Re: snd-hda-intel AD1981 - the microphone doesn't work

[Takashi Iwai <tiwai@suse.de>]

At Tue, 19 Feb 2008 15:12:50 +0100,
Grzegorz Chwesewicz wrote:
> 
> Thursday 14 of February 2008 13:14:18 Takashi Iwai napisał(a):
> > And, does a similar problem like Grzegorz's case occur on your laptop?
> > Some HP laptops (so far, AMD SB450-based) occasionally cause a hang-up
> > during recording from the mic when HD is accessed.  This is far more
> > nasty problem, and it's interesting to know whether this happens on
> > other chipset.
> 
> 	Hi Takashi. I've experimented a little with power saving in HD Intel Audio. 
> I've noticed that using power saving with timeout set to 10 sec. solves my 
> problem partially. When arecord stops recording I can turn it off, wait 10 
> seconds, and recording works back again. This is obvious, because power 
> saving simply powers off and on device. I was wondering that maybe we can 
> implement ioctl or something that will dump "state" (informations that will 
> help in debuging) of device, when recording is not working, and "state" just 
> after power on to compare what causes breakage ? Can you give me some tips 
> about what informations are userful in debuging ?

Well, the question is what "state".  Interesting thing is that it's
specific to ATI SB, and seems irrelevant with codec chip.  So, the
codec information is likely useless.  But, I don't think the broken
situation is reflected on PCI reg space.

BTW, Andi Kleen also has a similar problem on his machine and he used
noirqdebug boot option.  But, in his case, the irq line of HD audio
device isn't shared with others.


Takashi