[alsa-devel] 1.0.16->1.0.17 regression for some HDA NVidia's

[Ozan Çağlayan]

Takashi Iwai wrote:
> At Thu, 10 Jul 2008 19:41:45 +0300,
> Ozan Çağlayan wrote:
>   
>> Takashi Iwai wrote:
>>     
>>> At Tue, 08 Jul 2008 00:31:43 +0300,
>>> Ozan Çağlayan wrote:
>>>   
>>>       
>>>> Takashi Iwai wrote:
>>>>     
>>>>         
>>>>> At Sun, 06 Jul 2008 19:18:01 +0300,
>>>>> Ozan Çağlayan wrote:
>>>>>   
>>>>>       
>>>>>           
>>>>>> Everything seems OK with those onboard audio chipsets but there's 
>>>>>> absolutely no sound.
>>>>>> We have 2 bug reports on our bug tracking system indicating this 
>>>>>> regression for MCP51, MCP61 and MCP73.
>>>>>>
>>>>>> Here are the URL's for alsa-info.sh output's:
>>>>>>
>>>>>> MCP51: http://pastebin.ca/1062418
>>>>>> MCP61: http://pastebin.ca/1058482
>>>>>> MCP73: http://pastebin.ca/1058319
>>>>>>
>>>>>> Those users have recently reported that downgrading to 1.0.16 series 
>>>>>> solves the problem.
>>>>>>     
>>>>>>         
>>>>>>             
>>>>> Could you run alsa-info.sh on 1.0.16, too?  Then we can compare
>>>>> between working and non-working states.
>>>>>
>>>>>
>>>>> thanks,
>>>>>
>>>>> Takashi
>>>>>   
>>>>>       
>>>>>           
>>>> Hi,
>>>>
>>>> I've also entered a bug report for this problem which includes the two 
>>>> outputs:
>>>>
>>>> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4038
>>>>     
>>>>         
>>> The first thing to check is that you are using pulse plugin.
>>> Is it intended, and confirmed to work?
>>>
>>> The next thing to check is whether 1.0.16 driver works with the same 
>>> environment (1.0.17 lib, utils, etc) -- that is, your 1.0.16 alsa-info
>>> output.  If yes, get alsa-info.sh output again, then install 1.0.17
>>> driver again and re-test.  Get alsa-info.sh output on 1.0.17 again
>>> (for non-working) to compare exactly.
>>>   
>>>       
>> First the problem is not present when pulseaudio is stopped. And also, 
>> we can get sound with 1.0.17rc1 too. So briefly:
>>
>> 1.0.16 + pulse -> works well
>> 1.0.17rc1 + pulse -> works well
>> 1.0.17rc1,rc2 + pulse -> No sound.
>>
>> It seems that something is broken for these chipsets in either alsa or 
>> pulse tree. Any ideas for your alsa-{driver,lib,plugins} part?
>>     
>
> Try to Compare alsa-info.sh outputs between these versions.
> If there is a significant difference in codec#* proc files, then
> something was changed in the driver side.  If not, it's pulse
> problem.
>   
It seems that we've found the problem.
MCP51 and MCP61 works well when the 1.0.17 introduced module parameter 
bdl_pos_adj is explicitly set to 0 when probing the snd-hda-intel module.

-- 

Ozan Çağlayan
<ozan_at_pardus.org.tr>