[alsa-devel] Request for HW auto-recognition: HP Pavilion dv7 with STAC92xx

[Takashi Iwai]

At Thu, 13 Dec 2012 14:04:10 +0100,
Tomas Pospisek wrote:
> 
> On Wed, 12 Dec 2012 20:10:23 +0100, Takashi Iwai wrote:
> > At Wed, 12 Dec 2012 18:08:25 +0100, Tomas Pospisek wrote:
> >> 
> >> I have a HP Pavilion dv7 laptop (LS066EA#UUZ), with:
> >> 
> >>   # cat /proc/asound/card0/codec* | grep Codec
> >>   Codec: IDT 92HD81B1X5
> >>   Codec: Intel CougarPoint HDMI
> >> 
> >> when play sound without any tweaks, I get a very weak sound output. If
> I
> >> want to have "beats audio", that is subwoofer and all speakers, I need
> to
> >> set:
> >> 
> >>   # cat /etc/modprobe.d/alsa-base.conf 
> >>   options snd-hda-intel model=hp-dv7-4000
> >> 
> >> Could the snd_hda_intel alsa driver be updated to recognize and setup
> the
> >> driver accordingly?
> >> 
> >> alsa-info.sh output here:
> >>  
> >>
> http://www.alsa-project.org/db/?f=54ea874f43f12c1572aff3ee3113dc4b0ad0b398
> > 
> > Could you check whether the patch below works?
> > 
> > Takashi
> > 
> > ---
> > diff --git a/sound/pci/hda/patch_sigmatel.c
> > b/sound/pci/hda/patch_sigmatel.c
> > index df13c0f..a86547c 100644
> > --- a/sound/pci/hda/patch_sigmatel.c
> > +++ b/sound/pci/hda/patch_sigmatel.c
> > @@ -1725,7 +1725,7 @@ static const struct snd_pci_quirk
> > stac92hd83xxx_cfg_tbl[] = {
> >  	SND_PCI_QUIRK(PCI_VENDOR_ID_HP, 0x1658,
> >  			  "HP", STAC_92HD83XXX_HP_cNB11_INTQUAD),
> >  	SND_PCI_QUIRK(PCI_VENDOR_ID_HP, 0x1659,
> > -			  "HP", STAC_92HD83XXX_HP_cNB11_INTQUAD),
> > +			  "HP Pavilion dv7", STAC_HP_DV7_4000),
> >  	SND_PCI_QUIRK(PCI_VENDOR_ID_HP, 0x165A,
> >  			  "HP", STAC_92HD83XXX_HP_cNB11_INTQUAD),
> >  	SND_PCI_QUIRK(PCI_VENDOR_ID_HP, 0x165B,
> 
> It now indeed does switch on "beats audio" aka subwoofer et al.

OK, I'm going to queue the fix later.

> However now I got a different problem: When I play a video on youtube,
> then
> suspend to RAM, and then resume:
> 
> * if the music was playing when suspending, then the desktop gets stuck
>   after resume:
>   no mouse, no keyboard input gets through to the desktop. I can
>   Ctrl-Alt-F1 out of X and reboot and that's it. I guess it's the firefox
>   flash plugin getting blocked in the kernel while accessing the audio
>   device and thus blocking the whole desktop (that's my speculation).
> 
> * if the music was *not* playing (i.e. video paused), then after resume
>   there is no sound any more. "alsa force-reload" puts the audio in some
>   weird state where it gets in a loop playing some short bit of sound:
>   tak-tak-tak-tak-tak-tak... So again I reboot.

Is it a regression from 3.6 kernel?
If so, could you try to bisect?

Possibly it's a regression due to the runtime PM support.  Other than
that, I have no clue yet.


thanks,

Takashi