iomap: don't chain bios

Back in the days when a single bio could only be filled to the hardware
limits, and we scheduled a work item for each bio completion, chaining
multiple bios for a single ioend made a lot of sense to reduce the number
of completions.  But these days bios can be filled until we reach the
number of vectors or total size limit, which means we can always fit at
least 1 megabyte worth of data in the worst case, but usually a lot more
due to large folios.  The only thing bio chaining is buying us now is
to reduce the size of the allocation from an ioend with an embedded bio
into a plain bio, which is a 52 bytes differences on 64-bit systems.

This is not worth the added complexity, so remove the bio chaining and
only use the bio embedded into the ioend.  This will help to simplify
further changes to the iomap writeback code.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20231207072710.176093-10-hch@lst.de
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>

1 files changed, 23 insertions, 67 deletions

diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 3a3f3ebc070c..5f6affbe7056 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -1479,40 +1479,23 @@ static u32
 iomap_finish_ioend(struct iomap_ioend *ioend, int error)
 {
 	struct inode *inode = ioend->io_inode;
-	struct bio *bio = &ioend->io_inline_bio;
-	struct bio *last = ioend->io_bio, *next;
-	u64 start = bio->bi_iter.bi_sector;
-	loff_t offset = ioend->io_offset;
-	bool quiet = bio_flagged(bio, BIO_QUIET);
+	struct bio *bio = &ioend->io_bio;
+	struct folio_iter fi;
 	u32 folio_count = 0;
 
-	for (bio = &ioend->io_inline_bio; bio; bio = next) {
-		struct folio_iter fi;
-
-		/*
-		 * For the last bio, bi_private points to the ioend, so we
-		 * need to explicitly end the iteration here.
-		 */
-		if (bio == last)
-			next = NULL;
-		else
-			next = bio->bi_private;
-
-		/* walk all folios in bio, ending page IO on them */
-		bio_for_each_folio_all(fi, bio) {
-			iomap_finish_folio_write(inode, fi.folio, fi.length,
-					error);
-			folio_count++;
-		}
-		bio_put(bio);
+	/* walk all folios in bio, ending page IO on them */
+	bio_for_each_folio_all(fi, bio) {
+		iomap_finish_folio_write(inode, fi.folio, fi.length, error);
+		folio_count++;
 	}
-	/* The ioend has been freed by bio_put() */
 
-	if (unlikely(error && !quiet)) {
+	if (unlikely(error && !bio_flagged(bio, BIO_QUIET))) {
 		printk_ratelimited(KERN_ERR
 "%s: writeback error on inode %lu, offset %lld, sector %llu",
-			inode->i_sb->s_id, inode->i_ino, offset, start);
+			inode->i_sb->s_id, inode->i_ino,
+			ioend->io_offset, ioend->io_sector);
 	}
+	bio_put(bio);	/* frees the ioend */
 	return folio_count;
 }
 
@@ -1553,7 +1536,7 @@ EXPORT_SYMBOL_GPL(iomap_finish_ioends);
 static bool
 iomap_ioend_can_merge(struct iomap_ioend *ioend, struct iomap_ioend *next)
 {
-	if (ioend->io_bio->bi_status != next->io_bio->bi_status)
+	if (ioend->io_bio.bi_status != next->io_bio.bi_status)
 		return false;
 	if ((ioend->io_flags & IOMAP_F_SHARED) ^
 	    (next->io_flags & IOMAP_F_SHARED))
@@ -1618,9 +1601,8 @@ EXPORT_SYMBOL_GPL(iomap_sort_ioends);
 
 static void iomap_writepage_end_bio(struct bio *bio)
 {
-	struct iomap_ioend *ioend = bio->bi_private;
-
-	iomap_finish_ioend(ioend, blk_status_to_errno(bio->bi_status));
+	iomap_finish_ioend(iomap_ioend_from_bio(bio),
+			blk_status_to_errno(bio->bi_status));
 }
 
 /*
@@ -1635,9 +1617,6 @@ static int
 iomap_submit_ioend(struct iomap_writepage_ctx *wpc, struct iomap_ioend *ioend,
 		int error)
 {
-	ioend->io_bio->bi_private = ioend;
-	ioend->io_bio->bi_end_io = iomap_writepage_end_bio;
-
 	if (wpc->ops->prepare_ioend)
 		error = wpc->ops->prepare_ioend(ioend, error);
 	if (error) {
@@ -1647,12 +1626,12 @@ iomap_submit_ioend(struct iomap_writepage_ctx *wpc, struct iomap_ioend *ioend,
 		 * as there is only one reference to the ioend at this point in
 		 * time.
 		 */
-		ioend->io_bio->bi_status = errno_to_blk_status(error);
-		bio_endio(ioend->io_bio);
+		ioend->io_bio.bi_status = errno_to_blk_status(error);
+		bio_endio(&ioend->io_bio);
 		return error;
 	}
 
-	submit_bio(ioend->io_bio);
+	submit_bio(&ioend->io_bio);
 	return 0;
 }
 
@@ -1666,44 +1645,22 @@ static struct iomap_ioend *iomap_alloc_ioend(struct iomap_writepage_ctx *wpc,
 			       REQ_OP_WRITE | wbc_to_write_flags(wbc),
 			       GFP_NOFS, &iomap_ioend_bioset);
 	bio->bi_iter.bi_sector = iomap_sector(&wpc->iomap, pos);
+	bio->bi_end_io = iomap_writepage_end_bio;
 	wbc_init_bio(wbc, bio);
 
-	ioend = container_of(bio, struct iomap_ioend, io_inline_bio);
+	ioend = iomap_ioend_from_bio(bio);
 	INIT_LIST_HEAD(&ioend->io_list);
 	ioend->io_type = wpc->iomap.type;
 	ioend->io_flags = wpc->iomap.flags;
 	ioend->io_inode = inode;
 	ioend->io_size = 0;
 	ioend->io_offset = pos;
-	ioend->io_bio = bio;
 	ioend->io_sector = bio->bi_iter.bi_sector;
 
 	wpc->nr_folios = 0;
 	return ioend;
 }
 
-/*
- * Allocate a new bio, and chain the old bio to the new one.
- *
- * Note that we have to perform the chaining in this unintuitive order
- * so that the bi_private linkage is set up in the right direction for the
- * traversal in iomap_finish_ioend().
- */
-static struct bio *
-iomap_chain_bio(struct bio *prev)
-{
-	struct bio *new;
-
-	new = bio_alloc(prev->bi_bdev, BIO_MAX_VECS, prev->bi_opf, GFP_NOFS);
-	bio_clone_blkg_association(new, prev);
-	new->bi_iter.bi_sector = bio_end_sector(prev);
-
-	bio_chain(prev, new);
-	bio_get(prev);		/* for iomap_finish_ioend */
-	submit_bio(prev);
-	return new;
-}
-
 static bool iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t pos)
 {
 	if ((wpc->iomap.flags & IOMAP_F_SHARED) !=
@@ -1714,7 +1671,7 @@ static bool iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t pos)
 	if (pos != wpc->ioend->io_offset + wpc->ioend->io_size)
 		return false;
 	if (iomap_sector(&wpc->iomap, pos) !=
-	    bio_end_sector(wpc->ioend->io_bio))
+	    bio_end_sector(&wpc->ioend->io_bio))
 		return false;
 	/*
 	 * Limit ioend bio chain lengths to minimise IO completion latency. This
@@ -1739,15 +1696,14 @@ static void iomap_add_to_ioend(struct iomap_writepage_ctx *wpc,
 	size_t poff = offset_in_folio(folio, pos);
 
 	if (!wpc->ioend || !iomap_can_add_to_ioend(wpc, pos)) {
+new_ioend:
 		if (wpc->ioend)
 			list_add(&wpc->ioend->io_list, iolist);
 		wpc->ioend = iomap_alloc_ioend(wpc, wbc, inode, pos);
 	}
 
-	if (!bio_add_folio(wpc->ioend->io_bio, folio, len, poff)) {
-		wpc->ioend->io_bio = iomap_chain_bio(wpc->ioend->io_bio);
-		bio_add_folio_nofail(wpc->ioend->io_bio, folio, len, poff);
-	}
+	if (!bio_add_folio(&wpc->ioend->io_bio, folio, len, poff))
+		goto new_ioend;
 
 	if (ifs)
 		atomic_add(len, &ifs->write_bytes_pending);
@@ -1978,7 +1934,7 @@ EXPORT_SYMBOL_GPL(iomap_writepages);
 static int __init iomap_init(void)
 {
 	return bioset_init(&iomap_ioend_bioset, 4 * (PAGE_SIZE / SECTOR_SIZE),
-			   offsetof(struct iomap_ioend, io_inline_bio),
+			   offsetof(struct iomap_ioend, io_bio),
 			   BIOSET_NEED_BVECS);
 }
 fs_initcall(iomap_init);