The first commit cited below attempts to fix the off-by-one error that
appeared in some comparisons with an open range. Due to this error,
arithmetically equivalent pieces of code could get different verdicts
from the verifier, for example (pseudocode):
// 1. Passes the verifier:
if (data + 8 > data_end)
return early
read *(u64 *)data, i.e. [data; data+7]
// 2. Rejected by the verifier (should still pass):
if (data + 7 >= data_end)
return early
read *(u64 *)data, i.e. [data; data+7]
The attempted fix, however, shifts the range by one in a wrong
direction, so the bug not only remains, but also such piece of code
starts failing in the verifier:
// 3. Rejected by the verifier, but the check is stricter than in #1.
if (data + 8 >= data_end)
return early
read *(u64 *)data, i.e. [data; data+7]
The change performed by that fix converted an off-by-one bug into
off-by-two. The second commit cited below added the BPF selftests
written to ensure than code chunks like #3 are rejected, however,
they should be accepted.
This commit fixes the off-by-two error by adjusting new_range in the
right direction and fixes the tests by changing the range into the
one that should actually fail.
Fixes: fb2a311a31 ("bpf: fix off by one for range markings with L{T, E} patterns")
Fixes: b37242c773 ("bpf: add test cases to bpf selftests to cover all access tests")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211130181607.593149-1-maximmi@nvidia.com
Libbpf development version was bumped to 0.7 in c93faaaf2f
("libbpf: Deprecate bpf_prog_load_xattr() API"), activating a bunch of
previously scheduled deprecations. Most APIs are pretty straightforward
to replace with newer APIs, but perf has a complicated mixed setup with
libbpf used both as static and shared configurations, which makes it
non-trivial to migrate the APIs.
Further, bpf_program__set_prep() needs more involved refactoring, which
will require help from Arnaldo and/or Jiri.
So for now, mute deprecation warnings and work on migrating perf off of
deprecated APIs separately with the input from owners of the perf tool.
Fixes: c93faaaf2f ("libbpf: Deprecate bpf_prog_load_xattr() API")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20211203004640.2455717-1-andrii@kernel.org
Previously, the selftest framework always treats it as *ok* even though
some of them are failed actually. That's because the script always
returns 0.
It supports PASS/FAIL/SKIP exit code now.
CC: Philip Li <philip.li@intel.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
qdiscs/fq_pie requires CONFIG_NET_SCH_FQ_PIE, otherwise tc will fail
to create a fq_pie qdisc.
It fixes following issue:
# not ok 57 83be - Create FQ-PIE with invalid number of flows
# Command exited with 2, expected 0
# Error: Specified qdisc not found.
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mark the summary result as FAIL to prevent from confusing the selftest
framework if some of them are failed.
Previously, the selftest framework always treats it as *ok* even though
some of them are failed actually. That's because the script tdc.sh always
return 0.
# All test results:
#
# 1..97
# ok 1 83be - Create FQ-PIE with invalid number of flows
# ok 2 8b6e - Create RED with no flags
[...snip]
# ok 6 5f15 - Create RED with flags ECN, harddrop
# ok 7 53e8 - Create RED with flags ECN, nodrop
# ok 8 d091 - Fail to create RED with only nodrop flag
# ok 9 af8e - Create RED with flags ECN, nodrop, harddrop
# not ok 10 ce7d - Add mq Qdisc to multi-queue device (4 queues)
# Could not match regex pattern. Verify command output:
# qdisc mq 1: root
# qdisc fq_codel 0: parent 1:4 limit 10240p flows 1024 quantum 1514 target 5ms interval 100ms memory_limit 32Mb ecn drop_batch 64
# qdisc fq_codel 0: parent 1:3 limit 10240p flows 1024 quantum 1514 target 5ms interval 100ms memory_limit 32Mb ecn drop_batch 64
[...snip]
# ok 96 6979 - Change quantum of a strict ETS band
# ok 97 9a7d - Change ETS strict band without quantum
#
#
#
#
ok 1 selftests: tc-testing: tdc.sh <<< summary result
CC: Philip Li <philip.li@intel.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Acked-by: Davide Caratti <dcaratti@redhat.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Boris reported that in one of his randconfig builds, objtool got
infinitely stuck. Turns out there's trivial list corruption in the
pv_ops tracking when a function is both in a static table and in a code
assignment.
Avoid re-adding function to the pv_ops[] lists when they're already on
it.
Fixes: db2b0c5d7b ("objtool: Support pv_opsindirect calls for noinstr")
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Tested-by: Borislav Petkov <bp@alien8.de>
Link: https://lkml.kernel.org/r/20211202204534.GA16608@worktop.programming.kicks-ass.net
Currently rp_filter tests in fib_tests.sh:fib_rp_filter_test() are
failing. ping sockets are bound to dummy1 using the "-I" option
(SO_BINDTODEVICE), but socket lookup is failing when receiving ping
replies, since the routing table thinks they belong to dummy0.
For example, suppose ping is using a SOCK_RAW socket for ICMP messages.
When receiving ping replies, in __raw_v4_lookup(), sk->sk_bound_dev_if
is 3 (dummy1), but dif (skb_rtable(skb)->rt_iif) says 2 (dummy0), so the
raw_sk_bound_dev_eq() check fails. Similar things happen in
ping_lookup() for SOCK_DGRAM sockets.
These tests used to pass due to a bug [1] in iputils, where "ping -I"
actually did not bind ICMP message sockets to device. The bug has been
fixed by iputils commit f455fee41c07 ("ping: also bind the ICMP socket
to the specific device") in 2016, which is why our rp_filter tests
started to fail. See [2] .
Fixing the tests while keeping everything in one netns turns out to be
nontrivial. Rework the tests and build the following topology:
┌─────────────────────────────┐ ┌─────────────────────────────┐
│ network namespace 1 (ns1) │ │ network namespace 2 (ns2) │
│ │ │ │
│ ┌────┐ ┌─────┐ │ │ ┌─────┐ ┌────┐ │
│ │ lo │<───>│veth1│<────────┼────┼─>│veth2│<──────────>│ lo │ │
│ └────┘ ├─────┴──────┐ │ │ ├─────┴──────┐ └────┘ │
│ │192.0.2.1/24│ │ │ │192.0.2.1/24│ │
│ └────────────┘ │ │ └────────────┘ │
└─────────────────────────────┘ └─────────────────────────────┘
Consider sending an ICMP_ECHO packet A in ns2. Both source and
destination IP addresses are 192.0.2.1, and we use strict mode rp_filter
in both ns1 and ns2:
1. A is routed to lo since its destination IP address is one of ns2's
local addresses (veth2);
2. A is redirected from lo's egress to veth2's egress using mirred;
3. A arrives at veth1's ingress in ns1;
4. A is redirected from veth1's ingress to lo's ingress, again, using
mirred;
5. In __fib_validate_source(), fib_info_nh_uses_dev() returns false,
since A was received on lo, but reverse path lookup says veth1;
6. However A is not dropped since we have relaxed this check for lo in
commit 66f8209547 ("fib: relax source validation check for loopback
packets");
Making sure A is not dropped here in this corner case is the whole point
of having this test.
7. As A reaches the ICMP layer, an ICMP_ECHOREPLY packet, B, is
generated;
8. Similarly, B is redirected from lo's egress to veth1's egress (in
ns1), then redirected once again from veth2's ingress to lo's
ingress (in ns2), using mirred.
Also test "ping 127.0.0.1" from ns2. It does not trigger the relaxed
check in __fib_validate_source(), but just to make sure the topology
works with loopback addresses.
Tested with ping from iputils 20210722-41-gf9fb573:
$ ./fib_tests.sh -t rp_filter
IPv4 rp_filter tests
TEST: rp_filter passes local packets [ OK ]
TEST: rp_filter passes loopback packets [ OK ]
[1] https://github.com/iputils/iputils/issues/55
[2] f455fee41c
Reported-by: Hangbin Liu <liuhangbin@gmail.com>
Fixes: adb701d6cf ("selftests: add a test case for rp_filter")
Reviewed-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Peilin Ye <peilin.ye@bytedance.com>
Acked-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20211201004720.6357-1-yepeilin.cs@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bpf_prog_load_xattr() is high-level API that's named as a low-level
BPF_PROG_LOAD wrapper APIs, but it actually operates on struct
bpf_object. It's badly and confusingly misnamed as it will load all the
progs insige bpf_object, returning prog_fd of the very first BPF
program. It also has a bunch of ad-hoc things like log_level override,
map_ifindex auto-setting, etc. All this can be expressed more explicitly
and cleanly through existing libbpf APIs. This patch marks
bpf_prog_load_xattr() for deprecation in libbpf v0.8 ([0]).
[0] Closes: https://github.com/libbpf/libbpf/issues/308
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20211201232824.3166325-10-andrii@kernel.org
Add bpf_program__set_log_level() and bpf_program__log_level() to fetch
and adjust log_level sent during BPF_PROG_LOAD command. This allows to
selectively request more or less verbose output in BPF verifier log.
Also bump libbpf version to 0.7 and make these APIs the first in v0.7.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20211201232824.3166325-3-andrii@kernel.org
resolve_btfids prints a warning when it finds an unresolved symbol,
(id == 0) in id_patch. This can be the case for BTF sets that are empty
(due to disabled config options), hence printing warnings for certain
builds, most recently seen in [0].
The reason behind this is because id->cnt aliases id->id in btf_id
struct, leading to empty set showing up as ID 0 when we get to id_patch,
which triggers the warning. Since sets are an exception here, accomodate
by reusing hole in btf_id for bool is_set member, setting it to true for
BTF set when setting id->cnt, and use that to skip extraneous warning.
[0]: https://lore.kernel.org/all/1b99ae14-abb4-d18f-cc6a-d7e523b25542@gmail.com
Before:
; ./tools/bpf/resolve_btfids/resolve_btfids -v -b vmlinux net/ipv4/tcp_cubic.ko
adding symbol tcp_cubic_kfunc_ids
WARN: resolve_btfids: unresolved symbol tcp_cubic_kfunc_ids
patching addr 0: ID 0 [tcp_cubic_kfunc_ids]
sorting addr 4: cnt 0 [tcp_cubic_kfunc_ids]
update ok for net/ipv4/tcp_cubic.ko
After:
; ./tools/bpf/resolve_btfids/resolve_btfids -v -b vmlinux net/ipv4/tcp_cubic.ko
adding symbol tcp_cubic_kfunc_ids
patching addr 0: ID 0 [tcp_cubic_kfunc_ids]
sorting addr 4: cnt 0 [tcp_cubic_kfunc_ids]
update ok for net/ipv4/tcp_cubic.ko
Fixes: 0e32dfc80b ("bpf: Enable TCP congestion control kfunc from modules")
Reported-by: Pavel Skripkin <paskripkin@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20211122144742.477787-4-memxor@gmail.com
Add $(OUTPUT) prefix to testing_helpers.o, so it can be built out of
tree when necessary. At the moment, in addition to being built in-tree
even when out-of-tree is required, testing_helpers.o is not built with
the right recipe when cross-building.
For consistency the other helpers, cgroup_helpers and trace_helpers, can
also be passed as objects instead of source. Use *_HELPERS variable to
keep the Makefile readable.
Fixes: f87c1930ac ("selftests/bpf: Merge test_stub.c into testing_helpers.c")
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201145101.823159-1-jean-philippe@linaro.org
Pull networking fixes from Jakub Kicinski:
"Including fixes from wireless, and wireguard.
Mostly scattered driver changes this week, with one big clump in
mv88e6xxx. Nothing of note, really.
Current release - regressions:
- smc: keep smc_close_final()'s error code during active close
Current release - new code bugs:
- iwlwifi: various static checker fixes (int overflow, leaks, missing
error codes)
- rtw89: fix size of firmware header before transfer, avoid crash
- mt76: fix timestamp check in tx_status; fix pktid leak;
- mscc: ocelot: fix missing unlock on error in ocelot_hwstamp_set()
Previous releases - regressions:
- smc: fix list corruption in smc_lgr_cleanup_early
- ipv4: convert fib_num_tclassid_users to atomic_t
Previous releases - always broken:
- tls: fix authentication failure in CCM mode
- vrf: reset IPCB/IP6CB when processing outbound pkts, prevent
incorrect processing
- dsa: mv88e6xxx: fixes for various device errata
- rds: correct socket tunable error in rds_tcp_tune()
- ipv6: fix memory leak in fib6_rule_suppress
- wireguard: reset peer src endpoint when netns exits
- wireguard: improve resilience to DoS around incoming handshakes
- tcp: fix page frag corruption on page fault which involves TCP
- mpls: fix missing attributes in delete notifications
- mt7915: fix NULL pointer dereference with ad-hoc mode
Misc:
- rt2x00: be more lenient about EPROTO errors during start
- mlx4_en: update reported link modes for 1/10G"
* tag 'net-5.16-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (85 commits)
net: dsa: b53: Add SPI ID table
gro: Fix inconsistent indenting
selftests: net: Correct case name
net/rds: correct socket tunable error in rds_tcp_tune()
mctp: Don't let RTM_DELROUTE delete local routes
net/smc: Keep smc_close_final rc during active close
ibmvnic: drop bad optimization in reuse_tx_pools()
ibmvnic: drop bad optimization in reuse_rx_pools()
net/smc: fix wrong list_del in smc_lgr_cleanup_early
Fix Comment of ETH_P_802_3_MIN
ethernet: aquantia: Try MAC address from device tree
ipv4: convert fib_num_tclassid_users to atomic_t
net: avoid uninit-value from tcp_conn_request
net: annotate data-races on txq->xmit_lock_owner
octeontx2-af: Fix a memleak bug in rvu_mbox_init()
net/mlx4_en: Fix an use-after-free bug in mlx4_en_try_alloc_resources()
vrf: Reset IPCB/IP6CB when processing outbound pkts in vrf dev xmit
net: qlogic: qlcnic: Fix a NULL pointer dereference in qlcnic_83xx_add_rings()
net: dsa: mv88e6xxx: Link in pcs_get_state() if AN is bypassed
net: dsa: mv88e6xxx: Fix inband AN for 2500base-x on 88E6393X family
...
Add a test where randmap() function is appended to three different bpf
programs. That action checks struct bpf_core_relo replication logic
and offset adjustment in gen loader part of libbpf.
Fourth bpf program has 360 CO-RE relocations from vmlinux, bpf_testmod,
and non-existing type. It tests candidate cache logic.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-16-alexei.starovoitov@gmail.com
Without lskel the CO-RE relocations are processed by libbpf before any other
work is done. Instead, when lskel is needed, remember relocation as RELO_CORE
kind. Then when loader prog is generated for a given bpf program pass CO-RE
relos of that program to gen loader via bpf_gen__record_relo_core(). The gen
loader will remember them as-is and pass it later as-is into the kernel.
The normal libbpf flow is to process CO-RE early before call relos happen. In
case of gen_loader the core relos have to be added to other relos to be copied
together when bpf static function is appended in different places to other main
bpf progs. During the copy the append_subprog_relos() will adjust insn_idx for
normal relos and for RELO_CORE kind too. When that is done each struct
reloc_desc has good relos for specific main prog.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-10-alexei.starovoitov@gmail.com
struct bpf_core_relo is generated by llvm and processed by libbpf.
It's a de-facto uapi.
With CO-RE in the kernel the struct bpf_core_relo becomes uapi de-jure.
Add an ability to pass a set of 'struct bpf_core_relo' to prog_load command
and let the kernel perform CO-RE relocations.
Note the struct bpf_line_info and struct bpf_func_info have the same
layout when passed from LLVM to libbpf and from libbpf to the kernel
except "insn_off" fields means "byte offset" when LLVM generates it.
Then libbpf converts it to "insn index" to pass to the kernel.
The struct bpf_core_relo's "insn_off" field is always "byte offset".
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-6-alexei.starovoitov@gmail.com
enum bpf_core_relo_kind is generated by llvm and processed by libbpf.
It's a de-facto uapi.
With CO-RE in the kernel the bpf_core_relo_kind values become uapi de-jure.
Also rename them with BPF_CORE_ prefix to distinguish from conflicting names in
bpf_core_read.h. The enums bpf_field_info_kind, bpf_type_id_kind,
bpf_type_info_kind, bpf_enum_value_kind are passing different values from bpf
program into llvm.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-5-alexei.starovoitov@gmail.com
Make relo_core.c to be compiled for the kernel and for user space libbpf.
Note the patch is reducing BPF_CORE_SPEC_MAX_LEN from 64 to 32.
This is the maximum number of nested structs and arrays.
For example:
struct sample {
int a;
struct {
int b[10];
};
};
struct sample *s = ...;
int *y = &s->b[5];
This field access is encoded as "0:1:0:5" and spec len is 4.
The follow up patch might bump it back to 64.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-4-alexei.starovoitov@gmail.com
To prepare relo_core.c to be compiled in the kernel and the user space
replace btf__type_by_id with btf_type_by_id.
In libbpf btf__type_by_id and btf_type_by_id have different behavior.
bpf_core_apply_relo_insn() needs behavior of uapi btf__type_by_id
vs internal btf_type_by_id, but type_id range check is already done
in bpf_core_apply_relo(), so it's safe to replace it everywhere.
The kernel btf_type_by_id() does the check anyway. It doesn't hurt.
Suggested-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20211201181040.23337-2-alexei.starovoitov@gmail.com
$ ./fcnal-test.sh -t help
Test names: help
Looks it intent to list the available tests but it didn't do the right
thing. I will add another option the do that in the later patch.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6_addr_bind/ipv4_addr_bind are function names. Previously, bind test
would not be run by default due to the wrong case names
Fixes: 34d0302ab8 ("selftests: Add ipv6 address bind tests to fcnal-test")
Fixes: 75b2b2b3db ("selftests: Add ipv4 address bind tests to fcnal-test")
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead, jump directly to success case stores in case ret >= 0, else do
the default 0 value store and jump over the success case. This is better
in terms of readability. Readjust the code for kfunc relocation as well
to follow a similar pattern, also leads to easier to follow code now.
Suggested-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20211122235733.634914-3-memxor@gmail.com
Add benchmark to measure the throughput and latency of the bpf_loop
call.
Testing this on my dev machine on 1 thread, the data is as follows:
nr_loops: 10
bpf_loop - throughput: 198.519 ± 0.155 M ops/s, latency: 5.037 ns/op
nr_loops: 100
bpf_loop - throughput: 247.448 ± 0.305 M ops/s, latency: 4.041 ns/op
nr_loops: 500
bpf_loop - throughput: 260.839 ± 0.380 M ops/s, latency: 3.834 ns/op
nr_loops: 1000
bpf_loop - throughput: 262.806 ± 0.629 M ops/s, latency: 3.805 ns/op
nr_loops: 5000
bpf_loop - throughput: 264.211 ± 1.508 M ops/s, latency: 3.785 ns/op
nr_loops: 10000
bpf_loop - throughput: 265.366 ± 3.054 M ops/s, latency: 3.768 ns/op
nr_loops: 50000
bpf_loop - throughput: 235.986 ± 20.205 M ops/s, latency: 4.238 ns/op
nr_loops: 100000
bpf_loop - throughput: 264.482 ± 0.279 M ops/s, latency: 3.781 ns/op
nr_loops: 500000
bpf_loop - throughput: 309.773 ± 87.713 M ops/s, latency: 3.228 ns/op
nr_loops: 1000000
bpf_loop - throughput: 262.818 ± 4.143 M ops/s, latency: 3.805 ns/op
>From this data, we can see that the latency per loop decreases as the
number of loops increases. On this particular machine, each loop had an
overhead of about ~4 ns, and we were able to run ~250 million loops
per second.
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/20211130030622.4131246-5-joannekoong@fb.com
This patch tests bpf_loop in pyperf and strobemeta, and measures the
verifier performance of replacing the traditional for loop
with bpf_loop.
The results are as follows:
~strobemeta~
Baseline
verification time 6808200 usec
stack depth 496
processed 554252 insns (limit 1000000) max_states_per_insn 16
total_states 15878 peak_states 13489 mark_read 3110
#192 verif_scale_strobemeta:OK (unrolled loop)
Using bpf_loop
verification time 31589 usec
stack depth 96+400
processed 1513 insns (limit 1000000) max_states_per_insn 2
total_states 106 peak_states 106 mark_read 60
#193 verif_scale_strobemeta_bpf_loop:OK
~pyperf600~
Baseline
verification time 29702486 usec
stack depth 368
processed 626838 insns (limit 1000000) max_states_per_insn 7
total_states 30368 peak_states 30279 mark_read 748
#182 verif_scale_pyperf600:OK (unrolled loop)
Using bpf_loop
verification time 148488 usec
stack depth 320+40
processed 10518 insns (limit 1000000) max_states_per_insn 10
total_states 705 peak_states 517 mark_read 38
#183 verif_scale_pyperf600_bpf_loop:OK
Using the bpf_loop helper led to approximately a 99% decrease
in the verification time and in the number of instructions.
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/20211130030622.4131246-4-joannekoong@fb.com
This patch adds the kernel-side and API changes for a new helper
function, bpf_loop:
long bpf_loop(u32 nr_loops, void *callback_fn, void *callback_ctx,
u64 flags);
where long (*callback_fn)(u32 index, void *ctx);
bpf_loop invokes the "callback_fn" **nr_loops** times or until the
callback_fn returns 1. The callback_fn can only return 0 or 1, and
this is enforced by the verifier. The callback_fn index is zero-indexed.
A few things to please note:
~ The "u64 flags" parameter is currently unused but is included in
case a future use case for it arises.
~ In the kernel-side implementation of bpf_loop (kernel/bpf/bpf_iter.c),
bpf_callback_t is used as the callback function cast.
~ A program can have nested bpf_loop calls but the program must
still adhere to the verifier constraint of its stack depth (the stack depth
cannot exceed MAX_BPF_STACK))
~ Recursive callback_fns do not pass the verifier, due to the call stack
for these being too deep.
~ The next patch will include the tests and benchmark
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/20211130030622.4131246-2-joannekoong@fb.com
Pull kvm fixes from Paolo Bonzini:
"ARM64:
- Fix constant sign extension affecting TCR_EL2 and preventing
running on ARMv8.7 models due to spurious bits being set
- Fix use of helpers using PSTATE early on exit by always sampling it
as soon as the exit takes place
- Move pkvm's 32bit handling into a common helper
RISC-V:
- Fix incorrect KVM_MAX_VCPUS value
- Unmap stage2 mapping when deleting/moving a memslot
x86:
- Fix and downgrade BUG_ON due to uninitialized cache
- Many APICv and MOVE_ENC_CONTEXT_FROM fixes
- Correctly emulate TLB flushes around nested vmentry/vmexit and when
the nested hypervisor uses VPID
- Prevent modifications to CPUID after the VM has run
- Other smaller bugfixes
Generic:
- Memslot handling bugfixes"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (44 commits)
KVM: fix avic_set_running for preemptable kernels
KVM: VMX: clear vmx_x86_ops.sync_pir_to_irr if APICv is disabled
KVM: SEV: accept signals in sev_lock_two_vms
KVM: SEV: do not take kvm->lock when destroying
KVM: SEV: Prohibit migration of a VM that has mirrors
KVM: SEV: Do COPY_ENC_CONTEXT_FROM with both VMs locked
selftests: sev_migrate_tests: add tests for KVM_CAP_VM_COPY_ENC_CONTEXT_FROM
KVM: SEV: move mirror status to destination of KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM
KVM: SEV: initialize regions_list of a mirror VM
KVM: SEV: cleanup locking for KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM
KVM: SEV: do not use list_replace_init on an empty list
KVM: x86: Use a stable condition around all VT-d PI paths
KVM: x86: check PIR even for vCPUs with disabled APICv
KVM: VMX: prepare sync_pir_to_irr for running with APICv disabled
KVM: selftests: page_table_test: fix calculation of guest_test_phys_mem
KVM: x86/mmu: Handle "default" period when selectively waking kthread
KVM: MMU: shadow nested paging does not have PKU
KVM: x86/mmu: Remove spurious TLB flushes in TDP MMU zap collapsible path
KVM: x86/mmu: Use yield-safe TDP MMU root iter in MMU notifier unmapping
KVM: X86: Use vcpu->arch.walk_mmu for kvm_mmu_invlpg()
...
Commit 98e1385ef2 ("include/linux/radix-tree.h: replace kernel.h with
the necessary inclusions") broke the radix tree test suite in two
different ways; first by including math.h which didn't exist in the
tools directory, and second by removing an implicit include of
spinlock.h before lockdep.h. Fix both issues.
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently, we use hard code number to verify if we are in the
arp_interval timeslice. But some user may want to reduce/extend
the verify timeslice. With the similar team option 'missed_max'
the uers could change that number based on their own environment.
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
VMs that mirror an encryption context rely on the owner to keep the
ASID allocated. Performing a KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM
would cause a dangling ASID:
1. copy context from A to B (gets ref to A)
2. move context from A to L (moves ASID from A to L)
3. close L (releases ASID from L, B still references it)
The right way to do the handoff instead is to create a fresh mirror VM
on the destination first:
1. copy context from A to B (gets ref to A)
[later] 2. close B (releases ref to A)
3. move context from A to L (moves ASID from A to L)
4. copy context from L to M
So, catch the situation by adding a count of how many VMs are
mirroring this one's encryption context.
Fixes: 0b020f5af0 ("KVM: SEV: Add support for SEV-ES intra host migration")
Message-Id: <20211123005036.2954379-11-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
I am putting the tests in sev_migrate_tests because the failure conditions are
very similar and some of the setup code can be reused, too.
The tests cover both successful creation of a mirror VM, and error
conditions.
Cc: Peter Gonda <pgonda@google.com>
Cc: Sean Christopherson <seanjc@google.com>
Message-Id: <20211123005036.2954379-9-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
A kvm_page_table_test run with its default settings fails on VMX due to
memory region add failure:
> ==== Test Assertion Failure ====
> lib/kvm_util.c:952: ret == 0
> pid=10538 tid=10538 errno=17 - File exists
> 1 0x00000000004057d1: vm_userspace_mem_region_add at kvm_util.c:947
> 2 0x0000000000401ee9: pre_init_before_test at kvm_page_table_test.c:302
> 3 (inlined by) run_test at kvm_page_table_test.c:374
> 4 0x0000000000409754: for_each_guest_mode at guest_modes.c:53
> 5 0x0000000000401860: main at kvm_page_table_test.c:500
> 6 0x00007f82ae2d8554: ?? ??:0
> 7 0x0000000000401894: _start at ??:?
> KVM_SET_USER_MEMORY_REGION IOCTL failed,
> rc: -1 errno: 17
> slot: 1 flags: 0x0
> guest_phys_addr: 0xc0000000 size: 0x40000000
This is because the memory range that this test is trying to add
(0x0c0000000 - 0x100000000) conflicts with LAPIC mapping at 0x0fee00000.
Looking at the code it seems that guest_test_*phys*_mem variable gets
mistakenly overwritten with guest_test_*virt*_mem while trying to adjust
the former for alignment.
With the correct variable adjusted this test runs successfully.
Signed-off-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Message-Id: <52e487458c3172923549bbcf9dfccfbe6faea60b.1637940473.git.maciej.szmigiero@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Each peer's endpoint contains a dst_cache entry that takes a reference
to another netdev. When the containing namespace exits, we take down the
socket and prevent future sockets from being created (by setting
creating_net to NULL), which removes that potential reference on the
netns. However, it doesn't release references to the netns that a netdev
cached in dst_cache might be taking, so the netns still might fail to
exit. Since the socket is gimped anyway, we can simply clear all the
dst_caches (by way of clearing the endpoint src), which will release all
references.
However, the current dst_cache_reset function only releases those
references lazily. But it turns out that all of our usages of
wg_socket_clear_peer_endpoint_src are called from contexts that are not
exactly high-speed or bottle-necked. For example, when there's
connection difficulty, or when userspace is reconfiguring the interface.
And in particular for this patch, when the netns is exiting. So for
those cases, it makes more sense to call dst_release immediately. For
that, we add a small helper function to dst_cache.
This patch also adds a test to netns.sh from Hangbin Liu to ensure this
doesn't regress.
Tested-by: Hangbin Liu <liuhangbin@gmail.com>
Reported-by: Xiumei Mu <xmu@redhat.com>
Cc: Toke Høiland-Jørgensen <toke@redhat.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Fixes: 900575aa33 ("wireguard: device: avoid circular netns references")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
