Red Hat Bugzilla – Bug 53666
X11 server crashes with signial 11 caught
Last modified: 2007-04-18 12:37:07 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.2) Gecko/20010809
Description of problem:
The X11 server verion 4.x crashes randomly with signal #11 being caught by
the server and forcing the servier to exit.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
2.Wait, sometimes a day or two
3.What your windows disapear
It looks like this problem is coupled to the IRQ of the vga contoller card.
It also looks like an SMP problem. I have found a workaround. Currently I
have a dual intel SMP system which uses the VIA chipset. I observe the
following regarding this bug.
1) Running full SMP, with IRQ enabled in the card and "dri" loaded into X11
server. (This is my default way of running my system.) X11 would crash on
signal 11 when in active use. (i.e. my clicking around mozilla, reading
e-mail, listening to mp3s etc.) When I left the system alone and the screen
saver took over, the system would either hard freeze or the X11 server
2) Running in uniprocessor mode, with IRQ enabled and "dri" loaded, same
3) SMP running with IRQ *disabled* and "dri" not laoded, same result
4) Uniprocessor mode, with IRQ disabled and "dri" not loaded. Works!!!!
I have a strong suggestion. Can you please try running X11 on an SMP
system, and turn on the IRQ and make sure "dri" is loaded. You may be able
to reproduce this bug.
I run X 24/7 on 2 SMP servers and have never had a crash.
I'm unable to do anything at all without proper full bug report data however,
including knowing what video card you're using, as well as config files,
X server logs, kernel "messages" log (trimmed and privatized if necessary)
all submitted as file attachments using the link below.
Please supply requested info.
I'm very swamped with work right now. I'll try and get back to you. Currently
I'm running in non-smp mode, with video card interupts disabled and X11 is very
stable this way. My time scale to getting back to this problem is about 1 month.
Sorry for not being more forth coming with this.
Closing bug, no information to go on. Reopen bug when information
is available if you like.