HDA recording

[Wes Morgan]

On Sun, 22 Jul 2007, Ariff Abdullah wrote:

> On Sat, 21 Jul 2007 15:52:09 -0400 (EDT)
> Wes Morgan <morganw at chemikals.org> wrote:
>> On Sun, 22 Jul 2007, Ariff Abdullah wrote:
>>
>>> On Sat, 21 Jul 2007 14:40:29 -0400 (EDT)
>>> Wes Morgan <morganw at chemikals.org> wrote:
>>>> On Sat, 21 Jul 2007, Ariff Abdullah wrote:
>>>>
>>>>> On Sat, 21 Jul 2007 09:47:00 -0400 (EDT)
>>>>> Wes Morgan <morganw at chemikals.org> wrote:
>>>>>> Is recording via HDA working for anyone? When last I tried,
>>> back >> in > May, I  was able to record. However I'm not sure how
>>> old the >> kernel > I was running  was... As it stands now, I
>>> reverted hda >> back to > 20070320_0043 (not the  pcm driver
>>> though) and it still >> doesn't > work. I tried rolling back the
>>> entire sound >> infrastructure but it > would not compile.
>>>>>>
>>>>>> The error I'm seeing is:
>>>>>>
>>>>>> pcm0:virtual:dsp0.vr0: record interrupt timeout, channel dead
>>>>>>
>>>>>> /dev/sndstat is:
>>>>>>
>>>>>> FreeBSD Audio Driver (newpcm: 32bit 2007061600/i386)
>>>>>> Installed devices:
>>>>>> pcm0: <Intel 82801G High Definition Audio Controller> at memory
>>>>>> 0x80000000  irq 22 kld snd_hda [20070710_0047] [MPSAFE]
>>>>> (1p:1v/1r:1v > channels duplex  default)
>>>>>>
>>>>>
>>>>> Try disabling vrec. dev.pcm.%d.rec.vchans=0
>>>>
>>>> Now I get:
>>>>
>>>> pcm0:record:dsp0.r0: record interrupt timeout, channel dead
>>>>
>>>> instead.
>>>
>>> Try this patch:
>>>
>>> http://people.freebsd.org/~ariff/test/channel.c.diff
>>>
>>>
>>> else, try enabling polling mode.
>>
>> Still get timeouts, with polling on, vchan recording off,
>> with/without the  patch, and all permutations of the three I could
>> think of.
>
> No idea. Let see..
>
> 1) Try enabling msi with hint.pcm.%d.msi=1
> 2) Try releng6, where it is left unchanged as a control specimen.
> 3) Try moving backwards (not just sound), grab older sources, etc,
>   older cd snapshot,
> 4) ...

Ok, I've tried 1, no help. (3) is harder because of the gcc 4.2 import, 
but I tried all the kernel snapshots available for -current, none of which 
worked. I'll give 6.2 a try later.