size = min_t(u32, size, sizeof(attr));
|
2014-09-26 07:16:57 +00:00
|
|
|
|
|
|
|
|
/* copy attributes from user space, may be less than sizeof(bpf_attr) */
|
2020-03-20 09:48:13 +00:00
|
|
|
memset(&attr, 0, sizeof(attr));
|
2021-05-14 00:36:05 +00:00
|
|
|
if (copy_from_bpfptr(&attr, uattr, size) != 0)
|
2014-09-26 07:16:57 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
|
2017-10-18 20:00:24 +00:00
|
|
|
err = security_bpf(cmd, &attr, size);
|
|
|
|
|
if (err < 0)
|
|
|
|
|
return err;
|
|
|
|
|
|
2014-09-26 07:16:57 +00:00
|
|
|
switch (cmd) {
|
|
|
|
|
case BPF_MAP_CREATE:
|
|
|
|
|
err = map_create(&attr);
|
|
|
|
|
break;
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
case BPF_MAP_LOOKUP_ELEM:
|
|
|
|
|
err = map_lookup_elem(&attr);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_MAP_UPDATE_ELEM:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = map_update_elem(&attr, uattr);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_MAP_DELETE_ELEM:
|
2022-08-24 13:40:36 +00:00
|
|
|
err = map_delete_elem(&attr, uattr);
|
bpf: add lookup/update/delete/iterate methods to BPF maps
'maps' is a generic storage of different types for sharing data between kernel
and userspace.
The maps are accessed from user space via BPF syscall, which has commands:
- create a map with given type and attributes
fd = bpf(BPF_MAP_CREATE, union bpf_attr *attr, u32 size)
returns fd or negative error
- lookup key in a given map referenced by fd
err = bpf(BPF_MAP_LOOKUP_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero and stores found elem into value or negative error
- create or update key/value pair in a given map
err = bpf(BPF_MAP_UPDATE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->value
returns zero or negative error
- find and delete element by key in a given map
err = bpf(BPF_MAP_DELETE_ELEM, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key
- iterate map elements (based on input key return next_key)
err = bpf(BPF_MAP_GET_NEXT_KEY, union bpf_attr *attr, u32 size)
using attr->map_fd, attr->key, attr->next_key
- close(fd) deletes the map
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-09-26 07:16:59 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_MAP_GET_NEXT_KEY:
|
|
|
|
|
err = map_get_next_key(&attr);
|
|
|
|
|
break;
|
bpf: add syscall side map freeze support
This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.
Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.
Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.
A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-04-09 21:20:06 +00:00
|
|
|
case BPF_MAP_FREEZE:
|
|
|
|
|
err = map_freeze(&attr);
|
|
|
|
|
break;
|
2014-09-26 07:17:00 +00:00
|
|
|
case BPF_PROG_LOAD:
|
2023-04-06 23:41:58 +00:00
|
|
|
err = bpf_prog_load(&attr, uattr, size);
|
2014-09-26 07:17:00 +00:00
|
|
|
break;
|
2015-10-29 13:58:09 +00:00
|
|
|
case BPF_OBJ_PIN:
|
|
|
|
|
err = bpf_obj_pin(&attr);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_OBJ_GET:
|
|
|
|
|
err = bpf_obj_get(&attr);
|
|
|
|
|
break;
|
bpf: add BPF_PROG_ATTACH and BPF_PROG_DETACH commands
Extend the bpf(2) syscall by two new commands, BPF_PROG_ATTACH and
BPF_PROG_DETACH which allow attaching and detaching eBPF programs
to a target.
On the API level, the target could be anything that has an fd in
userspace, hence the name of the field in union bpf_attr is called
'target_fd'.
When called with BPF_ATTACH_TYPE_CGROUP_INET_{E,IN}GRESS, the target is
expected to be a valid file descriptor of a cgroup v2 directory which
has the bpf controller enabled. These are the only use-cases
implemented by this patch at this point, but more can be added.
If a program of the given type already exists in the given cgroup,
the program is swapped automically, so userspace does not have to drop
an existing program first before installing a new one, which would
otherwise leave a gap in which no program is attached.
For more information on the propagation logic to subcgroups, please
refer to the bpf cgroup controller implementation.
The API is guarded by CAP_NET_ADMIN.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-23 15:52:27 +00:00
|
|
|
case BPF_PROG_ATTACH:
|
|
|
|
|
err = bpf_prog_attach(&attr);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_PROG_DETACH:
|
|
|
|
|
err = bpf_prog_detach(&attr);
|
|
|
|
|
break;
|
2017-10-03 05:50:22 +00:00
|
|
|
case BPF_PROG_QUERY:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_prog_query(&attr, uattr.user);
|
2017-10-03 05:50:22 +00:00
|
|
|
break;
|
2017-03-31 04:45:38 +00:00
|
|
|
case BPF_PROG_TEST_RUN:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_prog_test_run(&attr, uattr.user);
|
2017-03-31 04:45:38 +00:00
|
|
|
break;
|
2017-06-05 19:15:48 +00:00
|
|
|
case BPF_PROG_GET_NEXT_ID:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_obj_get_next_id(&attr, uattr.user,
|
2017-06-05 19:15:48 +00:00
|
|
|
&prog_idr, &prog_idr_lock);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_MAP_GET_NEXT_ID:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_obj_get_next_id(&attr, uattr.user,
|
2017-06-05 19:15:48 +00:00
|
|
|
&map_idr, &map_idr_lock);
|
|
|
|
|
break;
|
2019-08-20 09:31:50 +00:00
|
|
|
case BPF_BTF_GET_NEXT_ID:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_obj_get_next_id(&attr, uattr.user,
|
2019-08-20 09:31:50 +00:00
|
|
|
&btf_idr, &btf_idr_lock);
|
|
|
|
|
break;
|
2017-06-05 19:15:49 +00:00
|
|
|
case BPF_PROG_GET_FD_BY_ID:
|
|
|
|
|
err = bpf_prog_get_fd_by_id(&attr);
|
|
|
|
|
break;
|
2017-06-05 19:15:50 +00:00
|
|
|
case BPF_MAP_GET_FD_BY_ID:
|
|
|
|
|
err = bpf_map_get_fd_by_id(&attr);
|
|
|
|
|
break;
|
2017-06-05 19:15:52 +00:00
|
|
|
case BPF_OBJ_GET_INFO_BY_FD:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_obj_get_info_by_fd(&attr, uattr.user);
|
2017-06-05 19:15:52 +00:00
|
|
|
break;
|
2018-03-28 19:05:37 +00:00
|
|
|
case BPF_RAW_TRACEPOINT_OPEN:
|
|
|
|
|
err = bpf_raw_tracepoint_open(&attr);
|
|
|
|
|
break;
|
2018-04-18 22:56:01 +00:00
|
|
|
case BPF_BTF_LOAD:
|
2023-04-06 23:41:58 +00:00
|
|
|
err = bpf_btf_load(&attr, uattr, size);
|
2018-04-18 22:56:01 +00:00
|
|
|
break;
|
2018-05-04 21:49:51 +00:00
|
|
|
case BPF_BTF_GET_FD_BY_ID:
|
|
|
|
|
err = bpf_btf_get_fd_by_id(&attr);
|
|
|
|
|
break;
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
case BPF_TASK_FD_QUERY:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_task_fd_query(&attr, uattr.user);
|
bpf: introduce bpf subcommand BPF_TASK_FD_QUERY
Currently, suppose a userspace application has loaded a bpf program
and attached it to a tracepoint/kprobe/uprobe, and a bpf
introspection tool, e.g., bpftool, wants to show which bpf program
is attached to which tracepoint/kprobe/uprobe. Such attachment
information will be really useful to understand the overall bpf
deployment in the system.
There is a name field (16 bytes) for each program, which could
be used to encode the attachment point. There are some drawbacks
for this approaches. First, bpftool user (e.g., an admin) may not
really understand the association between the name and the
attachment point. Second, if one program is attached to multiple
places, encoding a proper name which can imply all these
attachments becomes difficult.
This patch introduces a new bpf subcommand BPF_TASK_FD_QUERY.
Given a pid and fd, if the <pid, fd> is associated with a
tracepoint/kprobe/uprobe perf event, BPF_TASK_FD_QUERY will return
. prog_id
. tracepoint name, or
. k[ret]probe funcname + offset or kernel addr, or
. u[ret]probe filename + offset
to the userspace.
The user can use "bpftool prog" to find more information about
bpf program itself with prog_id.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-05-24 18:21:09 +00:00
|
|
|
break;
|
2018-10-18 13:16:30 +00:00
|
|
|
case BPF_MAP_LOOKUP_AND_DELETE_ELEM:
|
|
|
|
|
err = map_lookup_and_delete_elem(&attr);
|
|
|
|
|
break;
|
2020-01-15 18:43:01 +00:00
|
|
|
case BPF_MAP_LOOKUP_BATCH:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_map_do_batch(&attr, uattr.user, BPF_MAP_LOOKUP_BATCH);
|
2020-01-15 18:43:01 +00:00
|
|
|
break;
|
2020-01-15 18:43:04 +00:00
|
|
|
case BPF_MAP_LOOKUP_AND_DELETE_BATCH:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_map_do_batch(&attr, uattr.user,
|
2020-01-15 18:43:04 +00:00
|
|
|
BPF_MAP_LOOKUP_AND_DELETE_BATCH);
|
|
|
|
|
break;
|
2020-01-15 18:43:02 +00:00
|
|
|
case BPF_MAP_UPDATE_BATCH:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_map_do_batch(&attr, uattr.user, BPF_MAP_UPDATE_BATCH);
|
2020-01-15 18:43:02 +00:00
|
|
|
break;
|
|
|
|
|
case BPF_MAP_DELETE_BATCH:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_map_do_batch(&attr, uattr.user, BPF_MAP_DELETE_BATCH);
|
2020-01-15 18:43:02 +00:00
|
|
|
break;
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
case BPF_LINK_CREATE:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = link_create(&attr, uattr);
|
bpf: Implement bpf_link-based cgroup BPF program attachment
Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.
To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.
The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Roman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
2020-03-30 02:59:58 +00:00
|
|
|
break;
|
bpf: Implement bpf_prog replacement for an active bpf_cgroup_link
Add new operation (LINK_UPDATE), which allows to replace active bpf_prog from
under given bpf_link. Currently this is only supported for bpf_cgroup_link,
but will be extended to other kinds of bpf_links in follow-up patches.
For bpf_cgroup_link, implemented functionality matches existing semantics for
direct bpf_prog attachment (including BPF_F_REPLACE flag). User can either
unconditionally set new bpf_prog regardless of which bpf_prog is currently
active under given bpf_link, or, optionally, can specify expected active
bpf_prog. If active bpf_prog doesn't match expected one, no changes are
performed, old bpf_link stays intact and attached, operation returns
a failure.
cgroup_bpf_replace() operation is resolving race between auto-detachment and
bpf_prog update in the same fashion as it's done for bpf_link detachment,
except in this case update has no way of succeeding because of target cgroup
marked as dying. So in this case error is returned.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-3-andriin@fb.com
2020-03-30 02:59:59 +00:00
|
|
|
case BPF_LINK_UPDATE:
|
|
|
|
|
err = link_update(&attr);
|
|
|
|
|
break;
|
2020-04-29 00:16:07 +00:00
|
|
|
case BPF_LINK_GET_FD_BY_ID:
|
|
|
|
|
err = bpf_link_get_fd_by_id(&attr);
|
|
|
|
|
break;
|
|
|
|
|
case BPF_LINK_GET_NEXT_ID:
|
2021-05-14 00:36:05 +00:00
|
|
|
err = bpf_obj_get_next_id(&attr, uattr.user,
|
2020-04-29 00:16:07 +00:00
|
|
|
&link_idr, &link_idr_lock);
|
|
|
|
|
break;
|
bpf: Sharing bpf runtime stats with BPF_ENABLE_STATS
Currently, sysctl kernel.bpf_stats_enabled controls BPF runtime stats.
Typical userspace tools use kernel.bpf_stats_enabled as follows:
1. Enable kernel.bpf_stats_enabled;
2. Check program run_time_ns;
3. Sleep for the monitoring period;
4. Check program run_time_ns again, calculate the difference;
5. Disable kernel.bpf_stats_enabled.
The problem with this approach is that only one userspace tool can toggle
this sysctl. If multiple tools toggle the sysctl at the same time, the
measurement may be inaccurate.
To fix this problem while keep backward compatibility, introduce a new
bpf command BPF_ENABLE_STATS. On success, this command enables stats and
returns a valid fd. BPF_ENABLE_STATS takes argument "type". Currently,
only one type, BPF_STATS_RUN_TIME, is supported. We can extend the
command to support other types of stats in the future.
With BPF_ENABLE_STATS, user space tool would have the following flow:
1. Get a fd with BPF_ENABLE_STATS, and make sure it is valid;
2. Check program run_time_ns;
3. Sleep for the monitoring period;
4. Check program run_time_ns again, calculate the difference;
5. Close the fd.
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200430071506.1408910-2-songliubraving@fb.com
2020-04-30 07:15:04 +00:00
|
|
|
case BPF_ENABLE_STATS:
|
|
|
|
|
err = bpf_enable_stats(&attr);
|
|
|
|
|
break;
|
2020-05-09 17:59:05 +00:00
|
|
|
case BPF_ITER_CREATE:
|
|
|
|
|
err = bpf_iter_create(&attr);
|
|
|
|
|
break;
|
2020-07-31 18:28:26 +00:00
|
|
|
case BPF_LINK_DETACH:
|
|
|
|
|
err = link_detach(&attr);
|
|
|
|
|
break;
|
2020-09-15 23:45:40 +00:00
|
|
|
case BPF_PROG_BIND_MAP:
|
|
|
|
|
err = bpf_prog_bind_map(&attr);
|
|
|
|
|
break;
|
bpf: Introduce BPF token object
Add new kind of BPF kernel object, BPF token. BPF token is meant to
allow delegating privileged BPF functionality, like loading a BPF
program or creating a BPF map, from privileged process to a *trusted*
unprivileged process, all while having a good amount of control over which
privileged operations could be performed using provided BPF token.
This is achieved through mounting BPF FS instance with extra delegation
mount options, which determine what operations are delegatable, and also
constraining it to the owning user namespace (as mentioned in the
previous patch).
BPF token itself is just a derivative from BPF FS and can be created
through a new bpf() syscall command, BPF_TOKEN_CREATE, which accepts BPF
FS FD, which can be attained through open() API by opening BPF FS mount
point. Currently, BPF token "inherits" delegated command, map types,
prog type, and attach type bit sets from BPF FS as is. In the future,
having an BPF token as a separate object with its own FD, we can allow
to further restrict BPF token's allowable set of things either at the
creation time or after the fact, allowing the process to guard itself
further from unintentionally trying to load undesired kind of BPF
programs. But for now we keep things simple and just copy bit sets as is.
When BPF token is created from BPF FS mount, we take reference to the
BPF super block's owning user namespace, and then use that namespace for
checking all the {CAP_BPF, CAP_PERFMON, CAP_NET_ADMIN, CAP_SYS_ADMIN}
capabilities that are normally only checked against init userns (using
capable()), but now we check them using ns_capable() instead (if BPF
token is provided). See bpf_token_capable() for details.
Such setup means that BPF token in itself is not sufficient to grant BPF
functionality. User namespaced process has to *also* have necessary
combination of capabilities inside that user namespace. So while
previously CAP_BPF was useless when granted within user namespace, now
it gains a meaning and allows container managers and sys admins to have
a flexible control over which processes can and need to use BPF
functionality within the user namespace (i.e., container in practice).
And BPF FS delegation mount options and derived BPF tokens serve as
a per-container "flag" to grant overall ability to use bpf() (plus further
restrict on which parts of bpf() syscalls are treated as namespaced).
Note also, BPF_TOKEN_CREATE command itself requires ns_capable(CAP_BPF)
within the BPF FS owning user namespace, rounding up the ns_capable()
story of BPF token. Also creating BPF token in init user namespace is
currently not supported, given BPF token doesn't have any effect in init
user namespace anyways.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Christian Brauner <brauner@kernel.org>
Link: https://lore.kernel.org/bpf/20240124022127.2379740-4-andrii@kernel.org
2024-01-24 02:21:00 +00:00
|
|
|
case BPF_TOKEN_CREATE:
|
|
|
|
|
err = token_create(&attr);
|
|
|
|
|
break;
|
2014-09-26 07:16:57 +00:00
|
|
|
default:
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
2021-05-14 00:36:03 +00:00
|
|
|
|
2021-05-14 00:36:05 +00:00
|
|
|
SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size)
|
|
|
|
|
{
|
|
|
|
|
return __sys_bpf(cmd, USER_BPFPTR(uattr), size);
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-14 00:36:03 +00:00
|
|
|
static bool syscall_prog_is_valid_access(int off, int size,
|
|
|
|
|
enum bpf_access_type type,
|
|
|
|
|
const struct bpf_prog *prog,
|
|
|
|
|
struct bpf_insn_access_aux *info)
|
|
|
|
|
{
|
|
|
|
|
if (off < 0 || off >= U16_MAX)
|
|
|
|
|
return false;
|
|
|
|
|
if (off % size != 0)
|
|
|
|
|
return false;
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2022-02-09 23:19:57 +00:00
|
|
|
BPF_CALL_3(bpf_sys_bpf, int, cmd, union bpf_attr *, attr, u32, attr_size)
|
2021-05-14 00:36:03 +00:00
|
|
|
{
|
2021-05-14 00:36:05 +00:00
|
|
|
switch (cmd) {
|
|
|
|
|
case BPF_MAP_CREATE:
|
2022-08-24 13:40:36 +00:00
|
|
|
case BPF_MAP_DELETE_ELEM:
|
2021-05-14 00:36:05 +00:00
|
|
|
case BPF_MAP_UPDATE_ELEM:
|
|
|
|
|
case BPF_MAP_FREEZE:
|
2022-08-24 13:40:36 +00:00
|
|
|
case BPF_MAP_GET_FD_BY_ID:
|
2021-05-14 00:36:05 +00:00
|
|
|
case BPF_PROG_LOAD:
|
2021-05-14 00:36:08 +00:00
|
|
|
case BPF_BTF_LOAD:
|
2022-02-09 23:19:57 +00:00
|
|
|
case BPF_LINK_CREATE:
|
|
|
|
|
case BPF_RAW_TRACEPOINT_OPEN:
|
2021-05-14 00:36:05 +00:00
|
|
|
break;
|
2022-08-09 03:58:09 +00:00
|
|
|
default:
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
return __sys_bpf(cmd, KERNEL_BPFPTR(attr), attr_size);
|
|
|
|
|
}
|
|
|
|
|
|
2022-08-11 06:52:28 +00:00
|
|
|
|
|
|
|
|
/* To shut up -Wmissing-prototypes.
|
|
|
|
|
* This function is used by the kernel light skeleton
|
|
|
|
|
* to load bpf programs when modules are loaded or during kernel boot.
|
|
|
|
|
* See tools/lib/bpf/skel_internal.h
|
|
|
|
|
*/
|
|
|
|
|
int kern_sys_bpf(int cmd, union bpf_attr *attr, unsigned int size);
|
|
|
|
|
|
2022-08-09 03:58:09 +00:00
|
|
|
int kern_sys_bpf(int cmd, union bpf_attr *attr, unsigned int size)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog * __maybe_unused prog;
|
|
|
|
|
struct bpf_tramp_run_ctx __maybe_unused run_ctx;
|
|
|
|
|
|
|
|
|
|
switch (cmd) {
|
2022-02-09 23:19:57 +00:00
|
|
|
#ifdef CONFIG_BPF_JIT /* __bpf_prog_enter_sleepable used by trampoline and JIT */
|
|
|
|
|
case BPF_PROG_TEST_RUN:
|
|
|
|
|
if (attr->test.data_in || attr->test.data_out ||
|
|
|
|
|
attr->test.ctx_out || attr->test.duration ||
|
|
|
|
|
attr->test.repeat || attr->test.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
prog = bpf_prog_get_type(attr->test.prog_fd, BPF_PROG_TYPE_SYSCALL);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
|
|
|
|
if (attr->test.ctx_size_in < prog->aux->max_ctx_offset ||
|
|
|
|
|
attr->test.ctx_size_in > U16_MAX) {
|
|
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
2022-05-10 20:59:20 +00:00
|
|
|
run_ctx.bpf_cookie = 0;
|
2022-10-25 18:45:16 +00:00
|
|
|
if (!__bpf_prog_enter_sleepable_recur(prog, &run_ctx)) {
|
2022-02-09 23:19:57 +00:00
|
|
|
/* recursion detected */
|
2023-08-30 08:04:04 +00:00
|
|
|
__bpf_prog_exit_sleepable_recur(prog, 0, &run_ctx);
|
2022-02-09 23:19:57 +00:00
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return -EBUSY;
|
|
|
|
|
}
|
|
|
|
|
attr->test.retval = bpf_prog_run(prog, (void *) (long) attr->test.ctx_in);
|
2022-10-25 18:45:16 +00:00
|
|
|
__bpf_prog_exit_sleepable_recur(prog, 0 /* bpf_prog_run does runtime stats */,
|
|
|
|
|
&run_ctx);
|
2022-02-09 23:19:57 +00:00
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
return 0;
|
|
|
|
|
#endif
|
2021-05-14 00:36:05 +00:00
|
|
|
default:
|
2022-08-09 03:58:09 +00:00
|
|
|
return ____bpf_sys_bpf(cmd, attr, size);
|
2021-05-14 00:36:05 +00:00
|
|
|
}
|
2021-05-14 00:36:03 +00:00
|
|
|
}
|
2022-08-09 03:58:09 +00:00
|
|
|
EXPORT_SYMBOL(kern_sys_bpf);
|
2021-05-14 00:36:03 +00:00
|
|
|
|
2021-05-19 06:41:16 +00:00
|
|
|
static const struct bpf_func_proto bpf_sys_bpf_proto = {
|
2021-05-14 00:36:03 +00:00
|
|
|
.func = bpf_sys_bpf,
|
|
|
|
|
.gpl_only = false,
|
|
|
|
|
.ret_type = RET_INTEGER,
|
|
|
|
|
.arg1_type = ARG_ANYTHING,
|
2021-12-17 00:31:51 +00:00
|
|
|
.arg2_type = ARG_PTR_TO_MEM | MEM_RDONLY,
|
2021-05-14 00:36:03 +00:00
|
|
|
.arg3_type = ARG_CONST_SIZE,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
const struct bpf_func_proto * __weak
|
|
|
|
|
tracing_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
|
|
|
|
|
{
|
2024-01-24 02:21:04 +00:00
|
|
|
return bpf_base_func_proto(func_id, prog);
|
2021-05-14 00:36:03 +00:00
|
|
|
}
|
|
|
|
|
|
2021-05-14 00:36:12 +00:00
|
|
|
BPF_CALL_1(bpf_sys_close, u32, fd)
|
|
|
|
|
{
|
|
|
|
|
/* When bpf program calls this helper there should not be
|
|
|
|
|
* an fdget() without matching completed fdput().
|
|
|
|
|
* This helper is allowed in the following callchain only:
|
|
|
|
|
* sys_bpf->prog_test_run->bpf_prog->bpf_sys_close
|
|
|
|
|
*/
|
|
|
|
|
return close_fd(fd);
|
|
|
|
|
}
|
|
|
|
|
|
2021-05-19 06:41:16 +00:00
|
|
|
static const struct bpf_func_proto bpf_sys_close_proto = {
|
2021-05-14 00:36:12 +00:00
|
|
|
.func = bpf_sys_close,
|
|
|
|
|
.gpl_only = false,
|
|
|
|
|
.ret_type = RET_INTEGER,
|
|
|
|
|
.arg1_type = ARG_ANYTHING,
|
|
|
|
|
};
|
|
|
|
|
|
2021-10-28 06:34:54 +00:00
|
|
|
BPF_CALL_4(bpf_kallsyms_lookup_name, const char *, name, int, name_sz, int, flags, u64 *, res)
|
|
|
|
|
{
|
|
|
|
|
if (flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (name_sz <= 1 || name[name_sz - 1])
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
if (!bpf_dump_raw_ok(current_cred()))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
*res = kallsyms_lookup_name(name);
|
|
|
|
|
return *res ? 0 : -ENOENT;
|
|
|
|
|
}
|
|
|
|
|
|
2022-06-16 22:54:07 +00:00
|
|
|
static const struct bpf_func_proto bpf_kallsyms_lookup_name_proto = {
|
2021-10-28 06:34:54 +00:00
|
|
|
.func = bpf_kallsyms_lookup_name,
|
|
|
|
|
.gpl_only = false,
|
|
|
|
|
.ret_type = RET_INTEGER,
|
|
|
|
|
.arg1_type = ARG_PTR_TO_MEM,
|
2021-11-22 23:57:31 +00:00
|
|
|
.arg2_type = ARG_CONST_SIZE_OR_ZERO,
|
2021-10-28 06:34:54 +00:00
|
|
|
.arg3_type = ARG_ANYTHING,
|
|
|
|
|
.arg4_type = ARG_PTR_TO_LONG,
|
|
|
|
|
};
|
|
|
|
|
|
2021-05-14 00:36:03 +00:00
|
|
|
static const struct bpf_func_proto *
|
|
|
|
|
syscall_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
switch (func_id) {
|
|
|
|
|
case BPF_FUNC_sys_bpf:
|
2024-01-24 02:21:04 +00:00
|
|
|
return !bpf_token_capable(prog->aux->token, CAP_PERFMON)
|
|
|
|
|
? NULL : &bpf_sys_bpf_proto;
|
2021-05-14 00:36:11 +00:00
|
|
|
case BPF_FUNC_btf_find_by_name_kind:
|
|
|
|
|
return &bpf_btf_find_by_name_kind_proto;
|
2021-05-14 00:36:12 +00:00
|
|
|
case BPF_FUNC_sys_close:
|
|
|
|
|
return &bpf_sys_close_proto;
|
2021-10-28 06:34:54 +00:00
|
|
|
case BPF_FUNC_kallsyms_lookup_name:
|
|
|
|
|
return &bpf_kallsyms_lookup_name_proto;
|
2021-05-14 00:36:03 +00:00
|
|
|
default:
|
|
|
|
|
return tracing_prog_func_proto(func_id, prog);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
const struct bpf_verifier_ops bpf_syscall_verifier_ops = {
|
|
|
|
|
.get_func_proto = syscall_prog_func_proto,
|
|
|
|
|
.is_valid_access = syscall_prog_is_valid_access,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
const struct bpf_prog_ops bpf_syscall_prog_ops = {
|
|
|
|
|
.test_run = bpf_prog_test_run_syscall,
|
|
|
|
|
};
|
2022-04-07 07:07:59 +00:00
|
|
|
|
|
|
|
|
#ifdef CONFIG_SYSCTL
|
|
|
|
|
static int bpf_stats_handler(struct ctl_table *table, int write,
|
|
|
|
|
void *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
|
{
|
|
|
|
|
struct static_key *key = (struct static_key *)table->data;
|
|
|
|
|
static int saved_val;
|
|
|
|
|
int val, ret;
|
|
|
|
|
struct ctl_table tmp = {
|
|
|
|
|
.data = &val,
|
|
|
|
|
.maxlen = sizeof(val),
|
|
|
|
|
.mode = table->mode,
|
|
|
|
|
.extra1 = SYSCTL_ZERO,
|
|
|
|
|
.extra2 = SYSCTL_ONE,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
if (write && !capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
mutex_lock(&bpf_stats_enabled_mutex);
|
|
|
|
|
val = saved_val;
|
|
|
|
|
ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
|
|
|
|
|
if (write && !ret && val != saved_val) {
|
|
|
|
|
if (val)
|
|
|
|
|
static_key_slow_inc(key);
|
|
|
|
|
else
|
|
|
|
|
static_key_slow_dec(key);
|
|
|
|
|
saved_val = val;
|
|
|
|
|
}
|
|
|
|
|
mutex_unlock(&bpf_stats_enabled_mutex);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void __weak unpriv_ebpf_notify(int new_state)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_unpriv_handler(struct ctl_table *table, int write,
|
|
|
|
|
void *buffer, size_t *lenp, loff_t *ppos)
|
|
|
|
|
{
|
|
|
|
|
int ret, unpriv_enable = *(int *)table->data;
|
|
|
|
|
bool locked_state = unpriv_enable == 1;
|
|
|
|
|
struct ctl_table tmp = *table;
|
|
|
|
|
|
|
|
|
|
if (write && !capable(CAP_SYS_ADMIN))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
|
|
tmp.data = &unpriv_enable;
|
|
|
|
|
ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
|
|
|
|
|
if (write && !ret) {
|
|
|
|
|
if (locked_state && unpriv_enable != 1)
|
|
|
|
|
return -EPERM;
|
|
|
|
|
*(int *)table->data = unpriv_enable;
|
|
|
|
|
}
|
|
|
|
|
|
2023-05-02 18:14:18 +00:00
|
|
|
if (write)
|
|
|
|
|
unpriv_ebpf_notify(unpriv_enable);
|
2022-04-07 07:07:59 +00:00
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct ctl_table bpf_syscall_table[] = {
|
|
|
|
|
{
|
|
|
|
|
.procname = "unprivileged_bpf_disabled",
|
|
|
|
|
.data = &sysctl_unprivileged_bpf_disabled,
|
|
|
|
|
.maxlen = sizeof(sysctl_unprivileged_bpf_disabled),
|
|
|
|
|
.mode = 0644,
|
|
|
|
|
.proc_handler = bpf_unpriv_handler,
|
|
|
|
|
.extra1 = SYSCTL_ZERO,
|
|
|
|
|
.extra2 = SYSCTL_TWO,
|
|
|
|
|
},
|
|
|
|
|
{
|
|
|
|
|
.procname = "bpf_stats_enabled",
|
|
|
|
|
.data = &bpf_stats_enabled_key.key,
|
|
|
|
|
.mode = 0644,
|
|
|
|
|
.proc_handler = bpf_stats_handler,
|
|
|
|
|
},
|
|
|
|
|
{ }
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int __init bpf_syscall_sysctl_init(void)
|
|
|
|
|
{
|
|
|
|
|
register_sysctl_init("kernel", bpf_syscall_table);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
late_initcall(bpf_syscall_sysctl_init);
|
|
|
|
|
#endif /* CONFIG_SYSCTL */
|
