STICKY: Bullseye - Comments and bug reports thread.

[pidd]

by the way, i have no VNC server active on the RPi4 - i notices that when VNC server is enabled on the RPi4, some issues disappear. i think the VNC server forces the desktop to another mode.

If VNC is enabled in raspi-config the desktop is changes from mutter to openbox.

[Adam777]

VLC Player fullscreen broken

After an update few days back, video doesn't play in fullscreen mode.
In window mode no problems, but after activating fullscreen, window ui goes away as it should but video freezes, frame doesn't scale to fullscreen and stays still. Audio plays normally and controls work.
Affects all files no matter the resolution or format.

RPi3+, Bullseye, VLC 3.0.17.4 Vetinary
No additional software installed, all VLC settings at default

[twilightened]

VLC Player fullscreen broken

After an update few days back, video doesn't play in fullscreen mode.
In window mode no problems, but after activating fullscreen, window ui goes away as it should but video freezes, frame doesn't scale to fullscreen and stays still. Audio plays normally and controls work.
Affects all files no matter the resolution or format.

RPi3+, Bullseye, VLC 3.0.17.4 Vetinary
No additional software installed, all VLC settings at default

My problem is a little bit different. If i select "automatic" in the output mode for video in the settings (to get hardware acceleration, it has to stay in automatic), when i go to fullscreen mode in VLC, i can't go back. Fullscreen video is stuck on the screen no matter what key combination i use. What i can do to exit is to press ctrl+alt+F1 to go to terminal and type "pkill -9 vlc" and kill all instances of VLC to go back to the desktop. For the time being, i select "Opengl" as the output mode of VLC video, and it works. However, according to others, this output completely cancels out the hardware accelerated video playback.

Edit: RPI4 64bit.

[mrzottel]

Hi,

I have an Issue with Bullseye - or the latest updates that came today.

The package libldb2 version 2.2.3 contained the binaries of 2.2.4 - which breaks Samba.

I have no clue if this is the right place to post this, but could someone put the right binaries into that package?

Edit:
To be more exact: Raspi OS has a package called libldb2 (2:2.2.3-2~deb11u2) which contains the binaries for 2.2.4, despite its name.
I fixed it by manually installing libldb2 (2:2.2.3-2~deb11u1) from a debian repo.

Best regards
MrZottel

[hippie403]

Today's new Kernel seems to have messed up my mouse:

4GB Pi4B

Linux FLIRC 5.15.56-v8+ #1575 SMP PREEMPT Fri Jul 22 20:31:26 BST 2022 aarch64 GNU/Linux


This mouse needs to have usbhid.mousepoll=0 in /boot/cmdline.txt to work properly.
Tried another mouse, still jerking about uncontrollably. Both mice work fine on a system without the update.
Fiddled with the mouse preferences but still uncontrollable. Any ideas? Different mousepoll setting?

[dom]

Today's new Kernel seems to have messed up my mouse:

4GB Pi4B

Are you using the default kms driver? Or did you switch to fkms?

[hippie403]

Today's new Kernel seems to have messed up my mouse:

4GB Pi4B

Are you using the default kms driver? Or did you switch to fkms?

had to switch to fkms, HDMI audio wasn't working and something weird was going on with my screen being slightly blurry with kms.
(Running at the native res of the panel 1360 * 768 which is supported)

[ar51an]

Power LED is not switching off after regular kernel update to 5.15.56-v8+ through sudo apt full-upgrade.
I have below settings in config.txt for switching off power LED:

Code: Select all

dtparam=pwr_led_trigger=none
dtparam=pwr_led_activelow=off

I tried none, heartbeat and few more triggers on kernel 5.15.56-v8+. Only heartbeat worked.
When none is selected, cat /sys/class/leds/led1/trigger shows selected value as [none]

I downgraded kernel to 5.15.55-v8+ and now power LED is switching off.

[hippie403]

Today's new Kernel seems to have messed up my mouse:

4GB Pi4B

Are you using the default kms driver? Or did you switch to fkms?

had to switch to fkms, HDMI audio wasn't working and something weird was going on with my screen being slightly blurry with kms.
(Running at the native res of the panel 1360 * 768 which is supported)

Thanks for the gentle hint about kms. Switching back to the default kms fixed the mouse issue. HDMI sound is now working and the blurry video issue is also fine. Video playback, YouTube in Chromium is now super choppy, but I guess I can't have everything.

[hippie403]

Thanks for the gentle hint about kms. Switching back to the default kms fixed the mouse issue. HDMI sound is now working and the blurry video issue is also fine. Video playback, YouTube in Chromium is now super choppy, but I guess I can't have everything.

Cancel that, I messed up and was running without kms, hence the bad video performance. kms yields the usual result of no sound and blurry video, but the mouse works.

[dom]

Cancel that, I messed up and was running without kms, hence the bad video performance. kms yields the usual result of no sound and blurry video, but the mouse works.

Make sure overscan is disabled. That introduces scaling (hence blurriness) with kms. If you need the adjustment you should do this at TV end.

[hippie403]

Make sure overscan is disabled. That introduces scaling (hence blurriness) with kms. If you need the adjustment you should do this at TV end.

Thanks for trying to help. Definitely got it disabled.

# Disable compensation for displays with overscan
disable_overscan=1

The lack of HDMI audio as soon as kms is enabled is a real show stopper though. Is this a 64bit issue? Does it all work okay on a 32bit install?

[heymac]

Dom said "Are you using the default kms driver? Or did you switch to fkms?"

I don't know? How do I check?

[dom]

Dom said "Are you using the default kms driver? Or did you switch to fkms?"

I don't know? How do I check?

look in /boot/config.txt.
You should see a line like:
dtoverlay=vc4-kms-v3d
or
dtoverlay=vc4-fkms-v3d

The former is the default on bullseye and is known as the kms driver (display driven by arm code in kernel).
The latter is known as the fkms driver (display driven by firmware) and is now deprecated.

[heymac]

Thank you, Dom. It's set to fkms and I have now changed it to kms, so it now reads:
dtoverlay=vc4-kms-v3d

That fixed it, thanks again.

[redvli]

Another thing is that the timestamping (current pos. / file-length) is not displayed in the right upper corner in VLC when the file being played is appended to and fast-forward keys are pressed (right-arrow). This works in 32-bit Raspbian (note: that is with vlc 3.0.16, so mmal_vout, dtoverlay=vc4-fkms-v3d in config.txt instead of KMS).

Now running kernel 5.15.56 (apt update && apt full-upgrade done yesterday) and I see the timestamping is now working, so same as with old 32-bit Raspbian. Switching from windowed to full-screen (and vice-versa) takes a bit longer with this new kernel I notice.

What also works now is 2D accelerated windowed playing on both HDMI0 and HDMI1 at the same time (mirrored displays), although if I make the window bigger than 75% of full-size, still some sort of frame judder. It seems to me that the Pi4 rev 1.1 cannot handle it, maybe tweaking clocks helps or just 1 display.

[heymac]

Just a follow-up to the switch to km. I now have the following error in the dmesg event log -

Code: Select all

hdmi-audio-codec hdmi-audio-codec.3.auto: ASoC: error at snd_soc_dai_startup on i2s-hifi: -19
MAI: soc_pcm_open() failed (-19)

The above entries appear twelve times.

[trejan]

Just a follow-up to the switch to km. I now have the following error in the dmesg event log -

Code: Select all

hdmi-audio-codec hdmi-audio-codec.3.auto: ASoC: error at snd_soc_dai_startup on i2s-hifi: -19
MAI: soc_pcm_open() failed (-19)

The above entries appear twelve times.

It is a known issue and harmless. Ignore those errors if audio is working.

[andrebrait]

Power LED is not switching off after regular kernel update to 5.15.56-v8+ through sudo apt full-upgrade.
I have below settings in config.txt for switching off power LED:

Code: Select all

dtparam=pwr_led_trigger=none
dtparam=pwr_led_activelow=off

I tried none, heartbeat and few more triggers on kernel 5.15.56-v8+. Only heartbeat worked.
When none is selected, cat /sys/class/leds/led1/trigger shows selected value as [none]

I downgraded kernel to 5.15.55-v8+ and now power LED is switching off.

I can confirm this behavior. A bug, indeed.

[beta-tester]

Code: Select all

dtparam=pwr_led_trigger=none
dtparam=pwr_led_activelow=off

i didn't know that option. is it possible to turn off the LED when RPi is shutdown as well?
when i shutdown my RPi the power LED is always on
(or turns on again, in case the kernel was turning it off at booting)

[dom]

Power LED is not switching off after regular kernel update to 5.15.56-v8+ through sudo apt full-upgrade.
I have below settings in config.txt for switching off power LED:

Code: Select all

dtparam=pwr_led_trigger=none
dtparam=pwr_led_activelow=off

See https://github.com/raspberrypi/firmware/issues/1742

[fanoush]

2)When system freeze, is there a way to safely shutdown or restart the system without unplug the power?

Not sure anyone answered already but yes if the keyboard still works then there is the magic sysrq sequence, google "reisub" or "reisuo"

[StopBreakingMyPi4]

VLC Player fullscreen broken

After an update few days back, video doesn't play in fullscreen mode.
In window mode no problems, but after activating fullscreen, window ui goes away as it should but video freezes, frame doesn't scale to fullscreen and stays still. Audio plays normally and controls work.
Affects all files no matter the resolution or format.

RPi3+, Bullseye, VLC 3.0.17.4 Vetinary
No additional software installed, all VLC settings at default

Exactly this happened to me before & their solution was a beta firmware version, which worked for me, then it was repaired in the last update & now they've broken it again. This time I seem to have regained fullscreen functionality by changing "Video Output Module" to "OpenGL for Embedded Systems 2 video output". This hadn't been successful when it was broken before, so it seems they've broken it differently this time

[8809]

rpi4 + kingstone 128g sdxc
could not boot after upgrade
error message "blk_update_request i/o error dev mmcblk0......"

does anyone encounter the same problem??

[redvli]

rpi4 + kingstone 128g sdxc
could not boot after upgrade
error message "blk_update_request i/o error dev mmcblk0......"

does anyone encounter the same problem??

This looks like a hardware problem, not something to do with software Bullseye.
Might be powersupply is not good enough, dirty or bad contacts for your SD card or SD card itself has a problem. I would use another SD card brand to start with. But look around on the internet:
search.php?keywords=blk_update_request+ ... ev+mmcblk0