Bug 6704 - Xorg crashes with signal 11 after a while
Summary: Xorg crashes with signal 11 after a while
Status: RESOLVED DUPLICATE of bug 6827
Alias: None
Product: xorg
Classification: Unclassified
Component: * Other (show other bugs)
Version: 6.99.99.902 (7.0 RC2)
Hardware: x86 (IA32) Linux (All)
: high blocker
Assignee: Adam Jackson
QA Contact:
URL:
Whiteboard:
Keywords: have-backtrace
Depends on:
Blocks: 5041
  Show dependency treegraph
 
Reported: 2006-04-23 12:39 UTC by Dennis Jacobfeuerborn
Modified: 2006-05-15 11:16 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
xorg.conf and Xorg.0.log (52.57 KB, text/plain)
2006-04-23 20:48 UTC, Dennis Jacobfeuerborn
no flags Details
gdb backtrace (2.05 KB, text/plain)
2006-04-28 21:47 UTC, Dennis Jacobfeuerborn
no flags Details

Description Dennis Jacobfeuerborn 2006-04-23 12:39:25 UTC
After a while Xorg crashes with the following backtrace:

0: /usr/bin/Xorg(xf86SigHandler+0x87) [0x80b6d07]
1: [0x49a420]
2: /usr/lib/xorg/modules/libfb.so [0x155b81]
3: /usr/lib/xorg/modules/libfb.so [0x15d923]
4: /usr/lib/xorg/modules/libfb.so(fbCompositeGeneral+0x855) [0x15b4e5]
5: /usr/lib/xorg/modules/libfb.so(fbComposite+0x1cd) [0x1643ad]
6: /usr/lib/xorg/modules/libxaa.so(XAAComposite+0x261) [0xd56001]
7: /usr/bin/Xorg [0x8156d06]
8: /usr/bin/Xorg [0x8153915]
9: /usr/bin/Xorg(CompositePicture+0x153) [0x8140a13]
10: /usr/bin/Xorg [0x814687c]
11: /usr/bin/Xorg [0x8143bd5]
12: /usr/bin/Xorg(Dispatch+0x19a) [0x80869ba]
13: /usr/bin/Xorg(main+0x485) [0x806e3e5]
14: /lib/libc.so.6(__libc_start_main+0xdc) [0x4cd7e4]
15: /usr/bin/Xorg(FontFileCompleteXLFD+0xbd) [0x806d721]

Fatal server error:
Caught signal 11.  Server aborting

I'm running an up-to-date Fedora Core Rawhide system
(xorg-x11-server-Xorg-1.0.99.901-5) with the open source nv driver an no special
settings in the config.

This seems to get triggered after a random amount of time when using Firefox and
Firefox sometimes shows me random bits of the desktop or parts of images on
pages I visited minutes before instead of resized image so I'm wondering if this
"random access" of memory might be the reason for this.
Comment 1 Erik Andren 2006-04-23 18:26:53 UTC
Please post your xorg conf and your xorg.log 
Also please try to get a remote backtrace.
Comment 2 Dennis Jacobfeuerborn 2006-04-23 20:48:52 UTC
Created attachment 5430 [details]
xorg.conf and Xorg.0.log

This happened a couple of time now for me in the last couple of days and always
with the same trace. What exactly is a remote backtrace and how do I obtain it?
Comment 3 Dennis Jacobfeuerborn 2006-04-23 22:28:34 UTC
I just tried running Xorg using gdb which worked but when X crashed it caught
the  SIGSEGV, printed the backtrace and in gdb I only got the backtrace of a
followup SIGPIPE. Is there away to tell Xorg to let gdb catch the signal 11 instead?
Comment 4 Matthieu Herrb 2006-04-23 22:56:12 UTC
There is some documentation on running the X server under gdb on the wiki:
http://xorg.freedesktop.org/wiki/DebuggingTheXserver
Comment 5 Dennis Jacobfeuerborn 2006-04-28 21:47:04 UTC
Created attachment 5516 [details]
gdb backtrace

After a few days the crash happened again and this time I had a remote gdb
attached. See attachment.
Comment 6 Dennis Jacobfeuerborn 2006-05-03 04:14:08 UTC
Two observations:
1. XAAComposite() gets called with xSrc=51, ySrc=270 which then calls
fbComposite() with xSrc=109, ySrc=4578 which looks odd to me because I see no
instructions modifying these variables between those calls.

2. The call to fbComposite() happens right after the patch for bug #4320.
(called xorg-x11-server-1.0.1-composite-fastpath-fdo4320.patch in the Fedora
Core Rawhide SRPM).

Could this patch be responsible for the breakage?
Comment 7 Adam Jackson 2006-05-03 09:23:27 UTC
(In reply to comment #6)
> Two observations:
> 1. XAAComposite() gets called with xSrc=51, ySrc=270 which then calls
> fbComposite() with xSrc=109, ySrc=4578 which looks odd to me because I see no
> instructions modifying these variables between those calls.
> 
> 2. The call to fbComposite() happens right after the patch for bug #4320.
> (called xorg-x11-server-1.0.1-composite-fastpath-fdo4320.patch in the Fedora
> Core Rawhide SRPM).
> 
> Could this patch be responsible for the breakage?

Could be.  Should also be unnecessary now in rawhide.  The RC2 server build is
going through now, give it a try tomorrow.

I'd also be interested to know if this is a problem with Composite disabled.
Comment 8 Dennis Jacobfeuerborn 2006-05-04 02:36:46 UTC
This is still a problem with the package xorg-x11-server-Xorg-1.0.99.902-1. I've
disabled Composite now to see if that makes any difference.
Comment 9 Dennis Jacobfeuerborn 2006-05-05 06:12:04 UTC
Unfortunately I'm still getting hit by this even with Composite disabled.
Comment 10 Adam Jackson 2006-05-09 05:14:49 UTC
(In reply to comment #9)
> Unfortunately I'm still getting hit by this even with Composite disabled.

There's a good shot this is bug #6827.

I've built some rawhide RPMs with that patch included at:

http://people.freedesktop.org/~ajax/rawhide/

Could you test these?
Comment 11 Dennis Jacobfeuerborn 2006-05-09 05:57:12 UTC
Thanks, I've installed the RPMs an will check if Xorg runs more stable.

I can already report that this fixes the following Firefox bug I reported here:
https://bugzilla.mozilla.org/show_bug.cgi?id=334951 (Resized image displayed as
black rectangle)
Comment 12 Adam Jackson 2006-05-10 03:32:47 UTC
(In reply to comment #11)

> I can already report that this fixes the following Firefox bug I reported here:
> https://bugzilla.mozilla.org/show_bug.cgi?id=334951 (Resized image displayed as
> black rectangle)

That's probably justification enough to include this in the release.  If I don't
hear of any crashes from you before this Friday I'll land this for RC3.
Comment 13 Adam Jackson 2006-05-16 04:16:49 UTC
pretty clear that this is a dupe by now.

*** This bug has been marked as a duplicate of 6827 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.