Re: Failed to execute custom kernels which where build on a RPi 4 operated by 13.1-RELEASE

From: John Kennedy <warlock_at_phouka.net>
Date: Tue, 05 Jul 2022 12:47:12 UTC
On Mon, Jul 04, 2022 at 11:41:37PM -0700, Mark Millard wrote:
> Note: I do my own builds, installs, and configuration
> experiments. Do not expect my example outputs below to
> necessarily match those for official builds.
> 
> For reference from my root-on-UFS aarch64 environment that boots
> via U-Boot (not via EDK2 UEFI/ACPI):

  Gah.  There is so much more out there than RPI4 now for arm64 I didn't
stop to wonder if bsdinstall set it up right.  I just reformatted
/boot/efi and slapped the uboot stuff in there because that was the way
the SDcard image was set up.  I left it as /boot/efi (vs /boot/msdos).

> # sysctl hw.fdt # Likely a RPi* specific sysctl name?
> hw.fdt.serial-number: REPLACED
> hw.fdt.compatible: raspberrypi,4-model-b brcm,bcm2711
> hw.fdt.model: Raspberry Pi 4 Model B Rev 1.4

	# sysctl hw.fdt
	hw.fdt.serial-number: <REDACTED>
	hw.fdt.compatible: raspberrypi,4-model-b brcm,bcm2711
	hw.fdt.model: Raspberry Pi 4 Model B Rev 1.5

  Aha, that is what I was looking for.  Wasn't even seeing that with
hw-probe, and I wasn't going to interrupt the 12-hour llvm13 compile to
reboot and see what I could with uboot yet.

  There is this guy, Mark Millard, who writes these really good mailing
list emails that makes me paranoid about uboot version mismatches.  :D

  I haven't even looked at the EEPROM firmware stuff.  Reading the list,
it's all been such potential foot-cannon fodder seeing some people's
issues I just knew it would be a potential problem, but happily it just
worked as I would have hoped.

> I boot a RaspiOS64 microsd card for the related activity.

  Yeah, I haven't been doing that.