Re: [PATCH 11/12] iomap: add read_folio_range() handler for buffered writes
From: Brian Foster <hidden>
Date: 2025-06-27 19:15:09
Also in:
gfs2, linux-block, linux-fsdevel, linux-xfs
On Fri, Jun 27, 2025 at 09:02:44AM +0200, Christoph Hellwig wrote:
quoted hunk ↗ jump to hunk
From: Joanne Koong <joannelkoong@gmail.com> Add a read_folio_range() handler for buffered writes that filesystems may pass in if they wish to provide a custom handler for synchronously reading in the contents of a folio. Signed-off-by: Joanne Koong <joannelkoong@gmail.com> [hch: renamed to read_folio_range, pass less arguments] Signed-off-by: Christoph Hellwig <hch@lst.de> --- .../filesystems/iomap/operations.rst | 6 +++++ fs/iomap/buffered-io.c | 25 +++++++++++-------- include/linux/iomap.h | 10 ++++++++ 3 files changed, 31 insertions(+), 10 deletions(-)diff --git a/Documentation/filesystems/iomap/operations.rst b/Documentation/filesystems/iomap/operations.rst index 167d3ca7819c..04432f40e7a2 100644 --- a/Documentation/filesystems/iomap/operations.rst +++ b/Documentation/filesystems/iomap/operations.rst@@ -68,6 +68,8 @@ The following address space operations can be wrapped easily: void (*put_folio)(struct inode *inode, loff_t pos, unsigned copied, struct folio *folio); bool (*iomap_valid)(struct inode *inode, const struct iomap *iomap); + int (*read_folio_range)(const struct iomap_iter *iter, + struct folio *folio, loff_t pos, size_t len);
Whitespace ^
}; iomap calls these functions:
...
quoted hunk ↗ jump to hunk
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c index a77686977a2e..1a9ade77aeeb 100644 --- a/fs/iomap/buffered-io.c +++ b/fs/iomap/buffered-io.c@@ -667,22 +667,23 @@ iomap_write_failed(struct inode *inode, loff_t pos, unsigned len) pos + len - 1); } -static int iomap_read_folio_sync(loff_t block_start, struct folio *folio, - size_t poff, size_t plen, const struct iomap *iomap) +static int iomap_read_folio_range(const struct iomap_iter *iter, + struct folio *folio, loff_t pos, size_t len) { + const struct iomap *srcmap = iomap_iter_srcmap(iter); struct bio_vec bvec; struct bio bio; - bio_init(&bio, iomap->bdev, &bvec, 1, REQ_OP_READ); - bio.bi_iter.bi_sector = iomap_sector(iomap, block_start); - bio_add_folio_nofail(&bio, folio, plen, poff); + bio_init(&bio, srcmap->bdev, &bvec, 1, REQ_OP_READ); + bio.bi_iter.bi_sector = iomap_sector(srcmap, pos); + bio_add_folio_nofail(&bio, folio, len, offset_in_folio(folio, pos)); return submit_bio_wait(&bio); }
Hmm, so this kind of makes my brain hurt... pos here is now the old block_start and len is the old plen. We used to pass poff to the add_folio_nofail() call, and now that is dropped and instead we use offset_in_folio(..., pos). The old poff is an output of the previous iomap_adjust_read_range() call, which is initially set to offset_in_folio(folio, *pos), of which *pos is block_start and is bumped in that function in the same places that poff is. Therefore old poff and new offset_in_folio(folio, pos) are logically equivalent. Am I following that correctly? Brian
quoted hunk ↗ jump to hunk
-static int __iomap_write_begin(const struct iomap_iter *iter, size_t len, +static int __iomap_write_begin(const struct iomap_iter *iter, + const struct iomap_write_ops *write_ops, size_t len, struct folio *folio) { - const struct iomap *srcmap = iomap_iter_srcmap(iter); struct iomap_folio_state *ifs; loff_t pos = iter->pos; loff_t block_size = i_blocksize(iter->inode);@@ -731,8 +732,12 @@ static int __iomap_write_begin(const struct iomap_iter *iter, size_t len, if (iter->flags & IOMAP_NOWAIT) return -EAGAIN; - status = iomap_read_folio_sync(block_start, folio, - poff, plen, srcmap); + if (write_ops && write_ops->read_folio_range) + status = write_ops->read_folio_range(iter, + folio, block_start, plen); + else + status = iomap_read_folio_range(iter, + folio, block_start, plen); if (status) return status; }@@ -848,7 +853,7 @@ static int iomap_write_begin(struct iomap_iter *iter, else if (srcmap->flags & IOMAP_F_BUFFER_HEAD) status = __block_write_begin_int(folio, pos, len, NULL, srcmap); else - status = __iomap_write_begin(iter, len, folio); + status = __iomap_write_begin(iter, write_ops, len, folio); if (unlikely(status)) goto out_unlock;diff --git a/include/linux/iomap.h b/include/linux/iomap.h index 482787013ff7..b3588dd43105 100644 --- a/include/linux/iomap.h +++ b/include/linux/iomap.h@@ -166,6 +166,16 @@ struct iomap_write_ops { * locked by the iomap code. */ bool (*iomap_valid)(struct inode *inode, const struct iomap *iomap); + + /* + * Optional if the filesystem wishes to provide a custom handler for + * reading in the contents of a folio, otherwise iomap will default to + * submitting a bio read request. + * + * The read must be done synchronously. + */ + int (*read_folio_range)(const struct iomap_iter *iter, + struct folio *folio, loff_t pos, size_t len); }; /*-- 2.47.2