Bug 54359 - X crashes after resume (8600GT / NV84)
X crashes after resume (8600GT / NV84)
Status: RESOLVED FIXED
Product: xorg
Classification: Unclassified
Component: Driver/nouveau
unspecified
x86-64 (AMD64) Linux (All)
: medium normal
Assigned To: Nouveau Project
Xorg Project Team
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2012-09-01 19:26 UTC by Luca Wehrstedt
Modified: 2012-09-01 22:03 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
The dmesg output (62.14 KB, text/plain)
2012-09-01 19:26 UTC, Luca Wehrstedt
no flags Details
The X.org log of the first crash (73.21 KB, text/plain)
2012-09-01 19:26 UTC, Luca Wehrstedt
no flags Details
The X.org log of subsequent crashes (50.58 KB, text/plain)
2012-09-01 19:27 UTC, Luca Wehrstedt
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Luca Wehrstedt 2012-09-01 19:26:02 UTC
Created attachment 66458 [details]
The dmesg output

I have a NVIDIA GeForce 8600GT graphics card. I'm using Gentoo on a AMD64 machine with a custom built 3.5 kernel.

I've been using nouveau for more than a year by now without any kind of problems. Yet, since I enabled the last kernel options I needed to get 3D hardware acceleration, I'm having some issues.

When closing applications like web browsers or video players the X server suddenly hangs (I cannot even move the pointer) and, after a few seconds, it crashes and I see a TTY console. Then, I think, gdm tries to restart it and after a few seconds I have again some graphics on the screen which often is a corrupted image of what I was seeing before the crash. Then X crashes again, gdm tries to restart it and so on.

When I'm on the TTYs I'm having some keyboard issues (keys double-typed or missed) but otherwise they work fine, so I was able to log in and get some dmesg output and X logs, which I'm attaching.

This issue only appears after I suspend and resume the PC. I can trigger this behavior almost at will after a resume but I'm not able to do it before. It also seems to be related to Xorg drivers and/or 3D, since I didn't encounter it before activating hardware acceleration and it only affects the X server, even after a restart, while having no effect at all on TTYs.

If you need more detailed information just let me know.
Comment 1 Luca Wehrstedt 2012-09-01 19:26:50 UTC
Created attachment 66459 [details]
The X.org log of the first crash
Comment 2 Luca Wehrstedt 2012-09-01 19:27:25 UTC
Created attachment 66460 [details]
The X.org log of subsequent crashes
Comment 3 Sven Joachim 2012-09-01 20:02:16 UTC
This should be fixed in 3.6-rc2 by commit 2064db725cc6d4ea19a24c138bc37939b63e3ae6 (drm/nv86/fifo: suspend fix). Sadly, my attempts to get this into 3.5 have been ignored so far.
Comment 4 Luca Wehrstedt 2012-09-01 22:03:52 UTC
I tried out kernel 3.6_rc3 and indeed it seems the bug is fixed. Thanks!