Bug 95050 - i5-3230M: GPU HANG: ecode 7:-1:0x00000000, reason: Ring hung, action: reset
Summary: i5-3230M: GPU HANG: ecode 7:-1:0x00000000, reason: Ring hung, action: reset
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-04-21 13:06 UTC by blaffablaffa
Modified: 2017-07-24 22:42 UTC (History)
1 user (show)

See Also:
i915 platform: IVB
i915 features: GPU hang


Attachments
Xorg log and gpu dump (293.75 KB, text/plain)
2016-04-21 13:06 UTC, blaffablaffa
no flags Details

Description blaffablaffa 2016-04-21 13:06:07 UTC
Created attachment 123115 [details]
Xorg log and gpu dump

Using Mathematica to make a lot of plots concurrently. Apparently each subkernel creates a new window for each plot (I can see the window tab flicker). kernel 4.4.5 x86_64 Xorg 1.18.3 mesa 11.1.0 libdrm 2.4.66 xorg-x11-drv-intel 2.99.917
Comment 1 Chris Wilson 2016-04-21 13:15:34 UTC
Well that is curious. The GPU completed all the work we told it do, but we are waiting upon a batch that doesn't seem to have been submitted to the ring.
Comment 2 Chris Wilson 2016-04-21 13:24:35 UTC
It's not a direct match, but I could believe that

commit aa9b78104fe3210758fa9e6c644e9a108d371e8b
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Apr 13 17:35:15 2016 +0100

    drm/i915: Late request cancellations are harmful

is relevant here. Could you run a kernel from https://cgit.freedesktop.org/drm-intel/ #drm-intel-nightly or equivalent package?
Comment 3 blaffablaffa 2016-04-21 13:38:52 UTC
it just happened once in a series of run of the same thing and I've never compiled the kernel myself, but I'll see what i can do.
Comment 4 blaffablaffa 2016-04-21 13:44:00 UTC
mind that while the scree nwas frozen I hit sysrq-f (oom) a few times. Other times when I did so, sometimes the message "gpu: cannot release memory because of lock contention". in  case it's relevant here.
Comment 5 blaffablaffa 2016-05-05 14:17:39 UTC
I'm failing to reproduce this bug with the very same setup, so i wouldn't be able to conclude anything trying the testing kernel. I'll try when and if it happens again.
Comment 6 yann 2016-06-17 09:37:54 UTC
Reopen the bug if this is still occuring, even after applying Chris' patch


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.