Stefan, Erich, thank you for your suggestions. Am 24.04.21 um 18:54 schrieb Erich Eckner via arch-general:
On Sat, 24 Apr 2021, Stefan Göbel via arch-general wrote:
On 24.04.21 17:41, Uwe Sauter via arch-general wrote:
I've tried running 'pacstrap -C pacman.aarch64.conf /mnt base' where pacman.aarch64.conf is extracted from the Arch Linux Arm installation archive but after the packages are extracted, pacman tries to run something inside the installation environment which fails due to the mismatch in architectures.
You can setup QEMU to allow running foreign arch stuff, see [1].
I'm not sure if that will allow you to pacstrap an AArch64 root directly from the host. If it fails, you can grab the Arch Linux ARM AArch64 image and run pacstrap from within the extracted root dir, e.g. `arch-chroot </path/to/aarch64/root> pacstrap …`.
[1] <https://wiki.archlinux.org/index.php/QEMU#Chrooting_into_arm/arm64_environment_from_x86_64>
This method is good for recovering an unbootable raspi (that's, what I do in this case). IIRC, pacstrap from the host even works, too (mkinitcpio obviously fails in autodetect hooks and such stuff).
However, I did not yet get all the files necessary for boot into the right place (bootloader, device tree, etc.) to actually pacstrap onto a previously empty microsd card.
In case, that you are successful pacstrapping into an empty partition and makeing it bootable on a raspi, I would be really interested in the additional steps necessary. However, I think, this is rather a topic for the arch-ports mailing list, because it is too detached from x86_64 arch linux :-)
I decided to try out running an Aarch64 Qemu VM in which I can run pacstrap for the same architecture as my target and was able to get the Raspi to boot. @Erich: you may contact me off-list if you'd like to discuss further. Regards, Uwe
regards, Erich