[BPI-R4] unable to test 6.6

i checked out latest sources, enabled testing kernel in experimental features, made a dirclean and a build, all looks good. but when i flash, it shows me kernel 6.1.75.

root@OpenWrt:~# uname -a Linux OpenWrt 6.1.75 #0 SMP Fri Feb 2 02:03:01 2024 aarch64 GNU/Linux

my config.buildinfo contains: CONFIG_TESTING_KERNEL=y and in dl/ i can see linux-6.6.20.tar.xz, no kernel with 6.1

even more strange, /home/oli/openwrt/build_dir/toolchain-aarch64_cortex-a53_gcc-12.3.0_musl/ contains linux-6.6.20, also the target is 6.6.20: /home/oli/openwrt/build_dir/target-aarch64_cortex-a53_musl/linux-mediatek_filogic/linux-6.6.20/

i have 0 clue where 6.1 comes from?! anyone an idea what is wrong?

update: i was not able to install the testing system with luci-> firmware update, i had to reflash the sd, now it works. not sure if this is a bug? i assume that the update fails somehow and thats the reason why its falling back to the old version and never overwrites the sd.



@oli you need to set


Eg.: https://github.com/danpawlik/openwrt-builder/blob/master/configs/mediatek/mt7988a/extended#L59

He wrote about this option already

I guess either image/kernel in image is not updated or the wrong device is booted. Also the uname output suggests it is an older build

Fri Feb 2 02:03:01

I also assume that something prevents the update and therefore the old system is booted. But no idea how to figure out what is blocking it. The wrong device can’t be booted as there is only 1 device with 1 SD :wink:

Btw. the same image flashed to SD via usb sd writer works. When i connect via uart, can i get those information what its doing in the background?

With device i meant the bootdevice (nand/emmc) not the r4 itself…what hapens when you poweron with sdcard removed?

I do not know magic behind openwrt,possibly rescuesystem via initrd is booted which is at older state.

i dont think it boots from nand/emmc. never installed anything to that. when i remove the sd, it looks like the r4 does nothing, so i assume it hangs somewhere in the bootloader

Ok,then you should provide full bootlog to see whats happen exactly…maybe the recovery theory