Bug 51376

Summary: [GM45] crash in brw_update_renderbuffer_surface
Product: Mesa Reporter: Martin Koller <kollix>
Component: Drivers/DRI/i965Assignee: Ian Romanick <idr>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium CC: mgraesslin
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments: Xorg.0.log
/sys/kernel/debug/dri/0/i915_error_state
backtrace of kwin crash

Description Martin Koller 2012-06-24 03:59:09 UTC
Created attachment 63398 [details]
Xorg.0.log

First the problem, environment, hardware below:

I have a crash with KDE's kwin reported here
https://bugs.kde.org/show_bug.cgi?id=302169
and I was told it's a driver issue, therefore reporting it here.

In fact: I was away from the running Laptop, came back after about 30 minutes and saw that kwin had crashed.
I'll attach the complete backtrace. Probably the interesting part is:

Application: KWin (kwin), signal: Segmentation fault
[KCrash Handler]
#6  brw_update_renderbuffer_surface (brw=0x83f14f8, rb=0x826eac0, unit=0) at brw_wm_surface_state.c:959
#7  0xacb89343 in brw_update_renderbuffer_surfaces (brw=0x83f14f8) at brw_wm_surface_state.c:1052
#8  0xacb7347f in brw_upload_state (brw=0x83f14f8) at brw_state_upload.c:503
#9  0xacb5e807 in brw_try_draw_prims (max_index=77, min_index=0, ib=0x0, nr_prims=1, prim=0xbf944f28, arrays=0x8436d58, ctx=0x83f14f8) at brw_draw.c:482
#10 brw_draw_prims (ctx=0x83f14f8, arrays=0x8436d58, prim=0xbf944f28, nr_prims=1, ib=0x0, index_bounds_valid=1 '\001', min_index=0, max_index=77, tfb_vertcount=0x0) at brw_draw.c:566
#11 0xac89e7fe in vbo_draw_arrays (ctx=0x83f14f8, mode=4, start=0, count=78, numInstances=1) at vbo/vbo_exec_array.c:600
#12 0xb648b43d in KWin::GLVertexBufferPrivate::corePainting (this=0x821b5a8, region=..., primitiveMode=4, hardwareClipping=false) at /usr/src/debug/kde-workspace-4.8.90_20120617/kwin/libkwineffects/kwinglutils.cpp:1216

Now when I try to start kwin again and want to activate desktop effects, the crash happens immediately again.

environment, hardware:

I'm on a 32bit openSuse 12.1 on an Esprimo Mobile Laptop
with latest Xorg (7.6) packages installed
(rpms see here https://bugs.kde.org/show_bug.cgi?id=302169#c6 )

kwin output when I start it:
OpenGL vendor string:                   Tungsten Graphics, Inc
OpenGL renderer string:                 Mesa DRI Mobile Intel® GM45 Express Chipset x86/MMX/SSE2
OpenGL version string:                  2.1 Mesa 8.0.3
OpenGL shading language version string: 1.20
Driver:                                 Intel
GPU class:                              i965
OpenGL version:                         2.1
GLSL version:                           1.20
Mesa version:                           8.0.3
X server version:                       1.12.2
Linux kernel version:                   3.1.10
Direct rendering:                       yes
Requires strict binding:                yes
GLSL shaders:                           yes
Texture NPOT support:                   yes

From Xorg log:
X.Org X Server 1.12.2
[    34.385] (II) intel(0): Integrated Graphics Chipset: Intel(R) GM45
[    34.385] (--) intel(0): Chipset: "GM45"
Comment 1 Martin Koller 2012-06-24 04:00:27 UTC
Created attachment 63399 [details]
/sys/kernel/debug/dri/0/i915_error_state
Comment 2 Martin Koller 2012-06-24 04:01:12 UTC
Created attachment 63400 [details]
backtrace of kwin crash
Comment 3 Martin Koller 2012-06-24 04:02:45 UTC
dmesg|grep drm

[    1.918214] [drm] Initialized drm 1.1.0 20060810
[    1.955677] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[    1.955678] [drm] Driver supports precise vblank timestamp query.
[    2.350474] fb: conflicting fb hw usage inteldrmfb vs VESA VGA - removing generic driver
[    2.351126] fbcon: inteldrmfb (fb0) is primary device
[    2.781108] fb0: inteldrmfb frame buffer device
[    2.781110] drm: registered panic notifier
[    2.783392] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[17586.384246] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
[17586.384262] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
[17586.387587] [drm:i915_wait_request] *ERROR* i915_wait_request returns -11 (awaiting 163831 at 163827, next 163832)
Comment 4 Chris Wilson 2012-06-24 04:09:18 UTC
The GPU hang is fixed by xf86-video-intel-2.19.0 and kernel-3.5. However, mesa should not crash...
Comment 5 Eric Anholt 2012-07-30 21:02:56 UTC
Chris, are you still having the 2D driver give us no buffers sometimes when it feels like it?
Comment 6 Eric Anholt 2012-08-13 05:41:47 UTC
chris claims the bug is fixed.

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.