Re: Additional EFI definitions

[Alexandre Boeglin]

Le Sun, Dec 09, 2007 at 12:21:50PM +0100, Alexandre Boeglin a écrit :
> - the disk controller is still put back in IDE mode (whereas it's in AHCI mode
>   when the machine boots, and stays in AHCI mode when booting a EFI OS, so I
>   guesss one of the CSM drivers is responsible for it)

Actually, this sounds logical: to be able to boot, the legacy bios needs a
option rom that registers itself in the bios boot specification or hooks some
interrupts ...

And the Apple firmwares do not contain such option roms.

I wonder how (in)acceptable it is to get those option roms from a bois update
of a motherboard using the same chipset (not redistributing them, of course).

Do you think it is worth asking Intel about it ? As peole who might be
interrested in this already own the hardware, and the EFI software, I don't
really see how it could be bad for them to give us the BIOS software. But
maybe I'm just being too naive ...

And this doesn't even add new features, as the Linux driver, for instance,
currently switches the controller back to AHCI mode when initializing itself.
And to get the same in Windows, I just used grub-legacy and modified it to
write to the PCI config register before chainloading. It would just be a lot
nicer to have everything done in a single EFI loader.


Alex