Linux kernel -stable discussions
 help / color / mirror / Atom feed
* [REGRESSION] ALSA: hda/realtek: no speaker sound on Lenovo Legion 16IRX9H since f2293e7b2ddd (PCI SSID 17aa:3842 shared with ThinkBook X G2 IAH)
@ 2026-11-16  1:24 william
  2026-11-16  8:29 ` Thorsten Leemhuis
  0 siblings, 1 reply; 3+ messages in thread
From: william @ 2026-11-16  1:24 UTC (permalink / raw)
  To: linux-sound; +Cc: tiwai, evgeny, stable, regressions

Hi,
Commit f2293e7b2ddd ("ALSA: hda/realtek: Enable all four speaker amps on Lenovo
ThinkBook X G2 IAH", Cc: stable) matches the laptop by PCI SSID
17aa:3842, on the
assumption that this SSID had no quirk entry. The Lenovo Legion 16IRX9H (83DE)
has the same PCI SSID, so its internal speakers are silent on kernels that carry
this commit. Headphones work.
Hardware
Lenovo Legion 16IRX9H, product 83DE, BIOS N2CN32WW
Codec: Realtek ALC287, codec SSID 17aa:38cd, PCI SSID 17aa:3842
Amplifier: ACPI TIAS2781:00 (i2c-TIAS2781:00 on i2c_designware.2),
driver snd_hda_scodec_tas2781_i2c
Kernels
Works: 7.2.10-200.nobara.fc44, 7.1.6-1-cachyos (live ISO)
Fails: 7.3.4-201.nobara.fc44, 7.3.4-1-cachyos
The 3842 entry is absent from v7.2 and v7.3, present in v7.3.4, and still
present in v7.4-rc1 and master as of 2026-11-15 (checked via the GitHub mirror).
Analysis
snd_hda_pick_fixup() walks the table and matches PCI-SSID entries before the
codec-SSID fallback. Before the commit, 17aa:3842 had no entry, so this machine
fell through to the existing codec-SSID entry
SND_PCI_QUIRK(0x17aa, 0x38cd, "Y790 VECO DUAL", ALC287_FIXUP_TAS2781_I2C),
which works. With the commit, the 3842 entry wins and chains to
cs35l41_fixup_i2c_four(), so the codec master waits for four CS35L41 components
that do not exist on this machine.
Evidence on 7.3.4
dmesg: "ALC287: picked fixup for PCI SSID 17aa:3842"; no "bound i2c-TIAS2781:00"
line (on 7.2.10 that line is present).
/sys/kernel/debug/device_component/hdaudioC0D0: master "not bound",
four entries "(unknown) not registered".
With dyndbg, after "echo 1 > /sys/class/sound/hwC0D0/reconfig":
hdaudioC0D0: trying to bring up adev
hdaudioC0D0: Looking for component 0
hdaudioC0D0: master has incomplete components
After unbinding/binding i2c-TIAS2781:00, tas2781-hda logs "adding component"
but the master still reports incomplete components.
Workaround confirmed
Booting with snd_hda_intel.model=17aa:38cd (alias SSID, handled first by the
picker) gives:
ALC287: picked fixup for alias SSID 17aa:38cd
bound i2c-TIAS2781:00 (ops tas2781_hda_comp_ops [snd_hda_scodec_tas2781_i2c])
and the speakers work.
Possible direction
Since both laptops share the PCI SSID, the ThinkBook entry cannot be fixed by
PCI SSID alone. One option is a codec-SSID entry for this machine placed before
the 0x3842 line, e.g.
HDA_CODEC_QUIRK(0x17aa, 0x38cd, "Legion 16IRX9H / Y790 VECO DUAL",
ALC287_FIXUP_TAS2781_I2C),
I have not compiled or tested a patch, only the model= alias above. I am happy
to test patches on this hardware.
Separate observation
After the workaround, speakers still stopped after some idle time. It went away
with snd_hda_intel power_save=0 power_save_controller=N and keeping
i2c-TIAS2781:00 power/control at "on". I applied both at once, so I do not know
which is required. This looks unrelated to the regression above.

I'd really appreciate it if the issue is fixed, as I'm now having the
same problem when the laptop is turned on after being suspended for a
while, and I haven't been able to solve it.
Thanks,
William

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [REGRESSION] ALSA: hda/realtek: no speaker sound on Lenovo Legion 16IRX9H since f2293e7b2ddd (PCI SSID 17aa:3842 shared with ThinkBook X G2 IAH)
  2026-11-16  1:24 [REGRESSION] ALSA: hda/realtek: no speaker sound on Lenovo Legion 16IRX9H since f2293e7b2ddd (PCI SSID 17aa:3842 shared with ThinkBook X G2 IAH) william
@ 2026-11-16  8:29 ` Thorsten Leemhuis
  2026-11-16 12:34   ` Sasha Levin
  0 siblings, 1 reply; 3+ messages in thread
From: Thorsten Leemhuis @ 2026-11-16  8:29 UTC (permalink / raw)
  To: Greg KH, Sasha Levin
  Cc: tiwai, william, evgeny, stable, regressions, linux-sound

On 11/16/26 02:24, william wrote:
> Commit f2293e7b2ddd ("ALSA: hda/realtek: Enable all four speaker amps on Lenovo
> ThinkBook X G2 IAH", Cc: stable) matches the laptop by PCI SSID
> 17aa:3842, on the
> assumption that this SSID had no quirk entry. The Lenovo Legion 16IRX9H (83DE)
> has the same PCI SSID, so its internal speakers are silent on kernels that carry
> this commit. Headphones work.

Greg, Sasha, that change made it to 7.3.y already, but is in the current
6.18.y-rc afaics -- so might be something to drop until this is sorted out.

Ciao, Thorsten

> Hardware
> Lenovo Legion 16IRX9H, product 83DE, BIOS N2CN32WW
> Codec: Realtek ALC287, codec SSID 17aa:38cd, PCI SSID 17aa:3842
> Amplifier: ACPI TIAS2781:00 (i2c-TIAS2781:00 on i2c_designware.2),
> driver snd_hda_scodec_tas2781_i2c
> Kernels
> Works: 7.2.10-200.nobara.fc44, 7.1.6-1-cachyos (live ISO)
> Fails: 7.3.4-201.nobara.fc44, 7.3.4-1-cachyos
> The 3842 entry is absent from v7.2 and v7.3, present in v7.3.4, and still
> present in v7.4-rc1 and master as of 2026-11-15 (checked via the GitHub mirror).
> Analysis
> snd_hda_pick_fixup() walks the table and matches PCI-SSID entries before the
> codec-SSID fallback. Before the commit, 17aa:3842 had no entry, so this machine
> fell through to the existing codec-SSID entry
> SND_PCI_QUIRK(0x17aa, 0x38cd, "Y790 VECO DUAL", ALC287_FIXUP_TAS2781_I2C),
> which works. With the commit, the 3842 entry wins and chains to
> cs35l41_fixup_i2c_four(), so the codec master waits for four CS35L41 components
> that do not exist on this machine.
> Evidence on 7.3.4
> dmesg: "ALC287: picked fixup for PCI SSID 17aa:3842"; no "bound i2c-TIAS2781:00"
> line (on 7.2.10 that line is present).
> /sys/kernel/debug/device_component/hdaudioC0D0: master "not bound",
> four entries "(unknown) not registered".
> With dyndbg, after "echo 1 > /sys/class/sound/hwC0D0/reconfig":
> hdaudioC0D0: trying to bring up adev
> hdaudioC0D0: Looking for component 0
> hdaudioC0D0: master has incomplete components
> After unbinding/binding i2c-TIAS2781:00, tas2781-hda logs "adding component"
> but the master still reports incomplete components.
> Workaround confirmed
> Booting with snd_hda_intel.model=17aa:38cd (alias SSID, handled first by the
> picker) gives:
> ALC287: picked fixup for alias SSID 17aa:38cd
> bound i2c-TIAS2781:00 (ops tas2781_hda_comp_ops [snd_hda_scodec_tas2781_i2c])
> and the speakers work.
> Possible direction
> Since both laptops share the PCI SSID, the ThinkBook entry cannot be fixed by
> PCI SSID alone. One option is a codec-SSID entry for this machine placed before
> the 0x3842 line, e.g.
> HDA_CODEC_QUIRK(0x17aa, 0x38cd, "Legion 16IRX9H / Y790 VECO DUAL",
> ALC287_FIXUP_TAS2781_I2C),
> I have not compiled or tested a patch, only the model= alias above. I am happy
> to test patches on this hardware.
> Separate observation
> After the workaround, speakers still stopped after some idle time. It went away
> with snd_hda_intel power_save=0 power_save_controller=N and keeping
> i2c-TIAS2781:00 power/control at "on". I applied both at once, so I do not know
> which is required. This looks unrelated to the regression above.
> 
> I'd really appreciate it if the issue is fixed, as I'm now having the
> same problem when the laptop is turned on after being suspended for a
> while, and I haven't been able to solve it.
> Thanks,
> William


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [REGRESSION] ALSA: hda/realtek: no speaker sound on Lenovo Legion 16IRX9H since f2293e7b2ddd (PCI SSID 17aa:3842 shared with ThinkBook X G2 IAH)
  2026-11-16  8:29 ` Thorsten Leemhuis
@ 2026-11-16 12:34   ` Sasha Levin
  0 siblings, 0 replies; 3+ messages in thread
From: Sasha Levin @ 2026-11-16 12:34 UTC (permalink / raw)
  To: Thorsten Leemhuis, william
  Cc: Sasha Levin, Greg KH, tiwai, evgeny, stable, regressions,
	linux-sound

On Mon, Nov 16, 2026 at 09:29:45AM +0100, Thorsten Leemhuis wrote:
> Greg, Sasha, that change made it to 7.3.y already, but is in the current
> 6.18.y-rc afaics -- so might be something to drop until this is sorted out.

Thanks for the report, William and Thorsten. I've dropped f2293e7b2ddd ("ALSA:
hda/realtek: Enable all four speaker amps on Lenovo ThinkBook X G2 IAH") from
the 6.18 queue, so it will not be in 6.18.64.

It already shipped in 7.3.4, so 7.3.y will need an upstream fix (for example a
codec-SSID quirk for the Legion) that we can then backport.

-- 
Thanks,
Sasha

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-16 12:34 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16  1:24 [REGRESSION] ALSA: hda/realtek: no speaker sound on Lenovo Legion 16IRX9H since f2293e7b2ddd (PCI SSID 17aa:3842 shared with ThinkBook X G2 IAH) william
2026-11-16  8:29 ` Thorsten Leemhuis
2026-11-16 12:34   ` Sasha Levin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox