[alsa-devel] Getting board to work on 2.6.30

Tony Hoyle tmh at nodomain.org
Tue Jun 23 13:21:11 CEST 2009


I'm trying to get sound working on a board I'm working with.  The sound 
works correctly on but not 2.6.30.

I has an intel HDA (Poulson) chipset and a sigmatel 9202 codec. 
Physically there's only one output - the speakers, and no input, so from 
a testing point of view it either works or it doesn't.

Kernel output from both looks similar:

[   19.942029] ALSA sound/pci/hda/hda_intel.c:916: HDA snoop disabled, 
try to enable ... OK
[   20.157458] ALSA sound/pci/hda/hda_codec.c:2760: autoconfig: 
line_outs=1 (0xd/0x0/0x0/0x0/0x0)
[   20.157467] ALSA sound/pci/hda/hda_codec.c:2764:    speaker_outs=0 
[   20.157474] ALSA sound/pci/hda/hda_codec.c:2768:    hp_outs=1 
[   20.157481] ALSA sound/pci/hda/hda_codec.c:2776:    inputs: mic=0x0, 
fmic=0xb, line=0xc, fline=0x0, cd=0x11, aux=0x0

2.6.30 (running alsa snapshot from last night):

    19.379811] ALSA sound/pci/hda/hda_codec.c:3795: autoconfig: 
line_outs=1 (0x10/0x0/0x0/0x0/0x0)
[   19.379824] ALSA sound/pci/hda/hda_codec.c:3799:    speaker_outs=2 
[   19.379833] ALSA sound/pci/hda/hda_codec.c:3803:    hp_outs=1 
[   19.379842] ALSA sound/pci/hda/hda_codec.c:3804:    mono: mono_out=0x10
[   19.379849] ALSA sound/pci/hda/hda_codec.c:3807:    dig-out=0x8/0x0
[   19.379856] ALSA sound/pci/hda/hda_codec.c:3815:    inputs: mic=0xb, 
fmic=0x0, line=0xc, fline=0x0, cd=0x11, aux=0x0
[   19.390010] input: HDA Intel MID CD at Ext Rear Jack as 
[   19.401109] input: HDA Intel MID Line In at Ext Rear Jack as 
[   19.414726] input: HDA Intel MID Mic at Ext Rear Jack as 
[   19.430427] input: HDA Intel MID Line Out at Ext Rear Jack as 
[   19.447561] input: HDA Intel MID HP Out at Ext Front Jack as 

It seems to be creating the outputs as inputs?  This seems wrong.

I've tried every model setting I could find and reckon I need to create 
a new one.. but don't actually know what to create to get things working 
- there are lots of differences in the alsa between the two kernels and 
it isn't something that looks easy to pin down short of going nuts with 
git bisect.

Can someone give me a pointer where to start looking at this?

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: codec#0
Url: http://mailman.alsa-project.org/pipermail/alsa-devel/attachments/20090623/959a0a65/attachment.bat 

More information about the Alsa-devel mailing list