Re: Playing with SATA NCQ

[Erik Slagter]

On Sun, 2005-05-29 at 13:29 -0400, Jeff Garzik wrote:
> On Sun, May 29, 2005 at 07:23:52PM +0200, Erik Slagter wrote:
> > On Sun, 2005-05-29 at 12:59 -0400, Jeff Garzik wrote:
> > > > My question was if there is a fundamental reason why the AHCI mode of
> > > > the ICH6/7 must be enabled by the BIOS, is there a reason why the kernel
> > > > doesn't do it, or can't do it?
> > > 
> > > The BIOS sets up PCI resources necessary to use AHCI mode.
> > 
> > Ok. So there's absolutely no way to do that afterwards? It'd really be a
> > pity :-(
> 
> It is technically possible.  BIOS is just software, just like the OS.
> 
> It's just a huge pain in the butt, because the kernel might accidentally
> stomp on some resources the BIOS secretly set up, or somesuch.

Ah, ok. I wasn't aware of this large role of the BIOS in setting up
hardware, nowadays. Grmbl.

> > On the same subject: is there a reason why ICH6 gets "BAR0-3 ignored"
> > and always gets the legacy i/o ports and IRQ's assigned? I'd say there
> > is absolutely no need to be compatible in this way, the PCI code can
> > assign the IRQ and I/O ports as with any other PCI device?
> 
> IDE is special.
> 
> This is due to how the BIOS sets up an IDE PCI device in legacy mode.
> BAR0-3 are set to zero, which is a signal to the OS that the IDE PCI
> device is in legacy mode (io 0x1f0+0x170, irq 14+15).  Since the IDE I/O
> ports are in ISA space not PCI space, the PCI BARs reflect nothing.

Goodie. So we will be stuck with MS/DOS compatibility until somewhere in
the next millenium.

I guess the only way to have, for example the ICH6, not using legacy
IRQ/ports, is to switch it to AHCI, which only the BIOS can do (if
implemented).

Thrilling.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/