Cure found for kernel updates

Matthew Garrett mjg59 at srcf.ucam.org
Fri May 16 15:41:09 UTC 2014


On Fri, May 16, 2014 at 05:19:56PM +0200, Lennart Poettering wrote:
> On Fri, 16.05.14 15:55, Matthew Garrett (mjg59 at srcf.ucam.org) wrote:
> > If we want Thunderbolt then we have to perform an EFI boot rather than 
> > using the BIOS compatibility layer. The only way to perform an EFI boot 
> > and integrate with the OS X boot preferences is to have the bootloader 
> > on an HFS+ partition.
> 
> Sure, well, again, I don't see why we need to care about the boot prefs,
> if we have te firmware boot menu anyway, and if Windows doesn't care
> about the boot menu either...

Windows does - Apple provide a tool for setting the boot preferences 
under Windows.

> I just don't think the price you pay for HFS+ ESP madness and
> the special-casing for mac there is worth the effort. Without this
> everything works pretty fine too, and as good or bad as windows does on
> macs. And the panel, well, is not really worth the ugliness of the hack...

Hack? It's the way MacOS works. Their bootloader isn't on the ESP.

> > Yes, but
> > 
> > a) You need an HFS+ partition unless you want a generic "EFI Boot" 
> > option
> 
> which is totally OK, and what windows does too, right?

I haven't checked the EFI Windows install situation, but Apple 
special-case all BIOS boots and calls them Windows. They may have 
special-cased Windows in this case as well.

> > b) It's not obvious how to set the default from there (ctrl+click, it 
> > turns out)
> 
> Well, I really don't see why we need to be better than windows for this...

The special-casing means that Windows also shows up in the OS X boot 
preferences. We have to do this work to be as good.

> > On BIOS, just add the chain-drive number to 0x80 and make the 
> > appropriate BIOS call. The implementation at the bootloader level is 
> > trivial. Figuring out how to map the drive to the BIOS number is 
> > obviously a pain, but you don't have to care - that's left up to the 
> > installer, and we already have the code to do that. It requires no 
> > changes to gummiboot and only a small patch to grub.
> 
> The BLS is explicitly not about reinventing grub and all its complexity
> when referring to devices. I am sorry, but this really has no place in
> the BLS.

Well, that's cool. I'll look into implementing something that's 
almost-but-not-quite the BLS - it does seem like it'd be an improvement 
over what we currently have, but the spec as written introduces extra 
configuration complexity and would reduce the quality of our user 
experience on Apple hardware.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org


More information about the desktop mailing list