Bug 206085

On rc7 same
On rc1 sound is worked but system not logged

last version 5.5 same problem

Too little information: what did you test and what not working?

Also please give alsa-info.sh output at least.  Run the script with --no-upload option and attach the output.
Last but not least, give dmesg output, too.

I am install from this https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.5.1/, today v5.5.1 same problem

sudo alsa-info > alsa-info.log
log in attach but:
Your ALSA information is located at http://alsa-project.org/db/?f=e370494cdf0fc0dc6255b93db0cc2207d403e723

sudo dmesg >dmesg.log in attach

Created attachment 287059 [details]
alsa-info.log

Created attachment 287061 [details]
dmseg log

attached

sudo alsa-info --stdout >alsa-info5.5.2.log
sudo dmesg >dmesg5.5.2.log

Dummy Outout Test not work too

Created attachment 287181 [details]
sudo alsa-info --stdout >alsa-info5.5.2.log

sudo alsa-info --stdout >alsa-info5.5.2.log

Created attachment 287183 [details]
sudo dmesg >dmesg5.5.2.log

sudo dmesg >dmesg5.5.2.log

Created attachment 287301 [details]
sudo alsa-info --stdout >alsa-info5.6.0-rc1.log

sudo alsa-info --stdout >alsa-info5.6.0-rc1.log

Created attachment 287303 [details]
sudo dmesg >dmesg5.6.0-rc1.log

sudo dmesg >dmesg5.6.0-rc1.log

Could you give the patch below?

I meant try the patch below :)

Created attachment 287383 [details]
Test fix patch for 5.6

Note that you'll still need more user-space stuff to make the whole SOF working, including the SOF firmware.  Possibly the latest PulseAudio git version, alsa-lib, and alsa-ucm-conf would be needed as well.

Work to fix this case in SOF is under work in:
https://github.com/thesofproject/linux/issues/1658
https://github.com/thesofproject/linux/pull/1731

I apologize
If I put the version from here https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.6-rc2/
how to apply the patch correctly, can you give link to instructions?
Or can I wait for patch 5.6.0-rc3?

sealse wrote:
> how to apply the patch correctly, can you give link to instructions?
> Or can I wait for patch 5.6.0-rc3?

If you are not familiar with applying kernel patches yourself, maybe better wait until the patch is merged to some upstream tree. I've now sent the patches to alsa-devel for review:
https://mailman.alsa-project.org/pipermail/alsa-devel/2020-February/163603.html

Depending on how the reviews go, the patch might make it to the 5.6-rc series, but we'll see.

Thanks  Kai!
Tell me please - Where can I see if this patch is in the new version of kernel

@sealse: The patch was now merged to Mark Brown's ASoC tree and is making its way up. To try it, you can fetch latest "for-next" branch of Mark's tree:

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git/?h=for-next

FYI: A UCM fix is also needed to fully enable Pulseaudio on these systems. Solution submitted to ALSA UCM upstream:
https://github.com/alsa-project/alsa-ucm-conf/pull/15

Hi, I installed the version 5.6.0-rc5 but the problem is still there, I don’t quite understand how to track if there are any changes in the build, but the problem is still present

@sealse: This fix is not in 5.6.0-rc5 and as it's not a simple bugfix but enabling audio for a new category od devices, it is likely it is not going to be in 5.6.0 at at all -- i.e. 5.7.0 will be first official kernel with the fix. Distributions might backport the fix to older kernels, but this is not guaranteed.

You can partially enable audio with following module option (put this e.g. to /etc/modprobe.d/select-audio.conf):

#Force the DSP driver for Intel DSP (0=auto, 1=legacy, 2=SST, 3=SOF
options snd\_intel\_dspcfg dsp\_driver=1

... that should work with 5.6.0-rc5. Digital Mic will not work, but rest of audio should be ok. You need to remember to remove that when you upgrade to 5.7.0 (or backport the patch).

5.7.0-rc1 same problem

@sealse: 5.7.0-rc1 has the kernel patches now. So if it still fails, it's worth taking fresh set of logs. One known is issue is that latest alsa-ucm release (1.2.2) is still missing one patch that is needed on these machines, so Pulseaudio still gets confused. But the ALSA drivers should be good now.

Can you maybe post alsa-info output with 5.7.0-rc1?

Created attachment 288529 [details]
alsa-info.log

Created attachment 288531 [details]
dmseg.log

Hi! I also complained that Nvidia did not malfunction and as a result they claim that Alsa drivers are to blame https://forums.developer.nvidia.com/t/ubuntu-kernel-v5-6-v5-7-for-hp-omen-17t-bc000-gforce-rtx2080-unstable-driver-nvidia-linux-x86-64-440-82-run/120027/8

Sorry, the previous post wasn’t really, I read it wrong

@sealse: Thanks for the logs! Now failure to init HDMI audio in i915 is properly ignored, as shown by this dmesg:
[    4.038182] sof-audio-pci 0000:00:1f.3: init of i915 and HDMI codec failed

But audio driver load fails due to missing firmware now:
[    4.057638] sof-audio-pci 0000:00:1f.3: Direct firmware load for intel/sof/sof-cfl.ri failed with error -2
[    4.057640] sof-audio-pci 0000:00:1f.3: error: request firmware intel/sof/sof-cfl.ri failed err: -2

Please install the 1.4.2 SOF firmware
https://github.com/thesofproject/sof/releases/tag/v1.4.2

There are 4 processor options in the installation readme
Baytrail / Cherrytrail:
Haswell / Broadwell:
Apollolake:
Cannonlake:
But as I see it, I have Coffee Lake https://ark.intel.com/content/www/en/ru/ark/products/192987/intel-core-i9-9880h-processor-16m-cache-up-to- 4-80-ghz.html
What should I try to set?

Hi! I installed release ubuntu 20.04 kernel 5.4.0-26-generic
Ahd i have same problem
I can't install firmware becouse I have next message

cmake -DTOOLCHAIN=xtensa-hsw-elf -DROOT_DIR=`pwd`/../../xtensa-root/xtensa-hsw-elf ..
-- Preparing Xtensa toolchain
-- The C compiler identification is unknown
-- The ASM compiler identification is GNU
-- Didn't find assembler
CMake Error at CMakeLists.txt:39 (project):
  The CMAKE_C_COMPILER:

    xtensa-hsw-elf-gcc

  is not a full path and was not found in the PATH.

  Tell CMake where to find the compiler by setting either the environment
  variable "CC" or the CMake cache entry CMAKE_C_COMPILER to the full path to
  the compiler, or to the compiler name if it is in the PATH.


CMake Error at CMakeLists.txt:39 (project):
  No CMAKE_ASM_COMPILER could be found.

  Tell CMake where to find the compiler by setting either the environment
  variable "ASM" or the CMake cache entry CMAKE_ASM_COMPILER to the full path
  to the compiler, or to the compiler name if it is in the PATH.


-- Configuring incomplete, errors occurred!
See also "/home/sealse/Downloads/sof-1.4.3/build/CMakeFiles/CMakeOutput.log".
See also "/home/sealse/Downloads/sof-1.4.3/build/CMakeFiles/CMakeError.log".

Install the firmware binaries as Kai pointed in comment#30 .

I didn’t find the binary file from the link https://github.com/thesofproject/sof/releases/tag/v1.4.2

Scroll down to assets or install binaries from the sof-bin repo: https://github.com/thesofproject/sof-bin/tree/stable-v1.4.2

I installed sof-bin

>sudo ./go.sh 
>Installing Intel firmware and topology v1.4.2 to lib/firmware/intel
>Done installing Intel firmware and topology v1.4.2

I installed kernel 5.7.0-rc2 and again installed sod-bin
I rebooted but sound not working


Also I got next messade after command:
>sudo alsa-info --stdout >alsa-info5.7.0-rc2.log
>[sudo] password for sealse: 
>lspci: Unable to load libkmod resources: error -12
>lspci: Unable to load libkmod resources: error -12
>cat: '/sys/module/snd_soc_skl_hda_dsp/parameters/*': No such file or directory

Created attachment 288713 [details]
alsa-info.log

Created attachment 288715 [details]
dmseg.log

@sealse Thanks! The alsa-info output now looks good and sound should work at kernel level if you try to play and record sounds with ALSA aplay and arecord tools.

Based on the logs, pulseaudio does not seem to work and you are most probably hitting the issue I mentioned in comment #21 of this bug https://bugzilla.kernel.org/show_bug.cgi?id=206085#c21

In short the UCM file is missing rules to handle the missing HDMI/DP nodes for Intel ALSA cards when the HDMI/DP is connected to the NVidia card. The patch is in git master, but not yet in the released alsa-ucm-conf-1.2.2

A bit longer explanation at:
https://github.com/thesofproject/linux/issues/1985#issuecomment-615226041

I read the topic you proposed
But do not quite understand
Can I now compromise the kernel with the patch you have proposed? Or do I have to wait for the release of a new kernel version?
In which release is the patch planned to appear

(In reply to sealse from comment #40)
> Can I now compromise the kernel with the patch you have proposed? Or do I
> have to wait for the release of a new kernel version?
> In which release is the patch planned to appear

Kernel is now fine. 

The missing piece is update to user-space ALSA packages -> 1.2.2 doesn't have the fix, so whatever the next release is (1.2.3?) will have it. Jaroslav manages the ALSA releases, so he is best to comment on this.

@Jaroslav сan you help me install a beta version to fix this problem?

@sealse Here I've described a solution that worked for me
https://github.com/thesofproject/linux/issues/2065

@dzikages555  Thank's!!!
All I missed is:
wget https://raw.githubusercontent.com/alsa-project/alsa-ucm-conf/master/ucm2/sof-hda-dsp/Hdmi.conf
and
mv Hdmi.conf /usr/share/alsa/ucm2/sof-hda-dsp/Hdmi.conf

Excellent @sealse and @dzikages555! 

So Jaroslav, I think we can close this bug now -- definitely the kernel is ok now.