Skip to contentSkip to conversation

3.14-rc2 [PATCH] Revert "ALSA: hda/realtek - Avoid invalid COEFs for ALC271X"

BUG Martin Kepplinger 2 participants 5 messages Feb 12, 2014 Sound

lore
Share & tools
lore
#0 lore raw

This reverts commit d3c56568f43807135f2c2a09582a69f809f0d8b7.

The reverted commit breaks audio through headphone line out on the Acer TravelMate B113 (Type1Sku0) Notebook, my main work machine. I don't know much about it but this fixes my problem. Bisected and tested.

---
 sound/pci/hda/patch_realtek.c |    3 ---
 1 file changed, 3 deletions(-)
sound/pci/hda/patch_realtek.c +0 -3
index d9693ca..0f5af34 100644
@@ -4434,9 +4434,6 @@ static void alc269_fill_coef(struct hda_codec *codec)
 
 	if (spec->codec_variant != ALC269_TYPE_ALC269VB)
 		return;
-	/* ALC271X doesn't seem to support these COEFs (bko#52181) */
-	if (!strcmp(codec->chip_name, "ALC271X"))
-		return;
 
 	if ((alc_get_coef0(codec) & 0x00ff) < 0x015) {
 		alc_write_coef_idx(codec, 0xf, 0x960b);
Signature · 1 line
1.7.10.4
#1 lore raw

At Wed, 12 Feb 2014 17:09:23 +0100,

Quoted 10 lines · Martin Kepplinger

This reverts commit d3c56568f43807135f2c2a09582a69f809f0d8b7.

The reverted commit breaks audio through headphone line out on the Acer TravelMate B113 (Type1Sku0) Notebook, my main work machine. I don't know much about it but this fixes my problem. Bisected and tested.

Too bad, we need COEF for some machine but it breaks for some. Since reverting breaks obviously another machine, we need a different approach, e.g. checking the machine ID. Please give alsa-info.sh output of your machine.

thanks,

Takashi

Quoted 21 lines
---
 sound/pci/hda/patch_realtek.c |    3 ---
 1 file changed, 3 deletions(-)
sound/pci/hda/patch_realtek.c +0 -3
index d9693ca..0f5af34 100644
@@ -4434,9 +4434,6 @@ static void alc269_fill_coef(struct hda_codec *codec)
 
 	if (spec->codec_variant != ALC269_TYPE_ALC269VB)
 		return;
-	/* ALC271X doesn't seem to support these COEFs (bko#52181) */
-	if (!strcmp(codec->chip_name, "ALC271X"))
-		return;
 
 	if ((alc_get_coef0(codec) & 0x00ff) < 0x015) {
 		alc_write_coef_idx(codec, 0xf, 0x960b);
Signature · 1 line
1.7.10.4
#2 lore raw

At Wed, 12 Feb 2014 17:20:21 +0100,

Quoted 11 lines · Takashi Iwai

At Wed, 12 Feb 2014 17:09:23 +0100,

Quoted 8 lines · Martin Kepplinger

This reverts commit d3c56568f43807135f2c2a09582a69f809f0d8b7.

The reverted commit breaks audio through headphone line out on the Acer TravelMate B113 (Type1Sku0) Notebook, my main work machine. I don't know much about it but this fixes my problem. Bisected and tested.

Quoted 7 lines
Quoted 2 lines

Too bad, we need COEF for some machine but it breaks for some. Since reverting breaks obviously another machine, we need a different approach, e.g. checking the machine ID. Please give alsa-info.sh output of your machine.

Thinking it again, I'll take your patch as is, and put an additional fix for AO725 as below. Could you try it to see whether it brings any regressions? It's to be applied after your revert patch.

thanks,

Takashi

-- 8< --
From: Takashi Iwai <[email protected]>
Subject: [PATCH] ALSA: hda - Better fix for invalid COEF setup on Acer AO725

Instead of disabling the COEF setup for all ALC271X codec (like commit d3c56568), do it only if needed. Currently, Acer AO725 is known to show the problem, so clear the bad init_hook in the fixup.

The explicit call of alc269_fill_coef() in patch_alc269() is also removed, since the function will be called anyway at init callback.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=52181

---
 sound/pci/hda/patch_realtek.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)
sound/pci/hda/patch_realtek.c +16 -2
index a9a83b85517a..eaff10598f67 100644
@@ -3817,6 +3817,14 @@ static void alc290_fixup_mono_speakers(struct hda_codec *codec,
 	}
 }
 
+static void alc_fixup_clear_init_hook(struct hda_codec *codec,
+				      const struct hda_fixup *fix, int action)
+{
+	struct alc_spec *spec = codec->spec;
+	if (action == HDA_FIXUP_ACT_PROBE)
+		spec->init_hook = NULL;
+}
+
 /* for hda_fixup_thinkpad_acpi() */
 #include "thinkpad_helper.c"
 
@@ -3858,6 +3866,7 @@ enum {
 	ALC271_FIXUP_HP_GATE_MIC_JACK,
 	ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572,
 	ALC269_FIXUP_ACER_AC700,
+	ALC271_FIXUP_ACER_AO725,
 	ALC269_FIXUP_LIMIT_INT_MIC_BOOST,
 	ALC269VB_FIXUP_ASUS_ZENBOOK,
 	ALC269VB_FIXUP_ASUS_ZENBOOK_UX31A,
@@ -4250,13 +4259,19 @@ static const struct hda_fixup alc269_fixups[] = {
 		.type = HDA_FIXUP_FUNC,
 		.v.func = alc_fixup_headset_mode_alc255,
 	},
+	[ALC271_FIXUP_ACER_AO725] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_clear_init_hook,
+		.chained = true,
+		.chain_id = ALC271_FIXUP_HP_GATE_MIC_JACK,
+	},
 };
 
 static const struct snd_pci_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x1025, 0x029b, "Acer 1810TZ", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0349, "Acer AOD260", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x047c, "Acer AC700", ALC269_FIXUP_ACER_AC700),
-	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_HP_GATE_MIC_JACK),
+	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_ACER_AO725),
 	SND_PCI_QUIRK(0x1025, 0x0742, "Acer AO756", ALC271_FIXUP_HP_GATE_MIC_JACK),
 	SND_PCI_QUIRK_VENDOR(0x1025, "Acer Aspire", ALC271_FIXUP_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0775, "Acer Aspire E1-572", ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572),
@@ -4523,7 +4538,6 @@ static int patch_alc269(struct hda_codec *codec)
 		if (err < 0)
 			goto error;
 		spec->init_hook = alc269_fill_coef;
-		alc269_fill_coef(codec);
 		break;
 
 	case 0x10ec0280:
Signature · 1 line
1.8.5.2
#3 lore raw

Am 2014-02-12 17:45, schrieb Takashi Iwai:

Quoted 98 lines

At Wed, 12 Feb 2014 17:20:21 +0100,

Quoted 11 lines · Takashi Iwai

At Wed, 12 Feb 2014 17:09:23 +0100,

Quoted 8 lines · Martin Kepplinger

This reverts commit d3c56568f43807135f2c2a09582a69f809f0d8b7.

The reverted commit breaks audio through headphone line out on the Acer TravelMate B113 (Type1Sku0) Notebook, my main work machine. I don't know much about it but this fixes my problem. Bisected and tested.

Quoted 7 lines
Quoted 2 lines

Too bad, we need COEF for some machine but it breaks for some. Since reverting breaks obviously another machine, we need a different approach, e.g. checking the machine ID. Please give alsa-info.sh output of your machine.

Thinking it again, I'll take your patch as is, and put an additional fix for AO725 as below. Could you try it to see whether it brings any regressions? It's to be applied after your revert patch.

thanks,

Takashi

-- 8< --
From: Takashi Iwai <[email protected]>
Subject: [PATCH] ALSA: hda - Better fix for invalid COEF setup on Acer AO725

Instead of disabling the COEF setup for all ALC271X codec (like commit d3c56568), do it only if needed. Currently, Acer AO725 is known to show the problem, so clear the bad init_hook in the fixup.

The explicit call of alc269_fill_coef() in patch_alc269() is also removed, since the function will be called anyway at init callback.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=52181

---
 sound/pci/hda/patch_realtek.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)
sound/pci/hda/patch_realtek.c +16 -1
index a9a83b85517a..eaff10598f67 100644
@@ -3817,6 +3817,14 @@ static void alc290_fixup_mono_speakers(struct hda_codec *codec,
 	}
 }
 
+static void alc_fixup_clear_init_hook(struct hda_codec *codec,
+				      const struct hda_fixup *fix, int action)
+{
+	struct alc_spec *spec = codec->spec;
+	if (action == HDA_FIXUP_ACT_PROBE)
+		spec->init_hook = NULL;
+}
+
 /* for hda_fixup_thinkpad_acpi() */
 #include "thinkpad_helper.c"
 
@@ -3858,6 +3866,7 @@ enum {
 	ALC271_FIXUP_HP_GATE_MIC_JACK,
 	ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572,
 	ALC269_FIXUP_ACER_AC700,
+	ALC271_FIXUP_ACER_AO725,
 	ALC269_FIXUP_LIMIT_INT_MIC_BOOST,
 	ALC269VB_FIXUP_ASUS_ZENBOOK,
 	ALC269VB_FIXUP_ASUS_ZENBOOK_UX31A,
@@ -4250,13 +4259,19 @@ static const struct hda_fixup alc269_fixups[] = {
 		.type = HDA_FIXUP_FUNC,
 		.v.func = alc_fixup_headset_mode_alc255,
 	},
+	[ALC271_FIXUP_ACER_AO725] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_clear_init_hook,
+		.chained = true,
+		.chain_id = ALC271_FIXUP_HP_GATE_MIC_JACK,
+	},
 };
 
 static const struct snd_pci_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x1025, 0x029b, "Acer 1810TZ", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0349, "Acer AOD260", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x047c, "Acer AC700", ALC269_FIXUP_ACER_AC700),
-	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_HP_GATE_MIC_JACK),
+	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_ACER_AO725),
 	SND_PCI_QUIRK(0x1025, 0x0742, "Acer AO756", ALC271_FIXUP_HP_GATE_MIC_JACK),
 	SND_PCI_QUIRK_VENDOR(0x1025, "Acer Aspire", ALC271_FIXUP_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0775, "Acer Aspire E1-572", ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572),

sound/pci/hda/patch_realtek.c static int patch_alc269(struct hda_codec *codec) · line 4543

Quoted 7 lines
sound/pci/hda/patch_realtek.c +0 -1
@@ -4523,7 +4538,6 @@ static int patch_alc269(struct hda_codec *codec)
 		if (err < 0)
 			goto error;
 		spec->init_hook = alc269_fill_coef;
-		alc269_fill_coef(codec);
 		break;
 
 	case 0x10ec0280:

So my alsa-info.sh is
http://www.alsa-project.org/db/?f=3e9c5d39ff057106d6ae307f9b86c1e562056c1b
and I'm running your patch on top of mine without any problems.

thanks,

martin

#4 lore raw

At Wed, 12 Feb 2014 19:31:36 +0100,

Quoted 105 lines · Martin Kepplinger

Am 2014-02-12 17:45, schrieb Takashi Iwai:

Quoted 103 lines

At Wed, 12 Feb 2014 17:20:21 +0100,

Quoted 11 lines · Takashi Iwai

At Wed, 12 Feb 2014 17:09:23 +0100,

Quoted 8 lines · Martin Kepplinger

This reverts commit d3c56568f43807135f2c2a09582a69f809f0d8b7.

The reverted commit breaks audio through headphone line out on the Acer TravelMate B113 (Type1Sku0) Notebook, my main work machine. I don't know much about it but this fixes my problem. Bisected and tested.

Quoted 7 lines
Quoted 2 lines

Too bad, we need COEF for some machine but it breaks for some. Since reverting breaks obviously another machine, we need a different approach, e.g. checking the machine ID. Please give alsa-info.sh output of your machine.

Thinking it again, I'll take your patch as is, and put an additional fix for AO725 as below. Could you try it to see whether it brings any regressions? It's to be applied after your revert patch.

thanks,

Takashi

-- 8< --
From: Takashi Iwai <[email protected]>
Subject: [PATCH] ALSA: hda - Better fix for invalid COEF setup on Acer AO725

Instead of disabling the COEF setup for all ALC271X codec (like commit d3c56568), do it only if needed. Currently, Acer AO725 is known to show the problem, so clear the bad init_hook in the fixup.

The explicit call of alc269_fill_coef() in patch_alc269() is also removed, since the function will be called anyway at init callback.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=52181

---
 sound/pci/hda/patch_realtek.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)
sound/pci/hda/patch_realtek.c +16 -2
index a9a83b85517a..eaff10598f67 100644
@@ -3817,6 +3817,14 @@ static void alc290_fixup_mono_speakers(struct hda_codec *codec,
 	}
 }
 
+static void alc_fixup_clear_init_hook(struct hda_codec *codec,
+				      const struct hda_fixup *fix, int action)
+{
+	struct alc_spec *spec = codec->spec;
+	if (action == HDA_FIXUP_ACT_PROBE)
+		spec->init_hook = NULL;
+}
+
 /* for hda_fixup_thinkpad_acpi() */
 #include "thinkpad_helper.c"
 
@@ -3858,6 +3866,7 @@ enum {
 	ALC271_FIXUP_HP_GATE_MIC_JACK,
 	ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572,
 	ALC269_FIXUP_ACER_AC700,
+	ALC271_FIXUP_ACER_AO725,
 	ALC269_FIXUP_LIMIT_INT_MIC_BOOST,
 	ALC269VB_FIXUP_ASUS_ZENBOOK,
 	ALC269VB_FIXUP_ASUS_ZENBOOK_UX31A,
@@ -4250,13 +4259,19 @@ static const struct hda_fixup alc269_fixups[] = {
 		.type = HDA_FIXUP_FUNC,
 		.v.func = alc_fixup_headset_mode_alc255,
 	},
+	[ALC271_FIXUP_ACER_AO725] = {
+		.type = HDA_FIXUP_FUNC,
+		.v.func = alc_fixup_clear_init_hook,
+		.chained = true,
+		.chain_id = ALC271_FIXUP_HP_GATE_MIC_JACK,
+	},
 };
 
 static const struct snd_pci_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x1025, 0x029b, "Acer 1810TZ", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0349, "Acer AOD260", ALC269_FIXUP_INV_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x047c, "Acer AC700", ALC269_FIXUP_ACER_AC700),
-	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_HP_GATE_MIC_JACK),
+	SND_PCI_QUIRK(0x1025, 0x0740, "Acer AO725", ALC271_FIXUP_ACER_AO725),
 	SND_PCI_QUIRK(0x1025, 0x0742, "Acer AO756", ALC271_FIXUP_HP_GATE_MIC_JACK),
 	SND_PCI_QUIRK_VENDOR(0x1025, "Acer Aspire", ALC271_FIXUP_DMIC),
 	SND_PCI_QUIRK(0x1025, 0x0775, "Acer Aspire E1-572", ALC271_FIXUP_HP_GATE_MIC_JACK_E1_572),
@@ -4523,7 +4538,6 @@ static int patch_alc269(struct hda_codec *codec)
 		if (err < 0)
 			goto error;
 		spec->init_hook = alc269_fill_coef;
-		alc269_fill_coef(codec);
 		break;
 
Quoted 6 lines

sound/pci/hda/patch_realtek.c static int patch_alc269(struct hda_codec *codec) · line 4543

Quoted 2 lines
sound/pci/hda/patch_realtek.c +0 -0
 	case 0x10ec0280:

So my alsa-info.sh is
http://www.alsa-project.org/db/?f=3e9c5d39ff057106d6ae307f9b86c1e562056c1b
and I'm running your patch on top of mine without any problems.

OK. I queued the patches now.
Thanks!

Takashi