Summary: | sys-apps/devicekit-power: daemon segfaults | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Michael Edenfield <kutulu> |
Component: | [OLD] GNOME | Assignee: | Freedesktop bugs <freedesktop-bugs> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | alex_justes, hardened, pacho, underling |
Priority: | High | Keywords: | Inclusion |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://bugs.freedesktop.org/show_bug.cgi?id=27902 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
Output of gnome-power-bugreport.sh
dmesg output showing segfaults. ebuild diff to filter pie Backtrace about this bug Patch for devicekit-power backtrace |
Description
Michael Edenfield
2009-10-31 22:14:19 UTC
Created attachment 208902 [details]
Output of gnome-power-bugreport.sh
Created attachment 208904 [details]
dmesg output showing segfaults.
a backtrace of the crash would be nice (plus version of dk-p) http://www.gentoo.org/proj/en/qa/backtraces.xml (In reply to comment #3) > a backtrace of the crash would be nice (plus version of dk-p) > > http://www.gentoo.org/proj/en/qa/backtraces.xml > It's version 011. And.. it appears I was too quick to dismiss my hardened setup as the culprit. When I rebuild the daemon with -nopie on the final link stage (to get a usable backtrace), it doesn't segfault. Is there still a gdb version that understands PIE? CCing hardened since I have no clue about the specifics of hardened debugging. Please try dk-p 012 and provide a backtrace, thanks. (In reply to comment #6) > Please try dk-p 012 and provide a backtrace, thanks. > version 012 still segfaults. Working on the backtrace. Reopening a few of us on hardened dev team are looking into the problem Created attachment 215094 [details, diff]
ebuild diff to filter pie
This is only a bandage until the problem is located and fixed properly.
without a backtrace, there's little chance the problem will get fixed. Is using -fPIE in CFLAGS something that would make us non-hardened user able to reproduce the problem ? (In reply to comment #10) > without a backtrace, there's little chance the problem will get fixed. Is using > -fPIE in CFLAGS something that would make us non-hardened user able to > reproduce the problem ? > Not it won't, this is only effecting amd64 users from what I can see locally, my x86 hardened setup is fine with it. Soon as possible I will get the backtrace with a proper fix. Created attachment 224319 [details]
Backtrace about this bug
Created attachment 225071 [details, diff]
Patch for devicekit-power
Solve the problem for me, replacing 0 by NULL in g_object-set; seems more compliant with amd64.
sys-apps/devicekit-power-014 segfaults too. May be this is caused by the update to gnome-2.30, but i'am not sure. $dmesg | tail [11334.662974] devkit-power-da[9186]: segfault at 7f9600000000 ip 00007f963a0e31ba sp 00007fff5b5444c8 error 4 in libc-2.11.so[7f963a067000+152000] [11854.075345] devkit-power-da[7941]: segfault at 7f1d00000000 ip 00007f1d761bf1ba sp 00007fffd098b958 error 4 in libc-2.11.so[7f1d76143000+152000] [11854.130156] devkit-power-da[7943]: segfault at 7f9400000000 ip 00007f947f68b1ba sp 00007fff33aa06c8 error 4 in libc-2.11.so[7f947f60f000+152000] Created attachment 226659 [details]
backtrace
Great, recompiling it with -O2 flag (instead of -O3) solved the problem =) (In reply to comment #13) > Created an attachment (id=225071) [details] > Patch for devicekit-power > > Solve the problem for me, replacing 0 by NULL in g_object-set; seems more > compliant with amd64. > Have you reported this to upstream? It doesn't look like it since we have also bug 315199 now. I'm afraid anykind of auto mounting will be dead for hardened until this gets done. s/auto mounting/power management/ ;-) devicekit-power won't be fixed, but upower might (bug 315199) *** This bug has been marked as a duplicate of bug 315199 *** |