Re: [PATCH] libata-core More robust parsing for multi_count(v5)

[Tejun Heo <tj@kernel.org>]

Hello, Mark.

Mark Lord wrote:
> Alan Cox wrote:
>>> Oh, and zero means "don't care".  So it's really easy to add
>>> to libata without having to patch each and every driver.
>>
>> If its only one driver why not do it the way we do things like transfer
>> length limits for LBA48 (see pata_it821x and others)
> ..
> 
> Thanks.  I'll do that, and probably go that way for the sata_mv limit.

Yeap, for the time being, that seems to be the simplest thing to do.

> The bigger issue here is that libata really doesn't support multiple mode.
> 
> Sure, it will blindly use it if the drive appears to be already configured
> for it at startup.  But I don't see any where that we issue "SET MULTIPLE"
> commands or otherwise support turning it on/off or tuning it.
> 
> I wonder if we even snoop the SET MULTIPLE command from SG_IO /
> ATA_{12,16}?
> 
> This is more of an infrastructure thing than a driver-specific setting,
> so I'm hoping that one of you three (who know the code a lot better
> than I do) might think about that for the future.

Right, libata doesn't have any multiple sector related configuration
whatsoever.  Maybe I'll give it a shot.

Thanks.

-- 
tejun