| Age | Commit message (Collapse) | Author |
|---|
|
Calling generic selftests "make install" fails as rsync expects all
files from TEST_GEN_PROGS to be present. The binary is not generated
anymore (commit 3b09d27cc93d) so we can safely remove it from there
and also from gitignore.
Fixes: 3b09d27cc93d ("selftests/bpf: Move test_align under test_progs")
Signed-off-by: Veronika Kabatova <vkabatov@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Link: https://lore.kernel.org/bpf/20200819160710.1345956-1-vkabatov@redhat.com
|
|
99aacebecb75 ("selftests: do not use .ONESHELL") removed .ONESHELL, which
changes how Makefile "silences" multi-command target recipes. selftests/bpf's
Makefile relied (a somewhat unknowingly) on .ONESHELL behavior of silencing
all commands within the recipe if the first command contains @ symbol.
Removing .ONESHELL exposed this hack.
This patch fixes the issue by explicitly silencing each command with $(Q).
Also explicitly define fallback rule for building *.o from *.c, instead of
relying on non-silent inherited rule. This was causing a non-silent output for
bench.o object file.
Fixes: 92f7440ecc93 ("selftests/bpf: More succinct Makefile output")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200807033058.848677-1-andriin@fb.com
|
|
runqslower's Makefile is building/installing bpftool into
$(OUTPUT)/sbin/bpftool, which coincides with $(DEFAULT_BPFTOOL). In practice
this means that often when building selftests from scratch (after `make
clean`), selftests are racing with runqslower to simultaneously build bpftool
and one of the two processes fail due to file being busy. Prevent this race by
explicitly order-depending on $(BPFTOOL_DEFAULT).
Fixes: a2c9652f751e ("selftests: Refactor build to remove tools/lib/bpf from include path")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200805004757.2960750-1-andriin@fb.com
|
|
Adding resolve_btfids test under test_progs suite.
It's possible to use btf_ids.h header and its logic in
user space application, so we can add easy test for it.
The test defines BTF_ID_LIST and checks it gets properly
resolved.
For this reason the test_progs binary (and other binaries
that use TRUNNER* macros) is processed with resolve_btfids
tool, which resolves BTF IDs in .BTF_ids section. The BTF
data are taken from btf_data.o object rceated from
progs/btf_data.c.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200711215329.41165-10-jolsa@kernel.org
|
|
Similarly to bpftool Makefile, allow to specify custom location of vmlinux.h
to be used during the build. This allows simpler testing setups with
checked-in pre-generated vmlinux.h.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200630004759.521530-2-andriin@fb.com
|
|
When using make kselftest TARGETS=bpf, tools/bpf is built with
MAKEFLAGS=rR, which causes $(CXX) to be undefined, which in turn causes
the build to fail with
CXX test_cpp
/bin/sh: 2: g: not found
Fix by adding a default $(CXX) value, like tools/build/feature/Makefile
already does.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200602175649.2501580-3-iii@linux.ibm.com
|
|
Extend bench framework with ability to have benchmark-provided child argument
parser for custom benchmark-specific parameters. This makes bench generic code
modular and independent from any specific benchmark.
Also implement a set of benchmarks for new BPF ring buffer and existing perf
buffer. 4 benchmarks were implemented: 2 variations for each of BPF ringbuf
and perfbuf:,
- rb-libbpf utilizes stock libbpf ring_buffer manager for reading data;
- rb-custom implements custom ring buffer setup and reading code, to
eliminate overheads inherent in generic libbpf code due to callback
functions and the need to update consumer position after each consumed
record, instead of batching updates (due to pessimistic assumption that
user callback might take long time and thus could unnecessarily hold ring
buffer space for too long);
- pb-libbpf uses stock libbpf perf_buffer code with all the default
settings, though uses higher-performance raw event callback to minimize
unnecessary overhead;
- pb-custom implements its own custom consumer code to minimize any possible
overhead of generic libbpf implementation and indirect function calls.
All of the test support default, no data notification skipped, mode, as well
as sampled mode (with --rb-sampled flag), which allows to trigger epoll
notification less frequently and reduce overhead. As will be shown, this mode
is especially critical for perf buffer, which suffers from high overhead of
wakeups in kernel.
Otherwise, all benchamrks implement similar way to generate a batch of records
by using fentry/sys_getpgid BPF program, which pushes a bunch of records in
a tight loop and records number of successful and dropped samples. Each record
is a small 8-byte integer, to minimize the effect of memory copying with
bpf_perf_event_output() and bpf_ringbuf_output().
Benchmarks that have only one producer implement optional back-to-back mode,
in which record production and consumption is alternating on the same CPU.
This is the highest-throughput happy case, showing ultimate performance
achievable with either BPF ringbuf or perfbuf.
All the below scenarios are implemented in a script in
benchs/run_bench_ringbufs.sh. Tests were performed on 28-core/56-thread
Intel Xeon CPU E5-2680 v4 @ 2.40GHz CPU.
Single-producer, parallel producer
==================================
rb-libbpf 12.054 ± 0.320M/s (drops 0.000 ± 0.000M/s)
rb-custom 8.158 ± 0.118M/s (drops 0.001 ± 0.003M/s)
pb-libbpf 0.931 ± 0.007M/s (drops 0.000 ± 0.000M/s)
pb-custom 0.965 ± 0.003M/s (drops 0.000 ± 0.000M/s)
Single-producer, parallel producer, sampled notification
========================================================
rb-libbpf 11.563 ± 0.067M/s (drops 0.000 ± 0.000M/s)
rb-custom 15.895 ± 0.076M/s (drops 0.000 ± 0.000M/s)
pb-libbpf 9.889 ± 0.032M/s (drops 0.000 ± 0.000M/s)
pb-custom 9.866 ± 0.028M/s (drops 0.000 ± 0.000M/s)
Single producer on one CPU, consumer on another one, both running at full
speed. Curiously, rb-libbpf has higher throughput than objectively faster (due
to more lightweight consumer code path) rb-custom. It appears that faster
consumer causes kernel to send notifications more frequently, because consumer
appears to be caught up more frequently. Performance of perfbuf suffers from
default "no sampling" policy and huge overhead that causes.
In sampled mode, rb-custom is winning very significantly eliminating too
frequent in-kernel wakeups, the gain appears to be more than 2x.
Perf buffer achieves even more impressive wins, compared to stock perfbuf
settings, with 10x improvements in throughput with 1:500 sampling rate. The
trade-off is that with sampling, application might not get next X events until
X+1st arrives, which is not always acceptable. With steady influx of events,
though, this shouldn't be a problem.
Overall, single-producer performance of ring buffers seems to be better no
matter the sampled/non-sampled modes, but it especially beats ring buffer
without sampling due to its adaptive notification approach.
Single-producer, back-to-back mode
==================================
rb-libbpf 15.507 ± 0.247M/s (drops 0.000 ± 0.000M/s)
rb-libbpf-sampled 14.692 ± 0.195M/s (drops 0.000 ± 0.000M/s)
rb-custom 21.449 ± 0.157M/s (drops 0.000 ± 0.000M/s)
rb-custom-sampled 20.024 ± 0.386M/s (drops 0.000 ± 0.000M/s)
pb-libbpf 1.601 ± 0.015M/s (drops 0.000 ± 0.000M/s)
pb-libbpf-sampled 8.545 ± 0.064M/s (drops 0.000 ± 0.000M/s)
pb-custom 1.607 ± 0.022M/s (drops 0.000 ± 0.000M/s)
pb-custom-sampled 8.988 ± 0.144M/s (drops 0.000 ± 0.000M/s)
Here we test a back-to-back mode, which is arguably best-case scenario both
for BPF ringbuf and perfbuf, because there is no contention and for ringbuf
also no excessive notification, because consumer appears to be behind after
the first record. For ringbuf, custom consumer code clearly wins with 21.5 vs
16 million records per second exchanged between producer and consumer. Sampled
mode actually hurts a bit due to slightly slower producer logic (it needs to
fetch amount of data available to decide whether to skip or force notification).
Perfbuf with wakeup sampling gets 5.5x throughput increase, compared to
no-sampling version. There also doesn't seem to be noticeable overhead from
generic libbpf handling code.
Perfbuf back-to-back, effect of sample rate
===========================================
pb-sampled-1 1.035 ± 0.012M/s (drops 0.000 ± 0.000M/s)
pb-sampled-5 3.476 ± 0.087M/s (drops 0.000 ± 0.000M/s)
pb-sampled-10 5.094 ± 0.136M/s (drops 0.000 ± 0.000M/s)
pb-sampled-25 7.118 ± 0.153M/s (drops 0.000 ± 0.000M/s)
pb-sampled-50 8.169 ± 0.156M/s (drops 0.000 ± 0.000M/s)
pb-sampled-100 8.887 ± 0.136M/s (drops 0.000 ± 0.000M/s)
pb-sampled-250 9.180 ± 0.209M/s (drops 0.000 ± 0.000M/s)
pb-sampled-500 9.353 ± 0.281M/s (drops 0.000 ± 0.000M/s)
pb-sampled-1000 9.411 ± 0.217M/s (drops 0.000 ± 0.000M/s)
pb-sampled-2000 9.464 ± 0.167M/s (drops 0.000 ± 0.000M/s)
pb-sampled-3000 9.575 ± 0.273M/s (drops 0.000 ± 0.000M/s)
This benchmark shows the effect of event sampling for perfbuf. Back-to-back
mode for highest throughput. Just doing every 5th record notification gives
3.5x speed up. 250-500 appears to be the point of diminishing return, with
almost 9x speed up. Most benchmarks use 500 as the default sampling for pb-raw
and pb-custom.
Ringbuf back-to-back, effect of sample rate
===========================================
rb-sampled-1 1.106 ± 0.010M/s (drops 0.000 ± 0.000M/s)
rb-sampled-5 4.746 ± 0.149M/s (drops 0.000 ± 0.000M/s)
rb-sampled-10 7.706 ± 0.164M/s (drops 0.000 ± 0.000M/s)
rb-sampled-25 12.893 ± 0.273M/s (drops 0.000 ± 0.000M/s)
rb-sampled-50 15.961 ± 0.361M/s (drops 0.000 ± 0.000M/s)
rb-sampled-100 18.203 ± 0.445M/s (drops 0.000 ± 0.000M/s)
rb-sampled-250 19.962 ± 0.786M/s (drops 0.000 ± 0.000M/s)
rb-sampled-500 20.881 ± 0.551M/s (drops 0.000 ± 0.000M/s)
rb-sampled-1000 21.317 ± 0.532M/s (drops 0.000 ± 0.000M/s)
rb-sampled-2000 21.331 ± 0.535M/s (drops 0.000 ± 0.000M/s)
rb-sampled-3000 21.688 ± 0.392M/s (drops 0.000 ± 0.000M/s)
Similar benchmark for ring buffer also shows a great advantage (in terms of
throughput) of skipping notifications. Skipping every 5th one gives 4x boost.
Also similar to perfbuf case, 250-500 seems to be the point of diminishing
returns, giving roughly 20x better results.
Keep in mind, for this test, notifications are controlled manually with
BPF_RB_NO_WAKEUP and BPF_RB_FORCE_WAKEUP. As can be seen from previous
benchmarks, adaptive notifications based on consumer's positions provides same
(or even slightly better due to simpler load generator on BPF side) benefits in
favorable back-to-back scenario. Over zealous and fast consumer, which is
almost always caught up, will make thoughput numbers smaller. That's the case
when manual notification control might prove to be extremely beneficial.
Ringbuf back-to-back, reserve+commit vs output
==============================================
reserve 22.819 ± 0.503M/s (drops 0.000 ± 0.000M/s)
output 18.906 ± 0.433M/s (drops 0.000 ± 0.000M/s)
Ringbuf sampled, reserve+commit vs output
=========================================
reserve-sampled 15.350 ± 0.132M/s (drops 0.000 ± 0.000M/s)
output-sampled 14.195 ± 0.144M/s (drops 0.000 ± 0.000M/s)
BPF ringbuf supports two sets of APIs with various usability and performance
tradeoffs: bpf_ringbuf_reserve()+bpf_ringbuf_commit() vs bpf_ringbuf_output().
This benchmark clearly shows superiority of reserve+commit approach, despite
using a small 8-byte record size.
Single-producer, consumer/producer competing on the same CPU, low batch count
=============================================================================
rb-libbpf 3.045 ± 0.020M/s (drops 3.536 ± 0.148M/s)
rb-custom 3.055 ± 0.022M/s (drops 3.893 ± 0.066M/s)
pb-libbpf 1.393 ± 0.024M/s (drops 0.000 ± 0.000M/s)
pb-custom 1.407 ± 0.016M/s (drops 0.000 ± 0.000M/s)
This benchmark shows one of the worst-case scenarios, in which producer and
consumer do not coordinate *and* fight for the same CPU. No batch count and
sampling settings were able to eliminate drops for ringbuffer, producer is
just too fast for consumer to keep up. But ringbuf and perfbuf still able to
pass through quite a lot of messages, which is more than enough for a lot of
applications.
Ringbuf, multi-producer contention
==================================
rb-libbpf nr_prod 1 10.916 ± 0.399M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 2 4.931 ± 0.030M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 3 4.880 ± 0.006M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 4 3.926 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 8 4.011 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 12 3.967 ± 0.016M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 16 2.604 ± 0.030M/s (drops 0.001 ± 0.002M/s)
rb-libbpf nr_prod 20 2.233 ± 0.003M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 24 2.085 ± 0.015M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 28 2.055 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 32 1.962 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 36 2.089 ± 0.005M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 40 2.118 ± 0.006M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 44 2.105 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 48 2.120 ± 0.058M/s (drops 0.000 ± 0.001M/s)
rb-libbpf nr_prod 52 2.074 ± 0.024M/s (drops 0.007 ± 0.014M/s)
Ringbuf uses a very short-duration spinlock during reservation phase, to check
few invariants, increment producer count and set record header. This is the
biggest point of contention for ringbuf implementation. This benchmark
evaluates the effect of multiple competing writers on overall throughput of
a single shared ringbuffer.
Overall throughput drops almost 2x when going from single to two
highly-contended producers, gradually dropping with additional competing
producers. Performance drop stabilizes at around 20 producers and hovers
around 2mln even with 50+ fighting producers, which is a 5x drop compared to
non-contended case. Good kernel implementation in kernel helps maintain decent
performance here.
Note, that in the intended real-world scenarios, it's not expected to get even
close to such a high levels of contention. But if contention will become
a problem, there is always an option of sharding few ring buffers across a set
of CPUs.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200529075424.3139988-5-andriin@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
It is sometimes desirable to be able to trigger BPF program from user-space
with minimal overhead. sys_enter would seem to be a good candidate, yet in
a lot of cases there will be a lot of noise from syscalls triggered by other
processes on the system. So while searching for low-overhead alternative, I've
stumbled upon getpgid() syscall, which seems to be specific enough to not
suffer from accidental syscall by other apps.
This set of benchmarks compares tp, raw_tp w/ filtering by syscall ID, kprobe,
fentry and fmod_ret with returning error (so that syscall would not be
executed), to determine the lowest-overhead way. Here are results on my
machine (using benchs/run_bench_trigger.sh script):
base : 9.200 ± 0.319M/s
tp : 6.690 ± 0.125M/s
rawtp : 8.571 ± 0.214M/s
kprobe : 6.431 ± 0.048M/s
fentry : 8.955 ± 0.241M/s
fmodret : 8.903 ± 0.135M/s
So it seems like fmodret doesn't give much benefit for such lightweight
syscall. Raw tracepoint is pretty decent despite additional filtering logic,
but it will be called for any other syscall in the system, which rules it out.
Fentry, though, seems to be adding the least amoung of overhead and achieves
97.3% of performance of baseline no-BPF-attached syscall.
Using getpgid() seems to be preferable to set_task_comm() approach from
test_overhead, as it's about 2.35x faster in a baseline performance.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200512192445.2351848-5-andriin@fb.com
|
|
Add fmod_ret BPF program to existing test_overhead selftest. Also re-implement
user-space benchmarking part into benchmark runner to compare results. Results
with ./bench are consistently somewhat lower than test_overhead's, but relative
performance of various types of BPF programs stay consisten (e.g., kretprobe is
noticeably slower). This slowdown seems to be coming from the fact that
test_overhead is single-threaded, while benchmark always spins off at least
one thread for producer. This has been confirmed by hacking multi-threaded
test_overhead variant and also single-threaded bench variant. Resutls are
below. run_bench_rename.sh script from benchs/ subdirectory was used to
produce results for ./bench.
Single-threaded implementations
===============================
/* bench: single-threaded, atomics */
base : 4.622 ± 0.049M/s
kprobe : 3.673 ± 0.052M/s
kretprobe : 2.625 ± 0.052M/s
rawtp : 4.369 ± 0.089M/s
fentry : 4.201 ± 0.558M/s
fexit : 4.309 ± 0.148M/s
fmodret : 4.314 ± 0.203M/s
/* selftest: single-threaded, no atomics */
task_rename base 4555K events per sec
task_rename kprobe 3643K events per sec
task_rename kretprobe 2506K events per sec
task_rename raw_tp 4303K events per sec
task_rename fentry 4307K events per sec
task_rename fexit 4010K events per sec
task_rename fmod_ret 3984K events per sec
Multi-threaded implementations
==============================
/* bench: multi-threaded w/ atomics */
base : 3.910 ± 0.023M/s
kprobe : 3.048 ± 0.037M/s
kretprobe : 2.300 ± 0.015M/s
rawtp : 3.687 ± 0.034M/s
fentry : 3.740 ± 0.087M/s
fexit : 3.510 ± 0.009M/s
fmodret : 3.485 ± 0.050M/s
/* selftest: multi-threaded w/ atomics */
task_rename base 3872K events per sec
task_rename kprobe 3068K events per sec
task_rename kretprobe 2350K events per sec
task_rename raw_tp 3731K events per sec
task_rename fentry 3639K events per sec
task_rename fexit 3558K events per sec
task_rename fmod_ret 3511K events per sec
/* selftest: multi-threaded, no atomics */
task_rename base 3945K events per sec
task_rename kprobe 3298K events per sec
task_rename kretprobe 2451K events per sec
task_rename raw_tp 3718K events per sec
task_rename fentry 3782K events per sec
task_rename fexit 3543K events per sec
task_rename fmod_ret 3526K events per sec
Note that the fact that ./bench benchmark always uses atomic increments for
counting, while test_overhead doesn't, doesn't influence test results all that
much.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200512192445.2351848-4-andriin@fb.com
|
|
While working on BPF ringbuf implementation, testing, and benchmarking, I've
developed a pretty generic and modular benchmark runner, which seems to be
generically useful, as I've already used it for one more purpose (testing
fastest way to trigger BPF program, to minimize overhead of in-kernel code).
This patch adds generic part of benchmark runner and sets up Makefile for
extending it with more sets of benchmarks.
Benchmarker itself operates by spinning up specified number of producer and
consumer threads, setting up interval timer sending SIGALARM signal to
application once a second. Every second, current snapshot with hits/drops
counters are collected and stored in an array. Drops are useful for
producer/consumer benchmarks in which producer might overwhelm consumers.
Once test finishes after given amount of warm-up and testing seconds, mean and
stddev are calculated (ignoring warm-up results) and is printed out to stdout.
This setup seems to give consistent and accurate results.
To validate behavior, I added two atomic counting tests: global and local.
For global one, all the producer threads are atomically incrementing same
counter as fast as possible. This, of course, leads to huge drop of
performance once there is more than one producer thread due to CPUs fighting
for the same memory location.
Local counting, on the other hand, maintains one counter per each producer
thread, incremented independently. Once per second, all counters are read and
added together to form final "counting throughput" measurement. As expected,
such setup demonstrates linear scalability with number of producers (as long
as there are enough physical CPU cores, of course). See example output below.
Also, this setup can nicely demonstrate disastrous effects of false sharing,
if care is not taken to take those per-producer counters apart into
independent cache lines.
Demo output shows global counter first with 1 producer, then with 4. Both
total and per-producer performance significantly drop. The last run is local
counter with 4 producers, demonstrating near-perfect scalability.
$ ./bench -a -w1 -d2 -p1 count-global
Setting up benchmark 'count-global'...
Benchmark 'count-global' started.
Iter 0 ( 24.822us): hits 148.179M/s (148.179M/prod), drops 0.000M/s
Iter 1 ( 37.939us): hits 149.308M/s (149.308M/prod), drops 0.000M/s
Iter 2 (-10.774us): hits 150.717M/s (150.717M/prod), drops 0.000M/s
Iter 3 ( 3.807us): hits 151.435M/s (151.435M/prod), drops 0.000M/s
Summary: hits 150.488 ± 1.079M/s (150.488M/prod), drops 0.000 ± 0.000M/s
$ ./bench -a -w1 -d2 -p4 count-global
Setting up benchmark 'count-global'...
Benchmark 'count-global' started.
Iter 0 ( 60.659us): hits 53.910M/s ( 13.477M/prod), drops 0.000M/s
Iter 1 (-17.658us): hits 53.722M/s ( 13.431M/prod), drops 0.000M/s
Iter 2 ( 5.865us): hits 53.495M/s ( 13.374M/prod), drops 0.000M/s
Iter 3 ( 0.104us): hits 53.606M/s ( 13.402M/prod), drops 0.000M/s
Summary: hits 53.608 ± 0.113M/s ( 13.402M/prod), drops 0.000 ± 0.000M/s
$ ./bench -a -w1 -d2 -p4 count-local
Setting up benchmark 'count-local'...
Benchmark 'count-local' started.
Iter 0 ( 23.388us): hits 640.450M/s (160.113M/prod), drops 0.000M/s
Iter 1 ( 2.291us): hits 605.661M/s (151.415M/prod), drops 0.000M/s
Iter 2 ( -6.415us): hits 607.092M/s (151.773M/prod), drops 0.000M/s
Iter 3 ( -1.361us): hits 601.796M/s (150.449M/prod), drops 0.000M/s
Summary: hits 604.849 ± 2.739M/s (151.212M/prod), drops 0.000 ± 0.000M/s
Benchmark runner supports setting thread affinity for producer and consumer
threads. You can use -a flag for default CPU selection scheme, where first
consumer gets CPU #0, next one gets CPU #1, and so on. Then producer threads
pick up next CPU and increment one-by-one as well. But user can also specify
a set of CPUs independently for producers and consumers with --prod-affinity
1,2-10,15 and --cons-affinity <set-of-cpus>. The latter allows to force
producers and consumers to share same set of CPUs, if necessary.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200512192445.2351848-3-andriin@fb.com
|
|
Add testing_helpers.c, which will contain generic helpers for test runners and
tests needing some common generic functionality, like parsing a set of
numbers.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200512192445.2351848-2-andriin@fb.com
|
|
Before commit 74b5a5968fe8 ("selftests/bpf: Replace test_progs and
test_maps w/ general rule") selftests/bpf used generic install
target from selftests/lib.mk to install generated bpf test progs
by mentioning them in TEST_GEN_FILES variable.
Take that functionality back.
Fixes: 74b5a5968fe8 ("selftests/bpf: Replace test_progs and test_maps w/ general rule")
Signed-off-by: Yauheni Kaliuta <yauheni.kaliuta@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200513021722.7787-1-yauheni.kaliuta@redhat.com
|
|
Move the following routines that let us start a background listener
thread and connect to a server by fd to the test_prog:
* start_server - socket+bind+listen
* connect_to_fd - connect to the server identified by fd
These will be used in the next commit.
Also, extend these helpers to support AF_INET6 and accept the family
as an argument.
v5:
* drop pthread.h (Martin KaFai Lau)
* add SO_SNDTIMEO (Martin KaFai Lau)
v4:
* export extra helper to start server without a thread (Martin KaFai Lau)
* tcp_rtt is no longer starting background thread (Martin KaFai Lau)
v2:
* put helpers into network_helpers.c (Andrii Nakryiko)
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrey Ignatov <rdna@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200508174611.228805-2-sdf@google.com
|
|
Update bpf_sk_assign test to fetch the server socket from SOCKMAP, now that
map lookup from BPF in SOCKMAP is enabled. This way the test TC BPF program
doesn't need to know what address server socket is bound to.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200429181154.479310-4-jakub@cloudflare.com
|
|
Fold stand-alone test_hashmap test into test_progs.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200429012111.277390-4-andriin@fb.com
|
|
Add ability to specify extra compiler flags with SAN_CFLAGS for compilation of
all user-space C files. This allows to build all of selftest programs with,
e.g., custom sanitizer flags, without requiring support for such sanitizers
from anyone compiling selftest/bpf.
As an example, to compile everything with AddressSanitizer, one would do:
$ make clean && make SAN_CFLAGS="-fsanitize=address"
For AddressSanitizer to work, one needs appropriate libasan shared library
installed in the system, with version of libasan matching what GCC links
against. E.g., GCC8 needs libasan5, while GCC7 uses libasan4.
For CentOS 7, to build everything successfully one would need to:
$ sudo yum install devtoolset-8-gcc devtoolset-libasan-devel
$ scl enable devtoolset-8 bash # set up environment
For Arch Linux to run selftests, one would need to install gcc-libs package to
get libasan.so.5:
$ sudo pacman -S gcc-libs
N.B. EXTRA_CFLAGS name wasn't used, because it's also used by libbpf's
Makefile and this causes few issues:
1. default "-g -Wall" flags are overriden;
2. compiling shared library with AddressSanitizer generates a bunch of symbols
like: "_GLOBAL__sub_D_00099_0_btf_dump.c", "_GLOBAL__sub_D_00099_0_bpf.c",
etc, which screws up versioned symbols check.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Julia Kartseva <hex@fb.com>
Link: https://lore.kernel.org/bpf/20200429012111.277390-3-andriin@fb.com
|
|
Ensure that test runner flavors include their own skeletons from <flavor>/
directory. Previously, skeletons generated for no-flavor test_progs were used.
Apart from fixing correctness, this also makes it possible to compile only
flavors individually:
$ make clean && make test_progs-no_alu32
... now succeeds ...
Fixes: 74b5a5968fe8 ("selftests/bpf: Replace test_progs and test_maps w/ general rule")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200429012111.277390-2-andriin@fb.com
|
|
$(OUTPUT)/runqslower makefile target doesn't actually create runqslower
binary in the $(OUTPUT) directory. As lib.mk expects all
TEST_GEN_PROGS_EXTENDED (which runqslower is a part of) to be present in
the OUTPUT directory, this results in an error when running e.g. `make
install`:
rsync: link_stat "tools/testing/selftests/bpf/runqslower" failed: No
such file or directory (2)
Copy the binary into the OUTPUT directory after building it to fix the
error.
Fixes: 3a0d3092a4ed ("selftests/bpf: Build runqslower from selftests")
Signed-off-by: Veronika Kabatova <vkabatov@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200428173742.2988395-1-vkabatov@redhat.com
|
|
Add vmlinux.h generation to selftest/bpf's Makefile. Use it from newly added
test_vmlinux to trace nanosleep syscall using 5 different types of programs:
- tracepoint;
- raw tracepoint;
- raw tracepoint w/ direct memory reads (tp_btf);
- kprobe;
- fentry.
These programs are realistic variants of real-life tracing programs,
excercising vmlinux.h's usage with tracing applications.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20200313172336.1879637-5-andriin@fb.com
|
|
Commit fe4eb069edb7 ("bpftool: Use linux/types.h from source tree for
profiler build") added a build dependency on tools/testing/selftests/bpf
to tools/bpf/bpftool. This is suboptimal with respect to a possible
stand-alone build of bpftool.
Fix this by moving tools/testing/selftests/bpf/include/uapi/linux/types.h
to tools/include/uapi/linux/types.h.
This requires an adjustment in the include search path order for the
tests in tools/testing/selftests/bpf so that tools/include/linux/types.h
is selected when building host binaries and
tools/include/uapi/linux/types.h is selected when building bpf binaries.
Verified by compiling bpftool and the bpf selftests on x86_64 with this
change.
Fixes: fe4eb069edb7 ("bpftool: Use linux/types.h from source tree for profiler build")
Suggested-by: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200313113105.6918-1-tklauser@distanz.ch
|
|
bpf_get_ns_current_pid_tgid.
Self tests added for new helper bpf_get_ns_current_pid_tgid
Signed-off-by: Carlos Neira <cneirabustos@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200304204157.58695-4-cneirabustos@gmail.com
|
|
Add detection of out-of-tree built vmlinux image for the purpose of
VMLINUX_BTF detection. According to Documentation/kbuild/kbuild.rst, O takes
precedence over KBUILD_OUTPUT.
Also ensure ~/path/to/build/dir also works by relying on wildcard's resolution
first, but then applying $(abspath) at the end to also handle
O=../../whatever cases.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200304184336.165766-1-andriin@fb.com
|
|
Add Python module with tests for "bpftool feature" command, which mainly
checks whether the "full" option is working properly.
Signed-off-by: Michal Rostecki <mrostecki@opensuse.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/20200226165941.6379-6-mrostecki@opensuse.org
|
|
Due to various bugs in tests clean up code (usually), if host system is
misconfigured, it happens that test_progs will just crash in the middle of
running a test with little to no indication of where and why the crash
happened. For cases where coredump is not readily available (e.g., inside
a CI), it's very helpful to have a stack trace, which lead to crash, to be
printed out. This change adds a signal handler that will capture and print out
symbolized backtrace:
$ sudo ./test_progs -t mmap
test_mmap:PASS:skel_open_and_load 0 nsec
test_mmap:PASS:bss_mmap 0 nsec
test_mmap:PASS:data_mmap 0 nsec
Caught signal #11!
Stack trace:
./test_progs(crash_handler+0x18)[0x42a888]
/lib64/libpthread.so.0(+0xf5d0)[0x7f2aab5175d0]
./test_progs(test_mmap+0x3c0)[0x41f0a0]
./test_progs(main+0x160)[0x407d10]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f2aab15d3d5]
./test_progs[0x407ebc]
[1] 1988412 segmentation fault (core dumped) sudo ./test_progs -t mmap
Unfortunately, glibc's symbolization support is unable to symbolize static
functions, only global ones will be present in stack trace. But it's still a
step forward without adding extra libraries to get a better symbolization.
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/20200225000847.3965188-1-andriin@fb.com
|
|
The latest llvm supports cpu version v3, which is cpu version v1
plus some additional 64bit jmp insns and 32bit jmp insn support.
In selftests/bpf Makefile, the llvm flag -mcpu=probe did runtime
probe into the host system. Depending on compilation environments,
it is possible that runtime probe may fail, e.g., due to
memlock issue. This will cause generated code with cpu version v1.
This may cause confusion as the same compiler and the same C code
generates different byte codes in different environment.
Let us change the llvm flag -mcpu=probe to -mcpu=v3 so the
generated code will be the same regardless of the compilation
environment.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200219004236.2291125-1-yhs@fb.com
|
|
Detect when bpftool source code changes and trigger rebuild within
selftests/bpf Makefile. Also fix few small formatting problems.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20200124054148.2455060-1-andriin@fb.com
|
|
During cross-compilation, it was discovered that LDFLAGS and
LDLIBS were not being used while building binaries, leading
to defaults which were not necessarily correct.
OpenEmbedded reported this kind of problem:
ERROR: QA Issue: No GNU_HASH in the ELF binary [...], didn't pass LDFLAGS?
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
|
|
To make sure no new files are introduced that doesn't include the bpf/
prefix in its #include, remove tools/lib/bpf from the include path
entirely.
Instead, we introduce a new header files directory under the scratch tools/
dir, and add a rule to run the 'install_headers' rule from libbpf to have a
full set of consistent libbpf headers in $(OUTPUT)/tools/include/bpf, and
then use $(OUTPUT)/tools/include as the include path for selftests.
For consistency we also make sure we put all the scratch build files from
other bpftool and libbpf into tools/build/, so everything stays within
selftests/.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/157952561246.1683545.2762245552022369203.stgit@toke.dk
|
|
Fix all selftests to include libbpf header files with the bpf/ prefix, to
be consistent with external users of the library. Also ensure that all
includes of exported libbpf header files (those that are exported on 'make
install' of the library) use bracketed includes instead of quoted.
To not break the build, keep the old include path until everything has been
changed to the new one; a subsequent patch will remove that.
Fixes: 6910d7d3867a ("selftests/bpf: Ensure bpf_helper_defs.h are taken from selftests dir")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/157952560568.1683545.9649335788846513446.stgit@toke.dk
|
|
Add a VMLINUX_BTF variable with the locally-built path when calling the
runqslower Makefile from selftests. This makes sure a simple 'make'
invocation in the selftests dir works even when there is no BTF information
for the running kernel. Do a wildcard expansion and include the same paths
for BTF for the running kernel as in the runqslower Makefile, to make it
possible to build selftests without having a vmlinux in the local tree.
Also fix the make invocation to use $(OUTPUT)/tools as the destination
directory instead of $(CURDIR)/tools.
Fixes: 3a0d3092a4ed ("selftests/bpf: Build runqslower from selftests")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/157952560344.1683545.2723631988771664417.stgit@toke.dk
|
|
Ensure runqslower tool is built as part of selftests to prevent it from bit
rotting.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200113073143.1779940-7-andriin@fb.com
|
|
Bring selftest/bpf's Makefile output to the same format used by libbpf and
bpftool: 2 spaces of padding on the left + 8-character left-aligned build step
identifier.
Also, hide feature detection output by default. Can be enabled back by setting
V=1.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200113073143.1779940-4-andriin@fb.com
|
|
Further clean up Makefile output:
- hide "entering directory" messages;
- silvence sub-Make command echoing;
- succinct MKDIR messages.
Also remove few test binaries that are not produced anymore from .gitignore.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200110051716.1591485-4-andriin@fb.com
|
|
Reorder includes search path to ensure $(OUTPUT) and $(CURDIR) go before
libbpf's directory. Also fix bpf_helpers.h to include bpf_helper_defs.h in
such a way as to leverage includes search path. This allows selftests to not
use libbpf's local and potentially stale bpf_helper_defs.h. It's important
because selftests/bpf's Makefile only re-generates bpf_helper_defs.h in
seltests' output directory, not the one in libbpf's directory.
Also force regeneration of bpf_helper_defs.h when libbpf.a is updated to
reduce staleness.
Fixes: fa633a0f8919 ("libbpf: Fix build on read-only filesystems")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200110051716.1591485-3-andriin@fb.com
|
|
Libbpf's clean target should clean out generated files in $(OUTPUT) directory
and not make assumption that $(OUTPUT) directory is current working directory.
Selftest's Makefile should delegate cleaning of libbpf-generated files to
libbpf's Makefile. This ensures more robust clean up.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200110051716.1591485-2-andriin@fb.com
|
|
Simple overlapping changes in bpf land wrt. bpf_helper_defs.h
handling.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When auto-generated BPF skeleton C code is included from C++ application, it
triggers compilation error due to void * being implicitly casted to whatever
target pointer type. This is supported by C, but not C++. To solve this
problem, add explicit casts, where necessary.
To ensure issues like this are captured going forward, add skeleton usage in
test_cpp test.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191226210253.3132060-1-andriin@fb.com
|
|
I got the following error when I tried to build perf on a read-only
filesystem with O=dir option.
$ cd /some/where/ro/linux/tools/perf
$ make O=$HOME/build/perf
...
CC /home/namhyung/build/perf/lib.o
/bin/sh: bpf_helper_defs.h: Read-only file system
make[3]: *** [Makefile:184: bpf_helper_defs.h] Error 1
make[2]: *** [Makefile.perf:778: /home/namhyung/build/perf/libbpf.a] Error 2
make[2]: *** Waiting for unfinished jobs....
LD /home/namhyung/build/perf/libperf-in.o
AR /home/namhyung/build/perf/libperf.a
PERF_VERSION = 5.4.0
make[1]: *** [Makefile.perf:225: sub-make] Error 2
make: *** [Makefile:70: all] Error 2
It was becaused bpf_helper_defs.h was generated in current directory.
Move it to OUTPUT directory.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20191223061326.843366-1-namhyung@kernel.org
|
|
Convert test_cgroup_attach to prog_tests.
This change does a lot of things but in many cases it's pretty expensive
to separate them, so they go in one commit. Nevertheless the logic is
ketp as is and changes made are just moving things around, simplifying
them (w/o changing the meaning of the tests) and making prog_tests
compatible:
* split the 3 tests in the file into 3 separate files in prog_tests/;
* rename the test functions to test_<file_base_name>;
* remove unused includes, constants, variables and functions from every
test;
* replace `if`-s with or `if (CHECK())` where additional context should
be logged and with `if (CHECK_FAIL())` where line number is enough;
* switch from `log_err()` to logging via `CHECK()`;
* replace `assert`-s with `CHECK_FAIL()` to avoid crashing the whole
test_progs if one assertion fails;
* replace cgroup_helpers with test__join_cgroup() in
cgroup_attach_override only, other tests need more fine-grained
control for cgroup creation/deletion so cgroup_helpers are still used
there;
* simplify cgroup_attach_autodetach by switching to easiest possible
program since this test doesn't really need such a complicated program
as cgroup_attach_multi does;
* remove test_cgroup_attach.c itself.
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/0ff19cc64d2dc5cf404349f07131119480e10e32.1576741281.git.rdna@fb.com
|
|
Similarly to bpftool/libbpf output, make selftests/bpf output succinct
per-item output line. Output is roughly as follows:
$ make
...
CLANG-LLC [test_maps] pyperf600.o
CLANG-LLC [test_maps] strobemeta.o
CLANG-LLC [test_maps] pyperf100.o
EXTRA-OBJ [test_progs] cgroup_helpers.o
EXTRA-OBJ [test_progs] trace_helpers.o
BINARY test_align
BINARY test_verifier_log
GEN-SKEL [test_progs] fexit_bpf2bpf.skel.h
GEN-SKEL [test_progs] test_global_data.skel.h
GEN-SKEL [test_progs] sendmsg6_prog.skel.h
...
To see the actual command invocation, verbose mode can be turned on with V=1
argument:
$ make V=1
... very verbose output ...
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20191217061425.2346359-1-andriin@fb.com
|
