On Thu, Aug 13, 2026 at 08:03:57PM +0200, Jiri Olsa wrote:
> On Tue, Aug 04, 2026 at 03:40:30PM -0700, Andrii Nakryiko wrote:
> > On Tue, Aug 4, 2026 at 8:48 AM Leon Hwang <[email protected]> wrote:
> > >
> > > In CI, the tracing_multi_bench_attach test could cost >10s, which should
> > > be avoided in daily CI running.
> > >
> > >  ./test_progs -t tracing_multi_bench_attach -v
> > >  WATCHDOG: test case tracing_multi_bench_attach executes for 10 seconds...
> > >  serial_test_tracing_multi_bench_attach: found 40163 functions
> > >  serial_test_tracing_multi_bench_attach: attached in   9.367s
> > >  serial_test_tracing_multi_bench_attach: detached in   1.798s
> > >  #546     tracing_multi_bench_attach:OK
> > >
> > > Since it is for benchmark attachment of tracing_multi link, move it to
> > > the generic bench framework.
> > >
> > >  ./bench tracing-multi-attach
> > >  Setting up benchmark 'tracing-multi-attach'...
> > >  tracing-multi-attach: found 40166 functions
> > >  tracing-multi-attach: attached in   9.278s
> > >  tracing-multi-attach: detached in   1.836s
> > >
> > > Assisted-by: Codex:gpt-5.6-sol
> > > Signed-off-by: Leon Hwang <[email protected]>
> > > ---
> > >  tools/testing/selftests/bpf/Makefile          |   2 +
> > >  tools/testing/selftests/bpf/bench.c           |   2 +
> > >  .../bpf/benchs/bench_tracing_multi_attach.c   | 177 ++++++++++++++++++
> > >  .../selftests/bpf/prog_tests/tracing_multi.c  | 125 -------------
> > >  4 files changed, 181 insertions(+), 125 deletions(-)
> > >  create mode 100644 
> > > tools/testing/selftests/bpf/benchs/bench_tracing_multi_attach.c
> > >
> > 
> > Jiri, are you ok with not exercising this logic regularly?
> > 
> > but also hm, is 10 seconds expected?
> 
> sorry for late reply, I overlooked this one
> 
> the 'commited result' is around 1.2s.. will check
> 
>       serial_test_tracing_multi_bench_attach: found 51186 functions
>       serial_test_tracing_multi_bench_attach: attached in   1.295s
>       serial_test_tracing_multi_bench_attach: detached in   0.243s
> 
> thanks,
> jirka

fyi found the possible culprint:
  
https://lore.kernel.org/bpf/[email protected]/T/#m8184343c667a96efae72ae55f7a8679d51f6eaf6

jirka

Reply via email to