[Bug 105128] [CI] igt@* - dmesg-warn - *ERROR* dp aux hw did not signal timeout!

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=105128

--- Comment #17 from Matt Roper <matthew.d.roper at intel.com> ---
This failure seems to happen a lot more frequently on TGL than it did on
SKL/KBL, and not just on suspend/resume like it did before.  AFAICS, it's still
an "impossible" case if the hardware is actually behaving properly; even in
legitimate timeout cases the hardware is supposed to recognize the timeout and
clear the busy bit after a specific amount of time (gen12 has a 4ms hardware
timeout, gen6-11 had a 1.5ms timeout), and we only declare this timeout failure
and print the message after 10ms have elapsed.

I think the next step is to see if we're getting AUX completion interrupts even
though the "BUSY" flag never turns off like it's supposed to.  I'll probably
send a patch tomorrow to try to recognize cases where we received an aux
channel interrupt but still see the aux as busy; that should give us more
insight into why/how the hardware is misbehaving.  Then we can bring that data
to the hardware guys and see if they can think of any reason for this behavior
or have suggestions on how best to work around it.

-- 
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/20191024/915d3a21/attachment.html>