Lennart, That is how we're hoping it should work, so it's good to hear. I suppose I'm not sure that it's the firmware driving this process--I just assumed because I know that the UEFI spec has verbiage requiring EFI boot managers to try next options in case of certain failure cases. I think you're probably right in that sd-boot *should* be able to continue onwards down the list. We're seeing the following error message in red text: ---------------- Error loading \EFI\Linux\linux-5.15.0-unsigned.efi: Security Policy Violation Failed to execute [entry config name] (\EFI\Linux\linux-5.15.0-unsigned.efi): Security Policy Violation ------------ What I believe is happening based on these messages is that image_start() is returning an error here: https://github.com/systemd/systemd/blob/v252/src/boot/efi/boot.c#L2747 and the `goto out;` is being executed, ending/preventing any looping over boot options. If this is a bug, I'd be willing to attempt a pull request submission if a suggested fix is given. Overall we like the functionality sd-boot provides and the integration with systemd, but this is likely a hard requirement for our use case. Thanks, --Daniel On Wed, Nov 23, 2022 at 11:04 AM Lennart Poettering <lennart@xxxxxxxxxxxxxx> wrote: > > On Mi, 23.11.22 10:22, Daniel Harms (jdharms@xxxxxxxxx) wrote: > > > Hello, > > > > We are doing some experiments with booting self-signed Unified Kernel > > Images (UKIs) using systemd-boot. Our eventual use-case is edge/IoT > > devices, so no interactive user will be present for most OS upgrade > > flows. > > > > In doing some testing on the boot option fallback features (in a > > vmware vm) we’ve run into a snag—when we set up an unsigned UKI as the > > first option and a properly signed UKI as the second option, > > systemd-boot appears to attempt to boot the unsigned one (as > > expected), the system reports a security violation, but then the > > firmware kicks us to the next boot option. > > Hmm, are you sure this is the firmware? Normally a security violation > should just be returned as an error to sd-boot, and sd-boot should be > able to pick the next option then. Not entirely sure this works > correctly though. There might be a bug lurking somewhere. > > it's simply not a case we regular test for. But it should be a case > that just works. > > Lennart > > -- > Lennart Poettering, Berlin