Re: [PATCH] x86: Fix bug about glibc.cpu.hwcaps.

[Xi Ruoyao <xry111@xry111.site>]

On Tue, 2023-03-07 at 20:45 +0800, caiyinyu wrote:
> 
> 在 2023/3/2 下午12:22, H.J. Lu 写道:
> > On Wed, Mar 1, 2023 at 6:36 PM caiyinyu <caiyinyu@loongson.cn>
> > wrote:
> > > Recorded in [BZ #30183]:
> > > 
> > > 1. export GLIBC_TUNABLES=glibc.cpu.hwcaps=-AVX512
> > > 2. Add  _dl_printf("p -- %s\n", p); just before switch(nl) in
> > >     sysdeps/x86/cpu-tunables.c
> > > 3. compiled and run ./testrun.sh /usr/bin/ls
> > > you will get:
> > > 
> > > p -- -AVX512
> > > p -- LC_ADDRESS=en_US.UTF-8
> > > p -- LC_NUMERIC=C
> > > ...
> > > 
> > > The function, TUNABLE_CALLBACK (set_hwcaps)
> > > (tunable_val_t *valp), checks far more than it should and it
> > > should stop at end of "-AVX512".
> > > ---
> > >   sysdeps/x86/cpu-tunables.c | 6 +++---
> > >   1 file changed, 3 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/sysdeps/x86/cpu-tunables.c b/sysdeps/x86/cpu-
> > > tunables.c
> > > index d3e1367bda..772fb0c4c2 100644
> > > --- a/sysdeps/x86/cpu-tunables.c
> > > +++ b/sysdeps/x86/cpu-tunables.c
> > > @@ -107,13 +107,13 @@ TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t
> > > *valp)
> > >        NOTE: the IFUNC selection may change over time.  Please
> > > check all
> > >        multiarch implementations when experimenting.  */
> > > 
> > > -  const char *p = valp->strval;
> > > +  const char *p = valp->strval, *c;
> > >     struct cpu_features *cpu_features =
> > > &GLRO(dl_x86_cpu_features);
> > >     size_t len;
> > > 
> > >     do
> > >       {
> > > -      const char *c, *n;
> > > +      const char *n;
> > >         bool disable;
> > >         size_t nl;
> > > 
> > > @@ -283,7 +283,7 @@ TUNABLE_CALLBACK (set_hwcaps) (tunable_val_t
> > > *valp)
> > >          }
> > >         p += len + 1;
> > >       }
> > > -  while (*p != '\0');
> > > +  while (*c != '\0');
> > >   }
> > > 
> > >   # if CET_ENABLED
> > > --
> > > 2.31.1
> > > 
> > OK.
> > 
> > Thanks.
> 
> Could you help to push this patch?

I guess you can push it yourself.  A port maintainer has write access to
the entire Git repository, just use it carefully (i. e. always get a
permission before changing other subsystems).

-- 
Xi Ruoyao <xry111@xry111.site>
School of Aerospace Science and Technology, Xidian University