End of bios support on experimental platforms first? Also Alternative bootloaders

I have heard that the UEFI-only proposal has been stopped.
I can understand a lot of people still need BIOS either because they haven’t gotten around to migrating their systems or due to other reasons.
Obviously though these reasons over time will be harder to justify.
Would it be a better idea to end support for BIOS slowly on more experimental platforms first, like for example Silverblue.
It could also be used as a testing ground for moving away from grub and towards systemd-boot which is far more efficient at handling multiple entries as can be seen on he myriad of boot entries one can generate on another immutable distro, NixOS.

2 Likes

I have a system that I would VERY much like to retire (2006 IBM Intellistation). However, that isn’t going to happen for at least another year (probably longer). I would hate to be on an EOL version of Fedora while awaiting a hardware (and associated software) upgrade.

The choice of offering of BIOS support or no is really a red herring. Taking over from a booted system is the objective of the OS whether the hardware is X86_64 based with or without EFI, or AARCH64 based, or whatever. I can’t see a sound reason to support two divergent boot methods in the project, across different offerings.
Having said that, there is a reason to discuss what we use to boot EFI and what we use to boot BIOS. Grub2, has some difficulty with EFI hardware due to it’s design, systemd-boot does not. But systemd-boot currently does not support BIOS systems, and the likelihood of support for them landing there is next to nil. The logical approach seems to be provide both, and indeed you can use systemd-boot (not installed by default) or another boot manager other than Grub2 on Fedora Linux, just be prepared to roll up your sleeves and get into the dirty bits of booting a system. ATM, my understanding is that it isn’t also as simple as just picking another boot loader, since many users also multi-boot and often a non-linux alternative, and the secure boot can of worms adds another layer of complexity. Add disk encryption to the list of hurdles and you get to where we are now.