Re: [PATCH 14/14] libata: use PIO for misc ATAPI commands

[Tejun Heo <htejun@gmail.com>]

Alan Cox wrote:
>> polling for media change which involves reading maybe a few tens of
>> bytes.  The overhead of doing those using PIO wouldn't be too much more
>> than the cost of writing CDBs out.  It would help more if we can cut
>> down the number of commands used for testing media changed event (on my
>> to do list).
> 
> Actually the overhead is foul on some laptops as the command byte write
> wakes up the controller, which then whirrs away for a bit recovering its
> status before deciding to give us an answer.

Yeah, but that happens whether DMA is used or not.  Probably what's
needed is disabling media presence polling according to power profile.

>> Does module parameter / sysfs node sound good enough to you?
> 
> You mean like the existing "dma" one we already have which lets you turn
> on/off atapi DMA and could be extended with another bit so we went from
> 
> 1 - disk
> 2 - cd
> 4 - cf
> 
> to
> 
> 1 - disk
> 2 - cd (basic commands)
> 4 - cf
> 8 - cd (all)
> 
> 
> (The idea being that its ordered so troubleshooting almost always comes
> down to 1,3 - or 1,3,7  with this). So we have 99.9% of this

Basically, yes but I think it would be nice to be able to change the
value on the fly.

> As to the sysfs node for per device setup that makes sense as does fixing
> the support to allow users to force a mode of their choice, and putting
> in the geometry fields.

Yeap.

-- 
tejun