[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aOQPE33mYPVYnkl5@x1>
Date: Mon, 6 Oct 2025 15:48:51 -0300
From: Arnaldo Carvalho de Melo <acme@...nel.org>
To: Ian Rogers <irogers@...gle.com>
Cc: Leo Yan <leo.yan@....com>, Namhyung Kim <namhyung@...nel.org>,
Jiri Olsa <jolsa@...nel.org>,
Adrian Hunter <adrian.hunter@...el.com>,
Nathan Chancellor <nathan@...nel.org>,
Nick Desaulniers <nick.desaulniers+lkml@...il.com>,
Bill Wendling <morbo@...gle.com>,
Justin Stitt <justinstitt@...gle.com>,
Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>, Alexandre Ghiti <alex@...ti.fr>,
James Clark <james.clark@...aro.org>, linux-kernel@...r.kernel.org,
linux-perf-users@...r.kernel.org, llvm@...ts.linux.dev,
linux-riscv@...ts.infradead.org
Subject: Re: [PATCH v3 4/9] perf build: Disable thread safety analysis for
perl header
On Mon, Oct 06, 2025 at 10:56:43AM -0700, Ian Rogers wrote:
> On Mon, Oct 6, 2025 at 9:21 AM Leo Yan <leo.yan@....com> wrote:
> >
> > When build with perl5, it reports error:
> >
> > In file included from /usr/lib/perl5/5.42.0/x86_64-linux-thread-multi/CORE/perl.h:7933:
> > /usr/lib/perl5/5.42.0/x86_64-linux-thread-multi/CORE/inline.h:298:5: error:
> > mutex 'PL_env_mutex.lock' is not held on every path through
> > here [-Werror,-Wthread-safety-analysis]
> > 298 | ENV_UNLOCK;
> > | ^
> > /usr/lib/perl5/5.42.0/x86_64-linux-thread-multi/CORE/perl.h:7091:31: note:
> > expanded from macro 'ENV_UNLOCK'
> > 7091 | # define ENV_UNLOCK PERL_REENTRANT_UNLOCK("env"...
> > | ^
> > /usr/lib/perl5/5.42.0/x86_64-linux-thread-multi/CORE/perl.h:6465:7: note:
> > expanded from macro 'PERL_REENTRANT_UNLOCK'
> > 6465 | } STMT_END
> > | ^
> > /usr/lib/perl5/5.42.0/x86_64-linux-thread-multi/CORE/perl.h:865:28: note:
> > expanded from macro 'STMT_END'
> > 865 | # define STMT_END while (0)
> > | ^
> >
> > The error is caused by perl header but not perf code, disable thread
> > safety analysis if including the header.
> >
> > Though GCC does not support the thread safety analysis option, this
> > negative warning flag is silently ignored by it.
> >
> > Signed-off-by: Leo Yan <leo.yan@....com>
> > ---
> > tools/perf/scripts/perl/Perf-Trace-Util/Build | 2 +-
> > tools/perf/util/scripting-engines/Build | 2 +-
> > 2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/Build b/tools/perf/scripts/perl/Perf-Trace-Util/Build
> > index 9b0e5a8b5070f1a1640518fae75557f824ef21ee..01a1a0ed51aefd721b4bc7eba728c29a8ffcd551 100644
> > --- a/tools/perf/scripts/perl/Perf-Trace-Util/Build
> > +++ b/tools/perf/scripts/perl/Perf-Trace-Util/Build
> > @@ -2,7 +2,7 @@ perf-util-y += Context.o
> >
> > CFLAGS_Context.o += $(PERL_EMBED_CCOPTS) -Wno-redundant-decls -Wno-strict-prototypes -Wno-bad-function-cast -Wno-declaration-after-statement -Wno-switch-enum
> > CFLAGS_Context.o += -Wno-unused-parameter -Wno-nested-externs -Wno-undef
> > -CFLAGS_Context.o += -Wno-switch-default -Wno-shadow
> > +CFLAGS_Context.o += -Wno-switch-default -Wno-shadow -Wno-thread-safety-analysis
>
> I wasn't able to reproduce this on Debian with perf-tools-next and:
> make -C tools/perf O=/tmp/perf DEBUG=1 CC=clang CXX=clang++
> HOSTCC=clang LIBPERL=1 clean all
> but my libperl is 5.40. I'm a little concerned that the warning may
> cause an unrecognized compiler option warning/error with GCC given
> progress on thread-safety-analysis has ended there. Anyway, these
> aren't blockers.
>
> Reviewed-by: Ian Rogers <irogers@...gle.com>
I was a bit wary and saw this on only, IIRC Manjaro Linux, and since
libperl support is opt-in, I left it in the backburner, but having it
building there as well may help detecting some other stuff with things
only enabled/present there, so thanks Leo and Ian, for the patch and the
Reviewed-by, applying.
- Arnaldo
Powered by blists - more mailing lists