I noticed this ticket originally referenced openSUSE's sd-boot work. Their work is still considered experimental and non-default, and involves some evolving tooling (notably sdbootutil).

From the Fedora perspective, we still have a couple issues to work through:

• signed sd-boot (which requires re-review of shim to include usage of sd-boot)
• fixes to the installer stack so that sd-boot and grub2 can both exist on the image (for using it on live installs)

It also comes with a couple of serious downsides (at least for now):

• everything has to go on the ESP (unless we package up and sign filesystem drivers for sd-boot to load, which is possible), and larger ESPs run the risk of being incompatible with hardware and virtual machines alike (and space contention is a real problem that we keep trying to avoid)
• we must still use grub2 for legacy BIOS, so we have to maintain two codepaths for everything

I do not think yet the benefits outweigh the downsides. I've been poking at getting btrfs-efi packaged so that we have a reasonably featured and well developed UEFI fsdriver that can be combined with any boot manager, but it's not done yet.

Clearly we cannot do this yet.

But it's not clear to me whether we can do this at all. Can sd-boot handle dual booting Windows and macOS?

we must still use grub2 for legacy BIOS, so we have to maintain two codepaths for everything

I'm not sure how important this will be for our bootloader developers, but I suspect probably important?

It also comes with a couple of serious downsides (at least for now):

• everything has to go on the ESP (unless we package up and sign filesystem drivers for sd-boot to load, which is possible), and larger ESPs run the risk of being incompatible with hardware and virtual machines alike (and space contention is a real problem that we keep trying to avoid)
• we must still use grub2 for legacy BIOS, so we have to maintain two codepaths for everything

We're going to reject this proposal for now due to the serious downsides that Neal has mentioned above.

However, as time goes on, fewer and fewer people will need the BIOS boot support and these disadvantages might be become less relevant. So we are saying "not in the next couple years" rather than "no never." Eventually we will probably want to switch to systemd-boot when we are more comfortable with leaving BIOS users behind.

This issue has been migrated to Fedora Forge:
https://forge.fedoraproject.org/workstation/tickets/issues/419

Please continue any further discussion there.