You can login if you already have an account or register by clicking the button below.
Registering is free and all you need is a username and password. We never ask you for your e-mail.
[–]BaronHK0 points
1 point
1 point
(+1|-0)
ago
(edited ago)
Matthew Garrett is correct and not correct, throughout that article.
The fake RAID mode is designed to work around a limitation in Windows. Windows can't do driver overrides, so putting it in "RAID" forces it to load the Intel driver with better power management. That much is true.
But...
Removing the AHCI option and writing new code to write protect the RAID setting from being toggled with an EFI variable amount to a Linux lockout mechanism. Linux wouldn't benefit from having a driver for the RAID controller other than getting around the lockout. Power management would be the same in Linux in either mode.
If you read the comments, the better solution is probably having Linux kick the controller into AHCI mode after GRUB hands over. GRUB is using the uEFI disk access, but Linuxx can't use uEFI access because performance would suck. Putting the controller in AHCI mode with Linux before PCI device enumeration would take a few lines of code, while writing RAID controller support might take months. So ignoring the RAID controller and bringing the device back up in native AHCI mode is the better solution, if it can be done (Garrett isn't sure if the hardware allows that, but he said he'd look into it.).
view the rest of the comments →
[–] Stavon 1 point 0 points 1 point (+1|-1) ago
IMO this provides a better explanation.
[–] BaronHK 0 points 1 point 1 point (+1|-0) ago (edited ago)
Matthew Garrett is correct and not correct, throughout that article.
The fake RAID mode is designed to work around a limitation in Windows. Windows can't do driver overrides, so putting it in "RAID" forces it to load the Intel driver with better power management. That much is true.
But...
Removing the AHCI option and writing new code to write protect the RAID setting from being toggled with an EFI variable amount to a Linux lockout mechanism. Linux wouldn't benefit from having a driver for the RAID controller other than getting around the lockout. Power management would be the same in Linux in either mode.
If you read the comments, the better solution is probably having Linux kick the controller into AHCI mode after GRUB hands over. GRUB is using the uEFI disk access, but Linuxx can't use uEFI access because performance would suck. Putting the controller in AHCI mode with Linux before PCI device enumeration would take a few lines of code, while writing RAID controller support might take months. So ignoring the RAID controller and bringing the device back up in native AHCI mode is the better solution, if it can be done (Garrett isn't sure if the hardware allows that, but he said he'd look into it.).