Re: [PATCH v2 02/18] xfs: convert iomap ops to ->iomap_next()

[Joanne Koong]

On Fri, Jul 17, 2026 at 10:55 AM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
>
> On Fri, Jul 17, 2026 at 09:49:30AM -0700, Joanne Koong wrote:
> > On Thu, Jul 2, 2026 at 6:21 PM Joanne Koong <joannelkoong@xxxxxxxxx> wrote:
> > >
> > > Ahh okay, I think I see the point you were trying to make. The
> > > callsite functions are like:
> > >
> > > ssize_t iomap_file_buffered_write(struct kiocb *iocb, struct iov_iter *from,
> > >                 iomap_next_fn iomap_next,
> > >                 const struct iomap_write_ops *write_ops, void *private);
> > >
> > > int iomap_fsverity_write(struct file *file, loff_t pos, size_t length,
> > >                 const void *buf, iomap_next_fn iomap_next,
> > >                 const struct iomap_write_ops *write_ops);
> > >
> > > void iomap_read_folio(iomap_next_fn iomap_next,
> > >                 struct iomap_read_folio_ctx *ctx, void *private);
> > >
> > > void iomap_readahead(iomap_next_fn iomap_next,
> > >                 struct iomap_read_folio_ctx *ctx, void *private);
> > >
> > > etc, so it's not clear from the "iomap_next_fn iomap_next" naming that
> > > it handles per-iteration logic, whereas renaming it to "iomap_iter_fn
> > > iomap_iter" makes it more explicit. Am I interpreting this correctly?
> >
> > Going back to this... "iomap_iter_fn iomap_iter" doesn't feel quite
> > right. I think there's too much naming collision with iomap_iter() and
> > struct iomap_iter, and I like how in general ->next() matches the
> > orthodox naming scheme for advancing an iterator. I think
> > "iomap_iter_next_fn next" would be the best of both worlds - I'll
> > change it to this for v3. But if you don't agree with this, please let
> > me know.
>
> Fine by me :)
>
> --D
>
> >
> > Thanks,
> > Joanne
> >
> > >
> > > > > > +{
> > > > > > +     return iomap_process(iter, iomap, srcmap, xfs_direct_write_iomap_begin,
> > > > > > +                     NULL);
> > > > >
> > > > > But then "iomap_iter" sets up a new problem: should iomap_process have a
> > > > > new name that goes along with that?  iomap_iter is already taken for the

I'm starting not to like "iomap_process()" either, it seems too vague...

I think it'd be better renamed to "iomap_iter_next()". The callsites
would then just look something like:

static int
xfs_read_iomap_next(
       const struct iomap_iter *iter,
       struct iomap            *iomap,
       struct iomap            *srcmap)
{
       return iomap_iter_next(iter, iomap, srcmap, xfs_read_iomap_begin, NULL);
}

static int
xfs_seek_iomap_next(
       const struct iomap_iter *iter,
       struct iomap            *iomap,
       struct iomap            *srcmap)
{
       return iomap_iter_next(iter, iomap, srcmap, xfs_seek_iomap_begin, NULL);
}

I'll make this change for v3. But if you don't like this, please let me know :)

Thanks,
Joanne

> > > > > legacy path.  If we were emulating python I'd suggest iomap_iter_yield
> > > > > but this is C so we get to reinvent everything from scratch so who
> > > > > knows.
> > > > >
> > > > > --D
> > > > >
> >