Bug 80802

Summary: [HSW] GPU error state gathered while decoding 4K video with VA-API and displaying via an OpenGL compositor
Product: DRI Reporter: Simon Farnsworth <simon>
Component: DRM/IntelAssignee: Mika Kuoppala <mika.kuoppala>
Status: CLOSED DUPLICATE QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: intel-gfx-bugs
Version: unspecified   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Compressed error state none

Description Simon Farnsworth 2014-07-02 11:04:41 UTC 
I've no idea what's caused this error state, hence filing against the kernel as a placeholder.

I have code using VA-API to decode a 4K video and render to an X11 window with VA-API/X11. I then have an OpenGL compositor downscaling this window to 1920x1200 for scanout. Something in this stack is causing the GPU to hang and generate an error state.

I'm running Fedora 20 x86_64 with patches on a Haswell Intel(R) Core(TM) i3-4330 CPU.

Software versions are (I can supply SRPMs and x86_64 binary RPMs for all of this if needed):

xf86-video-intel: xorg-x11-drv-intel-2.21.15-5.fc20.x86_64
xserver: 1.14.4 patched to support high dotclocks on HDMI.
mesa: mesa-dri-drivers-10.1.3-1.20140509.fc20.x86_64
libdrm: libdrm-2.4.54-1.fc20.x86_64
kernel: kernel-3.14.4 patched with extras from bug #75502.

I'm connected via a DP to VGA adapter at the time:
xrandr 
Screen 0: minimum 320 x 200, current 1920 x 1200, maximum 8192 x 8192
VGA1 disconnected (normal left inverted right x axis y axis)
HDMI1 disconnected (normal left inverted right x axis y axis)
DP1 connected 1920x1200+0+0 (normal left inverted right x axis y axis) 518mm x 324mm
   Fallback       60.0 +
   1920x1200      60.0*+
   NTB-vesa_cvt-3840x2160   60.0  
   1600x1200      60.0  
   1280x1024      60.0  
   1024x768       60.0  
   800x600        60.3  
   640x480        60.0  
HDMI2 disconnected (normal left inverted right x axis y axis)

I'll attach a compressed error state shortly.
Comment 1 Simon Farnsworth 2014-07-02 11:05:47 UTC 
Created attachment 102128 [details]
Compressed error state

This is the error state; dmesg to go with it says:

[ 2964.222067] [drm] stuck on render ring
[ 2964.222090] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[ 2964.222102] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[ 2964.222114] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[ 2964.222119] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[ 2964.222125] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Comment 2 Chris Wilson 2014-07-02 11:12:44 UTC 

*** This bug has been marked as a duplicate of bug 77104 ***

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.