Re: [PATCH v2] ALSA: hda/conexant: fix FEVM FA880 PRO internal mic mux

On Wed, 10 Jun 2026 17:19:01 +0200,
Ocean wrote:
> 
> Hi Takashi,
> 
> Thank you for the suggestions. I have tested the stock driver as requested.
> 
> Here are the results:
> 
> 1. Does the jack detection work for the rear jack? Yes, the jack
> detection works properly for the rear jack.
> 
> 2. Does the output from the rear work as-is? Yes, the output works
> normally. I primarily use this jack for an external TRS speaker.
> 
> 3. Changing the pin control of 0x1c: I plugged a TRRS headset (with a
> microphone) into the rear jack. Both the audio output and the headset
> microphone worked properly without any modifications. When I ran sudo
> hda-verb /dev/snd/hwC1D0 0x1c SET_PIN_WIDGET_CONTROL 0x00, the
> headphone output stopped as expected, and the microphone continued to
> work normally.
> 
> Based on these tests, the rear jack is a fully functional TRRS combo
> jack, and there is no hardware interference between the output (0x1c)
> and input (0x1b) pins.
> 
> The loud buzzing noise I reported earlier occurs only when a 3-pole
> TRS plug (like my external speaker or a standard headphone without a
> mic) is connected to the rear jack. It appears the hardware lacks
> proper impedance sensing to distinguish between TRS and TRRS plugs,
> and blindly leaves the mic input pin (0x1b) active. Since the TRS plug
> shorts the mic ring to ground, the active 0x1b pin captures ground
> loop noise.
> 
> This behavior is consistent with what I observe on Windows on the same
> machine: when a TRS speaker is plugged in, Windows shows a microphone
> device that captures no sound, likely because the pin is shorted to
> ground.
> 
> Given that this machine is a Mini-PC and the rear jack is mostly used
> for TRS external speakers, would applying a pin quirk to disable 0x1b
> (setting it to 0x40000000) be the most appropriate upstream fix?

Have you tried ctx_fixup_headset_mic, BTW?  I'm not sure whether this
is applicable to your model, but worth for a try.

If this doesn't work, then disabling the mic would be the simplest
solution.


thanks,

Takashi