Re: The future of legacy BIOS support in Fedora.

[Solomon Peachy]

On Thu, Jul 02, 2020 at 01:16:43PM +0300, nickysn(a)gmail.com wrote:
>  The only Windows that no longer supports 32-bit systems is Windows
>  Server. So the obsolescence of Windows 7 and XP is irrelevant. 
I'm not talking about *old* hardware here, which obviously works as well 
now as it did before. I'm talking about *new* hardware, which is 
vanishingly unlikely to have been tested with anything other than the 
current Win10 release.

As Win10 certification does not mandate testing of BIOS/CSM modes, and 
all Windows versions that did require that have been completely EOL'd, 
new hardware purchased today is likely to only have had the most cursory 
testing in BIOS/CSM mode.. if it has BIOS/CSM support at all. (see: 
"Intel Dropping BIOS mode by 2020")

So, yes, BIOS/CSM mode works fine for older hardware, but the simple 
fact of the matter is that it is going away for new hardware, so it is 
no longer something that can be assumed to be there or be more 
functionally stable than UEFI.

In other words, if there are bugs/quirks/UX flaws/whatever in the 
Fedora's UEFI boot flow, we can no longer just say "oh, just switch to 
BIOS/CSM booting instead" as a functionally-acceptible workaround.

But take away BIOS/CSM, and the massive complexity of grub2 becomes 
simply unnecessary.  Simpler alternatives _should_ be considered.

(Of course, Fedora will need to continue supporting BIOS-based systems 
 for some time into the future.  Speaking for myseelf, I still have two 
 running systems that that lack UEFI; both are AMD-platform server 
 boards, and the newer of the two was first made in 2007 and was EOL'd 
 in 2011.  Plus a small pile of VMs..)

 - Solomon
-- 
Solomon Peachy			      pizza at shaftnet dot org (email&xmpp)
                                      @pizza:shaftnet dot org   (matrix)
High Springs, FL                      speachy (freenode)