"Upgrade" from 5.14.18-300 kernel to 5.15.4-201 kernel broke sound (pipewire)

Ask Fedora
, ,
1

I upgraded my kernel from 5.14.18-300 to 5.15.4.201 and my sound stopped working. If I boot the old kernel from grub, its OK. If I boot 5.15.4.201 or 5.15.5.200, I get no sound.
Journal show errors like:
Nov 30 09:29:25 book pipewire[1846]: spa.alsa: hw:0,0: Channels doesn’t match (requested 64, got 2)
After these errors, pipewire logs:
Nov 30 09:29:44 book pipewire[1846]: spa.alsa: hw:0,0: Channels doesn’t match (requested 64, got 2)
Nov 30 09:29:44 book kernel: snd_soc_skl 0000:00:1f.3: Blob NULL for id:0 type:3 dirn:0 ch:2, freq:48000, fmt:24
Nov 30 09:29:44 book kernel: snd_soc_skl 0000:00:1f.3: ASoC: error at snd_soc_dai_hw_params on SSP0 Pin: -22
Nov 30 09:29:44 book pipewire[1846]: spa.alsa: set_hw_params: Invalid argument
Nov 30 09:29:44 book kernel: SSP0-Codec: ASoC: soc_pcm_hw_params() failed (-22)
Nov 30 09:29:44 book kernel: Kbl Audio Port: ASoC: dpcm_fe_dai_hw_params failed (-22)
Nov 30 09:29:44 book pipewire[1846]: pw.node: (alsa_output.platform-kbl_da7219_mx98357a.playback.0.0-43) suspended → error (Start error: Invalid argument)
Nov 30 09:29:44 book pipewire[1846]: spa.alsa: hw:0,0: Channels doesn’t match (requested 64, got 2)
Nov 30 09:29:44 book pipewire[1846]: spa.audioadapter: params Spa:Enum:ParamId:EnumFormat: 1:0 (convert format) Success

This seems to be linked to the kernel upgrade, since I can boot into the old kernel with success. The upgrade release notes do not reflect any audio changes.

2

How is your speaker being connected?

Analog outputs? HDMI/Displayport? Bluetooth? USB?

3

This is a laptop. The internal speakers are connected to a max98357a amp. The internal audio uses the kbl_da7219_mx98357a card driver. When it was working, I also played sound through bluetooth speaker occasionally. Both the internal speakers and the BT worked well. I just tested the bluetooth. I get sounds out the bluetooth speakers! I would really like the internal speakers to start working again.

1 Like
4

I reported this problem to the pipewire team. They said it was probably a kernel driver issue.

1 Like
5

@xinato Nice to meet you. A new kernel driver is on its way (always a new one). New update is kernel 5.15.5-200.fc35
Is the problem still there when the kernel update is installed? Please let it know.
Did you write a bug report to bugzilla? Bugs and feature requests - Fedora Project Wiki

1 Like
6

Thank you for your interest. I tried 5.15.5-200.fc35 a few hours ago and I still have the same problem. I noted this fact above. Maybe I need another new driver.

I see the following in a verbose pipewire log that seems to point at a problem. It looks like the driver does not like the parameters that it is getting from pipewire, through alsa. I have not filed a bug report with Fedora, but I will happily do that this evening. I appreciate all the help I have been getting!!

Nov 30 16:44:11 book pipewire[17116]: spa.alsa: 0x5585224d0918: Stream parameters are 48000Hz fmt:S32_LE access:mmap-interleaved channels:2
Nov 30 16:44:11 book pipewire[17116]: spa.alsa: hw:0,0 (playback): format:S32_LE access:mmap-interleaved rate:48000 channels:2 buffer frames 32768, period frames 1024, periods 32, frame_size 8 headroom 0 start-delay:0
Nov 30 16:44:11 book pipewire[17116]: spa.alsa: set_hw_params: Invalid argument
Nov 30 16:44:11 book pipewire[17116]: spa.alsa: 0x5585224d0918: Device ‘hw:0,0’ closing
Nov 30 16:44:11 book kernel: SSP0-Codec: ASoC: soc_pcm_hw_params() failed (-22)
Nov 30 16:44:11 book kernel: Kbl Audio Port: ASoC: dpcm_fe_dai_hw_params failed (-22)
Nov 30 16:44:11 book pipewire[17116]: spa.system: 0x5585223ae748: close fd:82
Nov 30 16:44:11 book pipewire[17116]: pw.node: 0x55852246b8a0: (alsa_output.platform-kbl_da7219_mx98357a.playback.0.0) suspended → error (Start error: Invalid argument)
Nov 30 16:44:11 book pipewire[17116]: pw.node: (alsa_output.platform-kbl_da7219_mx98357a.playback.0.0-46) suspended → error (Start error: Invalid argument)

1 Like
7

Here is the same log clip from the old kernel that works. The parameters that pipewire passes are the same. snd_soc_skl spits out another error about “MCPS Budget Violation” that I have not investigated, but the sound plays through the built in speakers. Perhaps there was a change to the driver that catches the cause of the MCPS message and returns an error to alsa and pipewire.

Nov 30 20:38:05 book pipewire[2743]: spa.alsa: 0x55ec529ec9d8: Stream parameters are 48000Hz fmt:S32_LE access:mmap-interleaved channels:2
Nov 30 20:38:05 book pipewire[2743]: spa.alsa: hw:0,0 (playback): format:S32_LE access:mmap-interleaved rate:48000 channels:2 buffer frames 32768, period frames 1024, periods 32, frame_size 8 headroom 0 start-delay:0
Nov 30 20:38:05 book pipewire[2743]: pw.node: 0x55ec529eb120: flags:00000001 change_mask:00000002 max_in:1 max_out:0
Nov 30 20:38:05 book kernel: snd_soc_skl 0000:00:1f.3: MCPS Budget Violation: 9b050000
Nov 30 20:38:05 book kernel: snd_soc_skl 0000:00:1f.3: MCPS Budget Violation: 9b050000
Nov 30 20:38:05 book pipewire[2743]: pw.node: 0x55ec529eb120: updated 1 properties
Nov 30 20:38:05 book pipewire[2743]: pw.node: 0x55ec529eb120: priority driver 1000
Nov 30 20:38:05 book pipewire[2743]: pw.node: (alsa_output.platform-kbl_da7219_mx98357a.playback.0.0-0) max-latency:0/0 → 8192/48000

1 Like