Re: [PATCHv6] blk-integrity: support arbitrary buffer alignment

[Keith Busch <kbusch@kernel.org>]

On Mon, Nov 24, 2025 at 01:34:03PM -0800, Caleb Sander Mateos wrote:
> On Mon, Nov 24, 2025 at 8:18 AM Keith Busch <kbusch@meta.com> wrote:
> >
> > From: Keith Busch <kbusch@kernel.org>
> >
> > A bio segment may have partial interval block data with the rest continuing
> > into the next segments because direct-io data payloads only need to aligned in
> 
> "need to be aligned"?

In the original text, s/aligned/align
 
> > +       while (offset > 0) {
> > +               struct bio_vec pbv = mp_bvec_iter_bvec(bvec, iter->prot_iter);
> > +               unsigned int len = min(pbv.bv_len, offset);
> > +               void *prot_buf = bvec_kmap_local(&pbv);
> 
> Is it valid to use bvec_kmap_local() on a multi-page bvec? 

If it wasn't valid, there'd be a major problem with large folios.

> It calls
> kmap_local_page() internally, which will only map the first page,
> right?

Compare kmap_local_page() with kmap_local_folio(). They both resolve to
the same lower level mapping function, and folios have no problem
spanning pages.
 
> > @@ -1874,6 +1874,7 @@ static bool nvme_init_integrity(struct nvme_ns_head *head,
> >                 break;
> >         }
> >
> > +       bi->flags |= BLK_SPLIT_INTERVAL_CAPABLE;
> 
> Can just be = instead of |= since bi is zeroed above.

It can, but these kinds of syntax are a courtesty to future changes so
that you don't need to change this line later. You can also end a struct
initialization without a trailing "," too, but that's just mean.