Re: Bio pool & scsi scatter gather pool usage

[Stephen Lord <lord@sgi.com>]

Andrew Morton wrote:

>Alan Cox wrote:
>
>>>Perhaps, but calls are expensive. Repeated calls down stacked block
>>>devices will add up. In only the most unusually cases will there
>>>
>>You don't need to repeatedly query. At bind time you can compute the
>>limit for any device heirarchy and be done with it.
>>
>
>S'pose so.  The ideal request size is variable, based
>on the alignment.  So for exampe if the start block is
>halfway into a stripe, the ideal BIO size is half a stripe.
>
>But that's a pretty simple table to generate once-off,
>as you say.
>

But this gets you lowest common denominator sizes for the whole
volume, which is basically the buffer head approach, chop all I/O up
into a chunk size we know will always work. Any sort of nasty  boundary
condition at one spot in a volume means the whole thing is crippled
down to that level. It then becomes a black magic art to configure a
volume which is not restricted to a small request size.

Steve