[Bug 110848] Everything using GPU gets stuck after running+killing parallel Media loads (after running 3D benchmarks)

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=110848

--- Comment #35 from Chris Wilson <chris at chris-wilson.co.uk> ---
(In reply to David Weinehall from comment #34)
> Created attachment 144960 [details]
> Netconsole log with drm-tip from 2019y-08m-05d-17h-28m-24s

At the instant of that hungtask, it was waiting for

[ 2582.003387] R13: 0000000000000000 R14: 00007fc30c02be50 R15:
0000000000000008
[ 2582.003405] h264parse0:sink R  running task        0  6195      1 0x00004004
[ 2582.003418] Call Trace:
[ 2582.003429]  ? mark_held_locks+0x49/0x70
[ 2582.003437]  ? trace_hardirqs_on_thunk+0x1a/0x20
[ 2582.003445]  ? lockdep_hardirqs_on+0xeb/0x180
[ 2582.003454]  ? trace_hardirqs_on_thunk+0x1a/0x20
[ 2582.003465]  ? retint_kernel+0x2b/0x2b
[ 2582.003481]  ? drm_mm_scan_remove_block+0x23/0x50
[ 2582.003543]  ? i915_gem_evict_something+0x2fc/0x4e0 [i915]
[ 2582.003601]  ? i915_gem_gtt_insert+0x1ce/0x270 [i915]
[ 2582.003655]  ? __i915_vma_do_pin+0x40c/0x530 [i915]
[ 2582.003707]  ? i915_gem_object_ggtt_pin+0x11f/0x1e0 [i915]
[ 2582.003755]  ? i915_gem_fault+0x291/0x6d0 [i915]
[ 2582.003787]  ? __do_fault+0x31/0xb0
[ 2582.003801]  ? __handle_mm_fault+0x90d/0xfc0
[ 2582.003820]  ? handle_mm_fault+0x15a/0x360
[ 2582.003832]  ? __do_page_fault+0x24e/0x500
[ 2582.003845]  ? page_fault+0x34/0x40

Other than complaining about the stupidity of using mmap_gtt, I guess it means
we were badly fragmented and eviction was slow.

-- 
You are receiving this mail because:
You are the assignee for the bug.
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/20190806/6902605f/attachment.html>