At Tue, 26 Aug 2008 23:10:55 +0200, Guillaume Chazarain wrote:

On Tue, Aug 26, 2008 at 7:59 AM, Takashi Iwai tiwai@suse.de wrote:

...

Are you sure both are the same kernel but just one commit reverted?

I retested, yes.

Thanks for confirmation!

With or without the revert, model=auto disables sound completely, and I get this in dmesg:

HDA Intel 0000:00:1b.0: PCI INT A disabled HDA Intel 0000:00:1b.0: PCI INT A -> GSI 20 (level, low) -> IRQ 20 HDA Intel 0000:00:1b.0: setting latency timer to 64 ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:2325: hda_codec: model 'auto' is selected ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:3021: autoconfig: line_outs=1 (0x14/0x0/0x0/0x0/0x0) ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:3025: speaker_outs=0 (0x0/0x0/0x0/0x0/0x0) ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:3029: hp_outs=1 (0x15/0x0/0x0/0x0/0x0) ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:3030: mono: mono_out=0x0 ALSA /home/g/linux-2.6/sound/pci/hda/hda_codec.c:3038: inputs: mic=0x18, fmic=0x0, line=0x1a, fline=0x0, cd=0x0, aux=0x0

I played with model=intel-mac-auto and model=intel-mac-v3 but it had no effect, i.e. same as without the model= parameter.

So, without model=auto the problem is that I get the output on both the internal speakers and the headphones, and this is fixed by reverting the commit.

Then this is really a PCI SSID conflict, and reverting the patch is the right fix (and possibly addition of check of codec SID).

I fixed it now on my tree, and will include it in the next pull request.

thanks,

Takashi