Re: [PATCH v2 0/8] Update to C11, fix signal undefined behavior

From: Arnaldo Carvalho de Melo
Date: Thu Nov 03 2022 - 09:06:26 EST


Em Mon, Oct 24, 2022 at 11:19:05AM -0700, Ian Rogers escreveu:
> The use of C11 is mainstream in the kernel [1]. There was some
> confusion on volatile and signal handlers in [2] for which atomics
> (present in C11) make things clearer. Switch to using volatile
> sig_atomic_t as per [3]. Thanks to Leo Yan <leo.yan@xxxxxxxxxx> for
> the suggestions.

Thanks, applied and testing building it now.

- Arnaldo

> [1] https://lore.kernel.org/lkml/CAHk-=whWbENRz-vLY6vpESDLj6kGUTKO3khGtVfipHqwewh2HQ@xxxxxxxxxxxxxx/
> [2] https://lore.kernel.org/lkml/20221024011024.462518-1-irogers@xxxxxxxxxx/
> [3] https://wiki.sei.cmu.edu/confluence/display/c/SIG31-C.+Do+not+access+shared+objects+in+signal+handlers
>
> v2. Remove unneeded stdatomic.h includes for sig_atomic_t as it is
> declared in signal.h.
>
> Ian Rogers (8):
> perf build: Update to C standard to gnu11
> perf record: Use sig_atomic_t for signal handlers
> perf daemon: Use sig_atomic_t to avoid UB
> perf ftrace: Use sig_atomic_t to avoid UB
> perf session: Change type to avoid UB
> perf stat: Use sig_atomic_t to avoid UB
> perf top: Use sig_atomic_t to avoid UB
> perf trace: Use sig_atomic_t to avoid UB
>
> tools/perf/Makefile.config | 2 +-
> tools/perf/builtin-daemon.c | 2 +-
> tools/perf/builtin-ftrace.c | 4 ++--
> tools/perf/builtin-record.c | 8 ++++----
> tools/perf/builtin-stat.c | 8 ++++----
> tools/perf/builtin-top.c | 4 ++--
> tools/perf/builtin-trace.c | 4 ++--
> tools/perf/util/session.c | 3 ++-
> 8 files changed, 18 insertions(+), 17 deletions(-)
>
> --
> 2.38.0.135.g90850a2211-goog

--

- Arnaldo