Re: sata_sil24 broken since 2.6.23-rc4-mm1

[Tejun Heo <htejun@gmail.com>]

Jens Axboe wrote:
> This is the old ata_sg_is_last:
> 
> static inline int
> ata_sg_is_last(struct scatterlist *sg, struct ata_queued_cmd *qc)
> {
>         if (sg == &qc->pad_sgent)
>                 return 1;
>         if (qc->pad_len)
>                 return 0;
>         if (((sg - qc->__sg) + 1) == qc->n_elem)
>                 return 1;
>         return 0;
> }
> 
> and the new one:
> 
> static inline int
> ata_sg_is_last(struct scatterlist *sg, struct ata_queued_cmd *qc)
> {
>         if (sg == &qc->pad_sgent)
>                 return 1;
>         if (qc->pad_len)
>                 return 0;
>         if (qc->n_iter == qc->n_elem)
>                 return 1; 
>         return 0;
> }
> 
> ->n_iter is how ata_qc_next_sg() walks over the sglist, I don't
> understand your reference to why depending on that during iteration
> would be bad?

Because that makes ata_sg iterator macros have hidden side effects
(nothing in the interface suggests it can't be nested and when somebody
actually nests it, finding what went wrong can be pretty difficult).  I
think it would be better to have explicit ata_sg_iter passed around if
sg itself isn't enough to walk the sg list.

> So we could add a test for sg_last() there, but that would turn sg table
> iteration into an O(N^2) operation for those drivers that use
> ata_sg_is_last() with chained sg tables. I'd much rather just get rid of
> ata_sg_is_last(), it's only used to mark end-of-table entries for
> hardware. That logic can be performed cheaper.

Yeap, it can be removed but having "is this the last one?" test is just
nicer to low level drivers.  With ata_sg_iter, I think we can do it in
O(N) by looking up and caching the next entry.

Thanks.

-- 
tejun