[Bug 104952] [regression] Loss of i915 (Apollo Lake/Broxton) HDMI audio after display sleep v4.14+

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=104952

Imre Deak <imre.deak at intel.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |imre.deak at intel.com

--- Comment #5 from Imre Deak <imre.deak at intel.com> ---
Created attachment 137189
  --> https://bugs.freedesktop.org/attachment.cgi?id=137189&action=edit
Debug patches

In reply to Simon Casey from comment #3)
> So the issue does seems to be somehow triggered by 42d9366d41a992.

Ok, thanks for the effort.

I still suspect that some unintended dependency on timing causes this. Looking
at azx_rirb_get_response() it's obvious there is some magic involved when
accessing the audio codec. So based on that I'd like to ask you to try a few
debug patches on top of 42d9366d4 I attached to the bug to narrow down the root
cause.

First, could you reproduce the bug and get the stacktrace with
0001-sound-hda-Print-stack-dump-during-probe-failure.patch
?

One thing that 42d9366d4 changes is that intel_wait_for_register() is used
instead of wait_for() to wait for the power well enabled status. This means
we'll potentially complete the wait faster than before, as
intel_wait_for_register() uses a 2 usec atomic wait before doing a sleeping
wait. But that should be fine and nothing should depend on any delays there. To
check for such a dependency, could you try if the problem goes away with
0002-drm-i915-Don-t-spin-wait-during-power-well-enabling.patch
(without the other patches)?

If with the above you can't reproduce the problem, could you try the same with
0002-drm-i915-Don-t-spin-wait-during-power-well-enabling.patch
(again without the other patches)?

Thanks.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20180206/6e229e8b/attachment.html>