At Sat, 26 Jan 2008 00:51:58 +1100, Rick Measham wrote:
Takashi Iwai wrote:
After adding model=auto, then reboot and get a proc file. The result you attached is the overridden one.
Hmm .. that's what I did ..
OK, then it's indeed so. Interesting that the pincfg is identical with the *modified* preset.
BTW, please update to the very latest HG version. I see that you are still using the older version. The newer version should have more detailed proc output.
It's 24 hours old (or maybe 48?) but OK, I'll do that
Hm... is it from hg.alsa-project.org?
The newest changeset is: changeset: 5771:19029da61288 tag: tip user: tiwai date: Fri Jan 25 11:56:57 2008 +0100 summary: hda-codec - Add missing slave for AD1884 master switch
Check "hg log" whether it matches.
If I understand correctly, the speaker output on this laptop has never worked, so far, right?
Never. For anyone using Linux :-/
Then this is at least no regression. A good thing (let's think positively :)
Seriously, another missing thing can be either EAPD on this widget or EAPD via GPIO. In the former case, the patch like below might fix. In the latter case, you'd need to add the following line appropriately in patch_stac922x():
spec->gpio_mask = spec->gpio_data = 0x01;
The value 0x01 can vary depending on the machine implementation. Try different values if the above doesn't work.
Takashi
---
diff -r 19029da61288 pci/hda/patch_sigmatel.c --- a/pci/hda/patch_sigmatel.c Fri Jan 25 11:56:57 2008 +0100 +++ b/pci/hda/patch_sigmatel.c Fri Jan 25 15:18:03 2008 +0100 @@ -610,6 +610,7 @@ static struct hda_verb stac922x_core_ini static struct hda_verb stac922x_core_init[] = { /* set master volume and direct control */ { 0x16, AC_VERB_SET_VOLUME_KNOB_CONTROL, 0xff}, + { 0x0d, AC_VERB_SET_EAPD_BTLENABLE, 0x02 }, /* EAPD on speaker */ {} };