Re: Failures building glibc with mainline GCC

[Aldy Hernandez via Libc-alpha <libc-alpha@sourceware.org>]

There's a new jump threader in GCC which is much more aggressive, and
may trigger latent problems with other warning passes, especially
-Warray-bounds, -Woverflow, and -Wuninitialized.

Do your problems go away if you take out commit 2e96b5f14e?

I have notes throughout the commit analyzing variants of the above
warnings (for example my addendum to gcc.c-torture/compile/pr83510.c).
So far, all the warnings I've seen are legitimate jump threading
opportunities that we are now doing, but which throw off the warning
passes.

Before the inclusion of the new threader, I had warned that this could
happen.  Perhaps we'll have to come up a way to reduce the false
positives.

Aldy

On Fri, Jul 30, 2021 at 5:31 PM Joseph Myers <joseph@codesourcery.com> wrote:
>
> There are a lot of failures building glibc with mainline GCC right now
> <https://sourceware.org/pipermail/libc-testresults/2021q3/008335.html>
> (previously, there were ICEs building glibc on various architectures, so
> these might be hard to bisect):
>
>
> * x86_64-linux-gnu: "error: array subscript 0 is outside array bounds of
> '__seg_fs struct pthread * __seg_fs[0]' [-Werror=array-bounds]".  This is
> the one discussed in
> <https://gcc.gnu.org/pipermail/gcc/2021-July/236922.html>.
>
> In file included from ../sysdeps/generic/libc-tsd.h:44,
>                  from ../include/../locale/localeinfo.h:224,
>                  from ../include/ctype.h:26,
>                  from loadmsgcat.c:29:
> loadmsgcat.c: In function '_nl_load_domain':
> ../sysdeps/x86_64/nptl/tls.h:185:4: error: array subscript 0 is outside array bounds of '__seg_fs struct pthread * __seg_fs[0]' [-Werror=array-bounds]
>   185 |   (*(struct pthread *__seg_fs *) offsetof (struct pthread, header.self))
>       |   ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ../sysdeps/nptl/libc-lock.h:92:18: note: in expansion of macro 'THREAD_SELF'
>    92 |     void *self = THREAD_SELF;                                                 \
>       |                  ^~~~~~~~~~~
> loadmsgcat.c:770:3: note: in expansion of macro '__libc_lock_lock_recursive'
>   770 |   __libc_lock_lock_recursive (lock);
>       |   ^~~~~~~~~~~~~~~~~~~~~~~~~~
>
>
> * i686-gnu:
>
> hurdselect.c: In function '_hurd_select':
> hurdselect.c:555:7: error: 'ss' may be used uninitialized in this function [-Werror=maybe-uninitialized]
>   555 |       _hurd_sigstate_unlock (ss);
>       |       ^~~~~~~~~~~~~~~~~~~~~~~~~~
>
> As far as I can tell, this is a false positive from the compiler (this
> code is only reached if (sigport != MACH_PORT_NULL), in which case ss has
> been initialized).
>
>
> * Several architectures (all of them 32-bit), powerpc-linux-gnu for
> example:
>
> In file included from t.61.c:437:
> In function 'from_t_61_single',
>     inlined from 'gconv' at ../iconv/skeleton.c:568:15:
> ../iconv/loop.c:440:22: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
>   440 |     bytebuf[inlen++] = *inptr++;
>       |     ~~~~~~~~~~~~~~~~~^~~~~~~~~~
> ../iconv/loop.c: In function 'gconv':
> ../iconv/loop.c:382:17: note: at offset 2 into destination object 'bytebuf' of size 2
>   382 |   unsigned char bytebuf[MAX_NEEDED_INPUT];
>       |                 ^~~~~~~
>
> I don't know if this is a false positive or not.
>
>
> * powerpc64-linux-gnu:
>
> In file included from ../sysdeps/powerpc/dl-tls.c:20:
> In function '_dl_allocate_tls_init',
>     inlined from '_dl_allocate_tls' at ../elf/dl-tls.c:621:10:
> ../elf/dl-tls.c:529:10: error: array subscript -1 is outside array bounds of 'void[9223372036854775807]' [-Werror=array-bounds]
>   529 |   dtv_t *dtv = GET_DTV (result);
>       |          ^~~
> In file included from ../elf/dl-tls.c:28,
>                  from ../sysdeps/powerpc/dl-tls.c:20:
> ../sysdeps/powerpc/nptl/tls.h:136:34: error: array subscript -1 is outside array bounds of 'void[9223372036854775807]' [-Werror=array-bounds]
>   136 |   ((tcbhead_t *) (tcbp))[-1].dtv = dtvp + 1
>       |   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~
> ../elf/dl-tls.c:544:7: note: in expansion of macro 'INSTALL_DTV'
>   544 |       INSTALL_DTV (result, &dtv[-1]);
>       |       ^~~~~~~~~~~
>
>
> * powerpc64le-linux-gnu: "error: '-mabi=ibmlongdouble' requires
> '-mlong-double-128'".  See
> <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100909>.
>
>
> --
> Joseph S. Myers
> joseph@codesourcery.com
>