[alsa-devel] ALC892, Clevo P150HM: Input Source gets reset to "Mic" every resume

[Julian Sikorski]

W dniu 24.02.2012 10:43, Takashi Iwai pisze:
> At Thu, 23 Feb 2012 20:11:44 +0100,
> Julian Sikorski wrote:
>>
>> W dniu 09.12.2011 14:32, Raymond Yau pisze:
>>> 2011/12/4 Julian Sikorski <belegdol at gmail.com>:
>>>> Hi,
>>>>
>>>> whenever I resume the machine, the Input Source would get changed back
>>>> from "Internal Mic" to "Mic". This is a bit annoying, because pulseaudio
>>>> is unaware of that change. As a result, everything looks fine in
>>>> pavucontrol but no sound gets recorded. A quick look at alsamixer -c0
>>>> revealed the problem. My alsa info:
>>>>
>>>> http://www.alsa-project.org/db/?f=c5776b615a5f3b9ae624e4250fb0c2b69118ce26
>>>>
>>>> This is unrelated to the other problems I have been reporting, since
>>>> this happens on a mostly-stock Fedora 3.1.2 kernel (the patch I added
>>>> deals with xhci, not alsa).
>>>
>>> The problem seem can be reproduced by using hda-emu,
>>> it is init_capsrc_for_pin() and alc_auto_init_input_src() write value
>>> to cache before snd_hda_codec_resume_app() from cache
>>>
>>>
>>>> get 28
>>> 28 Input Source:0
>>> ITEM: 0:Mic, 1:Internal Mic, 2:Line, VAL: [Mic]
>>>> set 28 1
>>> send: NID=0x23, VERB=0x360(set_amp_gain_mute,I:L#0), PARM=0x80
>>> send: NID=0x23, VERB=0x350(set_amp_gain_mute,I:R#0), PARM=0x80
>>> send: NID=0x23, VERB=0x361(set_amp_gain_mute,I:L#1), PARM=0x0
>>> send: NID=0x23, VERB=0x351(set_amp_gain_mute,I:R#1), PARM=0x0
>>>> get 28
>>> 28 Input Source:0
>>> ITEM: 0:Mic, 1:Internal Mic, 2:Line, VAL: [Internal Mic]
>>>
>>> pm
>>> ** SUSPENDING **
>>> ...
>>> ** RESUMING **
>>> PM-Notified
>>> ...
>>> exec_init_verbs
>>> ...
>>>
>>> init_capsrc_for_pin 18
>>> send: NID=0x23, VERB=0xb20(get_amp_gain_mute,I:L#0), PARM=0x0
>>> receive: 0x0
>>> send: NID=0x23, VERB=0xb00(get_amp_gain_mute,I:R#0), PARM=0x0
>>> receive: 0x0
>>> init capsrc for pin 19
>>> send: NID=0x23, VERB=0xb20(get_amp_gain_mute,I:L#0), PARM=0x1
>>> receive: 0x80
>>> send: NID=0x23, VERB=0x361(set_amp_gain_mute,I:L#1), PARM=0x0
>>> send: NID=0x23, VERB=0xb00(get_amp_gain_mute,I:R#0), PARM=0x1
>>> receive: 0x80
>>> send: NID=0x23, VERB=0x351(set_amp_gain_mute,I:R#1), PARM=0x0
>>> init capsrc for pin 1a
>>> send: NID=0x23, VERB=0xb20(get_amp_gain_mute,I:L#0), PARM=0x2
>>> receive: 0x80
>>> send: NID=0x23, VERB=0x362(set_amp_gain_mute,I:L#2), PARM=0x0
>>>
>>> alc_auto_init_input_src
>>> send: NID=0x23, VERB=0x361(set_amp_gain_mute,I:L#1), PARM=0x80
>>> send: NID=0x23, VERB=0x351(set_amp_gain_mute,I:R#1), PARM=0x80
>>> send: NID=0x23, VERB=0x362(set_amp_gain_mute,I:L#2), PARM=0x80
>>> send: NID=0x23, VERB=0x352(set_amp_gain_mute,I:R#2), PARM=0x80
>>>
>>> snd_hda_codec_resume_app
>>> send: NID=0x23, VERB=0x363(set_amp_gain_mute,I:L#3), PARM=0x80
>>> send: NID=0x23, VERB=0x353(set_amp_gain_mute,I:R#3), PARM=0x80
>>> send: NID=0x23, VERB=0x364(set_amp_gain_mute,I:L#4), PARM=0x80
>>> send: NID=0x23, VERB=0x354(set_amp_gain_mute,I:R#4), PARM=0x80
>>> send: NID=0x23, VERB=0x365(set_amp_gain_mute,I:L#5), PARM=0x80
>>> send: NID=0x23, VERB=0x355(set_amp_gain_mute,I:R#5), PARM=0x80
>>> send: NID=0x23, VERB=0x366(set_amp_gain_mute,I:L#6), PARM=0x80
>>> send: NID=0x23, VERB=0x356(set_amp_gain_mute,I:R#6), PARM=0x80
>>> send: NID=0x23, VERB=0x367(set_amp_gain_mute,I:L#7), PARM=0x80
>>> send: NID=0x23, VERB=0x357(set_amp_gain_mute,I:R#7), PARM=0x80
>>> send: NID=0x23, VERB=0x368(set_amp_gain_mute,I:L#8), PARM=0x80
>>> send: NID=0x23, VERB=0x358(set_amp_gain_mute,I:R#8), PARM=0x80
>>> send: NID=0x23, VERB=0x369(set_amp_gain_mute,I:L#9), PARM=0x80
>>> send: NID=0x23, VERB=0x359(set_amp_gain_mute,I:R#9), PARM=0x80
>>> send: NID=0x23, VERB=0x36a(set_amp_gain_mute,I:L#10), PARM=0x80
>>> send: NID=0x23, VERB=0x35a(set_amp_gain_mute,I:R#10), PARM=0x80
>>> ...
>>>
>>>> get 28
>>> 28 Input Source:0
>>> ITEM: 0:Mic, 1:Internal Mic, 2:Line, VAL: [Mic]
>> The problem still exists in 3.2.7-1.fc16.x86_64 kernel, is there any
>> chance for a fix?
> 
> Does the patch below fix?
> 
> 
> Takashi

Yes, it seems so. I don't want to get ahead of myself, but two
suspend-resume cycles have passed and inputs are at Internal Mic and
Mic, as expected.

Julian