On 2025-07-31 11:43:52, Joanne Koong wrote:
> On Mon, Jul 28, 2025 at 1:31 PM Andrey Albershteyn <aalbersh@xxxxxxxxxx> wrote:
> >
> > From: Andrey Albershteyn <aalbersh@xxxxxxxxxx>
> >
> > Add iomap_writepages_unbound() without limit in form of EOF. XFS
> > will use this to write metadata (fs-verity Merkle tree) in range far
> > beyond EOF.
> >
> > Signed-off-by: Andrey Albershteyn <aalbersh@xxxxxxxxxx>
> > ---
> > fs/iomap/buffered-io.c | 51 +++++++++++++++++++++++++++++++++++++++-----------
> > include/linux/iomap.h | 3 +++
> > 2 files changed, 43 insertions(+), 11 deletions(-)
> >
> > diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> > index 3729391a18f3..7bef232254a3 100644
> > --- a/fs/iomap/buffered-io.c
> > +++ b/fs/iomap/buffered-io.c
> > @@ -1881,18 +1881,10 @@ static int iomap_writepage_map(struct iomap_writepage_ctx *wpc,
> > int error = 0;
> > u32 rlen;
> >
> > - WARN_ON_ONCE(!folio_test_locked(folio));
> > - WARN_ON_ONCE(folio_test_dirty(folio));
> > - WARN_ON_ONCE(folio_test_writeback(folio));
> > -
> > - trace_iomap_writepage(inode, pos, folio_size(folio));
> > -
> > - if (!iomap_writepage_handle_eof(folio, inode, &end_pos)) {
> > - folio_unlock(folio);
> > - return 0;
> > - }
> > WARN_ON_ONCE(end_pos <= pos);
> >
> > + trace_iomap_writepage(inode, pos, folio_size(folio));
> > +
> > if (i_blocks_per_folio(inode, folio) > 1) {
> > if (!ifs) {
> > ifs = ifs_alloc(inode, folio, 0);
> > @@ -1956,6 +1948,23 @@ static int iomap_writepage_map(struct iomap_writepage_ctx *wpc,
> > return error;
> > }
> >
> > +/* Map pages bound by EOF */
> > +static int iomap_writepage_map_eof(struct iomap_writepage_ctx *wpc,
> > + struct writeback_control *wbc, struct folio *folio)
> > +{
> > + int error;
> > + struct inode *inode = folio->mapping->host;
> > + u64 end_pos = folio_pos(folio) + folio_size(folio);
> > +
> > + if (!iomap_writepage_handle_eof(folio, inode, &end_pos)) {
> > + folio_unlock(folio);
> > + return 0;
> > + }
> > +
> > + error = iomap_writepage_map(wpc, wbc, folio);
> > + return error;
> > +}
> > +
> > int
> > iomap_writepages(struct address_space *mapping, struct writeback_control *wbc,
> > struct iomap_writepage_ctx *wpc,
> > @@ -1972,9 +1981,29 @@ iomap_writepages(struct address_space *mapping, struct writeback_control *wbc,
> > PF_MEMALLOC))
> > return -EIO;
> >
> > + wpc->ops = ops;
> > + while ((folio = writeback_iter(mapping, wbc, folio, &error))) {
> > + WARN_ON_ONCE(!folio_test_locked(folio));
> > + WARN_ON_ONCE(folio_test_dirty(folio));
> > + WARN_ON_ONCE(folio_test_writeback(folio));
> > +
> > + error = iomap_writepage_map_eof(wpc, wbc, folio);
> > + }
> > + return iomap_submit_ioend(wpc, error);
> > +}
> > +EXPORT_SYMBOL_GPL(iomap_writepages);
> > +
> > +int
> > +iomap_writepages_unbound(struct address_space *mapping, struct writeback_control *wbc,
> > + struct iomap_writepage_ctx *wpc,
> > + const struct iomap_writeback_ops *ops)
> > +{
> > + struct folio *folio = NULL;
> > + int error;
> > +
> > wpc->ops = ops;
> > while ((folio = writeback_iter(mapping, wbc, folio, &error)))
> > error = iomap_writepage_map(wpc, wbc, folio);
> > return iomap_submit_ioend(wpc, error);
> > }
> > -EXPORT_SYMBOL_GPL(iomap_writepages);
> > +EXPORT_SYMBOL_GPL(iomap_writepages_unbound);
> > diff --git a/include/linux/iomap.h b/include/linux/iomap.h
> > index 522644d62f30..4a0b5ebb79e9 100644
> > --- a/include/linux/iomap.h
> > +++ b/include/linux/iomap.h
> > @@ -464,6 +464,9 @@ void iomap_sort_ioends(struct list_head *ioend_list);
> > int iomap_writepages(struct address_space *mapping,
> > struct writeback_control *wbc, struct iomap_writepage_ctx *wpc,
> > const struct iomap_writeback_ops *ops);
> > +int iomap_writepages_unbound(struct address_space *mapping,
> > + struct writeback_control *wbc, struct iomap_writepage_ctx *wpc,
> > + const struct iomap_writeback_ops *ops);
> >
>
> Just curious, instead of having a new api for
> iomap_writepages_unbound, does adding a bitfield for unbound to the
> iomap_writepage_ctx struct suffice? afaict, the logic between the two
> paths is identical except for the iomap_writepage_handle_eof() call
> and some WARN_ONs - if that gets gated behind the bitfield check, then
> it seems like it does the same thing logically but imo is more
> straightforward to follow the code flow of. But maybe I"m missing some
> reason why this wouldn't work?
Yeah, that's another option. If others also prefer flag then I can
change it.
--
- Andrey
