Bug 100518

Summary:	[vulkan] GPU HANG: ecode 7:0:0x86f2fdfd, reason: Hang on render ring, action: reset
Product:	Mesa	Reporter:	github
Component:	Drivers/Vulkan/intel	Assignee:	Intel 3D Bugs Mailing List <intel-3d-bugs>
Status:	RESOLVED MOVED	QA Contact:	Intel 3D Bugs Mailing List <intel-3d-bugs>
Severity:	critical
Priority:	medium	CC:	intel-gfx-bugs, jason
Version:	unspecified
Hardware:	x86-64 (AMD64)
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description github 2017-03-31 16:29:49 UTC

This error occoured to me while playing around with Vulkan.

kernel: [drm] GPU HANG: ecode 7:0:0x86f2fdfd, in triangle [30010], reason: Hang on render ring, action: reset
kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error
kernel: drm/i915: Resetting chip after gpu hang
kernel: drm/i915: Resetting chip after gpu hang

After starting my program (triangle), my Xorg hung for several seconds, then crashed. I saw the above message in my dmesg. Error behaviour differs slightly, sometimes Xorg crashes, most of the time it hangs completely and I have to reboot.

The bug is reproducible under the following conditions:
- create device memory allocation with size=1MiB
- create a depth-stencil image with dimensions 800x600 and format VK_FORMAT_D32_SFLOAT_S8_UINT
- bind the memory allocation (which is obviously too small!) to the DS image
- draw something and watch it failing

If you really need it, I can attach the crashdump but I would like to avoid this since I'm not sure whether there is some sensitive data inside the render ring dump.

The error code sometimes seems to be 7:0:0x85ffffff, but unfortunately I
haven't been able to get a crashdump for this error code, because the machine
locked up and I had to reboot.

CPU is a Core i5-3210M.

Comment 1 Denis 2019-07-03 13:28:06 UTC

hi, could you please re-check on current vulkan version? If this still actual issue - provide please your application so we could reproduce the issue also.

Comment 2 GitLab Migration User 2019-09-18 19:48:22 UTC

-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/828.

Comment 3 Jason Ekstrand 2019-09-19 03:13:06 UTC

One of the core principals of Vulkan is that if you violate valid usage (including binding memory that's too small), you get to keep the pieces.  GPU hangs (of the recoverable kind) are considered a valid failure mode.  This is an app bug not a driver bug.

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.