Re: libata NCQ implementation questions

[Mark Lord <liml@rtr.ca>]

Tejun Heo wrote:
> Mikael Pettersson wrote:
>> Mark Lord writes:
>>  > Mikael Pettersson wrote:
>>  > > I've started working on NCQ support for sata_promise,
>>  > > and I have a few questions regarding NCQ and libata:
>>  > >  > > 1. Can I rely not seeing any non-NCQ commands while there
>>  > >    are uncompleted NCQ commands on a port?
>>  > ..
>>  >  > No, you cannot.  So you'll have to code a .qc_defer() method
>>  > to hold them off in that case.  For a *really* good example,
>>  > see my recently deployed mv_qc_defer() function in sata_mv.c
>>  > as of linux-2.6.26-rc2.
>>
>> Hmm, so I misinterpreted ata_std_qc_defer()? Ok, I'll look
>> at sata_mv's version.
> 
> ata_std_qc_defer() which is inherited from sata_port_ops does guarantee 
> that NCQ and non-NCQ commands don't mix on a single device.  Mark, 
> you're talking about PMP, right?
..

Ahh.. probably, yes.  I just saw it mixing PIO with NCQ and then
turfed it in favour of a more comprehensive mv_qc_defer().

But that was with a PM connected at the time, and I don't remember if
I also saw it without a PM.

Cheers