On Tue, 2020-06-30 at 13:34 +0000, Jóhann B. Guðmundsson wrote:
Given Hans proposal [1] introduced systemd/grub2/Gnome upstream
changes
it beg the question if now would not be the time to stop supporting
booting in legacy bios mode and move to uefi only supported boot
which
has been available on any common intel based x86 platform since
atleast
2005.
Now in 2017 Intel's technical marketing engineer Brian Richardson
revealed in a presentation that the company will require UEFI Class
3
and above as in it would remove legacy BIOS support from its client
and
datacenter platforms by 2020 and one might expect AMD to follow Intel
in
this regard.
So Intel platforms produced this year presumably will be unable
to run
32-bit operating systems, unable to use related software (at least
natively), and unable to use older hardware, such as RAID HBAs (and
therefore older hard drives that are connected to those HBAs),
network
cards, and even graphics cards that lack UEFI-compatible vBIOS
(launched
before 2012 – 2013) etc.
This post is just to gather feed back why Fedora should still
continue
to support legacy BIOS boot as opposed to stop supporting it and
potentially drop grub2 and use sd-boot instead.
Share your thoughts and comments on how such move might affect you
so
feedback can be collected for the future on why such a change might
be
bad, how it might affect the distribution and scope of such change
can
be determined for potential system wide proposal.
Reason 1: I don't see a reason for supporting the planned obsolescence
that hardware companies try to push on users.
Reason 2: I bought my latest laptop in the end of 2019. It's an HP
laptop with a Ryzen quad-core CPU. Yes, it has UEFI, but I bought it
without Windows, so it came preinstalled with FreeDOS, which runs in
legacy mode only. I installed Fedora on it, without switching to UEFI
mode. I believe many users who buy laptops without Windows would be
using Linux in legacy boot mode, just because it's the only way to boot
FreeDOS, and companies such as HP only offer you the option of having
Windows or FreeDOS preinstalled.
Overall, it is pain and breakage for a portion of the users, while I
don't see a particularly strong benefit - testing and maintaining
legacy BIOS boot support should be easy, because every VM out there
supports it, and it's even the default for most of them. In fact, I've
never tried UEFI in a VM, and I have no idea how stable it is. But even
if it's stable, I wouldn't want to have to migrate them to UEFI,
because that would require repartitioning, which could cause loss of
data in case something goes wrong, which leads us to reason 3:
Reason 3: Migration path to UEFI boot mode is difficult even for modern
systems, that were installed in legacy mode. It requires tinkering with
the BIOS settings, which is non-standard and therefore different for
every computer, and also repartitioning and reinstall of all operating
systems on the computer. Even if this is automated, there are many
dangerous scenarios that can cause loss of data. Think about multiboot
scenarios, for example, where a user is multibooting Windows 10 and
Fedora. We simply can't handle the upgrade in this case, without
destroying the Windows 10 installation, even if we are able to upgrade
a Fedora only install (and even that is dangerous).
Best regards,
Nikolay