At Tue, 3 Dec 2013 22:54:13 +0100, Bruno Prémont wrote:
Hi Takashi,
Since I started using Linux on a MBA2,1 I never had sound working. recently searching for some solution to get sound working I saw that some users had sound with 3.2.x kernels and reported it stopped working somewhere between 3.2 and 3.4.
Thus I tried bisecting this and found two steps of losing sound support, a first one being losing sound output on built-in speaker, a second point where sound through headphone jack stopped working.
Even with 3.2 non of the mute options is working (I did not try recording via build-in microphone)
First commit where speaker stops working: commit 3a93897ea37cbb8277f8a4232c12c0c18168a7db Author: Takashi Iwai tiwai@suse.de Date: Fri Oct 28 01:16:55 2011 +0200
ALSA: hda - Manage unsol tags in hda_jack.c Manage the tags assigned for unsolicited events dynamically together with the jack-detection routines. Basically this is almost same as what we've done in patch_sigmatel.c. Assign the new tag number for each new unsol event, associate with the given NID and the action type, etc. With this change, now all pins looked over in snd_hda_jack_add_kctls() are actually enabled for detection now even if the pins aren't used for jack-retasking by the driver. Signed-off-by: Takashi Iwai <tiwai@suse.de>First commit where headphone stops working (while speaker already does not work anymore): commit 1a97b7f22774b454531f013638b181803fba470f Author: Takashi Iwai tiwai@suse.de Date: Tue Feb 21 11:11:48 2012 +0100
ALSA: hda/realtek - Remove the last static quirks for ALC882 Resitance is futile. The remaining static model quirks for Apple machines with ALC882-compatible codecs are converted to the auto-parser now. We can remove all alc*_quirks.c finally. Signed-off-by: Takashi Iwai <tiwai@suse.de>Attached is the output of `alsa-info --with-devices` under current 3.12.2 kernel and under 3.2.53.
As this all seems related to quirks being replaced with auto-parsing. Trying to revert the commits on top of 3.12.x looks unworkable.
Is there some extra data I should try to provide or a workaround/fix I could try out? I'm booting in EFI mode using EFI stub (did backport it from 3.3 to 3.2 for bisecting/testing 3.2).
You seem to be passing model=3stack, and this must break things. To be sure, try sound.git tree for-linus branch without any options. git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git
If it still doesn't work, please give alsa-info.sh output with that kernel.
Takashi