Just wanted to say thank you to everybody that helps maintain the Intel HDA subsystem. My first Linux sound experience was running `modprobe cs46xx` or so on a SuSE installation and hearing a promising noise from the speakers, and these days it seems audio works out of the box on most hardware.
Thanks also to Takashi Iwai for `hda-emu`, which saved me a lot of rebooting while testing.
I have tested the patch on this hardware both by recompiling the relevant modules and running them with the 5.10 kernel distributed in Debian `testing`, and atop the 5.12-rc6 kernel built with `make deb-pkg`.
I could not decipher the ordering of the fixup tables, so I am of course happy to move these around to fit the organizational scheme as needed.
I'm not an EE, but I am inclined to think the upper two boost values can probably be made to work by some kind of software workaround, given that they do produce analog noise coming from something. If you have any hunches about a possible workaround based on experience with other similar codecs, please point me to those fixups on- or off-list and I can try them out on my hardware.
Phil
8------------------------------------------------------8<
Fix two bugs with the Intel HDA Realtek ALC233 sound codec present in Intel NUC NUC8i7BEH and probably a few other similar NUC models.
These codecs advertise a 4-level microphone input boost amplifier on pin 0x19, but the highest two boost settings do not work correctly, and produce only low analog noise that does not seem to contain any discernible signal. There is an existing fixup for this exact problem but for a different PCI subsystem ID, so we re-use that logic.
Changing the boost level also triggers a DC spike in the input signal that bleeds off over about a second and overwhelms any input during that time. Thankfully, the existing fixup has the side effect of making the boost control show up in userspace as a mute/unmute switch, and this keeps (e.g.) PulseAudio from fiddling with it during normal input volume adjustments.
Finally, the NUC hardware has built-in inverted stereo mics. This patch also enables the usual fixup for this so the two channels cancel noise instead of the actual signal.
Signed-off-by: Phil Calvin phil@philcalvin.com --- sound/pci/hda/patch_realtek.c | 13 +++++++++++++ 1 file changed, 13 insertions(+)
diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 58946d069ee5..e1fd4c81965a 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -6405,6 +6405,8 @@ enum { ALC269_FIXUP_LEMOTE_A1802, ALC269_FIXUP_LEMOTE_A190X, ALC256_FIXUP_INTEL_NUC8_RUGGED, + ALC233_FIXUP_INTEL_NUC8_DMIC, + ALC233_FIXUP_INTEL_NUC8_BOOST, ALC256_FIXUP_INTEL_NUC10, ALC255_FIXUP_XIAOMI_HEADSET_MIC, ALC274_FIXUP_HP_MIC, @@ -7122,6 +7124,16 @@ static const struct hda_fixup alc269_fixups[] = { .type = HDA_FIXUP_FUNC, .v.func = alc233_fixup_lenovo_line2_mic_hotkey, }, + [ALC233_FIXUP_INTEL_NUC8_DMIC] = { + .type = HDA_FIXUP_FUNC, + .v.func = alc_fixup_inv_dmic, + .chained = true, + .chain_id = ALC233_FIXUP_INTEL_NUC8_BOOST, + }, + [ALC233_FIXUP_INTEL_NUC8_BOOST] = { + .type = HDA_FIXUP_FUNC, + .v.func = alc269_fixup_limit_int_mic_boost + }, [ALC255_FIXUP_DELL_SPK_NOISE] = { .type = HDA_FIXUP_FUNC, .v.func = alc_fixup_disable_aamix, @@ -8265,6 +8277,7 @@ static const struct snd_pci_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x1c06, 0x2013, "Lemote A1802", ALC269_FIXUP_LEMOTE_A1802), SND_PCI_QUIRK(0x1c06, 0x2015, "Lemote A190X", ALC269_FIXUP_LEMOTE_A190X), SND_PCI_QUIRK(0x8086, 0x2080, "Intel NUC 8 Rugged", ALC256_FIXUP_INTEL_NUC8_RUGGED), + SND_PCI_QUIRK(0x8086, 0x2074, "Intel NUC 8", ALC233_FIXUP_INTEL_NUC8_DMIC), SND_PCI_QUIRK(0x8086, 0x2081, "Intel NUC 10", ALC256_FIXUP_INTEL_NUC10), #if 0