NOTE: The output directory when building the virtualization configuration for meta-aws-ewaol is meta-aws-ewaol/build/tmp_virtualization/deploy/images/generic-arm64 directory 1. The disk image generated from the virtualization build (using patch provided) is using systemd-boot and 3 partitions are created in the RAW disk image generated in the output directory /boot - systemd-boot EFI boot partition / - rootfs for dom0 Linux swap - swap partition for dom0 Linux 2. The EFI boot partition is set-up for systemd-boot and has the following directories/files: EFI/ BOOT/ bootaa64.efi Image loader/ entries/ boot.conf loader.conf 3. The EFI loader.conf is configured to boot the Image file (dom0 Linux image): loader/loader.conf ================== default boot timeout 5 loader/entries/boot.conf ======================== title boot linux /Image options LABEL=Boot root=PARTUUID=010d674c-1698-4cbf-a290-ada8c560733d rootwait rootfstype=ext4 4. With this configuration, the c6g.metal instance would boot into the dom0 Linux kernel and fail to interface to xen (fails to detect /proc/xen, etc, in Linux boot log) 5. So, my focus was to try and manually modify the EFI boot partition from the dom0 Linux instance as follows: * Add xen-generic-arm64.efi from the meta-aws-ewaol output directory into the EFI boot partition at same level as Linux kernel Image file * Add xen-generic-arm64.cfg created using the AVA platform as reference to EFI boot partition at same level as xen-generic-arm64.efi (see above): xen-generic-arm64.cfg ===================== [global] default=xen [xen] options=noreboot dom0_mem=2048M bootscrub=0 iommu=on loglvl=error guest_loglvl=error kernel=Image console=hvc0 earlycon=xenboot rootwait root=PARTUUID=010d674c-1698-4cbf-a290-ada8c560733d rootfstype=ext4 * Add new entry to EFI boot partition for xen loader/entries/xen.conf ======================= title Xen Hypervisor efi /xen-generic-arm64.efi * Modify loader.conf to boot xen loader/loader.conf ================== default xen timeout 5 * Stopped / Restarted the c6g.metal instance and no connectivity exists after these changes - No ssh - No EC2 serial console output - Instance status in management view showing at least 1 failure to connect to instance