Created attachment 137604 [details]
Anywhere between 1-2 hours of running a driving simulation using a proprietary Qt5 based navigation application we're observing the attached GPU hang.
CPU: Intel(R) Atom(TM) CPU E3840 (Baytrail)
Other: Weston 1.0.6 (proprietary DRM rendering backend)
Mesa 13.0.6 (stock i965 driver from this)
The issue wasn't previously seen on our 4.4.x based released, however we were using a mesa 9.x based proprietary user-space driver, and a similar driver in the kernel (though it was essentially a wrapper around i915)
Unfortunately due to the version of Weston we're using we won't be able to easily upgrade mesa from 13.0.6 (it's a binary blob)
Created attachment 137605 [details]
Ugh, it does look to be an unhappy mesa batch. I think you find yourself caught between a rock and a hard place.
(In reply to Chris Wilson from comment #2)
> Ugh, it does look to be an unhappy mesa batch. I think you find yourself
> caught between a rock and a hard place.
Thanks for the quick reply, is there anything I could take a look at here (short of moving forward to mesa 18 and a newer weston)?
James, can you at least test with Mesa 17.2 and 18.0rc4, to see if either of those branches resolves your issue?
-- 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/1697.