[PATCH] ALSA: hda/realtek: Fix cold-boot headset misdetection on Acer Aspire A515-57G

[PATCH] ALSA: hda/realtek: Fix cold-boot headset misdetection on Acer Aspire A515-57G

[Oleksandr Kovalov]

On the Acer Aspire A515-57G (PCI SSID 1025:1616), if headphones are
already inserted into the combo jack before the codec powers up (a cold
boot with the plug already seated), the impedance-based headset-type
sensing races and misclassifies the jack. This drives the wrong output
configuration and is audible as missing center-panned content (e.g.
vocals) while panned content plays normally.

A genuine physical unplug/replug after boot reliably fixes this by
forcing a fresh sense transient, which is a strong hint about the
underlying cause: the sensing hardware appears to need a settled,
freshly-triggered read rather than the one-shot classification done
during the normal HDA_FIXUP_ACT_INIT pass.

Add a machine-specific fixup that, on cold boot only (not S3/S4 resume,
which already gets its own re-check), waits briefly after the normal
init-time decision and then forces a fresh headset-mode classification
by resetting the cached mode and re-invoking the existing
alc_fixup_headset_mode() path -- mirroring what a manual replug already
does. The wait+recheck is skipped whenever the first pass already
determined nothing is plugged in, to avoid adding boot latency on the
common case.

Tested on kernel 7.1.9 by building the affected module standalone and
confirming cold boot with headphones pre-inserted plays correctly from
the very first sample, across multiple boots (including a full restart,
and headphones inserted mid-POST rather than before power-on), with no
crashes or warnings and no behavioral difference from a real
post-replug recovery.

Signed-off-by: Oleksandr Kovalov <oleksandr.kovalov.work@gmail.com>
---
 sound/hda/codecs/realtek/alc269.c | 33 +++++++++++++++++++++++++++++++
 1 file changed, 33 insertions(+)

diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c
index bfbdb88..51b0900 100644
--- a/sound/hda/codecs/realtek/alc269.c
+++ b/sound/hda/codecs/realtek/alc269.c
@@ -2309,6 +2309,33 @@ static void alc_fixup_headset_mode_alc255_no_hp_mic(struct hda_codec *codec,
 	}
 }
 
+/*
+ * On the Acer Aspire A515-57G (and possibly other models sharing this
+ * board), if headphones are already inserted into the combo jack before
+ * the codec powers up (cold boot), the impedance-based headset-type
+ * sensing races and misclassifies the jack, driving the wrong output
+ * configuration (audible as missing center-panned/vocal content). A
+ * genuine physical unplug/replug after boot fixes it by forcing a fresh
+ * sense transient. Mirror that here on cold boot only: give the sense
+ * hardware time to settle, then force a fresh classification.
+ */
+static void alc_fixup_headset_mode_acer_coldboot(struct hda_codec *codec,
+						 const struct hda_fixup *fix, int action)
+{
+	struct alc_spec *spec = codec->spec;
+
+	alc_fixup_headset_mode(codec, fix, action);
+
+	if (action == HDA_FIXUP_ACT_INIT &&
+	    !is_s3_resume(codec) && !is_s4_resume(codec) &&
+		spec->current_headset_mode != ALC_HEADSET_MODE_UNPLUGGED) {
+		msleep(500);
+		spec->current_headset_mode = ALC_HEADSET_MODE_UNKNOWN;
+		spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
+		alc_fixup_headset_mode(codec, fix, action);
+	}
+}
+
 static void alc288_update_headset_jack_cb(struct hda_codec *codec,
 				       struct hda_jack_callback *jack)
 {
@@ -4063,6 +4090,7 @@ enum {
 	ALC282_FIXUP_ACER_DISABLE_LINEOUT,
 	ALC255_FIXUP_ACER_LIMIT_INT_MIC_BOOST,
 	ALC256_FIXUP_ACER_HEADSET_MIC,
+	ALC256_FIXUP_ACER_COLDBOOT,
 	ALC285_FIXUP_IDEAPAD_S740_COEF,
 	ALC285_FIXUP_HP_LIMIT_INT_MIC_BOOST,
 	ALC295_FIXUP_ASUS_DACS,
@@ -6050,6 +6078,10 @@ static const struct hda_fixup alc269_fixups[] = {
 		.chained = true,
 		.chain_id = ALC269_FIXUP_HEADSET_MODE_NO_HP_MIC
 	},
+	[ALC256_FIXUP_ACER_COLDBOOT] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_headset_mode_acer_coldboot,
+	},
 	[ALC285_FIXUP_IDEAPAD_S740_COEF] = {
 		.type = HDA_FIXUP_FUNC,
 		.v.func = alc285_fixup_ideapad_s740_coef,
@@ -6796,6 +6828,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x1025, 0x159c, "Acer Nitro 5 AN515-58", ALC2XX_FIXUP_HEADSET_MIC),
 	SND_PCI_QUIRK(0x1025, 0x1597, "Acer Nitro 5 AN517-55", ALC2XX_FIXUP_HEADSET_MIC),
 	SND_PCI_QUIRK(0x1025, 0x160e, "Acer PT316-51S", ALC2XX_FIXUP_HEADSET_MIC),
+	SND_PCI_QUIRK(0x1025, 0x1616, "Acer Aspire A515-57G", ALC256_FIXUP_ACER_COLDBOOT),
 	SND_PCI_QUIRK(0x1025, 0x1640, "Acer Aspire A315-44P", ALC256_FIXUP_ACER_SFG16_MICMUTE_LED),
 	SND_PCI_QUIRK(0x1025, 0x1679, "Acer Nitro 16 AN16-41", ALC2XX_FIXUP_HEADSET_MIC),
 	SND_PCI_QUIRK(0x1025, 0x169a, "Acer Swift SFG16", ALC256_FIXUP_ACER_SFG16_MICMUTE_LED),
-- 
2.55.0

Re: [PATCH] ALSA: hda/realtek: Fix cold-boot headset misdetection on Acer Aspire A515-57G

[Takashi Iwai]

On Sun, 30 Aug 2026 23:10:38 +0200,
Oleksandr Kovalov wrote:
> 
> On the Acer Aspire A515-57G (PCI SSID 1025:1616), if headphones are
> already inserted into the combo jack before the codec powers up (a cold
> boot with the plug already seated), the impedance-based headset-type
> sensing races and misclassifies the jack. This drives the wrong output
> configuration and is audible as missing center-panned content (e.g.
> vocals) while panned content plays normally.
> 
> A genuine physical unplug/replug after boot reliably fixes this by
> forcing a fresh sense transient, which is a strong hint about the
> underlying cause: the sensing hardware appears to need a settled,
> freshly-triggered read rather than the one-shot classification done
> during the normal HDA_FIXUP_ACT_INIT pass.
> 
> Add a machine-specific fixup that, on cold boot only (not S3/S4 resume,
> which already gets its own re-check), waits briefly after the normal
> init-time decision and then forces a fresh headset-mode classification
> by resetting the cached mode and re-invoking the existing
> alc_fixup_headset_mode() path -- mirroring what a manual replug already
> does. The wait+recheck is skipped whenever the first pass already
> determined nothing is plugged in, to avoid adding boot latency on the
> common case.
> 
> Tested on kernel 7.1.9 by building the affected module standalone and
> confirming cold boot with headphones pre-inserted plays correctly from
> the very first sample, across multiple boots (including a full restart,
> and headphones inserted mid-POST rather than before power-on), with no
> crashes or warnings and no behavioral difference from a real
> post-replug recovery.
> 
> Signed-off-by: Oleksandr Kovalov <oleksandr.kovalov.work@gmail.com>

There is a recent fix-up for the same model to support the mic-mute
LED, and it conflicts with your fix.  Could you try to rebase to the
latest sound.git for-linus branch (or at least 7.3-rc1) and resubmit?


thanks,

Takashi

[PATCH v2] ALSA: hda/realtek: Fix cold-boot headset misdetection on Acer Aspire A515-57G

[Oleksandr Kovalov]

On the Acer Aspire A515-57G (PCI SSID 1025:1616), if headphones are
already inserted into the combo jack before the codec powers up (a cold
boot with the plug already seated), the impedance-based headset-type
sensing races and misclassifies the jack. This drives the wrong output
configuration and is audible as missing center-panned content (e.g.
vocals) while panned content plays normally.

A genuine physical unplug/replug after boot reliably fixes this by
forcing a fresh sense transient, which is a strong hint about the
underlying cause: the sensing hardware appears to need a settled,
freshly-triggered read rather than the one-shot classification done
during the normal HDA_FIXUP_ACT_INIT pass.

Add a machine-specific fixup that, on cold boot only (not S3/S4 resume,
which already gets its own re-check), waits briefly after the normal
init-time decision and then forces a fresh headset-mode classification
by resetting the cached mode and re-invoking the existing
alc_fixup_headset_mode() path -- mirroring what a manual replug already
does. The wait+recheck is skipped whenever the first pass already
determined nothing is plugged in, to avoid adding boot latency on the
common case.

Chain into the existing ALC256_FIXUP_ACER_SFG16_MICMUTE_LED fixup so
this quirk-table entry keeps providing mic-mute LED support alongside
the cold-boot headset fix.

Tested on kernel 7.1.9 by building the affected module standalone and
confirming cold boot with headphones pre-inserted plays correctly from
the very first sample, across multiple boots (including a full restart,
and headphones inserted mid-POST rather than before power-on), with no
crashes or warnings and no behavioral difference from a real
post-replug recovery.

Signed-off-by: Oleksandr Kovalov <oleksandr.kovalov.work@gmail.com>
---
v2: Rebase onto sound.git for-linus (2026-09-02, commit f4a23e17d).
    Resolve a conflict with the newly-added
    ALC256_FIXUP_ACER_SFG16_MICMUTE_LED quirk for the same SSID by
    chaining into it instead of overwriting the quirk-table entry, so
    both fixes apply to this laptop. No change to the fix logic itself,
    which was re-tested and confirmed working after the rebase.

 sound/hda/codecs/realtek/alc269.c | 36 ++++++++++++++++++++++++++++++-
 1 file changed, 35 insertions(+), 1 deletion(-)

diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c
index ecc9c6e..95b40a1 100644
--- a/sound/hda/codecs/realtek/alc269.c
+++ b/sound/hda/codecs/realtek/alc269.c
@@ -2379,6 +2379,33 @@ static void alc_fixup_headset_mode_alc255_no_hp_mic(struct hda_codec *codec,
 	}
 }
 
+/*
+ * On the Acer Aspire A515-57G (and possibly other models sharing this
+ * board), if headphones are already inserted into the combo jack before
+ * the codec powers up (cold boot), the impedance-based headset-type
+ * sensing races and misclassifies the jack, driving the wrong output
+ * configuration (audible as missing center-panned/vocal content). A
+ * genuine physical unplug/replug after boot fixes it by forcing a fresh
+ * sense transient. Mirror that here on cold boot only: give the sense
+ * hardware time to settle, then force a fresh classification.
+ */
+static void alc_fixup_headset_mode_acer_coldboot(struct hda_codec *codec,
+						 const struct hda_fixup *fix, int action)
+{
+	struct alc_spec *spec = codec->spec;
+
+	alc_fixup_headset_mode(codec, fix, action);
+
+	if (action == HDA_FIXUP_ACT_INIT &&
+	    !is_s3_resume(codec) && !is_s4_resume(codec) &&
+		spec->current_headset_mode != ALC_HEADSET_MODE_UNPLUGGED) {
+		msleep(500);
+		spec->current_headset_mode = ALC_HEADSET_MODE_UNKNOWN;
+		spec->current_headset_type = ALC_HEADSET_TYPE_UNKNOWN;
+		alc_fixup_headset_mode(codec, fix, action);
+	}
+}
+
 static void alc288_update_headset_jack_cb(struct hda_codec *codec,
 				       struct hda_jack_callback *jack)
 {
@@ -4248,6 +4275,7 @@ enum {
 	ALC282_FIXUP_ACER_DISABLE_LINEOUT,
 	ALC255_FIXUP_ACER_LIMIT_INT_MIC_BOOST,
 	ALC256_FIXUP_ACER_HEADSET_MIC,
+	ALC256_FIXUP_ACER_COLDBOOT,
 	ALC285_FIXUP_IDEAPAD_S740_COEF,
 	ALC285_FIXUP_HP_LIMIT_INT_MIC_BOOST,
 	ALC295_FIXUP_ASUS_DACS,
@@ -6311,6 +6339,12 @@ static const struct hda_fixup alc269_fixups[] = {
 		.chained = true,
 		.chain_id = ALC269_FIXUP_HEADSET_MODE_NO_HP_MIC
 	},
+	[ALC256_FIXUP_ACER_COLDBOOT] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_headset_mode_acer_coldboot,
+		.chained = true,
+		.chain_id = ALC256_FIXUP_ACER_SFG16_MICMUTE_LED,
+	},
 	[ALC285_FIXUP_IDEAPAD_S740_COEF] = {
 		.type = HDA_FIXUP_FUNC,
 		.v.func = alc285_fixup_ideapad_s740_coef,
@@ -7148,7 +7182,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x1025, 0x1597, "Acer Nitro 5 AN517-55", ALC2XX_FIXUP_HEADSET_MIC),
 	SND_PCI_QUIRK(0x1025, 0x159e, "Acer Nitro 5 AN515-46", ALC2XX_FIXUP_HEADSET_MIC),
 	SND_PCI_QUIRK(0x1025, 0x160e, "Acer PT316-51S", ALC2XX_FIXUP_HEADSET_MIC),
-	SND_PCI_QUIRK(0x1025, 0x1616, "Acer Aspire A515-57", ALC256_FIXUP_ACER_SFG16_MICMUTE_LED),
+	SND_PCI_QUIRK(0x1025, 0x1616, "Acer Aspire A515-57", ALC256_FIXUP_ACER_COLDBOOT),
 	SND_PCI_QUIRK(0x1025, 0x161f, "Acer S40-54", ALC256_FIXUP_ACER_MIC_NO_PRESENCE),
 	SND_PCI_QUIRK(0x1025, 0x1640, "Acer Aspire A315-44P", ALC256_FIXUP_ACER_SFG16_MICMUTE_LED),
 	SND_PCI_QUIRK(0x1025, 0x166c, "Acer Predator PH16-71", ALC2XX_FIXUP_HEADSET_MIC),
-- 
2.55.0

Re: [PATCH v2] ALSA: hda/realtek: Fix cold-boot headset misdetection on Acer Aspire A515-57G

[Takashi Iwai]

On Wed, 02 Sep 2026 19:13:05 +0200,
Oleksandr Kovalov wrote:
> 
> On the Acer Aspire A515-57G (PCI SSID 1025:1616), if headphones are
> already inserted into the combo jack before the codec powers up (a cold
> boot with the plug already seated), the impedance-based headset-type
> sensing races and misclassifies the jack. This drives the wrong output
> configuration and is audible as missing center-panned content (e.g.
> vocals) while panned content plays normally.
> 
> A genuine physical unplug/replug after boot reliably fixes this by
> forcing a fresh sense transient, which is a strong hint about the
> underlying cause: the sensing hardware appears to need a settled,
> freshly-triggered read rather than the one-shot classification done
> during the normal HDA_FIXUP_ACT_INIT pass.
> 
> Add a machine-specific fixup that, on cold boot only (not S3/S4 resume,
> which already gets its own re-check), waits briefly after the normal
> init-time decision and then forces a fresh headset-mode classification
> by resetting the cached mode and re-invoking the existing
> alc_fixup_headset_mode() path -- mirroring what a manual replug already
> does. The wait+recheck is skipped whenever the first pass already
> determined nothing is plugged in, to avoid adding boot latency on the
> common case.
> 
> Chain into the existing ALC256_FIXUP_ACER_SFG16_MICMUTE_LED fixup so
> this quirk-table entry keeps providing mic-mute LED support alongside
> the cold-boot headset fix.
> 
> Tested on kernel 7.1.9 by building the affected module standalone and
> confirming cold boot with headphones pre-inserted plays correctly from
> the very first sample, across multiple boots (including a full restart,
> and headphones inserted mid-POST rather than before power-on), with no
> crashes or warnings and no behavioral difference from a real
> post-replug recovery.
> 
> Signed-off-by: Oleksandr Kovalov <oleksandr.kovalov.work@gmail.com>

Thanks, applied.


Takashi