Re: [PATCH v4 00/10] Move uid filtering to BPF filters

[Namhyung Kim <namhyung@kernel.org>]

Hi Ian,

On Wed, Jun 04, 2025 at 10:45:34AM -0700, Ian Rogers wrote:
> Rather than scanning /proc and skipping PIDs based on their UIDs, use
> BPF filters for uid filtering. The /proc scanning in thread_map is
> racy as the PID may exit before the perf_event_open causing perf to
> abort. BPF UID filters are more robust as they avoid the race. The
> /proc scanning also misses processes starting after the perf
> command. Add a helper for commands that support UID filtering and wire
> up. Remove the non-BPF UID filtering support given it doesn't work.
> 
> v4: Add a warning message on top of Namhyung's BPF filter error message:
> https://lore.kernel.org/lkml/20250604054234.23608-1-namhyung@kernel.org/
>     in the parse_uid_filter helper. In TUI the warning is shown then
>     the BPF error shown, with stdio the warning appears below the BPF
>     errors.
> 
> v3: Add lengthier commit messages as requested by Arnaldo. Rebase on
>     tmp.perf-tools-next.
> 
> v2: Add a perf record uid test (Namhyung) and force setting
>     system-wide for perf trace and perf record (Namhyung). Ensure the
>     uid filter isn't set on tracepoint evsels.
> 
> v1: https://lore.kernel.org/lkml/20250111190143.1029906-1-irogers@google.com/
> 
> Ian Rogers (10):
>   perf parse-events filter: Use evsel__find_pmu
>   perf target: Separate parse_uid into its own function
>   perf parse-events: Add parse_uid_filter helper
>   perf record: Switch user option to use BPF filter
>   perf tests record: Add basic uid filtering test
>   perf top: Switch user option to use BPF filter
>   perf trace: Switch user option to use BPF filter
>   perf bench evlist-open-close: Switch user option to use BPF filter
>   perf target: Remove uid from target
>   perf thread_map: Remove uid options

I've noticed two things.

* it takes a quite long time to load the BPF filter.  Not sure what's
  the issue, but maybe annoying for users.

* normal users cannot use BPF filter even if the BPF is loaded and
  pinned already.  It works fine for perf record.

I don't think the issues are from this change though.

Thanks,
Namhyung

> 
>  tools/perf/bench/evlist-open-close.c        | 36 ++++++++------
>  tools/perf/builtin-ftrace.c                 |  1 -
>  tools/perf/builtin-kvm.c                    |  2 -
>  tools/perf/builtin-record.c                 | 27 ++++++-----
>  tools/perf/builtin-stat.c                   |  4 +-
>  tools/perf/builtin-top.c                    | 22 +++++----
>  tools/perf/builtin-trace.c                  | 27 +++++++----
>  tools/perf/tests/backward-ring-buffer.c     |  1 -
>  tools/perf/tests/event-times.c              |  8 ++-
>  tools/perf/tests/keep-tracking.c            |  2 +-
>  tools/perf/tests/mmap-basic.c               |  2 +-
>  tools/perf/tests/openat-syscall-all-cpus.c  |  2 +-
>  tools/perf/tests/openat-syscall-tp-fields.c |  1 -
>  tools/perf/tests/openat-syscall.c           |  2 +-
>  tools/perf/tests/perf-record.c              |  1 -
>  tools/perf/tests/perf-time-to-tsc.c         |  2 +-
>  tools/perf/tests/shell/record.sh            | 26 ++++++++++
>  tools/perf/tests/switch-tracking.c          |  2 +-
>  tools/perf/tests/task-exit.c                |  1 -
>  tools/perf/tests/thread-map.c               |  2 +-
>  tools/perf/util/bpf-filter.c                |  2 +-
>  tools/perf/util/evlist.c                    |  3 +-
>  tools/perf/util/parse-events.c              | 47 +++++++++++++-----
>  tools/perf/util/parse-events.h              |  1 +
>  tools/perf/util/python.c                    | 10 ++--
>  tools/perf/util/target.c                    | 54 +++------------------
>  tools/perf/util/target.h                    | 15 ++----
>  tools/perf/util/thread_map.c                | 32 ++----------
>  tools/perf/util/thread_map.h                |  6 +--
>  tools/perf/util/top.c                       |  4 +-
>  tools/perf/util/top.h                       |  1 +
>  31 files changed, 164 insertions(+), 182 deletions(-)
> 
> -- 
> 2.50.0.rc0.604.gd4ff7b7c86-goog
>