Re: [PATCH v3 5/5] avoid -Wuse-after-free [BZ #26779]

[Joseph Myers <joseph@codesourcery.com>]

On Mon, 31 Jan 2022, Carlos O'Donell via Libc-alpha wrote:

> On 1/28/22 12:51, Joseph Myers wrote:
> > On Fri, 28 Jan 2022, Carlos O'Donell via Libc-alpha wrote:
> > 
> >> I'm going to look at these right now. I want clean gcc 12 glibc 2.35 
> >> results before we release.
> > 
> > There are also three ICE bugs in GCC 12 preventing glibc from building, so 
> > you won't get fully clean build-many-glibcs.py results with just glibc 
> > fixes.  GCC bug 103722 on sh4, bug 104153 on or1k and 104154 on arc (the 
> > last two reported by Jeff based on ICEs building newlib, but the ICEs I 
> > see building glibc are in the same place in the compiler, so probably the 
> > same bugs).
>  
> Thanks.
> 
> I just fixed the last -Wuse-after-free bug.

I think there are still some more -Wuse-after-free issues building the 
testsuite.

https://sourceware.org/pipermail/libc-testresults/2022q1/009283.html

s390x-linux-gnu-O3 failures are as shown at 
<https://sourceware.org/pipermail/libc-alpha/2022-January/135797.html>.

hppa-linux-gnu, microblaze-linux-gnu, microblazeel-linux-gnu, 
glibcs-sparcv8-linux-gnu-leon3, sparcv9-linux-gnu, 
sparcv9-linux-gnu-disable-multi-arch:

In function 'do_test',
    inlined from 'legacy_test_function' at ../test-skeleton.c:55:10:tst-realloc.c:134:12: error: pointer 'p' may be used after 'realloc' [-Werror=use-after-free]
  134 |       if (c[i] != 0xff)
      |           ~^~~
tst-realloc.c:124:7: note: call to 'realloc' here
  124 |   c = realloc (p, -1);
      |       ^~~~~~~~~~~~~~~

-- 
Joseph S. Myers
joseph@codesourcery.com