[Bug 108456] userptr deadlock in Vulkan CTS: mmu_notifier vs workers vs struct_mutex

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=108456

--- Comment #24 from Tvrtko Ursulin <tvrtko.ursulin at linux.intel.com> ---
CI_DRM_5255 or newer runs contain the fix.

You look for the "success" in the summary block in the results email, but it is
not 100% accurate. There are false positives, and of course false negatives as
well, for all the issues which have no sufficient test coverage.

The one my fix handles is unfortunately one of those - which answers the second
part of your question. So you won't see the fail -> pass transition reliably
before and after this fix. It does include new tests so from this point onward
it will be caught.

Note that this is not fixing the userptr problem, just the system hang after
GPU hang which should enable more comfortable testing of Chris' proposed
userptr fixes.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20181204/4429984f/attachment.html>