Re: [PATCH 4/5] sata_mv new mv_sata_hardreset handler

[Tejun Heo <htejun@gmail.com>]

Mark Lord wrote:
> Mmm.. I still need a way to handle access to the port-multiplier registers.
> We never did resolve that one fully on the earlier pass.

Yeah, I thought about that during the API change.  I think we should one 
of the following two.

1. Separate out actual reset part from ata_sff_softreset() and use it in 
sata_mv so that sata_mv can do...

mv_srst()
{
	check onlineness;
	set up PMP reg;
	return ata_sff_do_softreset();
}

2. Move link onlineness test before softreset into EH.  It's common to 
all drivers anyway.  This will allow mv to do.

mv_srst()
{
	set up PMP reg;
	return ata_sff_softreset();
}

I like #2 better but there can be exotic cases where link onlineness 
can't be tested before the link is actually reset.  I don't think it 
will really matter because hardreset is present and preferred if link 
onlineness test is valid && hardreset doesn't care about link onlineness.

Thanks.

-- 
tejun