Legacy BIOS Support Remains Important For Some On Fedora, May Shift Responsibility To SIG

[clockwork]

If their UEFI support was good, this would be a bit less worrying. I'm sitting here on a laptop that I have to manually select a boot partition and file every time I boot because Fedora didn't install correctly and can't manage to fix things. Maybe make sure the new thing works before you rip out the already working older thing for once?

I'm currently using it on 3 different computers. 1 laptop with encryption, 2 Desktops out of which 1 is custom built. All EFI. I've never had that problem. It sounds like something went wrong either during the install or sometime after. I've also been using Fedora for many years and I've never encountered that issue. I don't think your issue is "lack of support". I think something went wrong for you.

[yump]

I'm currently using it on 3 different computers. 1 laptop with encryption, 2 Desktops out of which 1 is custom built. All EFI. I've never had that problem. It sounds like something went wrong either during the install or sometime after. I've also been using Fedora for many years and I've never encountered that issue. I don't think your issue is "lack of support". I think something went wrong for you.

Or something went wrong with the firmware. EFI booting gives the firmware more agency in the boot process, and so there is more variance in quality of implementation. And OEMs and motherboard vendors are likely to only test with Windows.

[stormcrow]

Or something went wrong with the firmware. EFI booting gives the firmware more agency in the boot process, and so there is more variance in quality of implementation. And OEMs and motherboard vendors are likely to only test with Windows.

Yeah, seen this happen with Dell Laptops even with Windows. For some reason it's not enough to point it at a Windows UEFI entry in the boot partition, it has to be a specific one usually 'rediscovered' by the Dell diagnostic system. I've also found a bug in some MSI motherboards where they automatically drop to UEFI boot selection menu if more than one UEFI entry is in the boot partition. I reported the bug to MSI only to get a "We don't support Linux" mantra. I pointed out it effects Windows users, too and miraculously it went away 2 firmware updates later.

Lesson here: Don't tell MSI support your using Linux/BSD to report firmware bugs. Just tell them it's there.

[CommunityMember]

All they have to do is adopt the Clover Bootloader for x86_64 platforms.

And that has been proposed, although it should be noted that some vendors BIOS implementations require adding quirks (which clover does support, at least to a point), so that can be far from a seamless experience.

[ezst036]

And that has been proposed, although it should be noted that some vendors BIOS implementations require adding quirks (which clover does support, at least to a point), so that can be far from a seamless experience.

This is true, but 70% of something is better than 100% of nothing.

[willmore]

I'm currently using it on 3 different computers. 1 laptop with encryption, 2 Desktops out of which 1 is custom built. All EFI. I've never had that problem. It sounds like something went wrong either during the install or sometime after. I've also been using Fedora for many years and I've never encountered that issue. I don't think your issue is "lack of support". I think something went wrong for you.

I have reached out to them in various ways and no one has yet been able to solve the problem, so that is exactly a lack of support in both meanings of that phrase. I also have it working fine on many other machines, but that doesn't help the one machine where it doesn't work.

[stormcrow]

I have reached out to them in various ways and no one has yet been able to solve the problem, so that is exactly a lack of support in both meanings of that phrase. I also have it working fine on many other machines, but that doesn't help the one machine where it doesn't work.

Funny part of the promise of UEFI was a standardization to fix the unholy mess of broken BIOS implementations had become... only to end up with an unholy mess of non-compliant, broken, UEFI implementations. The problem likely isn't with Fedora. It's probably a firmware boot loader bug. At least, that's what it sounds like to me.

[willmore]

Funny part of the promise of UEFI was a standardization to fix the unholy mess of broken BIOS implementations had become... only to end up with an unholy mess of non-compliant, broken, UEFI implementations. The problem likely isn't with Fedora. It's probably a firmware boot loader bug. At least, that's what it sounds like to me.

Well, I don't want to hijack this comment section with my debugging, but the laptop says something like "cannot find a configured boot path" and then halts. If I hit F9 for boot options before that happens, I can select one of the EFI boot files from the partition that Fedora setup and it will boot into Fedora. So, I'm assuming there must be some kind of environment variable that Fedora should set to say what file the EFI is supposed to load to boot it, but that communication never happened or fedora set the value in a way the EFI doesn't like.

And, yes, the situation with EFI is exactly like BIOS was. I'm going to place some of that blame on Apple because they pushed it early on and did things Their Way(tm). I'd love to see some major manufacturer get behind one of the open UEFI firmwares and use it for all of their products. Maybe that would push the industry towards the right direction.

[davidhendricks]

Or something went wrong with the firmware. EFI booting gives the firmware more agency in the boot process, and so there is more variance in quality of implementation. And OEMs and motherboard vendors are likely to only test with Windows.

Yep, this is exactly right. Back in the 80's and 90's the legacy BIOS was used as a driver layer for operating systems such as DOS (https://en.wikipedia.org/wiki/BIOS_i...nterrupt_table has some good examples). It had terrible performance and different vendors had different quality/bugs.

UEFI is more standardized in many ways, but with very few exceptions the implementations consumers get with their off-the-shelf hardware are all proprietary to each vendor. OSes having any dependency on firmware, whether it's legacy BIOS or UEFI we're talking about, is history is repeating itself. We can expect much of the same in terms of problems from past decades.

[hikingpete]

I have very conflicted feelings about this. There are some things coming down the pipe that look extremely valuable, and systemd-boot really helps clean up the mess that is grub. On the other hand, I long for the simplicity of BIOS, I feel like UEFI is far too complex for the value it provides. When my VMs require UEFI to boot I still sad face. I'm not anxious to switch to managing kernels out of band, but that may be the future.