Bug 17767

Summary: Random crashes on Mobility x1400
Product: xorg Reporter: Friedrich Gräter <graeter>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED DUPLICATE QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
X.org.0.log none

Description Friedrich Gräter 2008-09-25 03:52:25 UTC
My X.org crashes randomly when using the GIT version of the radeon driver with my Mobility x1400 and compiz. Mostly it crashes when using Inkscape.

Today it crashed, when I tried to increase the font size of Xfce. Unfortunately mouse and keyboard got locked and I had no access to another computer to save the dmesg log via ssh. However, I have these two logging messages in /var/log/messages about the crash:

Sep 25 20:29:57 friedi [drm:r300_emit_carefully_checked_packet0] *ERROR* Offset failed range check (reg=4540 sz=1)
Sep 25 20:29:57 friedi [drm:r300_do_cp_cmdbuf] *ERROR* r300_emit_packet0 failed
Sep 25 20:29:57 friedi [drm] Num pipes: 1
Comment 1 Alex Deucher 2008-09-25 07:28:11 UTC

*** This bug has been marked as a duplicate of bug 17723 ***
Comment 2 Michel Dänzer 2008-09-25 07:57:05 UTC
The symptoms indeed look similar to bug 17723, but X1400 doesn't sound like something that would match my analysis from there... Friedrich, please attach your full Xorg.0.log file.
Comment 3 Alex Deucher 2008-09-25 08:02:55 UTC
I was thinking perhaps this related to the latest changes in mesa.  Friedrich, what versions of mesa and the drm are you using?
Comment 4 Michel Dänzer 2008-09-25 08:10:34 UTC
(In reply to comment #3)
> I was thinking perhaps this related to the latest changes in mesa.

The submitter of bug 17723 has tried down to the Mesa 7.2 release with no change, and the Mesa r300 driver warning he's seeing indicates the invalid texture offset can't originate from the r300 driver, so I suspect it's a bogus offset getting passed from the 2D driver to the 3D driver for zero-copy texture-from-pixmap. I think it should only be possible for that to happen if a pixmap can't fit into the EXA offscreen area, unless maybe this is an xf86-video-ati Git regression?
Comment 5 Alex Deucher 2008-09-25 14:25:42 UTC
Can you try xf86-video-ati 6.9.0 rather than git and see if it works there with 32 MB of vram?  If 6.9.0 works, can you git bisect to see where it broke?
Comment 6 Friedrich Gräter 2008-09-25 18:34:31 UTC
I currently use mesa 7.1_rc3 and the x11-drm GIT version (18.09.2008, 10:23) and the xf86-video-ati GIT version (18.9.2008 10:21). I have these package from the gentoo portage (x11 overlay). Im using the X.org server 1.4.99.905.

The problem ist, that I can't really reproduce these crashes. I had them also with 6.9.0 (most time when using Inkscape - probably when working with larger vector graphics). But most times, I got a full system lock-up, so I have no further logs. 

There is another strange phenomenon: sometimes when using Inkscape Compiz uses its window decorators and got other strange graphical problems (but I don't know whether it is a bug of Compiz, Inkscape or X11).

I've attached the log file, too. 

Thank you so far!
Comment 7 Friedrich Gräter 2008-09-25 18:35:07 UTC
Created attachment 19217 [details]
X.org.0.log
Comment 8 Michel Dänzer 2008-10-21 10:14:21 UTC
Friedrich, do you still get the same kernel messages when this happens? If so, is there also something in the X server's stderr output? If you're using a display manager like gdm, the X server stderr should be captured in its log file.

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.