possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Samantha Glocker]

Hi, I'm a kernel maintainer from nobara and ASUS-Linux.

I've bifurcated an issue with sound on Lenovo Legion laptop(s)
starting from mainline 6.12.5, verified on Fedora proper, Nobara, and
CachyOS.

From the changelog,
https://patch.msgid.link/20241008120233.7154-4-tiwai@suse.de looks
like a likely candidate due to touching TAS2781 / TIA2781.

The issue will possibly affect sound on any or all Lenovo Legion Gen 8 and 9.

The obvious difference in dmesg on my Legion Pro 7 Gen 8 (16ARX8H) on
6.12.5, is the following line is completely missing:
[    4.761665] snd_hda_codec_realtek hdaudioC1D0: bound
i2c-TIAS2781:00 (ops tas2781_hda_comp_ops
[snd_hda_scodec_tas2781_i2c])

This suggests the TI amplifier is never identified.

(A second issue is LenovoLegionLinux DKMS not loading at all as of
6.12.5. I believe this is a separate issue, but it could be related to
how that project detects laptop models. Mentioning it here for
completeness.)

Best regards,
Samantha Glocker (SammiLucia)

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Takashi Iwai]

On Wed, 22 Jan 2025 01:58:16 +0100,
Samantha Glocker wrote:
> 
> Hi, I'm a kernel maintainer from nobara and ASUS-Linux.
> 
> I've bifurcated an issue with sound on Lenovo Legion laptop(s)
> starting from mainline 6.12.5, verified on Fedora proper, Nobara, and
> CachyOS.
> 
> From the changelog,
> https://patch.msgid.link/20241008120233.7154-4-tiwai@suse.de looks
> like a likely candidate due to touching TAS2781 / TIA2781.
> 
> The issue will possibly affect sound on any or all Lenovo Legion Gen 8 and 9.
> 
> The obvious difference in dmesg on my Legion Pro 7 Gen 8 (16ARX8H) on
> 6.12.5, is the following line is completely missing:
> [    4.761665] snd_hda_codec_realtek hdaudioC1D0: bound
> i2c-TIAS2781:00 (ops tas2781_hda_comp_ops
> [snd_hda_scodec_tas2781_i2c])
> 
> This suggests the TI amplifier is never identified.

What's the exact problem?  Is it a regression?  Any original report on
the bug tracker or such?


thanks,

Takashi

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Takashi Iwai]

On Wed, 22 Jan 2025 08:21:26 +0100,
Takashi Iwai wrote:
> 
> On Wed, 22 Jan 2025 01:58:16 +0100,
> Samantha Glocker wrote:
> > 
> > Hi, I'm a kernel maintainer from nobara and ASUS-Linux.
> > 
> > I've bifurcated an issue with sound on Lenovo Legion laptop(s)
> > starting from mainline 6.12.5, verified on Fedora proper, Nobara, and
> > CachyOS.
> > 
> > From the changelog,
> > https://patch.msgid.link/20241008120233.7154-4-tiwai@suse.de looks
> > like a likely candidate due to touching TAS2781 / TIA2781.
> > 
> > The issue will possibly affect sound on any or all Lenovo Legion Gen 8 and 9.
> > 
> > The obvious difference in dmesg on my Legion Pro 7 Gen 8 (16ARX8H) on
> > 6.12.5, is the following line is completely missing:
> > [    4.761665] snd_hda_codec_realtek hdaudioC1D0: bound
> > i2c-TIAS2781:00 (ops tas2781_hda_comp_ops
> > [snd_hda_scodec_tas2781_i2c])
> > 
> > This suggests the TI amplifier is never identified.
> 
> What's the exact problem?  Is it a regression?  Any original report on
> the bug tracker or such?

... and if it's a regression, please give alsa-info.sh outputs from
both working and non-working cases.  Run the script with --no-upload
option and attach the outputs.


Takashi

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Takashi Iwai]

On Wed, 22 Jan 2025 08:23:56 +0100,
Takashi Iwai wrote:
> 
> On Wed, 22 Jan 2025 08:21:26 +0100,
> Takashi Iwai wrote:
> > 
> > On Wed, 22 Jan 2025 01:58:16 +0100,
> > Samantha Glocker wrote:
> > > 
> > > Hi, I'm a kernel maintainer from nobara and ASUS-Linux.
> > > 
> > > I've bifurcated an issue with sound on Lenovo Legion laptop(s)
> > > starting from mainline 6.12.5, verified on Fedora proper, Nobara, and
> > > CachyOS.
> > > 
> > > From the changelog,
> > > https://patch.msgid.link/20241008120233.7154-4-tiwai@suse.de looks
> > > like a likely candidate due to touching TAS2781 / TIA2781.
> > > 
> > > The issue will possibly affect sound on any or all Lenovo Legion Gen 8 and 9.
> > > 
> > > The obvious difference in dmesg on my Legion Pro 7 Gen 8 (16ARX8H) on
> > > 6.12.5, is the following line is completely missing:
> > > [    4.761665] snd_hda_codec_realtek hdaudioC1D0: bound
> > > i2c-TIAS2781:00 (ops tas2781_hda_comp_ops
> > > [snd_hda_scodec_tas2781_i2c])
> > > 
> > > This suggests the TI amplifier is never identified.
> > 
> > What's the exact problem?  Is it a regression?  Any original report on
> > the bug tracker or such?
> 
> ... and if it's a regression, please give alsa-info.sh outputs from
> both working and non-working cases.  Run the script with --no-upload
> option and attach the outputs.

If it's a device with PCI SSID 17aa:386f, the patch below might fix.
Let me know if it is.


thanks,

Takashi

-- 8< --
--- a/sound/pci/hda/patch_realtek.c
+++ b/sound/pci/hda/patch_realtek.c
@@ -10918,7 +10918,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x17aa, 0x3869, "Lenovo Yoga7 14IAL7", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
 	HDA_CODEC_QUIRK(0x17aa, 0x386e, "Legion Y9000X 2022 IAH7", ALC287_FIXUP_CS35L41_I2C_2),
 	SND_PCI_QUIRK(0x17aa, 0x386e, "Yoga Pro 7 14ARP8", ALC285_FIXUP_SPEAKER2_TO_DAC1),
-	HDA_CODEC_QUIRK(0x17aa, 0x386f, "Legion Pro 7 16ARX8H", ALC287_FIXUP_TAS2781_I2C),
+	HDA_CODEC_QUIRK(0x17aa, 0x38a8, "Legion Pro 7 16ARX8H", ALC287_FIXUP_TAS2781_I2C), /* this must match before PCI SSID 17aa:386f below */
 	SND_PCI_QUIRK(0x17aa, 0x386f, "Legion Pro 7i 16IAX7", ALC287_FIXUP_CS35L41_I2C_2),
 	SND_PCI_QUIRK(0x17aa, 0x3870, "Lenovo Yoga 7 14ARB7", ALC287_FIXUP_YOGA7_14ARB7_I2C),
 	SND_PCI_QUIRK(0x17aa, 0x3877, "Lenovo Legion 7 Slim 16ARHA7", ALC287_FIXUP_CS35L41_I2C_2),

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Samantha Glocker]

[-- Attachment #1: Type: text/plain, Size: 3188 bytes --]

Hi Takashi,

Yes confirming it's a regression. I haven't raised it on bug tracker
because I was pretty sure I'd located the regression as above.

Sound was working on 6.12.4, and non-working from 6.12.5 (thru 6.12.10)

I've attached alsa-info from 6.12.4-nobara (working) and 6.12.10
(non-working), which should be similar output to 6.12.5, and should be
similar to stock Fedora.

Confirming this is a PCI SSID 17aa:386f... I'll try the patch - thank you!

I'll come back to you with the outcome.

Best regards,

Samantha Glocker

On Wed, Jan 22, 2025 at 2:53 AM Takashi Iwai <tiwai@suse.de> wrote:
>
> On Wed, 22 Jan 2025 08:23:56 +0100,
> Takashi Iwai wrote:
> >
> > On Wed, 22 Jan 2025 08:21:26 +0100,
> > Takashi Iwai wrote:
> > >
> > > On Wed, 22 Jan 2025 01:58:16 +0100,
> > > Samantha Glocker wrote:
> > > >
> > > > Hi, I'm a kernel maintainer from nobara and ASUS-Linux.
> > > >
> > > > I've bifurcated an issue with sound on Lenovo Legion laptop(s)
> > > > starting from mainline 6.12.5, verified on Fedora proper, Nobara, and
> > > > CachyOS.
> > > >
> > > > From the changelog,
> > > > https://patch.msgid.link/20241008120233.7154-4-tiwai@suse.de looks
> > > > like a likely candidate due to touching TAS2781 / TIA2781.
> > > >
> > > > The issue will possibly affect sound on any or all Lenovo Legion Gen 8 and 9.
> > > >
> > > > The obvious difference in dmesg on my Legion Pro 7 Gen 8 (16ARX8H) on
> > > > 6.12.5, is the following line is completely missing:
> > > > [    4.761665] snd_hda_codec_realtek hdaudioC1D0: bound
> > > > i2c-TIAS2781:00 (ops tas2781_hda_comp_ops
> > > > [snd_hda_scodec_tas2781_i2c])
> > > >
> > > > This suggests the TI amplifier is never identified.
> > >
> > > What's the exact problem?  Is it a regression?  Any original report on
> > > the bug tracker or such?
> >
> > ... and if it's a regression, please give alsa-info.sh outputs from
> > both working and non-working cases.  Run the script with --no-upload
> > option and attach the outputs.
>
> If it's a device with PCI SSID 17aa:386f, the patch below might fix.
> Let me know if it is.
>
>
> thanks,
>
> Takashi
>
> -- 8< --
> --- a/sound/pci/hda/patch_realtek.c
> +++ b/sound/pci/hda/patch_realtek.c
> @@ -10918,7 +10918,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
>         SND_PCI_QUIRK(0x17aa, 0x3869, "Lenovo Yoga7 14IAL7", ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
>         HDA_CODEC_QUIRK(0x17aa, 0x386e, "Legion Y9000X 2022 IAH7", ALC287_FIXUP_CS35L41_I2C_2),
>         SND_PCI_QUIRK(0x17aa, 0x386e, "Yoga Pro 7 14ARP8", ALC285_FIXUP_SPEAKER2_TO_DAC1),
> -       HDA_CODEC_QUIRK(0x17aa, 0x386f, "Legion Pro 7 16ARX8H", ALC287_FIXUP_TAS2781_I2C),
> +       HDA_CODEC_QUIRK(0x17aa, 0x38a8, "Legion Pro 7 16ARX8H", ALC287_FIXUP_TAS2781_I2C), /* this must match before PCI SSID 17aa:386f below */
>         SND_PCI_QUIRK(0x17aa, 0x386f, "Legion Pro 7i 16IAX7", ALC287_FIXUP_CS35L41_I2C_2),
>         SND_PCI_QUIRK(0x17aa, 0x3870, "Lenovo Yoga 7 14ARB7", ALC287_FIXUP_YOGA7_14ARB7_I2C),
>         SND_PCI_QUIRK(0x17aa, 0x3877, "Lenovo Legion 7 Slim 16ARHA7", ALC287_FIXUP_CS35L41_I2C_2),
>

[-- Attachment #2: alsa-info-16ARX8H.zip --]
[-- Type: application/zip, Size: 15634 bytes --]

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Takashi Iwai]

On Thu, 23 Jan 2025 03:00:29 +0100,
Samantha Glocker wrote:
> 
> Hi Takashi,
> 
> Yes confirming it's a regression. I haven't raised it on bug tracker
> because I was pretty sure I'd located the regression as above.
> 
> Sound was working on 6.12.4, and non-working from 6.12.5 (thru 6.12.10)

What exactly do you mean as "sound was working (or non-working)"?
I suppose it's about the speaker output.  Or anything else broken?
The symptom needs to be reported more clearly -- only few people own
crystal balls.


thanks,

Takashi

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Samantha Glocker]

[-- Attachment #1: Type: text/plain, Size: 2340 bytes --]

Hi Takashi,

Sorry for the delay I had a build pipeline problem.

Confirming your TIA patch works perfectly on my 16ARX8H. The TI amp is
detected and enabled correctly again.

We're including it in the nobara and cachyOS kernels.

Thank you,
Samantha Glocker

---
diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c
index 4210bc8..453835b 100644
--- a/sound/pci/hda/patch_realtek.c
+++ b/sound/pci/hda/patch_realtek.c
@@ -10888,7 +10888,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = {
  SND_PCI_QUIRK(0x17aa, 0x3869, "Lenovo Yoga7 14IAL7",
ALC287_FIXUP_YOGA9_14IAP7_BASS_SPK_PIN),
  HDA_CODEC_QUIRK(0x17aa, 0x386e, "Legion Y9000X 2022 IAH7",
ALC287_FIXUP_CS35L41_I2C_2),
  SND_PCI_QUIRK(0x17aa, 0x386e, "Yoga Pro 7 14ARP8",
ALC285_FIXUP_SPEAKER2_TO_DAC1),
- HDA_CODEC_QUIRK(0x17aa, 0x386f, "Legion Pro 7 16ARX8H",
ALC287_FIXUP_TAS2781_I2C),
+ HDA_CODEC_QUIRK(0x17aa, 0x38a8, "Legion Pro 7 16ARX8H",
ALC287_FIXUP_TAS2781_I2C), /* this must match before PCI SSID
17aa:386f below */
  SND_PCI_QUIRK(0x17aa, 0x386f, "Legion Pro 7i 16IAX7",
ALC287_FIXUP_CS35L41_I2C_2),
  SND_PCI_QUIRK(0x17aa, 0x3870, "Lenovo Yoga 7 14ARB7",
ALC287_FIXUP_YOGA7_14ARB7_I2C),
  SND_PCI_QUIRK(0x17aa, 0x3877, "Lenovo Legion 7 Slim 16ARHA7",
ALC287_FIXUP_CS35L41_I2C_2),
---

On Fri, Jan 24, 2025 at 3:04 AM Samantha Glocker
<iam@anislandsomewhere.com> wrote:
>
> The Texas Instruments amp was not detected on kernel 6.12.5.
>
> You've understood perfectly. The patch you provided also looks good.
>
> I just finished compiling a test kernel, I will report back tomorrow.
>
> Thank you,
>
> Samantha Glocker
>
> On Thu, Jan 23, 2025, 05:07 Takashi Iwai <tiwai@suse.de> wrote:
>>
>> On Thu, 23 Jan 2025 03:00:29 +0100,
>> Samantha Glocker wrote:
>> >
>> > Hi Takashi,
>> >
>> > Yes confirming it's a regression. I haven't raised it on bug tracker
>> > because I was pretty sure I'd located the regression as above.
>> >
>> > Sound was working on 6.12.4, and non-working from 6.12.5 (thru 6.12.10)
>>
>> What exactly do you mean as "sound was working (or non-working)"?
>> I suppose it's about the speaker output.  Or anything else broken?
>> The symptom needs to be reported more clearly -- only few people own
>> crystal balls.
>>
>>
>> thanks,
>>
>> Takashi

[-- Attachment #2: alsa-info.patched.working.6.12.10.zip --]
[-- Type: application/zip, Size: 8051 bytes --]

Re: possible issue with ALSA: hda/realtek: Use codec SSID matching for Lenovo devices

[Takashi Iwai]

On Fri, 24 Jan 2025 23:58:29 +0100,
Samantha Glocker wrote:
> 
> Hi Takashi,
> 
> Sorry for the delay I had a build pipeline problem.
> 
> Confirming your TIA patch works perfectly on my 16ARX8H. The TI amp is
> detected and enabled correctly again.
> 
> We're including it in the nobara and cachyOS kernels.

Good to hear.  I'm going to submit the fix patch for the upstream.


Takashi