On Tue, 27 Dec 2022 at 22:27, Michael Grunditz <michael.grunditz@gmail.com> wrote:

Hi!

I finally managed to setup my RK3588 board again. I also have built my own u-boot. The error I was experiencing before was because of wrong address of binary. bootm uImage still doesn't work , fails in some dt_header thing in u-boot. If I can get things moving I will try to isolate that problem.

Now to current status:
Complete hang in switch to hypervisor. That eret never returns... unless Genode::raw stops working after it. Another thing in cortex_a53_cpu.cc that doesn't work is: Cpu::current_core_id(). I get a compile time error about not being in Board::Cpu. I think those two errors are connected.

Perhaps someone can tell me about the last error (core-id), and what might cause it?

I rebased it and it compiles now.

But it still halts in the jump to hypervisor. The entrypoint is set here https://github.com/genodelabs/genode/blob/master/repos/base-hw/src/bootstrap/spec/arm_64/cortex_a53_mmu.cc#L131

Is the uart mapped one to one and is it usable at all during mmu setiup? I am (still) a bit new to how Genode sets up things with c++ and also on mmu setup on aarch64.

Which is the next part after mmu enabling? I assume that a bad write to phys address when the mmu is on can cause various issues.

Thanks again.