Re: support for SATA port multipliers

[David Brownlee <abs%netbsd.org@localhost>]

On 2 February 2012 10:41, Manuel Bouyer <bouyer%antioche.eu.org@localhost> 
wrote:
> On Wed, Feb 01, 2012 at 10:33:13PM -0500, Thor Lancelot Simon wrote:
>> On Wed, Feb 01, 2012 at 08:09:18PM -0700, Brook Milligan wrote:
>> > It appears that the siisata driver currently does not support
>> > attachment to port multiplier boards even though the underlying
>> > hardware does in fact support that configuration.
>>
>> All SATA hardware should support that configuration, but there is
>> no support for it at all in our ATA code.
>>
>> Perhaps more urgent, though, if someone wants to hack the SATA
>> code out of the sheer goodness of his or her heart, is tagged
>> queueing support!
>
> I think both are equally urgent. As most SATA drives comes with a
> large write-back cache enabled by default, I'm not sure NCQ will give
> a lot of performances improvements in the common case.


What sort of hardware are we looking at here for the port multiplier?
Something like http://www.cooldrives.com/siseata5pomu.html ?