* [PATCH] ALSA: hda/realtek: Fix hiss/crackle on Lenovo Xiaoxin 14 GT (17aa:3912)
@ 2026-10-15 13:13 Viktor Menshin
0 siblings, 0 replies; 2+ messages in thread
From: Viktor Menshin @ 2026-10-15 13:13 UTC (permalink / raw)
To: alsa-devel; +Cc: tiwai, Viktor Menshin
Commit 3a3e810e9108 ("ALSA: hda/realtek: Add quirk for Lenovo Xiaoxin
14 GT") added a quirk mapping SSID 17aa:3912 to
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN, chosen by analogy with
neighboring SSIDs (0x390d, 0x3911, 0x391a) without being verified on
hardware, as noted in that commit's own message.
I have since tested it on the actual hardware (Lenovo Xiaoxin Pro 14
ASP10, same 17aa:3912 SSID) and confirmed it does not fix the reported
issue. ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN only repoints pin 0x17
and the preferred DAC pairs for the bass speaker; it does not touch
the codec's vendor processing coefficients. The actual symptom -
constant electrical hiss/crackle on both the headphone jack and the
internal speakers, independent of volume - is unrelated to pin/DAC
routing and is also present on stock Windows before the Realtek
driver is installed, which points at those coefficients being left
at their power-on default.
To find the correct values, I passed the physical HD Audio controller
through to a Windows 10 VM (VFIO, with QEMU patched to log CORB verb
traffic) and captured the coefficient writes the vendor Realtek driver
performs at init. Replaying a single one of them on Linux via hda-verb
reproducibly eliminates the headphone jack noise:
hda-verb /dev/snd/hwC#D0 0x20 SET_COEF_INDEX 0x36
hda-verb /dev/snd/hwC#D0 0x20 SET_PROC_COEF 0x77d7
This has held up across reboots and kernel updates for several weeks
(applied via a small systemd oneshot unit while testing).
Add a dedicated fixup that performs this coefficient write and chains
it with the existing ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN (which may
still be relevant for the bass speaker pin wiring on this board), and
point the 17aa:3912 quirk entry at it instead.
Honesty note: this fixes the headphone jack. The internal speakers
still have noise under certain conditions (notably worse on AC power)
that I have not been able to isolate via coefficient-table replay;
that part remains open. Sending what is verified to work rather than
holding this back.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=221203
Fixes: 3a3e810e9108 ("ALSA: hda/realtek: Add quirk for Lenovo Xiaoxin 14 GT")
Signed-off-by: Viktor Menshin <ripeeerr@gmail.com>
---
sound/hda/codecs/realtek/alc269.c | 31 ++++++++++++++++++++++++++++++-
1 file changed, 30 insertions(+), 1 deletion(-)
diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c
index 1623cf926..b72e04dc9 100644
--- a/sound/hda/codecs/realtek/alc269.c
+++ b/sound/hda/codecs/realtek/alc269.c
@@ -3661,6 +3661,28 @@ static void alc287_fixup_yoga9_14iap7_bass_spk_pin(struct hda_codec *codec,
}
}
+/*
+ * The Lenovo Xiaoxin 14 GT (SSID 17aa:3912) uses the same bass-speaker
+ * pin wiring as its 0x390d/0x3911/0x391a siblings, but on this board the
+ * codec's vendor processing coefficients (node 0x20) are left at their
+ * power-on default instead of being programmed the way the vendor
+ * (Windows) driver programs them. That leaves audible hiss/crackle on
+ * both the headphone jack and the internal speakers, independent of
+ * volume, regardless of which pin/DAC routing is used.
+ *
+ * This single coefficient write was captured from the vendor driver's
+ * own init sequence (via a VFIO pass-through of the physical codec to a
+ * Windows VM) and confirmed to eliminate the headphone jack noise.
+ */
+static void alc287_fixup_lenovo_xiaoxin_14gt_hp_noise(struct hda_codec *codec,
+ const struct hda_fixup *fix, int action)
+{
+ if (action != HDA_FIXUP_ACT_INIT)
+ return;
+
+ alc_write_coef_idx(codec, 0x36, 0x77d7);
+}
+
static void alc_fixup_yoga_pro7_audio(struct hda_codec *codec,
const struct hda_fixup *fix, int action)
{
@@ -4320,6 +4342,7 @@ enum {
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK,
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN,
ALC287_FIXUP_YOGA9_14IMH9_BASS_SPK_PIN,
+ ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE,
ALC295_FIXUP_DELL_INSPIRON_TOP_SPEAKERS,
ALC236_FIXUP_DELL_DUAL_CODECS,
ALC287_FIXUP_CS35L41_I2C_2_THINKPAD_ACPI,
@@ -6787,6 +6810,12 @@ static const struct hda_fixup alc269_fixups[] = {
.chained = true,
.chain_id = ALC287_FIXUP_CS35L41_I2C_2,
},
+ [ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE] = {
+ .type = HDA_FIXUP_FUNC,
+ .v.func = alc287_fixup_lenovo_xiaoxin_14gt_hp_noise,
+ .chained = true,
+ .chain_id = ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN,
+ },
[ALC295_FIXUP_DELL_INSPIRON_TOP_SPEAKERS] = {
.type = HDA_FIXUP_FUNC,
.v.func = alc295_fixup_dell_inspiron_top_speakers,
@@ -8309,7 +8338,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
HDA_CODEC_QUIRK(0x17aa, 0x3907, "Legion Pro 7i 16IAX10H / Y9000P IAX10", ALC287_FIXUP_LENOVO_LEGION_AW88399),
SND_PCI_QUIRK(0x17aa, 0x390d, "Lenovo Yoga Pro 7 14ASP10", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN_HEADSET),
SND_PCI_QUIRK(0x17aa, 0x3911, "Lenovo Yoga Pro 7 14IAH10", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
- SND_PCI_QUIRK(0x17aa, 0x3912, "Lenovo Xiaoxin 14 GT", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
+ SND_PCI_QUIRK(0x17aa, 0x3912, "Lenovo Xiaoxin 14 GT", ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE),
SND_PCI_QUIRK(0x17aa, 0x3913, "Lenovo 145", ALC236_FIXUP_LENOVO_INV_DMIC),
SND_PCI_QUIRK(0x17aa, 0x391f, "Yoga S990-16 pro Quad YC Quad", ALC287_FIXUP_TXNW2781_I2C),
SND_PCI_QUIRK(0x17aa, 0x3920, "Yoga S990-16 pro Quad VECO Quad", ALC287_FIXUP_YOGA9_SPEAKER2_TO_DAC1),
--
2.43.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* [PATCH] ALSA: hda/realtek: Fix hiss/crackle on Lenovo Xiaoxin 14 GT (17aa:3912)
@ 2026-10-15 13:21 Viktor Menshin
0 siblings, 0 replies; 2+ messages in thread
From: Viktor Menshin @ 2026-10-15 13:21 UTC (permalink / raw)
To: linux-sound; +Cc: tiwai, alsa-devel, Viktor Menshin
Commit 3a3e810e9108 ("ALSA: hda/realtek: Add quirk for Lenovo Xiaoxin
14 GT") added a quirk mapping SSID 17aa:3912 to
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN, chosen by analogy with
neighboring SSIDs (0x390d, 0x3911, 0x391a) without being verified on
hardware, as noted in that commit's own message.
I have since tested it on the actual hardware (Lenovo Xiaoxin Pro 14
ASP10, same 17aa:3912 SSID) and confirmed it does not fix the reported
issue. ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN only repoints pin 0x17
and the preferred DAC pairs for the bass speaker; it does not touch
the codec's vendor processing coefficients. The actual symptom -
constant electrical hiss/crackle on both the headphone jack and the
internal speakers, independent of volume - is unrelated to pin/DAC
routing and is also present on stock Windows before the Realtek
driver is installed, which points at those coefficients being left
at their power-on default.
To find the correct values, I passed the physical HD Audio controller
through to a Windows 10 VM (VFIO, with QEMU patched to log CORB verb
traffic) and captured the coefficient writes the vendor Realtek driver
performs at init. Replaying a single one of them on Linux via hda-verb
reproducibly eliminates the headphone jack noise:
hda-verb /dev/snd/hwC#D0 0x20 SET_COEF_INDEX 0x36
hda-verb /dev/snd/hwC#D0 0x20 SET_PROC_COEF 0x77d7
This has held up across reboots and kernel updates for several weeks
(applied via a small systemd oneshot unit while testing).
Add a dedicated fixup that performs this coefficient write and chains
it with the existing ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN (which may
still be relevant for the bass speaker pin wiring on this board), and
point the 17aa:3912 quirk entry at it instead.
Honesty note: this fixes the headphone jack. The internal speakers
still have noise under certain conditions (notably worse on AC power)
that I have not been able to isolate via coefficient-table replay;
that part remains open. Sending what is verified to work rather than
holding this back.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=221203
Fixes: 3a3e810e9108 ("ALSA: hda/realtek: Add quirk for Lenovo Xiaoxin 14 GT")
Signed-off-by: Viktor Menshin <ripeeerr@gmail.com>
---
sound/hda/codecs/realtek/alc269.c | 31 ++++++++++++++++++++++++++++++-
1 file changed, 30 insertions(+), 1 deletion(-)
diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c
index 1623cf926..b72e04dc9 100644
--- a/sound/hda/codecs/realtek/alc269.c
+++ b/sound/hda/codecs/realtek/alc269.c
@@ -3661,6 +3661,28 @@ static void alc287_fixup_yoga9_14iap7_bass_spk_pin(struct hda_codec *codec,
}
}
+/*
+ * The Lenovo Xiaoxin 14 GT (SSID 17aa:3912) uses the same bass-speaker
+ * pin wiring as its 0x390d/0x3911/0x391a siblings, but on this board the
+ * codec's vendor processing coefficients (node 0x20) are left at their
+ * power-on default instead of being programmed the way the vendor
+ * (Windows) driver programs them. That leaves audible hiss/crackle on
+ * both the headphone jack and the internal speakers, independent of
+ * volume, regardless of which pin/DAC routing is used.
+ *
+ * This single coefficient write was captured from the vendor driver's
+ * own init sequence (via a VFIO pass-through of the physical codec to a
+ * Windows VM) and confirmed to eliminate the headphone jack noise.
+ */
+static void alc287_fixup_lenovo_xiaoxin_14gt_hp_noise(struct hda_codec *codec,
+ const struct hda_fixup *fix, int action)
+{
+ if (action != HDA_FIXUP_ACT_INIT)
+ return;
+
+ alc_write_coef_idx(codec, 0x36, 0x77d7);
+}
+
static void alc_fixup_yoga_pro7_audio(struct hda_codec *codec,
const struct hda_fixup *fix, int action)
{
@@ -4320,6 +4342,7 @@ enum {
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK,
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN,
ALC287_FIXUP_YOGA9_14IMH9_BASS_SPK_PIN,
+ ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE,
ALC295_FIXUP_DELL_INSPIRON_TOP_SPEAKERS,
ALC236_FIXUP_DELL_DUAL_CODECS,
ALC287_FIXUP_CS35L41_I2C_2_THINKPAD_ACPI,
@@ -6787,6 +6810,12 @@ static const struct hda_fixup alc269_fixups[] = {
.chained = true,
.chain_id = ALC287_FIXUP_CS35L41_I2C_2,
},
+ [ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE] = {
+ .type = HDA_FIXUP_FUNC,
+ .v.func = alc287_fixup_lenovo_xiaoxin_14gt_hp_noise,
+ .chained = true,
+ .chain_id = ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN,
+ },
[ALC295_FIXUP_DELL_INSPIRON_TOP_SPEAKERS] = {
.type = HDA_FIXUP_FUNC,
.v.func = alc295_fixup_dell_inspiron_top_speakers,
@@ -8309,7 +8338,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
HDA_CODEC_QUIRK(0x17aa, 0x3907, "Legion Pro 7i 16IAX10H / Y9000P IAX10", ALC287_FIXUP_LENOVO_LEGION_AW88399),
SND_PCI_QUIRK(0x17aa, 0x390d, "Lenovo Yoga Pro 7 14ASP10", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN_HEADSET),
SND_PCI_QUIRK(0x17aa, 0x3911, "Lenovo Yoga Pro 7 14IAH10", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
- SND_PCI_QUIRK(0x17aa, 0x3912, "Lenovo Xiaoxin 14 GT", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
+ SND_PCI_QUIRK(0x17aa, 0x3912, "Lenovo Xiaoxin 14 GT", ALC287_FIXUP_LENOVO_XIAOXIN_14GT_HP_NOISE),
SND_PCI_QUIRK(0x17aa, 0x3913, "Lenovo 145", ALC236_FIXUP_LENOVO_INV_DMIC),
SND_PCI_QUIRK(0x17aa, 0x391f, "Yoga S990-16 pro Quad YC Quad", ALC287_FIXUP_TXNW2781_I2C),
SND_PCI_QUIRK(0x17aa, 0x3920, "Yoga S990-16 pro Quad VECO Quad", ALC287_FIXUP_YOGA9_SPEAKER2_TO_DAC1),
--
2.43.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-10-15 13:23 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-15 13:13 [PATCH] ALSA: hda/realtek: Fix hiss/crackle on Lenovo Xiaoxin 14 GT (17aa:3912) Viktor Menshin
-- strict thread matches above, loose matches on Subject: below --
2026-10-15 13:21 Viktor Menshin
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.