Backport: fix(vm): avoid false no bootable device events from EFI shell

[eofff]

Description

Add a downstream EDK2 shell patch that moves the No bootable device. debug-port marker from unconditional shell startup to the interactive fallback shell path only.

The patch now checks that the EFI shell was started without a target file and without a Startup.nsh script before writing the marker to debug port 0x403. The patches README was updated to document the new downstream patch and patch ordering.

Why do we need it, and what problem does it solve?

Some operating systems legitimately launch the EFI shell with startup parameters or scripts. Emitting No bootable device. at every shell entry makes QEMU report a NO_BOOTABLE_DEVICE event even when the VM boot flow is valid.

This change keeps real firmware fallback detection intact while avoiding false no-bootable-device reports for normal parameterized EFI shell launches.

What is the expected result?

A VM that really falls back to the EFI shell without a bootable device still emits No bootable device. and produces the expected NO_BOOTABLE_DEVICE event.

A VM or OS flow that starts the EFI shell with a file, options, or Startup.nsh no longer emits the marker and is not treated as a no-bootable-device failure.

Checklist

• The code is covered by unit tests.
• e2e tests passed.
• Documentation updated according to the changes.
• Changes were tested in the Kubernetes cluster manually.

Changelog entries

section: core
type: fix
summary: "False no-bootable-device events are no longer reported when EFI shell is launched with startup parameters."