Re: [PATCH v2] ALSA: hda/realtek: Add inverted LED quirk for HP ZBook 8 G2a

[Takashi Iwai]

On Tue, 14 Jul 2026 03:55:53 +0200,
Chris Chiu wrote:
> 
> On Mon, Jul 13, 2026 at 6:09 PM Takashi Iwai <tiwai@xxxxxxx> wrote:
> >
> > On Mon, 13 Jul 2026 04:08:39 +0200,
> > Chris Chiu wrote:
> > >
> > > HP ZBook 8 G2a 14 and 16 (SSIDs 0x103c:0x8f94, 0x103c:0x8f95) use
> > > Realtek ALC245 codec with TAS2781 amplifier via I2C. These models
> > > require inverted speaker mute LED handling compared to the existing
> > > ALC245_FIXUP_HP_TAS2781_I2C_MUTE_LED quirk.
> > >
> > > The existing quirk produces opposite LED behavior: when speaker is
> > > muted, the LED turns off (should be on), and when unmuted, the LED
> > > turns on (should be off). This patch adds a dedicated quirk that
> > > properly handles the inverted LED polarity.
> > >
> > > Code flow explanation:
> > > - With HP pin (hp_pin != 0): Uses standard HDA LED framework via
> > >   alc245_fixup_hp_mute_led_coefbit() for full control interface
> > >   support, enabling both manual control and direct hardware updates.
> > >
> > > - Without HP pin (hp_pin == 0, speaker-only laptops): Registers
> > >   direct vmaster_mute hook (alc245_hp_spk_mute_led_update_inverted)
> > >   to control LED during mute operations, ensuring LED feedback
> > >   without headphone jack detection capability.
> > >
> > > Both paths use inverted LED update logic to correct polarity and
> > > maintain backward compatibility with existing systems.
> >
> > So there can be multiple models with the same SSID and they have
> > different hardware configurations?
> >
> 
> The new SSIDs 0x8f94 and 0x8f95 actually have the same hardware
> configuration as the old 0x8f40, 0x8f41, 0x8f42, 0x8f62 but different
> GPIO polarity. I tried to request that they remain aligned but received no
> positive response.

I guess v2 patch is broken in anyway, and that's the reason I asked
about the same SSID -- this one adds the entry for only 103c:8f95, so
I thought it's a change only to a single SSID.  But the patch showed
that it already had 103c:8f94 with the non-existing inverted quirk...

I saw that v3 corrected it, but then the question remains: why do you
need to figure out the hp_pin dynamically?  It's a static
configuration, and you can know beforehand, no?


thanks,

Takashi