[alsa-devel] [PING!] Re: HDMI audio is not working on my new desktop

[Alexander E. Patrakov]

2013/9/5 Alexander E. Patrakov <patrakov at gmail.com>:
> 2013/8/22 Alexander E. Patrakov <patrakov at gmail.com>:
>> 2013/8/22 Wang Xingchao <wangxingchao2011 at gmail.com>:
>>> cc Mengdong and Liam, we're in the same PRC audio Team.
>>>
>>> 2013/8/21 Alexander E. Patrakov <patrakov at gmail.com>:
>>>> 2013/8/21 Wang Xingchao <wangxingchao2011 at gmail.com>:
>>>>> Hi Takashi/Alexander E. Patrakov,
>>>>>
>>>>> 2013/8/20 Takashi Iwai <tiwai at suse.de>:
>>>>>> Xingchao, could you take a look at this bug?
>>>>>
>>>>> sure, i will take a look at it.
>>>>> But i'm afraid i cannot continue tracking from next week, i'm under
>>>>> leaving process at Intel.
>>>>
>>>> Best wishes to you, and good luck with a new job :)
>>>>
>>>> But then please tell us the contact information of the person at Intel
>>>> who should be contacted instead of you for any further questions
>>>> related to HDMI audio on Haswell.
>>>>
>>>>> At first glance, i saw a lot of errors:
>>>>>
>>>>> [  772.822663] playback write error (DMA or IRQ trouble?)
>>>>> [  782.912776] playback write error (DMA or IRQ trouble?)
>>>>> [  793.002904] playback write error (DMA or IRQ trouble?)
>>>>> [  803.093052] playback write error (DMA or IRQ trouble?)
>>>>> [  813.183205] playback write error (DMA or IRQ trouble?)
>>>>> [  823.273217] playback write error (DMA or IRQ trouble?)
>>>>>
>>>>> I know this issue on some Haswell ULT boards(STP/WTM1) because they
>>>>> are not stable, and it's better to use latest BIOS image(V131 or
>>>>> later).
>>>
>>> btw, did you do any suspend/resume operation before the audio playback test?
>>> if not, it's a new case. i found the issue only on some particular
>>> boards after suspend/resume.
>>
>> The issue is 100% reproducible without suspend/resume. Just in case, I
>> tested after one suspend-resume cycle, it didn't fix the problem.
>
> On IRC (#intel-gfx) ohsix said that the bug
> (https://bugzilla.kernel.org/show_bug.cgi?id=60769) may be graphics
> related after all, but was too scared with the interrupt handling code
> there. The contents of /proc/interrupts are attached to the bug, but
> he said there is nothing interesting there.
>
> BTW I am surprised that the bug is still assigned to Jaroslav Kysela
> and not to Intel specialists (who are not even in the CC list of the
> bug).
>
> --
> Alexander E. Patrakov

broonie at kernel.org asked to CC: him to the thread. As far as I know,
Intel tried to reproduce this bug on their hardware and on some OEM
boards, but didn't try exactly the same motherboard, and reproduced
another bug (files play silently but IRQs seem to come correctly).

It is interesting to know that today I found two people on FreeNode
IRC with the same board, and only one of them (rvalles) can reproduce
the IRQ problem. The other one (broonie) uses an older kernel
(Debian's 3.10-3, current unstable) and sees the "files play silently"
problem instead of the IRQ one.

-- 
Alexander E. Patrakov