Re: [PATCH] sata_via: PATA support, resubmit

[Jeff Garzik <jgarzik@pobox.com>]

Alan wrote:
> This is a clean version of the PATA support for the sata_via hardware.
> I'm resubmitting it since nothing has happened since the last submission
> despite promises of libata core changes. Given users actually need to use
> this stuff today and the code is clean it should get merged irrespective
> of any longer term plans for per channel operations structs and the like.
> 
> Signed-off-by: Alan Cox <alan@redhat.com>

Re-explanation since you missed it the first time:

Per-channel operations structs work /today/.

The problem you need to fix or work around is ata_probe_ent, which 
doesn't properly fill in ata_port info for this situation.  Tejun has 
posted patches that kill ata_probe_ent, which you were pointed to. 
Mikael Pettersson just posted a sata_promise example that uses 
->port_start to work around this problem, setting the cable type and 
ata_port::ops properly at runtime, based on SATA or PATA.  See "[RFC] 
sata_promise: handle TX2plus PATA locally", I believe you were CC'd on 
my response.

So, working code for both the short term workaround and long term fix 
exist /today/.

If you get the setup right, you don't bloat each hook with "is this port 
PATA?" tests.  At present, your sata_via patch introduces these needless 
tests.

	Jeff