[PATCH -next v3 0/2] perf stat: Support inherit events for bperf

[Tengda Wu]

Hi,

Here is the 3th version of the series to support inherit events for bperf.
This version add pid or tgid selection based on filter type in new_task
prog to avoid memory waste and potential count loss.


bperf (perf-stat --bpf-counter) has not supported inherit events
during fork() since it was first introduced.

This patch series tries to add this support by:
 1) adding two new bpf programs to monitor task lifecycle;
 2) recording new tasks in the filter map dynamically;
 3) reusing `accum_key` of parent task for new tasks.

Thanks,
Tengda


Changelog:
---------
v3: (Address comments from Namhyung, thanks)
 * Use pid or tgid based on filter type in new_task prog
 * Add comments to explain pid usage for TGID type in exit_task prog

v2: https://lore.kernel.org/all/20240905115918.772234-1-wutengda@xxxxxxxxxxxxxxx/
 * Remove the unused init_filter_entries in follower bpf, declare
   a global filter_entry_count in bpf_counter instead
 * Attach on_newtask and on_exittask progs only if the filter type
   is either PID or TGID

v1: https://lore.kernel.org/all/20240904123103.732507-1-wutengda@xxxxxxxxxxxxxxx/


Tengda Wu (2):
  perf stat: Support inherit events during fork() for bperf
  perf test: Use sqrtloop workload to test bperf event

 tools/perf/tests/shell/stat_bpf_counters.sh   |  2 +-
 tools/perf/util/bpf_counter.c                 | 32 +++++--
 tools/perf/util/bpf_skel/bperf_follower.bpf.c | 87 +++++++++++++++++--
 tools/perf/util/bpf_skel/bperf_u.h            |  5 ++
 4 files changed, 112 insertions(+), 14 deletions(-)

-- 
2.34.1