Re: bio_chain: proposed solution for bio_alloc failure and large IO

[Andrew Morton]

"Adam J. Richter" wrote:
> 
> >- It's tricky to determine how many bvecs are available in
> >  a bio.  There is no straightforward "how big is it" field
> >  in struct bio.
> 
>         Can't I make a macro to do a table lookup from bio->bi_max?

Not really.  If I do

	bio_alloc(GFP_KERNEL, 27);

then I'll get a 32-slot bvec.  But presumably, I don't
want to put more than 27 pages into it.

-