Summary: | net-print/cups-1.6.2-r5 - cupsd: segfault in libdbus-1.so.3.7.3 at startup | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Bob Johnson <bob> |
Component: | [OLD] Printing | Assignee: | Printing Team <printing> |
Status: | RESOLVED NEEDINFO | ||
Severity: | normal | ||
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 463014 |
Description
Bob Johnson
2013-05-15 20:30:35 UTC
We can't do very much about this without further information (i.e. a backtrace of the segfault). Please have a look at the relevant documentation: http://www.gentoo.org/proj/en/qa/backtraces.xml Differences between -r3 and -r5 are minimal, basically only prefix support has been added - is it reproducible that the switch between -r3 and -r5 fixes/recreates the segfault? (In reply to comment #1) > We can't do very much about this without further information (i.e. a > backtrace of the segfault). > Please have a look at the relevant documentation: > http://www.gentoo.org/proj/en/qa/backtraces.xml > > Differences between -r3 and -r5 are minimal, basically only prefix support > has been added - is it reproducible that the switch between -r3 and -r5 > fixes/recreates the segfault? I think you might as well close this bug. It seems to be a temporal bug that is going to be almost impossible to find, and may no longer 'exist'. Here's what I did: 1) I unmasked net-print/cups-1.6.2-r5 and emerged it. I no longer see the cupsd segfault at startup with the current stable kernel 3.8.13-gentoo. 2) I re-booted to the previous kernel 3.7.10-gentoo-r1. cupsd *did* crash at startup with this kernel. 3) I created a non-stripped debug build of cups, enabled core dumps in my /etc/security/limits.conf file and added the following two lines in /etc/init.d/cupsd: echo "/tmp/core" > /proc/sys/kernel/core_pattern echo 1 > /proc/sys/kernel/core_uses_pid just after the 'ebegin "Staring cupsd" line. 4) When I rebooted with the above changes into the 3.7.10-gentoo-r1 kernel, cupsd started normally with no segfault. 5) I then removed the /etc/portage/env/net-print/cups splitdebug file, and re-built cups without debug information. When I rebooted into the 3.7.10-gentoo-r1 kernel, cupsd still started normally. 6) I then commented out the two lines I had added to the /etc/init.d/cupsd startup script, and rebooted again into 3.7.10-gentoo-r1. This time, cupsd segfaulted. So, it appears that this is a very sensitive race bug that 1) is going to be almost impossible to get a backtrace on and 2) no longer manifests itself with the current stable kernel. I'm not sure it's worth any more effort at this point. Wow. You put in a lot of work there. Thank you very much. This is what's called a heisenbug. It's only there if you don't look closely enough... Let's close it now, and if it reappears with the new kernel, reopen. |