Re: [REGRESSION][BISECTED] Audio volume issues since 4178d78cd7a8

[Werner Sembach <wse@xxxxxxxxxxxxxxxxxxx>]

Hi Takashi,

Am 19.09.24 um 10:57 schrieb Takashi Iwai:
> On Wed, 18 Sep 2024 16:10:00 +0200,
> Christoffer Sandberg wrote:
> > On 18.9.2024 15:49, Takashi Iwai wrote:
> > > On Wed, 18 Sep 2024 15:39:28 +0200,
> > > Werner Sembach wrote:
> > > > Hi
> > > >
> > > > Am 18.09.24 um 10:09 schrieb Takashi Iwai:
> > > > > On Wed, 18 Sep 2024 10:06:01 +0200,
> > > > > Takashi Iwai wrote:
> > > > > > [...]
> > > > > > I don't see any relevant about the incorrect volumes by the suggested
> > > > > > commit, but at least we should avoid applying the quirk for a
> > > > > > non-existing speaker pin.
> > > > > >
> > > > > > Jerry, yours is with CX11970 (codec id 0x14f120d0), right?
> > > > Yes, at least for the preproduction sample of the Sirius Gen1 I have
> > > > at hand atm.
> > > >
> > > > pciid is 2782:12c3
> > > >
> > > > @Christoffer: I guess 2782:12c5 is for the Gen2?
> > Yes the other is for Sirius Gen2
> >
> > > Hm, and one of those PCI SSID conflicts with System76 Pangolin
> > > (pang14)?  All seem to have the very same codec CX11970.
> > I don't think I saw the pci id explicitly listed, but that was the
> > assumption.
> >
> > > If they really conflict in both PCI and HD-audio codec IDs, the only
> > > way would be to check DMI string, I'm afraid.
> > I asked about the volumes since it could be a hint that a second
> > speaker is activated but stuck on a fixed volume. This would make a
> > low general volume setting sound louder but still somewhat adjustable.
> >
> > However, if we can not verify second speaker pair, then I agree, DMI
> > limit would be the safe choice.
> So, just to make clear: the device reported for a regression and the
> device you tried to tweak are different models?  The former is
> Pangolin and the latter Sirius Gen 2.  I'm confused by models.

Yes, the regression report is for the Pangolin, which seems to be affected by a 
patch that should only affect the Sirius 16 Gen1 & Gen2.

> With the quirk, you'll get an individual mixer element controlling the
> top speaker volume?
Yes (for the Sirius devices at least, we ofc don't have a Pangolin laying around 
we could test), unless you use this 
https://github.com/alsa-project/alsa-ucm-conf/pull/410 alsa patch Christoffer 
wrote for TUXEDO_OS and sent to upstream which merges them down into one 
controlling both speaker pairs.
> Takashi

Kind regards,

Werner