This is the mail archive of the libc-help@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: strange threads problem using glibc-2.10.1


On Wed, 13 May 2009, Allin Cottrell wrote:

> I don't suppose this is a glibc bug, but it just emerged since
> updating glibc and I wonder if anyone has any idea what's going
> on.
>
> I updated from glibc-2.9 to 2.10.1, build from the git sources,
> glibc-2_10-branch.  Some programs that previously ran OK have
> started segfaulting...

Although gdb failed, valgrind produces something, for the segfault
in ld-2.10.1.so when I run wine via make:

waverley:~/src/winbuild/windist$ valgrind --tool=memcheck \
 make -f testmk

==8990== Invalid read of size 4
==8990==    at 0x4001507: security_init (dl-osinfo.h:89)
==8990==    by 0x400347D: dl_main (rtld.c:1857)
==8990==    by 0x40147E2: _dl_sysdep_start (dl-sysdep.c:243)
==8990==    by 0x40011F0: _dl_start (rtld.c:328)
==8990==    by 0x4000806: (within /lib/ld-2.10.1.so)
==8990==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
==8990==
==8990== Process terminating with default action of signal 11 (SIGSEGV)
==8990==  Access not within mapped region at address 0x0
==8990==    at 0x4001507: security_init (dl-osinfo.h:89)
==8990==    by 0x400347D: dl_main (rtld.c:1857)
==8990==    by 0x40147E2: _dl_sysdep_start (dl-sysdep.c:243)
==8990==    by 0x40011F0: _dl_start (rtld.c:328)
==8990==    by 0x4000806: (within /lib/ld-2.10.1.so)

and dl-osinfo.h:89 reads:

memcpy (&ret, dl_random, sizeof (ret));

So it seems that the new version of _dl_setup_stack_chk_guard
must, in this case, be taking dl_random to be non-NULL when it is
in fact NULL?

Allin Cottrell


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]