[FEATURE] Xiaomi Redmibook 16 2024 mic is not working

[OrelSokolov]

So can anyone help with implementing firmware for this model?
What should I provide?

options snd-hda-intel dmic_detect=0 currently works for sound, but internal microphone does not work.
3.5 headset works.

Thanks.

UPD:
Windows works fine with internal mic, device info:

INTELAUDIO\CTLR_DEV_7E28&LINKTYPE_02&DEVTYPE_00&VEN_8086&DEV_AE20&SUBSYS_1D722309&REV_10EC\5&1ed5312d&0&0000

[plbossart]

please share the output of 'alsa-info', it's not clear that it's an SOF problem but more likely a Linux hardware detection issue.

[OrelSokolov]

@plbossart https://alsa-project.org/db/?f=d737cae0b969ade16eb2ee0cc55ef12687a840ff

[lgirdwood]

Looks like this issue is unrelated to SOF and is a platform issue with HDA mic integration.

[plbossart]

please share alsa-info without this "option snd_intel_dspcfg: dsp_driver=1"

[OrelSokolov]

@plbossart how to do this? I ran just alsa-info without any options.

[plbossart]

you have this option somewhere in /etc/modprobe.d/*.conf. Remove it, reboot and send the results of alsa-info.

[plbossart]

double check the results,

!!Modprobe options (Sound related)
!!--------------------------------

snd_pcsp: index=-2
snd_atiixp_modem: index=-2
snd_intel8x0m: index=-2
snd_via82xx_modem: index=-2
snd_atiixp_modem: index=-2
snd_intel8x0m: index=-2
snd_via82xx_modem: index=-2
snd_usb_audio: index=-2
snd_usb_caiaq: index=-2
snd_usb_ua101: index=-2
snd_usb_us122l: index=-2
snd_usb_usx2y: index=-2
snd_cmipci: mpu_port=0x330 fm_port=0x388
snd_pcsp: index=-2
snd_usb_audio: index=-2
snd_intel_dspcfg: dsp_driver=1 <<<< THIS NEEDS TO BE GONE

[OrelSokolov]

@plbossart done
https://alsa-project.org/db/?f=b1f3f31e1135aa62303b8ac4ed4a30c40d382bad

[plbossart]

Ha yes, we've seen this before

[    6.000606] sof-audio-pci-intel-mtl 0000:00:1f.3: no matching blob for sample rate: 48000 sample width: 32 channels: 4
[    6.000623] sof-audio-pci-intel-mtl 0000:00:1f.3: failed to prepare widget dai-copier.DMIC.dmic01.capture
[    6.000629] sof-audio-pci-intel-mtl 0000:00:1f.3: Failed to prepare connected widgets

This was reported in #4973 and fixed by 806f52f

The patch was submitted upstream and is already in maintainer branches, commit d556f80, and already in Linus' tree.

Please try to apply this commit or wait until it shows up in your distribution.

[OrelSokolov]

@plbossart so how can I fix this on Ubuntu? Kernel did not help some weeks ago.

[plbossart]

you can compile your own kernel or wait for the Ubuntu kernel to have this supported. Presumably the fix will be applied to Linux stable so all distros will get the fix at some point.

[OrelSokolov]

@plbossart thanks a lot!

[OrelSokolov]

@plbossart this is not Huawei, this is Xiaomi

[OrelSokolov]

@plbossart anyway, I am running 6.10rc2, looks like it includes this commit. But still no any microphone input.

Can you please give advice how to configure alsa now?

https://alsa-project.org/db/?f=ebfc9d57a9e98132909114151d1f70e9109b8eaa

[plbossart]

are you sure this is the right log? There's the same error

[    5.310431] sof-audio-pci-intel-mtl 0000:00:1f.3: no matching blob for sample rate: 48000 sample width: 32 channels: 4
[    5.310450] sof-audio-pci-intel-mtl 0000:00:1f.3: failed to prepare widget dai-copier.DMIC.dmic01.capture
[    5.310457] sof-audio-pci-intel-mtl 0000:00:1f.3: Failed to prepare connected widgets

I can't recall if we need a topology update as well. @ujfalusi might know?

[OrelSokolov]

@plbossart as you can see this is within new kernel 6.10 which includes your patch. And yes, I have no idea how to setup this thing properly.
I turned off again dsp driver, I tried to install some repos from the SOF project earlier. So I have no clue what is the correct combination of these.
That's why I created and issue to ask for help to detect correct combination of kernel, driver, config and etc...

[plbossart]

Can you also extract the NHLT?

cat /sys/firmware/acpi/tables/NHLT > nhlt.dat; gzip nhlt.dat

and attach the zipped file, thanks!

[OrelSokolov]

nhlt.dat.gz

[plbossart]

nhlt.dsl.txt

This still looks like the 16-bit issue corrected by the commit d556f80

my money is on the topology now. Please try to add (backup what you have of course!) from this comment: #4973 (comment)

[OrelSokolov]

@plbossart locate command gives this

/usr/lib/firmware/intel/sof-ipc4-tplg/sof-hda-generic-4ch.tplg
/usr/lib/firmware/intel/sof-ipc4-tplg-v2.9/sof-hda-generic-4ch.tplg
/usr/lib/firmware/intel/sof-ipc4-tplg-v2.9/sof-hda-generic-4ch.tplg.zst
/usr/lib/firmware/intel/sof-tplg/sof-hda-generic-4ch.tplg
/usr/lib/firmware/intel/sof-tplg/sof-hda-generic-4ch.tplg.zst

[OrelSokolov]

@plbossart OMG it works!!!! Without any additional options

Thanks. What is futher steps to fix this in codebase?

https://alsa-project.org/db/?f=71084485390ea3ac2dab60932cd23e0ee6b91e16

[plbossart]

Excellent.

What is futher steps to fix this in codebase?

If you meant "when is this going to be supported by distros", there will be an 'sof-bin' 2.10 release soonish - in the coming weeks - where the required topology will be included. The kernel patch should also be propagated to stable kernels but that takes time and depends on stable maintainers.

[OrelSokolov]

@plbossart Yes, I mean when this update will be in the linux kernel or somewhere downloadable.

[yw662]

Fedora is now on Linux 6.10.7 SOF 2024.06, sound and microphone both work.

[johndoe3535]

I'm on Fedora KDE now and can confirm that sound and microphone is working now. However something in the sound pipepline seems to crash somehow after some time. I get robotic sounding voices, everything sounds distored. It stops when I switch to pro audio in the sound settings. When I get it on pro audio, it stops again when switching to play hi fi quality music. Maybe some buffering issues. It seems I can trigger it faster when skipping trough a youtube video with the arrow keys, and often start and stop playing sounds. But yeah, maybe the issue is somewhere in pipewire idk. Anyone else with a redmibook 16 that has this now ? I was on bazzite before (fedora atomic spin with custom kernel) with kernel 6.10, and the sof files from march. So for me its hard to say what introduced this on my side...

[malvery]

I'm on Fedora KDE now and can confirm that sound and microphone is working now. However something in the sound pipepline seems to crash somehow after some time. I get robotic sounding voices, everything sounds distored. It stops when I switch to pro audio in the sound settings. When I get it on pro audio, it stops again when switching to play hi fi quality music. Maybe some buffering issues. It seems I can trigger it faster when skipping trough a youtube video with the arrow keys, and often start and stop playing sounds. But yeah, maybe the issue is somewhere in pipewire idk. Anyone else with a redmibook 16 that has this now ? I was on bazzite before (fedora atomic spin with custom kernel) with kernel 6.10, and the sof files from march. So for me its hard to say what introduced this on my side...

Same issues on Arch Linux (linux=6.10.8, sof-firmware=2024.06, pipewire 1:1.2.3).

Steps to reproduce robotic sound:

• change microphone level => gets robotic sound until toggle card on/off or switch card profile

Steps to reproduce missed sound:

• start play sound on user1 and tty1
• switch to the user2 on tty2 and play something sound too
• return to the user1 on tt1 => gets missed sound in any application until toggle card profile

[ujfalusi]

@johndoe3535, @malvery, can you describe or better provide a recording of the 'robotic' sound?
The robotic sound is on the playback or capture direction?

[OrelSokolov]

@ujfalusi I can confirm this issue. This is completely random issue. I have this issue while using Chrome browser. Restarting application process helps.
Bug just adds glitches to the original sound, so it sounds similar to original sound just with glitches.

[ujfalusi]

@OrelSokolov, so it is with playback?
Pipewire/Pulseaudio keeps the audio open for few more seconds after the last user is gone, stopping/restarting application 'fast' enough will still use the same stream the audio was opened with (the sound servers are playing silence when no client is around).

If the issue is easy to reproduce, can you make a recording of it with a phone for example for us to hear it? A video is also good so we could see what is going on in the UI.

I think we will need to collect firmware logs to have cleaner picture and more data points.

[malvery]

@johndoe3535, @malvery, can you describe or better provide a recording of the 'robotic' sound? The robotic sound is on the playback or capture direction?

Recorded "robotic" sound on the attached video
https://github.com/user-attachments/assets/6221b976-dbba-4026-ba7d-f3011afe3804

How can I help in collecting firmware logs ?