Re: libata - success with 82801AA IDE

[Jeff Garzik <jgarzik@pobox.com>]

Pavel Roskin wrote:
> Hello!
> 
> I just want to report that Intel 82801AA IDE controller (8086:2411) is
> working with libata from 2.6.0-test9-bk3.  I applied following patch:

Cool.



> I have a CD-ROM as master on the the secondary channel (normally seen as
> /dev/hdc), but it doesn't support DMA, so it wasn't detected.

Well, even so, ATAPI support is very incomplete at the moment.  That's a 
bigger reason why PATA support isn't receiving more love, at the moment.


> I don't know if there is any interest in developing the "parallel IDE"
> support.  I see that the option to enable it is not present in Linux
> configuration.  Anyway, that's a device that can be added to the
> device table if PATA support is here to stay.

I find it quite useful locally, and will be finishing up cable detection 
(currently PATA is limited to UDMA/33) soon.  However, with my 
apologies, I won't be applying your patch -- for the moment.  Before any 
more PATA PCI ids are added, I'll need to go through the docs and 
drivers/ide/pci/piix.c, to determine precisely PATA chipset speed 
limits:  UDMA/133 for some, UDMA/100, etc.  Your patch is a bit 
incorrect in that it does not truly reflect the maximum speed of the 
82801AA chipset.

In any case, the executive decision has been made "drivers/ide for PATA, 
libata for SATA", so that's the overall direction.

Thanks,

	Jeff