Re: [PATCH] ALSA: hda/realtek: Enable speakers on further Samsung Galaxy Book2 Pro

[Takashi Iwai <tiwai@suse.de>]

On Thu, 27 Mar 2025 11:26:20 +0100,
Benjamin Berg wrote:
> 
> On Thu, 2025-03-27 at 11:09 +0100, Takashi Iwai wrote:
> > On Thu, 27 Mar 2025 10:55:04 +0100,
> > Benjamin Berg wrote:
> > > 
> > > On Thu, 2025-03-27 at 10:05 +0100, Takashi Iwai wrote:
> > > > On Wed, 26 Mar 2025 21:19:53 +0100,
> > > > Benjamin Berg wrote:
> > > > > Note that I didn't really manage to test this directly. Trying to
> > > > > do a
> > > > > reconfigure resulted in a hang and I didn't figure out how to use
> > > > > the
> > > > > patch= option with a firmware file.
> > > > > However, the machine does work fine when configuring the two
> > > > > speakers
> > > > > from userspace, so I am confident that this patch is correct.
> > > > 
> > > > FYI, the equivalent can be achieved with model option, too.
> > > > e.g. passing model=144d:c870 will apply the same existing quirk as
> > > > SSID 144d:c870.  If your target is the secondary codec, pass like
> > > > "model=,144d:c870" instead.
> > > 
> > > Hmm, it is probably really straight forward, but I didn't grasp the
> > > documentation in Documentation/sound/hd-audio/notes.rst on a quick
> > > read.
> > 
> > It's mentioned in "Model Option" subsection of that document.
> > ```
> > A new style for the model option that was introduced since 5.15 kernel
> > is to pass the PCI or codec SSID in the form of ``model=XXXX:YYYY``
> > where XXXX and YYYY are the sub-vendor and sub-device IDs in hex
> > numbers, respectively.  This is a kind of aliasing to another device;
> > when this form is given, the driver will refer to that SSID as a
> > reference to the quirk table.  It'd be useful especially when the
> > target quirk isn't listed in the model table.  For example, passing
> > model=103c:8862 will apply the quirk for HP ProBook 445 G8 (which
> > isn't found in the model table as of writing) as long as the device is
> > handled equivalently by the same driver.
> > ```
> > 
> > > I am not sure anymore what I did (and I don't have access to the
> > > machine anymore). It might be I tried putting the model= line into a
> > > firmware file instead of just passing it on the kernel command line.
> > 
> > > How do I even know whether it is the first or second codec? Is that
> > > what the D0 in hwC0D0 is?
> > 
> > I meant the device probe order, so card#0 device#0 must be the very
> > first one :)
> 
> Right, that make sense, but wasn't obvious to me then.
> 
> Unfortunately, it seems that I messed something up and I cannot fully
> put it back together now.
> 
> I have one photo of the sysfs entries, and the subsystem_id is
> 0x144dc1ac there. So now I am not entirely certain that this is the
> correct ID, because I might have successfully overridden it to an
> incorrect value.
> Lkely, it is correct, but I do not know where I got the 0xc01d from.

144d:c01d can be PCI SSID while 144d:c1ac is the codec SSID.
Both are matched by the driver (PCI SSID is matched primarily).
In many cases the vendor sets different ID numbers.


Takashi