Latest working kernel version: 2.6.23.17 Earliest failing kernel version: all 2.6.24 and all 2.6.25 Distribution: Debian Etch 4.0 stable Hardware Environment: Intel Celeron M 1,5 Ghz processor and Intel 915 chipset Software Environment: KDE 3.5.5.a Problem Description: Hello Using kde-3.5.5a in association with kdm (Debian Etch 4.0 stable release), I get the following segfault when I enter in a root text console (tty1): # /etc/init.d/kdm stop KDE and kdm stop normally as expected, but a segfault appears on the screen. When I restart kde with: # /etc/init.d/kdm start There's no segfault at all, and the KDE desktop starts normally. This bug is reproductible each time I enter # /etc/init.d/kdm stop Here is the log file: Apr 21 20:59:37 localhost kdm: :0[1985]: (pam_unix) session closed for user damien Apr 21 20:59:37 localhost kdm: :0[1985]: PAM unable to dispatch function Apr 21 20:59:37 localhost kdm: :0[1985]: pam_setcred(DELETE_CRED) for damien failed: Bad item passed to pam_*_item() Apr 21 20:59:37 localhost kdm: :0[1985]: PAM unable to dispatch function Apr 21 20:59:37 localhost kernel: kdm[1985]: segfault at b7d68db0 ip b7d68db0 sp bfffbadc error 4 in libnss_compat-2.3.6.so[b7d7e000+7000] Apr 21 20:59:37 localhost kdm[1980]: Unknown session exit code 0 (sig 11) from manager process Apr 21 20:59:49 localhost kdm: :0[2156]: (pam_unix) session opened for user damien by (uid=0) There seems to be a conflict between the kernel pam_unix module and the libc6 library Important note : this bug appears with all 2.6.24 and 2.6.25 kernels, but not with the 2.6.23.17 kernel. Can this bug been resolved within the kernel? Thanks Damien BRUCKER Steps to reproduce: see above
The kernel's segfault message is newly-added. Probably the userspace code was always generating a segfault, only we didn't know about it. Are any other problems observeable?
There's no other problem observable in the log files, located under the /var/log directory, about his subject. Damien -----Message d'origine----- De : bugme-daemon@bugzilla.kernel.org [mailto:bugme-daemon@bugzilla.kernel.org] Envoy
ERRATUM: This bug also appears with the 2.6.23.17 kernel, as opposed to what was stated above.
Closing as invalid: Not a kernel bug but user observing a userspace bug now being usefully recorded