Skip to contentSkip to conversation

ALSA: hda: cs35l41: Fix Speakers on HP OmniBook 7 16-bh0xxx

PATCH In review Yufan 3 participants 5 messages Jul 16 – Jul 19 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore

Yufan

author
#0 lore raw

PATCH 1/1 ALSA: hda: cs35l41: Fix Speakers on HP OmniBook 7 16-bh0xxx

From: Yufan Kuang <[email protected]>

The HP OmniBook 7 Laptop 16-bh0xxx (SSID 103C8E60) has two CS35L41 amplifiers connected via I2C with internal boost, but its _DSD provides cirrus,dev-index as a Package while the driver expects per-device scalar values, causing the probe to fail with -EINVAL, leaving only the left speaker working.

Add the SSID to the configuration and prop-model override tables.

Tested on a HP OmniBook 7 Laptop 16-bh0xxx (BIOS F.12) with CachyOS 7.1.3-2; speaker-test confirms both channels work correctly.

Assisted-by Copilot:DeepSeek-V4-Pro
---
 sound/hda/codecs/side-codecs/cs35l41_hda_property.c | 2 ++
 1 file changed, 2 insertions(+)
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +2 -0
index 416d7bf3e289..e14a87336478 100644
@@ -85,6 +85,7 @@ static const struct cs35l41_config cs35l41_config_table[] = {
 	{ "103C8C51", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4100, 24 },
 	{ "103C8CDD", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4100, 24 },
 	{ "103C8CDE", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 3900, 24 },
+	{ "103C8E60", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
 	{ "104312AF", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 1, 2, 0, 1000, 4500, 24 },
 	{ "10431433", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
 	{ "10431463", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
@@ -511,6 +512,7 @@ static const struct cs35l41_prop_model cs35l41_prop_model_table[] = {
 	{ "CSC3551", "103C8C6A", hp_i2c_int_2amp_dual_spkid },
 	{ "CSC3551", "103C8CDD", generic_dsd_config },
 	{ "CSC3551", "103C8CDE", generic_dsd_config },
+	{ "CSC3551", "103C8E60", generic_dsd_config },
 	{ "CSC3551", "104312AF", generic_dsd_config },
 	{ "CSC3551", "10431433", generic_dsd_config },
 	{ "CSC3551", "10431463", generic_dsd_config },
Signature · 1 line
2.55.0
#1 lore raw
Quoted 9 lines · Yufan

From: Yufan Kuang <[email protected]>

The HP OmniBook 7 Laptop 16-bh0xxx (SSID 103C8E60) has two CS35L41 amplifiers connected via I2C with internal boost, but its _DSD provides cirrus,dev-index as a Package while the driver expects per-device scalar values, causing the probe to fail with -EINVAL, leaving only the left speaker working.

Quoted 7 lines

Add the SSID to the configuration and prop-model override tables.

Tested on a HP OmniBook 7 Laptop 16-bh0xxx (BIOS F.12) with CachyOS 7.1.3-2; speaker-test confirms both channels work correctly.

Assisted-by Copilot:DeepSeek-V4-Pro

The very same change has been already submitted (maybe multiple times), e.g.
https://lore.kernel.org/[email protected]

The problem is known, and it's basically an issue of the firmware. But obviously we need to address *something* in the driver side.

Cirrus people -- do you have ideas to fix those HP laptops?

If it's not feasible in a short term, maybe it'd be better to take those quirk additions; although they aren't 100% correct, but they do work somehow in a certain level.

thanks,

Takashi

Quoted 27 lines
---
 sound/hda/codecs/side-codecs/cs35l41_hda_property.c | 2 ++
 1 file changed, 2 insertions(+)
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +2 -0
index 416d7bf3e289..e14a87336478 100644
@@ -85,6 +85,7 @@ static const struct cs35l41_config cs35l41_config_table[] = {
 	{ "103C8C51", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4100, 24 },
 	{ "103C8CDD", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4100, 24 },
 	{ "103C8CDE", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 3900, 24 },
+	{ "103C8E60", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
 	{ "104312AF", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 1, 2, 0, 1000, 4500, 24 },
 	{ "10431433", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
 	{ "10431463", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0, 1, -1, 1000, 4500, 24 },
@@ -511,6 +512,7 @@ static const struct cs35l41_prop_model cs35l41_prop_model_table[] = {
 	{ "CSC3551", "103C8C6A", hp_i2c_int_2amp_dual_spkid },
 	{ "CSC3551", "103C8CDD", generic_dsd_config },
 	{ "CSC3551", "103C8CDE", generic_dsd_config },
+	{ "CSC3551", "103C8E60", generic_dsd_config },
 	{ "CSC3551", "104312AF", generic_dsd_config },
 	{ "CSC3551", "10431433", generic_dsd_config },
 	{ "CSC3551", "10431463", generic_dsd_config },
Signature · 1 line
2.55.0
#2 lore raw

Hi Takashi,

We should be able to add entries for this laptop and others in cs35l41_hda_property.c to override all the missing properties from ACPI. The ACPI in those laptops is malformed, which means the entire _DSD for the CS35L41 is corrupted, and appears to be missing from the kernels perspective.
In most cases the malformed ACPI is also not correct for that laptop, so its not a case of taking the settings from the ACPI on the laptop itself and putting them in cs35l41_hda_property.c.

However, we should be able to determine these settings ourselves, but end users will not be able to.
We are actively looking at doing so for this laptop, and several others.

Thanks,
Stefan

Quoted 7 lines

-----Original Message-----
From: Takashi Iwai <[email protected]>
Sent: Thursday, July 16, 2026 11:30 AM
To: Yufan <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected];

[email protected]

Quoted 16 lines

Subject: Re: [PATCH] ALSA: hda: cs35l41: Fix Speakers on HP OmniBook 7 16- bh0xxx

Quoted 11 lines · Yufan

From: Yufan Kuang <[email protected]>

The HP OmniBook 7 Laptop 16-bh0xxx (SSID 103C8E60) has two CS35L41 amplifiers connected via I2C with internal boost, but its _DSD provides cirrus,dev-index as a Package while the driver expects per-device scalar values, causing the probe to fail with -EINVAL, leaving only the left speaker working.

Add the SSID to the configuration and prop-model override tables.

Quoted 7 lines
Quoted 5 lines

Tested on a HP OmniBook 7 Laptop 16-bh0xxx (BIOS F.12) with CachyOS 7.1.3-2; speaker-test confirms both channels work correctly.

Assisted-by Copilot:DeepSeek-V4-Pro

The very same change has been already submitted (maybe multiple times),

e.g.

Quoted 8 lines

https://lore.kernel.org/[email protected]

The problem is known, and it's basically an issue of the firmware. But obviously we need to address *something* in the driver side.

Cirrus people -- do you have ideas to fix those HP laptops?

If it's not feasible in a short term, maybe it'd be better to take those

quirk

Quoted 1 line

additions; although they aren't 100% correct, but they do work somehow in

a

Quoted 12 lines

certain level.

thanks,

Takashi

Quoted 4 lines
---
 sound/hda/codecs/side-codecs/cs35l41_hda_property.c | 2 ++
 1 file changed, 2 insertions(+)
Quoted 6 lines

sound/hda/codecs/side-codecs/cs35l41_hda_property.c static const struct cs35l41_config

Quoted 6 lines
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +0 -0
index 416d7bf3e289..e14a87336478 100644
@@ -85,6 +85,7 @@ static const struct cs35l41_config

cs35l41_config_table[]

Quoted 2 lines

= {

Quoted 1 line
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +0 -0
 	{ "103C8C51", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,

1,

Quoted 2 lines

-1, 1000, 4100, 24 },

Quoted 1 line
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +0 -0
 	{ "103C8CDD", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,

1,

Quoted 6 lines

-1, 1000, 4100, 24 },

sound/hda/codecs/side-codecs/cs35l41_hda_property.c static const struct cs35l41_config · line 91

Quoted 5 lines
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +2 -0
 	{ "103C8CDE", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,
 1, -1, 1000, 3900, 24 },
+	{ "103C8E60", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,
+1, -1, 1000, 4500, 24 },
 	{ "104312AF", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 1,

2,

Quoted 2 lines

0, 1000, 4500, 24 },

Quoted 1 line

{ "10431433", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,

1,

Quoted 14 lines

-1, 1000, 4500, 24 },

Quoted 2 lines

{ "10431463", 2, INTERNAL, { CS35L41_LEFT, CS35L41_RIGHT, 0, 0 }, 0,
1, -1, 1000, 4500, 24 }, @@ -511,6 +512,7 @@ static const struct

cs35l41_prop_model cs35l41_prop_model_table[] = {

Quoted 10 lines
sound/hda/codecs/side-codecs/cs35l41_hda_property.c +1 -1
 	{ "CSC3551", "103C8C6A", hp_i2c_int_2amp_dual_spkid },
 	{ "CSC3551", "103C8CDD", generic_dsd_config },
 	{ "CSC3551", "103C8CDE", generic_dsd_config },
+	{ "CSC3551", "103C8E60", generic_dsd_config },
 	{ "CSC3551", "104312AF", generic_dsd_config },
 	{ "CSC3551", "10431433", generic_dsd_config },
 	{ "CSC3551", "10431463", generic_dsd_config },
--

2.55.0

#3 lore raw
Quoted 12 lines · Stefan Binding (Opensource)

Hi Takashi,

We should be able to add entries for this laptop and others in cs35l41_hda_property.c to override all the missing properties from ACPI. The ACPI in those laptops is malformed, which means the entire _DSD for the CS35L41 is corrupted, and appears to be missing from the kernels perspective.
In most cases the malformed ACPI is also not correct for that laptop, so its not a case of taking the settings from the ACPI on the laptop itself and putting them in cs35l41_hda_property.c.

Quoted 3 lines

However, we should be able to determine these settings ourselves, but end users will not be able to.
We are actively looking at doing so for this laptop, and several others.

OK, then I'll keep waiting for your solution at first.

Thanks!

Takashi

A message that is not in the archive (sent off-list or not yet received): [email protected]

author Tested-by
#5 lore raw

From: Yufan Kuang <[email protected]>

Hi Takashi and all,

Thank you for the review. I missed Sven's earlier submission — I only searched the source tree and commit history for the SSID, not the mailing list.

Since his v2 patch is identical and predates mine, please drop my submission.

I have tested the identical quirk on my HP OmniBook 7 (BIOS F.12) and both speakers work correctly. If it helps, you can add:

to Sven's patch.

Thanks,
Yufan