Here is my bootlog:
IPL is Rev.2.0.4 in my environment, but Rev.2.0.3 is used in your environment.
So, please update Android and IPL, then try to boot yocto v3.21.0 again.
Thank you very much for the boot-log of your system. I tried your suggestions with the new IPL and I'm now on the same version as you are.
Comparing the two logs (https://elinux.org/File:H3ulcb_yocto_with_android_ipl_kh.txt) they look very similar - except some memory addresses during IPL which I don't understand why they are different. Even the Kernel Size is identical - but the device tree is not.
$ md5sum *
$ ls -l
-rw-r--r-- 1 root root 16497152 Apr 7 09:53 Image
-rw-r--r-- 1 root root 92809 Apr 7 09:53 Image-r8a7795-h3ulcb-kf.dtb
I did a recompile of my yocto to rule out any problems there. Also there are no local modifications done.
If you still have the files, would you mind uploading your Image.dtb (if it really differs) and also your srec files? This way I would be able to rule out some hardware issue on my side.
Sorry, I noticed that I had used upstream dtb instead of KF BSP's one...
Thus, I confirmed again whether yocto can be booted or not.
As the result, I got kernel panic as same as you did.
I investigated the log, and noticed kernel is stopped at around rpc.
Below is a part of your log.
[ 1.574607] renesas-rpc ee200000.rpc0: using DMA read (dma0chan0)
[ 1.581125] Bad mode in Error handler detected on CPU1, code 0xbf000000 -- SError
I think it may be caused by a mismatch configuration for rpc between Android IPL and kernel .
Finally, I confirmed to boot yocto after I disabled rpc driver in kernel.
Here is patch of modification.
I hope that you can boot yocto as same as I do.
note) If you are not going to use rpc, please ignore after this.
I found patch about RPC for IPL:
If this patch is applied to Android IPL, Yocto may be booted without any modification.
However, I don't have build environment for android so that I cannot confirmed it.