linux-kernel - Re: [PATCH v1] tools build: Fix a number of Wconversion warnings

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAP-5=fUZ2QCocFKdLfBoNYC-CQfSAcdbA05OhegKmTt_PLR1WA@mail.gmail.com>
Date: Tue, 7 Jan 2025 08:11:23 -0800
From: Ian Rogers <irogers@...gle.com>
To: James Clark <james.clark@...aro.org>
Cc: Leo Yan <leo.yan@....com>, Peter Zijlstra <peterz@...radead.org>, 
	Ingo Molnar <mingo@...hat.com>, Arnaldo Carvalho de Melo <acme@...nel.org>, Namhyung Kim <namhyung@...nel.org>, 
	Mark Rutland <mark.rutland@....com>, 
	Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>, 
	Adrian Hunter <adrian.hunter@...el.com>, Kan Liang <kan.liang@...ux.intel.com>, 
	linux-perf-users@...r.kernel.org, bpf@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1] tools build: Fix a number of Wconversion warnings

On Tue, Jan 7, 2025 at 2:33 AM James Clark <james.clark@...aro.org> wrote:
>
> On 06/01/2025 9:54 pm, Ian Rogers wrote:
> > There's some expressed interest in having the compiler flag
> > -Wconversion detect at build time certain kinds of potential problems:
> > https://lore.kernel.org/lkml/20250103182532.GB781381@e132581.arm.com/
> >
> > As feature detection passes -Wconversion from CFLAGS when set, the
> > feature detection compile tests need to not fail because of
> > -Wconversion as the failure will be interpretted as a missing
> > feature. Switch various types to avoid the -Wconversion issue, the
> > exact meaning of the code is unimportant as it is typically looking
> > for header file definitions.
> >
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
>
> What's the plan for errors in #includes that we can't modify? I noticed
> the Perl feature test fails with -Wconversion but can be fixed by
> disabling the warning:
>
>    #pragma GCC diagnostic push
>    #pragma GCC diagnostic ignored "-Wsign-conversion"
>    #pragma GCC diagnostic ignored "-Wconversion"
>    #include <EXTERN.h>
>    #include <perl.h>
>    #pragma GCC diagnostic pop
>
> Not sure why it needs both those things to be disabled when I only
> enabled -Wconversion, but it does.

This change lgtm, I'm not sure how others feel. I don't have a plan, I
was just following up on Leo's Wconversion comment to see what state
things were in. The feature tests without these changes pretty much
break the build (I can live without perl support :-) ) so I thought I
could move things forward there and then see the state of Wconversion
with the patch I was working on.

I'm not sure how others feel about fixing Wconversion in perf, the
errors are quite noisy imo. The biggest issue imo will be with headers
shared by tools and the kernel, where kernel people may be vocal on
the merits of Wconversion.

> > ---
> >   tools/build/feature/test-backtrace.c           | 2 +-
> >   tools/build/feature/test-bpf.c                 | 2 +-
> >   tools/build/feature/test-glibc.c               | 2 +-
> >   tools/build/feature/test-libdebuginfod.c       | 2 +-
> >   tools/build/feature/test-libdw.c               | 2 +-
> >   tools/build/feature/test-libelf-gelf_getnote.c | 2 +-
> >   tools/build/feature/test-libelf.c              | 2 +-
> >   tools/build/feature/test-lzma.c                | 2 +-
> >   8 files changed, 8 insertions(+), 8 deletions(-)
> >
> > diff --git a/tools/build/feature/test-backtrace.c b/tools/build/feature/test-backtrace.c
> > index e9ddd27c69c3..7962fbad6401 100644
> > --- a/tools/build/feature/test-backtrace.c
> > +++ b/tools/build/feature/test-backtrace.c
> > @@ -5,7 +5,7 @@
> >   int main(void)
> >   {
> >       void *backtrace_fns[10];
> > -     size_t entries;
> > +     int entries;
> >
> >       entries = backtrace(backtrace_fns, 10);
> >       backtrace_symbols_fd(backtrace_fns, entries, 1);
> > diff --git a/tools/build/feature/test-bpf.c b/tools/build/feature/test-bpf.c
> > index 727d22e34a6e..e7a405f83af6 100644
> > --- a/tools/build/feature/test-bpf.c
> > +++ b/tools/build/feature/test-bpf.c
> > @@ -44,5 +44,5 @@ int main(void)
> >        * Test existence of __NR_bpf and BPF_PROG_LOAD.
> >        * This call should fail if we run the testcase.
> >        */
> > -     return syscall(__NR_bpf, BPF_PROG_LOAD, &attr, sizeof(attr));
> > +     return syscall(__NR_bpf, BPF_PROG_LOAD, &attr, sizeof(attr)) == 0;
>
> Seems a bit weird to invert some of the return values rather than doing
> != 0, but as you say, the actual values seem to be unimportant.
>
> Reviewed-by: James Clark <james.clark@...aro.org>

Yeah it was arbitrary and I didn't want to add a stdlib.h dependency
in a bunch of places for the sake of a definition of NULL. I'm happy
for things to be done differently if people like.

Thanks,
Ian