The selftest build fails when trying to install the scripts:
rsync: [sender] link_stat "tools/testing/selftests/bpf/test_docs_build.sh" failed: No such file or directory (2)
Fix the filename.
Fixes: a01d935b2e ("tools/bpf: Remove bpf-helpers from bpftool docs")
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210308182830.155784-1-jean-philippe@linaro.org
Björn Töpel says:
====================
This two-patch series introduces load-acquire/store-release barriers
for the AF_XDP rings.
For most contemporary architectures, this is more effective than a
SPSC ring based on smp_{r,w,}mb() barriers. More importantly,
load-acquire/store-release semantics make the ring code easier to
follow.
This is effectively the change done in commit 6c43c091bd
("documentation: Update circular buffer for
load-acquire/store-release"), but for the AF_XDP rings.
Both libbpf and the kernel-side are updated.
Full details are outlined in the commits!
Thanks to the LKMM-folks (Paul/Alan/Will) for helping me out in this
complicated matter!
Changelog
v1[1]->v2:
* Expanded the commit message for patch 1, and included the LKMM
litmus tests. Hopefully this clear things up. (Daniel)
* Clarified why the smp_mb()/smp_load_acquire() is not needed in (A);
control dependency with load to store. (Toke)
[1] https://lore.kernel.org/bpf/20210301104318.263262-1-bjorn.topel@gmail.com/
Thanks,
Björn
====================
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Now that the AF_XDP rings have load-acquire/store-release semantics,
move libbpf to that as well.
The library-internal libbpf_smp_{load_acquire,store_release} are only
valid for 32-bit words on ARM64.
Also, remove the barriers that are no longer in use.
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210305094113.413544-3-bjorn.topel@gmail.com
Currently, the AF_XDP rings uses general smp_{r,w,}mb() barriers on
the kernel-side. On most modern architectures
load-acquire/store-release barriers perform better, and results in
simpler code for circular ring buffers.
This change updates the XDP socket rings to use
load-acquire/store-release barriers.
It is important to note that changing from the old smp_{r,w,}mb()
barriers, to load-acquire/store-release barriers does not break
compatibility. The old semantics work with the new one, and vice
versa.
As pointed out by "Documentation/memory-barriers.txt" in the "SMP
BARRIER PAIRING" section:
"General barriers pair with each other, though they also pair with
most other types of barriers, albeit without multicopy atomicity.
An acquire barrier pairs with a release barrier, but both may also
pair with other barriers, including of course general barriers."
How different barriers behaves and pairs is outlined in
"tools/memory-model/Documentation/cheatsheet.txt".
In order to make sure that compatibility is not broken, LKMM herd7
based litmus tests can be constructed and verified.
We generalize the XDP socket ring to a one entry ring, and create two
scenarios; One where the ring is full, where only the consumer can
proceed, followed by the producer. One where the ring is empty, where
only the producer can proceed, followed by the consumer. Each scenario
is then expanded to four different tests: general producer/general
consumer, general producer/acqrel consumer, acqrel producer/general
consumer, acqrel producer/acqrel consumer. In total eight tests.
The empty ring test:
C spsc-rb+empty
// Simple one entry ring:
// prod cons allowed action prod cons
// 0 0 => prod => 1 0
// 0 1 => cons => 0 0
// 1 0 => cons => 1 1
// 1 1 => prod => 0 1
{}
// We start at prod==0, cons==0, data==0, i.e. nothing has been
// written to the ring. From here only the producer can start, and
// should write 1. Afterwards, consumer can continue and read 1 to
// data. Can we enter state prod==1, cons==1, but consumer observed
// the incorrect value of 0?
P0(int *prod, int *cons, int *data)
{
... producer
}
P1(int *prod, int *cons, int *data)
{
... consumer
}
exists( 1:d=0 /\ prod=1 /\ cons=1 );
The full ring test:
C spsc-rb+full
// Simple one entry ring:
// prod cons allowed action prod cons
// 0 0 => prod => 1 0
// 0 1 => cons => 0 0
// 1 0 => cons => 1 1
// 1 1 => prod => 0 1
{ prod = 1; }
// We start at prod==1, cons==0, data==1, i.e. producer has
// written 0, so from here only the consumer can start, and should
// consume 0. Afterwards, producer can continue and write 1 to
// data. Can we enter state prod==0, cons==1, but consumer observed
// the write of 1?
P0(int *prod, int *cons, int *data)
{
... producer
}
P1(int *prod, int *cons, int *data)
{
... consumer
}
exists( 1:d=1 /\ prod=0 /\ cons=1 );
where P0 and P1 are:
P0(int *prod, int *cons, int *data)
{
int p;
p = READ_ONCE(*prod);
if (READ_ONCE(*cons) == p) {
WRITE_ONCE(*data, 1);
smp_wmb();
WRITE_ONCE(*prod, p ^ 1);
}
}
P0(int *prod, int *cons, int *data)
{
int p;
p = READ_ONCE(*prod);
if (READ_ONCE(*cons) == p) {
WRITE_ONCE(*data, 1);
smp_store_release(prod, p ^ 1);
}
}
P1(int *prod, int *cons, int *data)
{
int c;
int d = -1;
c = READ_ONCE(*cons);
if (READ_ONCE(*prod) != c) {
smp_rmb();
d = READ_ONCE(*data);
smp_mb();
WRITE_ONCE(*cons, c ^ 1);
}
}
P1(int *prod, int *cons, int *data)
{
int c;
int d = -1;
c = READ_ONCE(*cons);
if (smp_load_acquire(prod) != c) {
d = READ_ONCE(*data);
smp_store_release(cons, c ^ 1);
}
}
The full LKMM litmus tests are found at [1].
On x86-64 systems the l2fwd AF_XDP xdpsock sample performance
increases by 1%. This is mostly due to that the smp_mb() is removed,
which is a relatively expensive operation on these
platforms. Weakly-ordered platforms, such as ARM64 might benefit even
more.
[1] https://github.com/bjoto/litmus-xsk
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210305094113.413544-2-bjorn.topel@gmail.com
When testing uprobes we the test gets GEP (Global Entry Point)
address from kallsyms, but then the function is called locally
so the uprobe is not triggered.
Fixing this by adjusting the address to LEP (Local Entry Point)
for powerpc arch plus instruction check stolen from ppc_function_entry
function pointed out and explained by Michael and Naveen.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Link: https://lore.kernel.org/bpf/20210305134050.139840-1-jolsa@kernel.org
Add BPF_F_ADJ_ROOM_ENCAP_L2_ETH flag to the existing tests which
encapsulates the ethernet as the inner l2 header.
Update a vxlan encapsulation test case.
Signed-off-by: Xuesen Huang <huangxuesen@kuaishou.com>
Signed-off-by: Li Wang <wangli09@kuaishou.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/bpf/20210305123347.15311-1-hxseverything@gmail.com
bpf_skb_adjust_room sets the inner_protocol as skb->protocol for packets
encapsulation. But that is not appropriate when pushing Ethernet header.
Add an option to further specify encap L2 type and set the inner_protocol
as ETH_P_TEB.
Suggested-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Xuesen Huang <huangxuesen@kuaishou.com>
Signed-off-by: Zhiyong Cheng <chengzhiyong@kuaishou.com>
Signed-off-by: Li Wang <wangli09@kuaishou.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/bpf/20210304064046.6232-1-hxseverything@gmail.com
Fix the following coccicheck warnings:
./tools/testing/selftests/bpf/test_sockmap.c:735:35-37: WARNING !A || A
&& B is equivalent to !A || B.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/1614757930-17197-1-git-send-email-jiapeng.chong@linux.alibaba.com
Fix the following coccicheck warnings:
./tools/bpf/bpf_dbg.c:1201:55-57: WARNING !A || A && B is equivalent to
!A || B.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/1614756035-111280-1-git-send-email-jiapeng.chong@linux.alibaba.com
Lorenz Bauer says:
====================
We don't have PROG_TEST_RUN support for sk_lookup programs at the
moment. So far this hasn't been a problem, since we can run our
tests in a separate network namespace. For benchmarking it's nice
to have PROG_TEST_RUN, so I've gone and implemented it.
Based on discussion on the v1 I've dropped support for testing multiple
programs at once.
Changes since v3:
- Use bpf_test_timer prefix (Andrii)
Changes since v2:
- Fix test_verifier failure (Alexei)
Changes since v1:
- Add sparse annotations to the t_* functions
- Add appropriate type casts in bpf_prog_test_run_sk_lookup
- Drop running multiple programs
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
sk_lookup doesn't allow setting data_in for bpf_prog_run. This doesn't
play well with the verifier tests, since they always set a 64 byte
input buffer. Allow not running verifier tests by setting
bpf_test.runs to a negative value and don't run the ctx access case
for sk_lookup. We have dedicated ctx access tests so skipping here
doesn't reduce coverage.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210303101816.36774-6-lmb@cloudflare.com
Extend a simple prog_run test to check that PROG_TEST_RUN adheres
to the requested repetitions. Convert it to use BPF skeleton.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210303101816.36774-5-lmb@cloudflare.com
Convert the selftests for sk_lookup narrow context access to use
PROG_TEST_RUN instead of creating actual sockets. This ensures that
ctx is populated correctly when using PROG_TEST_RUN.
Assert concrete values since we now control remote_ip and remote_port.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210303101816.36774-4-lmb@cloudflare.com
Allow to pass sk_lookup programs to PROG_TEST_RUN. User space
provides the full bpf_sk_lookup struct as context. Since the
context includes a socket pointer that can't be exposed
to user space we define that PROG_TEST_RUN returns the cookie
of the selected socket or zero in place of the socket pointer.
We don't support testing programs that select a reuseport socket,
since this would mean running another (unrelated) BPF program
from the sk_lookup test handler.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210303101816.36774-3-lmb@cloudflare.com
Share the timing / signal interruption logic between different
implementations of PROG_TEST_RUN. There is a change in behaviour
as well. We check the loop exit condition before checking for
pending signals. This resolves an edge case where a signal
arrives during the last iteration. Instead of aborting with
EINTR we return the successful result to user space.
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210303101816.36774-2-lmb@cloudflare.com
Joe Stringer says:
====================
The state of bpf(2) manual pages today is not exactly ideal. For the
most part, it was written several years ago and has not kept up with the
pace of development in the kernel tree. For instance, out of a total of
~35 commands to the BPF syscall available today, when I pull the
kernel-man-pages tree today I find just 6 documented commands: The very
basics of map interaction and program load.
In contrast, looking at bpf-helpers(7), I am able today to run one
command[0] to fetch API documentation of the very latest eBPF helpers
that have been added to the kernel. This documentation is up to date
because kernel maintainers enforce documenting the APIs as part of
the feature submission process. As far as I can tell, we rely on manual
synchronization from the kernel tree to the kernel-man-pages tree to
distribute these more widely, so all locations may not be completely up
to date. That said, the documentation does in fact exist in the first
place which is a major initial hurdle to overcome.
Given the relative success of the process around bpf-helpers(7) to
encourage developers to document their user-facing changes, in this
patch series I explore applying this technique to bpf(2) as well.
Unfortunately, even with bpf(2) being so out-of-date, there is still a
lot of content to convert over. In particular, the following aspects of
the bpf syscall could also be individually be generated from separate
documentation in the header:
* BPF syscall commands
* BPF map types
* BPF program types
* BPF program subtypes (aka expected_attach_type)
* BPF attachment points
Rather than tackle everything at once, I have focused in this series on
the syscall commands, "enum bpf_cmd". This series is structured to first
import what useful descriptions there are from the kernel-man-pages
tree, then piece-by-piece document a few of the syscalls in more detail
in cases where I could find useful documentation from the git tree or
from a casual read of the code. Not all documentation is comprehensive
at this point, but a basis is provided with examples that can be further
enhanced with subsequent follow-up patches. Note, the series in its
current state only includes documentation around the syscall commands
themselves, so in the short term it doesn't allow us to automate bpf(2)
generation; Only one section of the man page could be replaced. Though
if there is appetite for this approach, this should be trivial to
improve on, even if just by importing the remaining static text from the
kernel-man-pages tree.
Following that, the series enhances the python scripting around parsing
the descriptions from the header files and generating dedicated
ReStructured Text and troff output. Finally, to expose the new text and
reduce the likelihood of having it get out of date or break the docs
parser, it is added to the selftests and exposed through the kernel
documentation web pages.
The eventual goal of this effort would be to extend the kernel UAPI
headers such that each of the categories I had listed above (commands,
maps, progs, hooks) have dedicated documentation in the kernel tree, and
that developers must update the comments in the headers to document the
APIs prior to patch acceptance, and that we could auto-generate the
latest version of the bpf(2) manual pages based on a few static
description sections combined with the dynamically-generated output from
the header.
This patch series can also be found at the following location on GitHub:
https://github.com/joestringer/linux/tree/submit/bpf-command-docs_v2
Thanks also to Quentin Monnet for initial review.
[0]: make -C tools/testing/selftests/bpf docs
v2:
* Remove build infrastructure in favor of kernel-doc directives
* Shift userspace-api docs under Documentation/userspace-api/ebpf
* Fix scripts/bpf_doc.py syscall --header (throw unsupported error)
* Improve .gitignore handling of newly autogenerated files
====================
Acked-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Synchronize the header after all of the recent changes.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-16-joe@cilium.io
Generate the syscall command reference from the UAPI header file and
include it in the main bpf docs page.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-15-joe@cilium.io
Add building of the bpf(2) syscall commands documentation as part of the
docs building step in the build. This allows us to pick up on potential
parse errors from the docs generator script as part of selftests.
The generated manual pages here are not intended for distribution, they
are just a fragment that can be integrated into the other static text of
bpf(2) to form the full manual page.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-14-joe@cilium.io
Previously, the Makefile here was only targeting a single manual page so
it just hardcoded a bunch of individual rules to specifically handle
build, clean, install, uninstall for that particular page.
Upcoming commits will generate manual pages for an additional section,
so this commit prepares the makefile first by converting the existing
targets into an evaluated set of targets based on the manual page name
and section.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-13-joe@cilium.io
This logic is used for validating the manual pages from selftests, so
move the infra under tools/testing/selftests/bpf/ and rely on selftests
for validation rather than tying it into the bpftool build.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-12-joe@cilium.io
Add a new target to bpf_doc.py to support generating the list of syscall
commands directly from the UAPI headers. Assuming that developer
submissions keep the main header up to date, this should allow the man
pages to be automatically generated based on the latest API changes
rather than requiring someone to separately go back through the API and
describe each command.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-11-joe@cilium.io
Abstract out the target parameter so that upcoming commits, more than
just the existing "helpers" target can be called to generate specific
portions of docs from the eBPF UAPI headers.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-10-joe@cilium.io
Based roughly on the following commits:
* Commit cb4d03ab49 ("bpf: Add generic support for lookup batch op")
* Commit 057996380a ("bpf: Add batch ops to all htab bpf map")
* Commit aa2e93b8e5 ("bpf: Add generic support for update and delete
batch ops")
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Brian Vazquez <brianvv@google.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-9-joe@cilium.io
Commit 468e2f64d2 ("bpf: introduce BPF_PROG_QUERY command") originally
introduced this, but there have been several additions since then.
Unlike BPF_PROG_ATTACH, it appears that the sockmap progs are not able
to be queried so far.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-8-joe@cilium.io
Based on a brief read of the corresponding source code.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-7-joe@cilium.io
Commit b2197755b2 ("bpf: add support for persistent maps/progs")
contains the original implementation and git logs, used as reference for
this documentation.
Also pull in the filename restriction as documented in commit 6d8cb045cd
("bpf: comment why dots in filenames under BPF virtual FS are not allowed")
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-5-joe@cilium.io
Document the meaning of the BPF_F_LOCK flag for the map lookup/update
descriptions. Based on commit 96049f3afd ("bpf: introduce BPF_F_LOCK
flag").
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-4-joe@cilium.io
Introduce high-level descriptions of the intent and return codes of the
bpf() syscall commands. Subsequent patches may further flesh out the
content to provide a more useful programming reference.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-3-joe@cilium.io
These descriptions are present in the man-pages project from the
original submissions around 2015-2016. Import them so that they can be
kept up to date as developers extend the bpf syscall commands.
These descriptions follow the pattern used by scripts/bpf_helpers_doc.py
so that we can take advantage of the parser to generate more up-to-date
man page writing based upon these headers.
Some minor wording adjustments were made to make the descriptions
more consistent for the description / return format.
Signed-off-by: Joe Stringer <joe@cilium.io>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210302171947.2268128-2-joe@cilium.io
Co-authored-by: Alexei Starovoitov <ast@kernel.org>
Co-authored-by: Michael Kerrisk <mtk.manpages@gmail.com>
Ilya Leoshkevich says:
====================
Some BPF programs compiled on s390 fail to load, because s390
arch-specific linux headers contain float and double types.
Introduce support for such types by representing them using the new
BTF_KIND_FLOAT. This series deals with libbpf, bpftool, in-kernel BTF
parser as well as selftests and documentation.
There are also pahole and LLVM parts:
* https://github.com/iii-i/dwarves/commit/btf-kind-float-v2
* https://reviews.llvm.org/D83289
but they should go in after the libbpf part is integrated.
---
v0: https://lore.kernel.org/bpf/20210210030317.78820-1-iii@linux.ibm.com/
v0 -> v1: Per Andrii's suggestion, remove the unnecessary trailing u32.
v1: https://lore.kernel.org/bpf/20210216011216.3168-1-iii@linux.ibm.com/
v1 -> v2: John noticed that sanitization corrupts BTF, because new and
old sizes don't match. Per Yonghong's suggestion, use a
modifier type (which has the same size as the float type) as
a replacement.
Per Yonghong's suggestion, add size and alignment checks to
the kernel BTF parser.
v2: https://lore.kernel.org/bpf/20210219022543.20893-1-iii@linux.ibm.com/
v2 -> v3: Based on Yonghong's suggestions: Use BTF_KIND_CONST instead of
BTF_KIND_TYPEDEF and make sure that the C code generated from
the sanitized BTF is well-formed; fix size calculation in
tests and use NAME_TBD everywhere; limit allowed sizes to 2,
4, 8, 12 and 16 (this should also fix m68k and nds32le
builds).
v3: https://lore.kernel.org/bpf/20210220034959.27006-1-iii@linux.ibm.com/
v3 -> v4: More fixes for the Yonghong's findings: fix the outdated
comment in bpf_object__sanitize_btf() and add the error
handling there (I've decided to check uint_id and uchar_id
too in order to simplify debugging); add bpftool output
example; use div64_u64_rem() instead of % in order to fix the
linker error.
Also fix the "invalid BTF_INFO" test (new commit, #4).
v4: https://lore.kernel.org/bpf/20210222214917.83629-1-iii@linux.ibm.com/
v4 -> v5: Fixes for the Andrii's findings: Use BTF_KIND_STRUCT instead
of BTF_KIND_TYPEDEF for sanitization; check byte_sz in
libbpf; move btf__add_float; remove relo support; add a dedup
test (new commit, #7).
v5: https://lore.kernel.org/bpf/20210223231459.99664-1-iii@linux.ibm.com/
v5 -> v6: Fixes for further findings by Andrii: split whitespace issue
fix into a separate patch; add 12-byte float to "float test #1,
well-formed".
v6: https://lore.kernel.org/bpf/20210224234535.106970-1-iii@linux.ibm.com/
v6 -> v7: John suggested to add a comment explaining why sanitization
does not preserve the type name, as well as what effect it has
on running the code on the older kernels.
Yonghong has asked to add a comment explaining why we are not
checking the alignment very precisely in the kernel.
John suggested to add a bpf_core_field_size test (commit #9).
Based on Alexei's feedback [1] I'm proceeding with the BTF_KIND_FLOAT
approach.
[1] https://lore.kernel.org/bpf/CAADnVQKWPODWZ2RSJ5FJhfYpxkuV0cvSAL1O+FSr9oP1ercoBg@mail.gmail.com/
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Check that floats don't interfere with struct deduplication, that they
are not merged with another kinds and that floats of different sizes are
not merged with each other.
Suggested-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210226202256.116518-9-iii@linux.ibm.com
Test the good variants as well as the potential malformed ones.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210226202256.116518-8-iii@linux.ibm.com
On the kernel side, introduce a new btf_kind_operations. It is
similar to that of BTF_KIND_INT, however, it does not need to
handle encodings and bit offsets. Do not implement printing, since
the kernel does not know how to format floating-point values.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210226202256.116518-7-iii@linux.ibm.com
The bit being checked by this test is no longer reserved after
introducing BTF_KIND_FLOAT, so use the next one instead.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210226202256.116518-6-iii@linux.ibm.com
Only dumping support needs to be adjusted, the code structure follows
that of BTF_KIND_INT. Example plain and JSON outputs:
[4] FLOAT 'float' size=4
{"id":4,"kind":"FLOAT","name":"float","size":4}
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210226202256.116518-5-iii@linux.ibm.com
The logic follows that of BTF_KIND_INT most of the time. Sanitization
replaces BTF_KIND_FLOATs with equally-sized empty BTF_KIND_STRUCTs on
older kernels, for example, the following:
[4] FLOAT 'float' size=4
becomes the following:
[4] STRUCT '(anon)' size=4 vlen=0
With dwarves patch [1] and this patch, the older kernels, which were
failing with the floating-point-related errors, will now start working
correctly.
[1] https://github.com/iii-i/dwarves/commit/btf-kind-float-v2
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210226202256.116518-4-iii@linux.ibm.com
The original bcc pull request https://github.com/iovisor/bcc/pull/3270 exposed
a verifier failure with Clang 12/13 while Clang 4 works fine.
Further investigation exposed two issues:
Issue 1: LLVM may generate code which uses less refined value. The issue is
fixed in LLVM patch: https://reviews.llvm.org/D97479
Issue 2: Spills with initial value 0 are marked as precise which makes later
state pruning less effective. This is my rough initial analysis and
further investigation is needed to find how to improve verifier
pruning in such cases.
With the above LLVM patch, for the new loop6.c test, which has smaller loop
bound compared to original test, I got:
$ test_progs -s -n 10/16
...
stack depth 64
processed 390735 insns (limit 1000000) max_states_per_insn 87
total_states 8658 peak_states 964 mark_read 6
#10/16 loop6.o:OK
Use the original loop bound, i.e., commenting out "#define WORKAROUND", I got:
$ test_progs -s -n 10/16
...
BPF program is too large. Processed 1000001 insn
stack depth 64
processed 1000001 insns (limit 1000000) max_states_per_insn 91
total_states 23176 peak_states 5069 mark_read 6
...
#10/16 loop6.o:FAIL
The purpose of this patch is to provide a regression test for the above LLVM fix
and also provide a test case for further analyzing the verifier pruning issue.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Zhenwei Pi <pizhenwei@bytedance.com>
Link: https://lore.kernel.org/bpf/20210226223810.236472-1-yhs@fb.com
This should fix the following warning:
include/linux/skbuff.h:932: warning: Function parameter or member
'_sk_redir' not described in 'sk_buff'
Reported-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Lorenz Bauer <lmb@cloudflare.com>
Link: https://lore.kernel.org/bpf/20210301184805.8174-1-xiyou.wangcong@gmail.com
Just like was done for bpftool and selftests in ec23eb7056 ("tools/bpftool:
Allow substituting custom vmlinux.h for the build") and ca4db6389d
("selftests/bpf: Allow substituting custom vmlinux.h for selftests build"),
allow to provide pre-generated vmlinux.h for runqslower build.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210303004010.653954-1-andrii@kernel.org
Yonghong Song says:
====================
This patch set introduced bpf_for_each_map_elem() helper.
The helper permits bpf program iterates through all elements
for a particular map.
The work originally inspired by an internal discussion where
firewall rules are kept in a map and bpf prog wants to
check packet 5 tuples against all rules in the map.
A bounded loop can be used but it has a few drawbacks.
As the loop iteration goes up, verification time goes up too.
For really large maps, verification may fail.
A helper which abstracts out the loop itself will not have
verification time issue.
A recent discussion in [1] involves to iterate all hash map
elements in bpf program. Currently iterating all hashmap elements
in bpf program is not easy if key space is really big.
Having a helper to abstract out the loop itself is even more
meaningful.
The proposed helper signature looks like:
long bpf_for_each_map_elem(map, callback_fn, callback_ctx, flags)
where callback_fn is a static function and callback_ctx is
a piece of data allocated on the caller stack which can be
accessed by the callback_fn. The callback_fn signature might be
different for different maps. For example, for hash/array maps,
the signature is
long callback_fn(map, key, val, callback_ctx)
In the rest of series, Patches 1/2/3/4 did some refactoring. Patch 5
implemented core kernel support for the helper. Patches 6 and 7
added hashmap and arraymap support. Patches 8/9 added libbpf
support. Patch 10 added bpftool support. Patches 11 and 12 added
selftests for hashmap and arraymap.
[1]: https://lore.kernel.org/bpf/20210122205415.113822-1-xiyou.wangcong@gmail.com/
Changelogs:
v4 -> v5:
- rebase on top of bpf-next.
v3 -> v4:
- better refactoring of check_func_call(), calculate subprogno outside
of __check_func_call() helper. (Andrii)
- better documentation (like the list of supported maps and their
callback signatures) in uapi header. (Andrii)
- implement and use ASSERT_LT in selftests. (Andrii)
- a few other minor changes.
v2 -> v3:
- add comments in retrieve_ptr_limit(), which is in sanitize_ptr_alu(),
to clarify the code is not executed for PTR_TO_MAP_KEY handling,
but code is manually tested. (Alexei)
- require BTF for callback function. (Alexei)
- simplify hashmap/arraymap callback return handling as return value
[0, 1] has been enforced by the verifier. (Alexei)
- also mark global subprog (if used in ld_imm64) as RELO_SUBPROG_ADDR. (Andrii)
- handle the condition to mark RELO_SUBPROG_ADDR properly. (Andrii)
- make bpftool subprog insn offset dumping consist with pcrel calls. (Andrii)
v1 -> v2:
- setup callee frame in check_helper_call() and then proceed to verify
helper return value as normal (Alexei)
- use meta data to keep track of map/func pointer to avoid hard coding
the register number (Alexei)
- verify callback_fn return value range [0, 1]. (Alexei)
- add migrate_{disable, enable} to ensure percpu value is the one
bpf program expects to see. (Alexei)
- change bpf_for_each_map_elem() return value to the number of iterated
elements. (Andrii)
- Change libbpf pseudo_func relo name to RELO_SUBPROG_ADDR and use
more rigid checking for the relocation. (Andrii)
- Better format to print out subprog address with bpftool. (Andrii)
- Use bpf_prog_test_run to trigger bpf run, instead of bpf_iter. (Andrii)
- Other misc changes.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
A test is added for arraymap and percpu arraymap. The test also
exercises the early return for the helper which does not
traverse all elements.
$ ./test_progs -n 45
#45/1 hash_map:OK
#45/2 array_map:OK
#45 for_each:OK
Summary: 1/2 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210226204934.3885756-1-yhs@fb.com
A test case is added for hashmap and percpu hashmap. The test
also exercises nested bpf_for_each_map_elem() calls like
bpf_prog:
bpf_for_each_map_elem(func1)
func1:
bpf_for_each_map_elem(func2)
func2:
$ ./test_progs -n 45
#45/1 hash_map:OK
#45 for_each:OK
Summary: 1/1 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210226204933.3885657-1-yhs@fb.com
