Raspberry Pi Kernels (2.0)

v6.9.2 is out but v.6.10.x is not even present :frowning:
Kernel 6.10 seems promising !

New packages pushed to unstable when the mirrors sync:

linux-rpi4-mainline 6.9.2-1
linux-rpi4-mainline-headers 6.9.2-1
linux-rpi5-mainline 6.9.2-1
linux-rpi5-mainline-headers 6.9.2-1
raspberrypi-bootloader 20240524-1

so, when we will have new img?
plasma-5 sdcard broken after recent huge update :rofl:

v6.10 is out.

It’s compiling… I’m a little concerned about the last commit though. Looks like it could be a work in progress looking at the commit message.


Well it compiled and seems to test ok. New packages pushed.

linux-rpi5 rc-6.10.rc1-1
linux-rpi5-rc-headers 6.10.rc1-1

They have the linux-rpi4-rc 6.10.rc1-1 broke right now

Pi4/sway, 6.10-rc reboot fail.
plasma-5 broken, now back to PiOS rescue :rofl:

ok, how to fix this, did it once but forgot?
copy initramfs8 & /lib/modules from bootable kernel?

1 Like


ok, done thanks.
but you said test RC ok?

I only tested on the pi5. Too many pi devices to test all of them. I did just test on the pi4 and they defiantly have something messed up with it so I removed linux-rpi4-rc from the repo until they get it fixed. I get a kernel panic here.

btw, while back to PiOS-6.6, keyboard did not repeat, but your build did.

I do not know what is different in the OS but the kernels use their exact same config as a base using make bcm2711_defconfig like they do. I never have any keyboard repeats here.

These latest kernel packages has been pushed to the unstable branch when the mirrors sync.

The pi4 -rc still does not boot for me. It seems to be worse. All I get is a rainbow screen which usually means a missing/corrupt /boot file. I do not see any missing files so looks like something is amiss with the existing files regarding the pi4.

They seem to be having issues with this kernel getting things right on the pi4. They have yet been able to get through all of the compile tests also without it failing.

The pi5 -rc kernel still seems to be ok. I am guessing the reason is that it has not been upstreamed yet so they do not have to patch any of the upstream’s existing code because of it’s new changes to make things work the way they want.

linux-rpi5-rc 6.10.rc3-1
linux-rpi5-rc-headers 6.10.rc3-1

I need to use a snap package and apparmor is not enabled.

 ConditionSecurity=apparmor was not met
 Load AppArmor profiles managed internally by snapd was skipped because of an unmet condition check (ConditionSecurity=apparmor)

Does the kernel have the options to use apparmor or i just have to enable lsm options in cmdline ?

When compiling the kernel, it is required to set at least the following options:
To enable the AppArmor Linux security model by default and omit the need to set kernel parameters, additionally set the CONFIG_LSM option and specify apparmor as the first “major” module in the list:

The kernel modules should be enabled. I have this in my cmdline.txt and have had no issues with snaps.

1 Like

I have pushed the latest linux-rpi4-rc to the unstable repo.

I have messed with this kernel off and on most of the day today on my pi4 after their latest commits. It was booting on my test image on the stable branch but was having an issue with getting a Kernel Stack Trace when trying to play youtube vids. I had not updated the eeprom in a couple of years and flash it with the latest and still no change. Then I switched to the unstable branch and updated with no joy.

I then looked at the Kernel Stack Trace and saw that it looked like v4lm2m was involved so I decided to try chromium and had no issues playing youtube vids. Then I went back to firefox Edit → Settings → General → Performance and unchecked the 2 boxes and and it seemed to solve the Kernel Stack Trace problem.

I did not have jc-kynesim’s ffmpeg installed to test but it is just for mp4 HW so not much use for the vp9 stuff youtube defaults to. I am going to open an issue with the rpi github tomorrow but am too tired to mess with it today.

linux-rpi4-rc 6.10.rc3-2
linux-rpi4-rc-headers 6.10.rc3-2

thought, V3D/VC4 in kernel side is the end?
only look forward is CPU-jobs, but didn’t feel anything help?
so, desktop user should stay 6.6 follow PiOS?

v6.9.4 is out

@Dulbi Now pushed to unstable when the mirrors sync.

linux-rpi4 6.6.33-1
linux-rpi4-headers 6.6.33-1
linux-rpi4-mainline 6.9.4-1
linux-rpi4-mainline-headers 6.9.4-1
linux-rpi5 6.6.33-1
linux-rpi5-headers 6.6.33-1
linux-rpi5-mainline 6.9.4-1
linux-rpi5-mainline-headers 6.9.4-1
raspberrypi-bootloader 20240612-1
1 Like

Pi4, 6.9.4 test OK.

1 Like