Warnins on glibc dlopen()

[João Pinto]

I have already reported this and provided a fix at:
https://launchpad.net/distros/ubuntu/+source/valgrind/+bug/2958
However the fix that was selected and applied was only partial !

I still got:
==14336==
==14336== Conditional jump or move depends on uninitialised value(s)
==14336== at 0x4010BEE: (within /lib/ld-2.3.6.so)
==14336== by 0x400624D: (within /lib/ld-2.3.6.so)
==14336== by 0x427CC75: (within /lib/tls/i686/cmov/libc-2.3.6.so)
==14336== by 0x400B105: (within /lib/ld-2.3.6.so)
==14336== by 0x427D937: _dl_open (in /lib/tls/i686/cmov/libc-2.3.6.so)
==14336== by 0x402ECB7: (within /lib/tls/i686/cmov/libdl-2.3.6.so)
==14336== by 0x400B105: (within /lib/ld-2.3.6.so)
==14336== by 0x402F2BA: (within /lib/tls/i686/cmov/libdl-2.3.6.so)
==14336== by 0x402ED10: dlopen (in /lib/tls/i686/cmov/libdl-2.3.6.so)
==14336== by 0x804FE2C: module_load (modules.c:65)
==14336== by 0x8050206: load_modules_file (modules.c:185)
==14336== by 0x8050165: module_load (modules.c:161)

Please read the other bug info on the details for how to reproduce the bug.

[João Pinto (joaopinto)]

No one cares about the bug, neither do I any longer.

[Andy Buckley (andy-insectnation)]

Shame. I noticed a dlopen leak in my valgrind trace and I think this may also be responsible for gdb dying during module loading when I test my program on an Ubuntu system... it's a real problem for debugging/testing, even if it seems "sort of safe" for uninstrumented running.