Re: hda_intel : Patch + Regression in 2.6.18 -> 2.6.22

[Takashi Iwai <tiwai@suse.de>]

At Wed, 5 Sep 2007 08:38:44 -0700,
Andrew Morton wrote:
> 
> > On Wed, 22 Aug 2007 23:29:02 +0100 Roger Gammans <roger@computer-surgery.co.uk> wrote:
> > Hi
> 
> A distinct regression, a confirmed fix, yet no reply after more than two weeks.

Sorry, I've overseen this report (as I'm obviously no akpm ;-)


> > I'm fighting a problem with my (new) Acer laptop.  (Aspire 9303)
> > 
> > In 2.6.18 the audio worked fine, but now I've upgraded to 2.6.22
> > it has stopped working. 
> > 
> > I note that there is a acer specific quirk mode in patch_realtek.c
> > enabled for all acer devices introduced in the timeframe. Disabling
> > the quirk so that the driver uses it auot mode makes playback work
> > fine again.
> > 
> > To be clear this pathc seems to work for me:-
> > 
> > --- /usr/src/linux-source-2.6.22/sound/pci/hda/patch_realtek.c 2007-07-09 00:32:17.000000000 +0100
> > +++ linux-source-2.6.22/sound/pci/hda/patch_realtek.c   2007-08-20 22:21:37.000000000 +0100
> > @@ -6392,7 +6392,7 @@
> >         SND_PCI_QUIRK(0x1462, 0x4319, "MSI", ALC883_TARGA_DIG),
> >         SND_PCI_QUIRK(0x1462, 0x4324, "MSI", ALC883_TARGA_DIG),
> >         SND_PCI_QUIRK(0x1462, 0xa422, "MSI", ALC883_TARGA_2ch_DIG),
> > -       SND_PCI_QUIRK(0x1025, 0, "Acer laptop", ALC883_ACER),
> > +/*     SND_PCI_QUIRK(0x1025, 0, "Acer laptop", ALC883_ACER) , */
> >         SND_PCI_QUIRK(0x15d9, 0x8780, "Supermicro PDSBA", ALC883_3ST_6ch),
> >         SND_PCI_QUIRK(0x161f, 0x2054, "Medion laptop", ALC883_MEDION),
> >         SND_PCI_QUIRK(0x1071, 0x8258, "Evesham Voyaeger", ALC883_LAPTOP_EAPD), 
> > 
> > 
> > Unfortunately I think it's pretty clear it will break for others.
> > 
> > The lspci reports the following on the appropriate device:-
> >   00:10.1 0403: 10de:026c (rev a2)
> >         Subsystem: 1025:0112
> >         Flags: 66MHz, fast devsel, IRQ 19
> >         Memory at c0000000 (32-bit, non-prefetchable) [size=16K]
> >         Capabilities: <access denied>
> > 
> 
> Jaroslav, Takashi: can you please comment on this?

The patch would definitely break many other machines, so no, it can't
be applied as is.  Note that there is an equivalent workaround by
adding "model=auto" module option.

The support for Acer Aspire was already improved in mm tree.  At
least, we have the support for Aspire 9810 and co.  If it's the
compatible one, the chance is high.

Roger, could you try git-alsa patch in the latest mm together with
model=acer-aspiore option?  If it works, I can easily add your device
ID to the table.


thanks,

Takashi