[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: libc6 and X11 seem not to work together



Hans-J. Ullrich wrote:
Dear maintainers, since the last upgrade (debian-amd64/sid) it is not more possible to start X. It seems, libc6 is incompatible to the xserver-packages in debiansid

Backtrace:
0: /usr/bin/X11/X(xorg_backtrace+0x26) [0x4edfe6]
1: /usr/bin/X11/X(xf86SigHandler+0x39) [0x483699]
2: /lib/libc.so.6 [0x7f7e17a41db0] 3: /usr/bin/X11/X(DGAAvailable+0x39) [0x4823c9] 4: /usr/bin/X11/X(xf86DiDGAInit+0x30) [0x4ae200] 5: /usr/bin/X11/X(xf86CrtcScreenInit+0x10d) [0x4aafad]
6: /usr/lib/xorg/modules/drivers//radeon_drv.so [0x7f7e15e26a49]
7: /usr/bin/X11/X(AddScreen+0x1c6) [0x432ac6] 8: /usr/bin/X11/X(InitOutput+0x241) [0x46d361] 9: /usr/bin/X11/X(main+0x200) [0x4331d0] 10: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f7e17a2e5c6] 11: /usr/bin/X11/X [0x432819]

For me it looks like Xorg gets a problem with the latest libc6-version from sid. Can someone confirm this ?

I see something very similar on a testing/squeeze amd64 machine. But I am seeing several problems on 2 i386 and amd64 testing machines which all seem to be related to libc.so.6. On the i386 boxes this seems to show up in very different circumstances. In "cdrom" which I have yet to confirm is part of apt; & also in dd access to floppy discs. None of these machines use a tainted kernel. Also it happens with both debian kernels and also with the latest git stock kernels. So a kernel problem is unlikely.

I have yet to investigate properly, partly because it was not obvious how to track this down. I now have aptitude-dbg installed, and may find time (very busy with other things) to see if I can use gdb to catch the cdrom module and its segfault. Then I will have something useful to report.

In summary, I suspect a libc6 problem which is causing a variety of programs to fail across a range of architectures and programs. Or a huge coincidence. What mystified me was why I found no other reports of this sort of thing. I am not sure when libc6 was updated in squeezew/testing, but it was maybe over a month ago, and I have been seeing the cdrom segfault for that sort of time. I didn't notice at first because it only appeared in /var/log/messages.

The report from Hans gives me some assurance that I am not going mad :-)

ael


Reply to: