Moving Linux Kernel to 5.10

[busywait]

Why remove pulseaudio?

No good reason, and off-topic for this thread probably

I haven't uninstalled PulseAudio from the 64-bit OS - ideally I'll be able to get it working, or an update will fix up the config.

[Bluestang]

There is definitely some 5.10-rc7 issues with the new pulseaudio changes in RPiOS.

I was also getting EmulationStation crashes with either FKMS/firmware audio or KMS/audio. It would happen when a video snap played (using VLC).

Reverting back to the 5.4 kernel fixed the EmulationStation issues, but since KMS is broken under 5.4 well that’s not ideal.

[DKERL]

hello everyone, after the update from Kernal, audio is no longer possible for me via the 3.5 "connection, which is a shame.

Br detlef

[pasman1]

Kernel 5.10 is released.

https://www.phoronix.com/scan.php?page= ... 0-Released

[Bluestang]

@6by9
What is the magic formula to get the KMS audio driver working with the new pulse audio updates?

Both my 32-bit and 64-bit images no longer will output sound with the KMS driver. I had to revert back to the firmware audio to get sound back. Sound worked previously with both images before the update, and the info in this thread.

[fruitoftheloom]

Kernel 5.10 is released.

https://www.phoronix.com/scan.php?page= ... 0-Released

Support added for Raspberry Pi 4 / BCM2711 display with the VC4 DRM driver.

https://www.debugpoint.com/2020/12/linu ... nouncement

[dom]

@6by9
What is the magic formula to get the KMS audio driver working with the new pulse audio updates?

Both my 32-bit and 64-bit images no longer will output sound with the KMS driver. I had to revert back to the firmware audio to get sound back. Sound worked previously with both images before the update, and the info in this thread.

I haven't found the proper solution, but:

Code: Select all

systemctl --user stop pulseaudio.socket
systemctl --user stop pulseaudio.service

and make sure vc4-hdmi.conf matches version here and audio worked again.

[CaptainMidnight]

@6by9
What is the magic formula to get the KMS audio driver working with the new pulse audio updates?

Both my 32-bit and 64-bit images no longer will output sound with the KMS driver. I had to revert back to the firmware audio to get sound back. Sound worked previously with both images before the update, and the info in this thread.

I haven't found the proper solution, but:

Code: Select all

systemctl --user stop pulseaudio.socket
systemctl --user stop pulseaudio.service

and make sure vc4-hdmi.conf matches version here and audio worked again.

Checked vc4-hdmi.conf matches version as linked, but found stopping the pulseaudio.service kills the menu bar?

With making the pulseaudio changes above does the menu panel sound applet 'Volume Control (PulseAudio)' not become unusable?

[CaptainMidnight]

Configured with the new kms driver and audio disabled in /boot/config.txt

Code: Select all

# Enable audio (loads snd_bcm2835)
#dtparam=audio=on

# Enable DRM VC4 V3D driver on top of the dispmanx display stack
#dtoverlay=vc4-fkms-v3d
dtoverlay=vc4-kms-v3d

And with the vc4-hdmi.conf as specified, no audio hardware device is listed in PulseAudio control panel and the only thing I can find in the dmesg is :-

Code: Select all

[    2.574719] debugfs: Directory 'fef00700.hdmi' with parent 'vc4-hdmi-0' already present!
[    2.575172] vc4_hdmi fef00700.hdmi: ASoC: no DMI vendor name!
[    2.575464] vc4-drm gpu: bound fef00700.hdmi (ops vc4_hdmi_ops [vc4])
[    2.576171] debugfs: Directory 'fef05700.hdmi' with parent 'vc4-hdmi-1' already present!
[    2.576865] vc4_hdmi fef05700.hdmi: ASoC: no DMI vendor name!
[    2.585378] vc4-drm gpu: bound fef05700.hdmi (ops vc4_hdmi_ops [vc4])
[   50.404719] vc4_hdmi fef05700.hdmi: ASoC: error at snd_soc_dai_startup on fef05700.hdmi: -19

Only 'Dummy Output' is available in PulseAudio.

[busywait]

Code: Select all

systemctl --user stop pulseaudio.socket
systemctl --user stop pulseaudio.service

and make sure vc4-hdmi.conf matches version here and audio worked again.

Not for me
[Edit: Oh, you mean switch PulseAudio off? Yes, in that case the sound is available again, eg in RetroArch, but obviously if I startx then the things that are expecting to use PulseAudio don't work.]

In my 64-bit RPi OS I've grabbed the "released" 5.10.0 kernel with rpi-update.

/usr/share/alsa/cards/vc4-hdmi.conf already matched (and was sitting next to my careful backup of the original). I rewrote it anyway with the same content, just to be sure.

Code: Select all

systemctl --user stop pulseaudio.socket
systemctl --user stop pulseaudio.service

systemctl --user start pulseaudio.service
systemctl --user start pulseaudio.socket

journalctl -xe | grep snd -B 5 -A 5
-- 
-- The job identifier is 943.
Dec 14 22:00:43 rpi-tv pulseaudio[897]: E: [pulseaudio] module-alsa-card.c: Failed to find a working profile.
Dec 14 22:00:43 rpi-tv pulseaudio[897]: E: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="0" name="platform-fef00700.hdmi" card_name="alsa_card.platform-fef00700.hdmi" namereg_fail=false tsched=no fixed_latency_range=no ignore_dB=no deferred_volume=yes use_ucm=yes card_properties="module-udev-detect.discovered=1""): initialization failed.
Dec 14 22:00:43 rpi-tv pulseaudio[897]: E: [pulseaudio] module-alsa-card.c: Failed to find a working profile.
Dec 14 22:00:43 rpi-tv kernel: vc4_hdmi fef05700.hdmi: ASoC: error at snd_soc_dai_startup on fef05700.hdmi: -19
Dec 14 22:00:43 rpi-tv pulseaudio[897]: E: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="1" name="platform-fef05700.hdmi" card_name="alsa_card.platform-fef05700.hdmi" namereg_fail=false tsched=no fixed_latency_range=no ignore_dB=no deferred_volume=yes use_ucm=yes card_properties="module-udev-detect.discovered=1""): initialization failed.
Dec 14 22:00:43 rpi-tv udisksd[918]: udisks daemon version 2.8.1 starting
Dec 14 22:00:43 rpi-tv udisksd[918]: failed to load module crypto: libbd_crypto.so.2: cannot open shared object file: No such file or directory
Dec 14 22:00:43 rpi-tv udisksd[918]: failed to load module mdraid: libbd_mdraid.so.2: cannot open shared object file: No such file or directory
Dec 14 22:00:44 rpi-tv udisksd[918]: Failed to load the 'mdraid' libblockdev plugin
--
Dec 14 22:14:05 rpi-tv rtkit-daemon[911]: Supervising 3 threads of 3 processes of 1 users.
Dec 14 22:14:05 rpi-tv pulseaudio[1988]: E: [pulseaudio] module-alsa-card.c: Failed to find a working profile.
Dec 14 22:14:05 rpi-tv pulseaudio[1988]: E: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="0" name="platform-fef00700.hdmi" card_name="alsa_card.platform-fef00700.hdmi" namereg_fail=false tsched=no fixed_latency_range=no ignore_dB=no deferred_volume=yes use_ucm=yes card_properties="module-udev-detect.discovered=1""): initialization failed.
Dec 14 22:14:05 rpi-tv pulseaudio[1988]: E: [pulseaudio] module-alsa-card.c: Failed to find a working profile.
Dec 14 22:14:05 rpi-tv pulseaudio[1988]: E: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="1" name="platform-fef05700.hdmi" card_name="alsa_card.platform-fef05700.hdmi" namereg_fail=false tsched=no fixed_latency_range=no ignore_dB=no deferred_volume=yes use_ucm=yes card_properties="module-udev-detect.discovered=1""): initialization failed.
Dec 14 22:14:05 rpi-tv kernel: vc4_hdmi fef05700.hdmi: ASoC: error at snd_soc_dai_startup on fef05700.hdmi: -19
Dec 14 22:14:05 rpi-tv systemd[586]: Started Sound Service.
-- Subject: A start job for unit UNIT has finished successfully
-- Defined-By: systemd
-- Support: https://www.debian.org/support
-- 

[DougieLawson]

Something bad has happened with 5.10. Every one of my Zeros and RPI1Bs has failed to boot with that latest BRANCH=next rpi-update.

[MrEngman]

Something bad has happened with 5.10. Every one of my Zeros and RPI1Bs has failed to boot with that latest BRANCH=next rpi-update.

I'm seeing the same problem, but not just with kernel 5.10.0 but also with the latest 5.4.83 kernel from rpi-update.

Only the Pi4 4GB and 8GB and Pi400 work. All other Pi failed to boot, Pi0, Pi0W, Pi1B, Pi2B, Pi3A. More info viewtopic.php?p=1779825#p1779825


MrEngman

[DougieLawson]

Something bad has happened with 5.10. Every one of my Zeros and RPI1Bs has failed to boot with that latest BRANCH=next rpi-update.

I'm seeing the same problem, but not just with kernel 5.10.0 but also with the latest 5.4.83 kernel from rpi-update.

Only the Pi4 4GB and 8GB and Pi400 work. All other Pi failed to boot, Pi0, Pi0W, Pi1B, Pi2B, Pi3A. More info viewtopic.php?p=1779825#p1779825


MrEngman

My 2B failed.
My 3B (with 5.4.83-v7+ worked OK)
My 3A+ failed
My B+ failed

There's five left to fix tomorrow.

So it is exceedingly poisonous - which should greatly discourage the folks who run rpi-update without understanding what it does (and don't have a viable plan for a failed boot).

[graysky]

https://github.com/raspberrypi/firmware/issues/1511

[lee-yunseok]

After update 5.10.0-v8+ in RPiOS64bit, set the theme in the Openbox Configuration Manager, there is no action with click the titlebar and side in any programs. Both of minimize/maximize/exit/resize/move doesn't work.

Edited: solved with Set Defaults in Appearance Settings.

[thatchunkylad198966]

5.10.1 is stable now. when is RPI going to release?

[dom]

A firmware fix for the sdcard boot failure on pi0-pi3 has been pushed.

[DougieLawson]

A firmware fix for the sdcard boot failure on pi0-pi3 has been pushed.

That was quick thank you. I'll give that a go (on a subset of machines first - so I don't lose another four hours).

[MrEngman]

A firmware fix for the sdcard boot failure on pi0-pi3 has been pushed.

That was quick thank you. I'll give that a go (on a subset of machines first - so I don't lose another four hours).

Fix works for me on all the Pi I had problems with.


MrEngman

[dom]

The master branch of rpi-update will now get you 5.10 kernel.
No need to use BRANCH=next which will now be dormant.

[Bluestang]

Great news on 5.10...but is a fix being worked for the KMS audio and PulseAudio compatibility?

Its more of a nuisance but just curious if there is a timeline on this?

[graysky]

The master branch of rpi-update will now get you 5.10 kernel.
No need to use BRANCH=next which will now be dormant.

Is there a roadmap for replacing 5.4.y with 5.10.y as the default branch?

[Aetallx]

Updated today and after a successful boot it failed to login properly.
It accepted the password but it would only display the desktop background after, and when trying to login in text mode with ctrl+alt+f1 similar thing, it accepted the password and then nothing.
Shutting down was possible so it did not freeze.
Also there where no errors while booting up or shutting down.
I suspect it's because I am using an encrypted home with ecryptfs-utils but I didn't have time to further debug to confirm this.
I replaced the kernel with the previous one from a backup sd card and it booted successfully again.

Does this mean that ecryptfs-utils is no longer supported with the new kernel or it's just a temporary bug?

[dom]

Does this mean that ecryptfs-utils is no longer supported with the new kernel or it's just a temporary bug?

Nothing has been changed in this regard downstream. I'm not sure what may have changed upstream.
The 5.10 kernel is still built with CONFIG_ECRYPT_FS=m

[dom]

Is there a roadmap for replacing 5.4.y with 5.10.y as the default branch?

As described in the first post, when it's ready.
I'd expect a couple of months, but basically depends on us being confident there are enough people testing it,
and there are not major issues expected in the bump.