In message <20051219105220.GA24011@vaasje.org>,
Frank van der Linden writes:

>On Wed, Dec 14, 2005 at 12:58:12PM -0800, jonathan@dsg.stanford.edu wrote:
>> I should have two PCI-e cards with SiI 3132 chips delivered today.
>> AFAIK, NetBSD doesn't have a driver for the SiI3124 or 3132, but
>> FreeBSD has a 3124 driver; and the Linux sil24 driver supports both
>> PCI-X and PCI-e cards.
>> 
>> I've not seen any mention of open-source ACHI drivers which support
>> these chips.  Have you?
>
>Having written an SiI3124 driver for NetBSD (which won't be open-sourced as
>far as I know) in my previous job, I can tell you that the SiI3124 interface
>and the AHCI interface are different.

Yes, that's what I was tryihng to say. Frank, if may I quote again the
immediately-following text, which you snipped without comment (perhaps
without reading it?):

	JS>I've not seen any mention of open-source ACHI drivers which support
	JS>these chips.  Have you?

you can see that I was already saying, politely, that I didn't beleive
AHCI drivers would support the SiI3214 or SiI3132.

As it happens, I can't find a FreeBSD source driver for the SiI chips.
Linux does has a 3124 driver, quite separate from the linux 3112/3114
driver, but I don't know enough about the Linux libata to confidently
write a driver based on the Linux code.

A shame the SiI3124 driver for NetBSD you mention (Wasabi, I assume?)
will never see the light of day.


>The good news is, that the NetBSD ATA code has been changed to not have
>any dependencies on IDE registers in the controller-independent layer,
>so new host interfaces like AHCI can be fit into the ATA framework
>without having to change it.

Well, that at least is good news.