[Bug 111937] [CI][BAT] igt at i915_selftest@live_execlists - incomplete - GEM_BUG_ON(i915_active_is_idle(&ce->active))

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=111937

--- Comment #4 from Chris Wilson <chris at chris-wilson.co.uk> ---
(In reply to Chris Wilson from comment #2)
> During reset, the context idled. Which also shouldn't have happened -- I
> think the engine parked, but we are after the set-wedged (and after the
> GEM_TRACE) so it is retired immediately. Hmm. Seems possible.

intel_gt_set_wedged() takes the intel_engine_pm before marking it as wedged, so
we should not be able to discard the idle barrier so easily. Except for

switch_to_kernel_context:
        rq = __i915_request_create(engine->kernel_context, GFP_NOWAIT);
        if (IS_ERR(rq))
                /* Context switch failed, hope for the best! Maybe reset? */
                goto out_unlock;

Otherwise there should have been an idle request in flight keeping the retired
context active.

It might interesting to stick an intel_engine_pm_is_awake() into the reset
code, but first need to think how that upsets the replay or note.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee 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/20191010/a6647602/attachment-0001.html>