Merge tag 'vfs-6.9.iomap' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

Pull iomap updates from Christian Brauner:

 - Restore read-write hints in struct bio through the bi_write_hint
   member for the sake of UFS devices in mobile applications. This can
   result in up to 40% lower write amplification in UFS devices. The
   patch series that builds on this will be coming in via the SCSI
   maintainers (Bart)

 - Overhaul the iomap writeback code. Afterwards ->map_blocks() is able
   to map multiple blocks at once as long as they're in the same folio.
   This reduces CPU usage for buffered write workloads on e.g., xfs on
   systems with lots of cores (Christoph)

 - Record processed bytes in iomap_iter() trace event (Kassey)

 - Extend iomap_writepage_map() trace event after Christoph's
   ->map_block() changes to map mutliple blocks at once (Zhang)

* tag 'vfs-6.9.iomap' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (22 commits)
  iomap: Add processed for iomap_iter
  iomap: add pos and dirty_len into trace_iomap_writepage_map
  block, fs: Restore the per-bio/request data lifetime fields
  fs: Propagate write hints to the struct block_device inode
  fs: Move enum rw_hint into a new header file
  fs: Split fcntl_rw_hint()
  fs: Verify write lifetime constants at compile time
  fs: Fix rw_hint validation
  iomap: pass the length of the dirty region to ->map_blocks
  iomap: map multiple blocks at a time
  iomap: submit ioends immediately
  iomap: factor out a iomap_writepage_map_block helper
  iomap: only call mapping_set_error once for each failed bio
  iomap: don't chain bios
  iomap: move the iomap_sector sector calculation out of iomap_add_to_ioend
  iomap: clean up the iomap_alloc_ioend calling convention
  iomap: move all remaining per-folio logic into iomap_writepage_map
  iomap: factor out a iomap_writepage_handle_eof helper
  iomap: move the PF_MEMALLOC check to iomap_writepages
  iomap: move the io_folios field out of struct iomap_ioend
  ...

12 files changed, 392 insertions, 331 deletions

diff --git a/fs/buffer.c b/fs/buffer.c
index 9a54077de87d..4f73d23c2c46 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -55,7 +55,7 @@
 
 static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);
 static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
-			  struct writeback_control *wbc);
+			  enum rw_hint hint, struct writeback_control *wbc);
 
 #define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers)
 
@@ -1889,7 +1889,8 @@ int __block_write_full_folio(struct inode *inode, struct folio *folio,
 	do {
 		struct buffer_head *next = bh->b_this_page;
 		if (buffer_async_write(bh)) {
-			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh, wbc);
+			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh,
+				      inode->i_write_hint, wbc);
 			nr_underway++;
 		}
 		bh = next;
@@ -1944,7 +1945,8 @@ recover:
 		struct buffer_head *next = bh->b_this_page;
 		if (buffer_async_write(bh)) {
 			clear_buffer_dirty(bh);
-			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh, wbc);
+			submit_bh_wbc(REQ_OP_WRITE | write_flags, bh,
+				      inode->i_write_hint, wbc);
 			nr_underway++;
 		}
 		bh = next;
@@ -2756,6 +2758,7 @@ static void end_bio_bh_io_sync(struct bio *bio)
 }
 
 static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
+			  enum rw_hint write_hint,
 			  struct writeback_control *wbc)
 {
 	const enum req_op op = opf & REQ_OP_MASK;
@@ -2783,6 +2786,7 @@ static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
 	fscrypt_set_bio_crypt_ctx_bh(bio, bh, GFP_NOIO);
 
 	bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
+	bio->bi_write_hint = write_hint;
 
 	__bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
 
@@ -2802,7 +2806,7 @@ static void submit_bh_wbc(blk_opf_t opf, struct buffer_head *bh,
 
 void submit_bh(blk_opf_t opf, struct buffer_head *bh)
 {
-	submit_bh_wbc(opf, bh, NULL);
+	submit_bh_wbc(opf, bh, WRITE_LIFE_NOT_SET, NULL);
 }
 EXPORT_SYMBOL(submit_bh);
 
diff --git a/fs/direct-io.c b/fs/direct-io.c
index 60456263a338..62c97ff9e852 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -410,6 +410,8 @@ dio_bio_alloc(struct dio *dio, struct dio_submit *sdio,
 		bio->bi_end_io = dio_bio_end_io;
 	if (dio->is_pinned)
 		bio_set_flag(bio, BIO_PAGE_PINNED);
+	bio->bi_write_hint = file_inode(dio->iocb->ki_filp)->i_write_hint;
+
 	sdio->bio = bio;
 	sdio->logical_offset_in_bio = sdio->cur_page_fs_offset;
 }
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index dda0aed95464..4a4e60cdac4e 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -24,6 +24,7 @@
 #include <linux/blkdev.h>
 #include <linux/quotaops.h>
 #include <linux/part_stat.h>
+#include <linux/rw_hint.h>
 #include <crypto/hash.h>
 
 #include <linux/fscrypt.h>
diff --git a/fs/fcntl.c b/fs/fcntl.c
index c3e342eb74af..54cc85d3338e 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -27,6 +27,7 @@
 #include <linux/memfd.h>
 #include <linux/compat.h>
 #include <linux/mount.h>
+#include <linux/rw_hint.h>
 
 #include <linux/poll.h>
 #include <asm/siginfo.h>
@@ -268,8 +269,15 @@ static int f_getowner_uids(struct file *filp, unsigned long arg)
 }
 #endif
 
-static bool rw_hint_valid(enum rw_hint hint)
+static bool rw_hint_valid(u64 hint)
 {
+	BUILD_BUG_ON(WRITE_LIFE_NOT_SET != RWH_WRITE_LIFE_NOT_SET);
+	BUILD_BUG_ON(WRITE_LIFE_NONE != RWH_WRITE_LIFE_NONE);
+	BUILD_BUG_ON(WRITE_LIFE_SHORT != RWH_WRITE_LIFE_SHORT);
+	BUILD_BUG_ON(WRITE_LIFE_MEDIUM != RWH_WRITE_LIFE_MEDIUM);
+	BUILD_BUG_ON(WRITE_LIFE_LONG != RWH_WRITE_LIFE_LONG);
+	BUILD_BUG_ON(WRITE_LIFE_EXTREME != RWH_WRITE_LIFE_EXTREME);
+
 	switch (hint) {
 	case RWH_WRITE_LIFE_NOT_SET:
 	case RWH_WRITE_LIFE_NONE:
@@ -283,34 +291,40 @@ static bool rw_hint_valid(enum rw_hint hint)
 	}
 }
 
-static long fcntl_rw_hint(struct file *file, unsigned int cmd,
-			  unsigned long arg)
+static long fcntl_get_rw_hint(struct file *file, unsigned int cmd,
+			      unsigned long arg)
 {
 	struct inode *inode = file_inode(file);
 	u64 __user *argp = (u64 __user *)arg;
-	enum rw_hint hint;
-	u64 h;
+	u64 hint = READ_ONCE(inode->i_write_hint);
 
-	switch (cmd) {
-	case F_GET_RW_HINT:
-		h = inode->i_write_hint;
-		if (copy_to_user(argp, &h, sizeof(*argp)))
-			return -EFAULT;
-		return 0;
-	case F_SET_RW_HINT:
-		if (copy_from_user(&h, argp, sizeof(h)))
-			return -EFAULT;
-		hint = (enum rw_hint) h;
-		if (!rw_hint_valid(hint))
-			return -EINVAL;
+	if (copy_to_user(argp, &hint, sizeof(*argp)))
+		return -EFAULT;
+	return 0;
+}
 
-		inode_lock(inode);
-		inode->i_write_hint = hint;
-		inode_unlock(inode);
-		return 0;
-	default:
+static long fcntl_set_rw_hint(struct file *file, unsigned int cmd,
+			      unsigned long arg)
+{
+	struct inode *inode = file_inode(file);
+	u64 __user *argp = (u64 __user *)arg;
+	u64 hint;
+
+	if (copy_from_user(&hint, argp, sizeof(hint)))
+		return -EFAULT;
+	if (!rw_hint_valid(hint))
 		return -EINVAL;
-	}
+
+	WRITE_ONCE(inode->i_write_hint, hint);
+
+	/*
+	 * file->f_mapping->host may differ from inode. As an example,
+	 * blkdev_open() modifies file->f_mapping.
+	 */
+	if (file->f_mapping->host != inode)
+		WRITE_ONCE(file->f_mapping->host->i_write_hint, hint);
+
+	return 0;
 }
 
 static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
@@ -416,8 +430,10 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 		err = memfd_fcntl(filp, cmd, argi);
 		break;
 	case F_GET_RW_HINT:
+		err = fcntl_get_rw_hint(filp, cmd, arg);
+		break;
 	case F_SET_RW_HINT:
-		err = fcntl_rw_hint(filp, cmd, arg);
+		err = fcntl_set_rw_hint(filp, cmd, arg);
 		break;
 	default:
 		break;
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index d9ccfd27e4f1..789af5c8fade 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -2465,7 +2465,7 @@ out:
 }
 
 static int gfs2_map_blocks(struct iomap_writepage_ctx *wpc, struct inode *inode,
-		loff_t offset)
+		loff_t offset, unsigned int len)
 {
 	int ret;
 
diff --git a/fs/inode.c b/fs/inode.c
index 5c8a5250d0ac..d290f007b3d1 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -20,6 +20,7 @@
 #include <linux/ratelimit.h>
 #include <linux/list_lru.h>
 #include <linux/iversion.h>
+#include <linux/rw_hint.h>
 #include <trace/events/writeback.h>
 #include "internal.h"
 
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index 093c4515b22a..4e8e41c8b3c0 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 /*
  * Copyright (C) 2010 Red Hat, Inc.
- * Copyright (C) 2016-2019 Christoph Hellwig.
+ * Copyright (C) 2016-2023 Christoph Hellwig.
  */
 #include <linux/module.h>
 #include <linux/compiler.h>
@@ -95,6 +95,44 @@ static inline bool ifs_block_is_dirty(struct folio *folio,
 	return test_bit(block + blks_per_folio, ifs->state);
 }
 
+static unsigned ifs_find_dirty_range(struct folio *folio,
+		struct iomap_folio_state *ifs, u64 *range_start, u64 range_end)
+{
+	struct inode *inode = folio->mapping->host;
+	unsigned start_blk =
+		offset_in_folio(folio, *range_start) >> inode->i_blkbits;
+	unsigned end_blk = min_not_zero(
+		offset_in_folio(folio, range_end) >> inode->i_blkbits,
+		i_blocks_per_folio(inode, folio));
+	unsigned nblks = 1;
+
+	while (!ifs_block_is_dirty(folio, ifs, start_blk))
+		if (++start_blk == end_blk)
+			return 0;
+
+	while (start_blk + nblks < end_blk) {
+		if (!ifs_block_is_dirty(folio, ifs, start_blk + nblks))
+			break;
+		nblks++;
+	}
+
+	*range_start = folio_pos(folio) + (start_blk << inode->i_blkbits);
+	return nblks << inode->i_blkbits;
+}
+
+static unsigned iomap_find_dirty_range(struct folio *folio, u64 *range_start,
+		u64 range_end)
+{
+	struct iomap_folio_state *ifs = folio->private;
+
+	if (*range_start >= range_end)
+		return 0;
+
+	if (ifs)
+		return ifs_find_dirty_range(folio, ifs, range_start, range_end);
+	return range_end - *range_start;
+}
+
 static void ifs_clear_range_dirty(struct folio *folio,
 		struct iomap_folio_state *ifs, size_t off, size_t len)
 {
@@ -1454,15 +1492,10 @@ out_unlock:
 EXPORT_SYMBOL_GPL(iomap_page_mkwrite);
 
 static void iomap_finish_folio_write(struct inode *inode, struct folio *folio,
-		size_t len, int error)
+		size_t len)
 {
 	struct iomap_folio_state *ifs = folio->private;
 
-	if (error) {
-		folio_set_error(folio);
-		mapping_set_error(inode->i_mapping, error);
-	}
-
 	WARN_ON_ONCE(i_blocks_per_folio(inode, folio) > 1 && !ifs);
 	WARN_ON_ONCE(ifs && atomic_read(&ifs->write_bytes_pending) <= 0);
 
@@ -1479,40 +1512,29 @@ 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++;
+	if (error) {
+		mapping_set_error(inode->i_mapping, error);
+		if (!bio_flagged(bio, BIO_QUIET)) {
+			pr_err_ratelimited(
+"%s: writeback error on inode %lu, offset %lld, sector %llu",
+				inode->i_sb->s_id, inode->i_ino,
+				ioend->io_offset, ioend->io_sector);
 		}
-		bio_put(bio);
 	}
-	/* The ioend has been freed by bio_put() */
 
-	if (unlikely(error && !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);
+	/* walk all folios in bio, ending page IO on them */
+	bio_for_each_folio_all(fi, bio) {
+		if (error)
+			folio_set_error(fi.folio);
+		iomap_finish_folio_write(inode, fi.folio, fi.length);
+		folio_count++;
 	}
+
+	bio_put(bio);	/* frees the ioend */
 	return folio_count;
 }
 
@@ -1553,7 +1575,7 @@ 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))
@@ -1618,47 +1640,46 @@ 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));
 }
 
 /*
  * Submit the final bio for an ioend.
  *
  * If @error is non-zero, it means that we have a situation where some part of
- * the submission process has failed after we've marked pages for writeback
- * and unlocked them.  In this situation, we need to fail the bio instead of
- * submitting it.  This typically only happens on a filesystem shutdown.
+ * the submission process has failed after we've marked pages for writeback.
+ * We cannot cancel ioend directly in that case, so call the bio end I/O handler
+ * with the error status here to run the normal I/O completion handler to clear
+ * the writeback bit and let the file system proess the errors.
  */
-static int
-iomap_submit_ioend(struct iomap_writepage_ctx *wpc, struct iomap_ioend *ioend,
-		int error)
+static int iomap_submit_ioend(struct iomap_writepage_ctx *wpc, int error)
 {
-	ioend->io_bio->bi_private = ioend;
-	ioend->io_bio->bi_end_io = iomap_writepage_end_bio;
+	if (!wpc->ioend)
+		return error;
 
+	/*
+	 * Let the file systems prepare the I/O submission and hook in an I/O
+	 * comletion handler.  This also needs to happen in case after a
+	 * failure happened so that the file system end I/O handler gets called
+	 * to clean up.
+	 */
 	if (wpc->ops->prepare_ioend)
-		error = wpc->ops->prepare_ioend(ioend, error);
+		error = wpc->ops->prepare_ioend(wpc->ioend, error);
+
 	if (error) {
-		/*
-		 * If we're failing the IO now, just mark the ioend with an
-		 * error and finish it.  This will run IO completion immediately
-		 * 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);
-		return error;
+		wpc->ioend->io_bio.bi_status = errno_to_blk_status(error);
+		bio_endio(&wpc->ioend->io_bio);
+	} else {
+		submit_bio(&wpc->ioend->io_bio);
 	}
 
-	submit_bio(ioend->io_bio);
-	return 0;
+	wpc->ioend = NULL;
+	return error;
 }
 
-static struct iomap_ioend *
-iomap_alloc_ioend(struct inode *inode, struct iomap_writepage_ctx *wpc,
-		loff_t offset, sector_t sector, struct writeback_control *wbc)
+static struct iomap_ioend *iomap_alloc_ioend(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct inode *inode, loff_t pos)
 {
 	struct iomap_ioend *ioend;
 	struct bio *bio;
@@ -1666,63 +1687,42 @@ iomap_alloc_ioend(struct inode *inode, struct iomap_writepage_ctx *wpc,
 	bio = bio_alloc_bioset(wpc->iomap.bdev, BIO_MAX_VECS,
 			       REQ_OP_WRITE | wbc_to_write_flags(wbc),
 			       GFP_NOFS, &iomap_ioend_bioset);
-	bio->bi_iter.bi_sector = sector;
+	bio->bi_iter.bi_sector = iomap_sector(&wpc->iomap, pos);
+	bio->bi_end_io = iomap_writepage_end_bio;
 	wbc_init_bio(wbc, bio);
+	bio->bi_write_hint = inode->i_write_hint;
 
-	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_folios = 0;
-	ioend->io_offset = offset;
-	ioend->io_bio = bio;
-	ioend->io_sector = sector;
-	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);
+	ioend->io_offset = pos;
+	ioend->io_sector = bio->bi_iter.bi_sector;
 
-	bio_chain(prev, new);
-	bio_get(prev);		/* for iomap_finish_ioend */
-	submit_bio(prev);
-	return new;
+	wpc->nr_folios = 0;
+	return ioend;
 }
 
-static bool
-iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t offset,
-		sector_t sector)
+static bool iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t pos)
 {
 	if ((wpc->iomap.flags & IOMAP_F_SHARED) !=
 	    (wpc->ioend->io_flags & IOMAP_F_SHARED))
 		return false;
 	if (wpc->iomap.type != wpc->ioend->io_type)
 		return false;
-	if (offset != wpc->ioend->io_offset + wpc->ioend->io_size)
+	if (pos != wpc->ioend->io_offset + wpc->ioend->io_size)
 		return false;
-	if (sector != bio_end_sector(wpc->ioend->io_bio))
+	if (iomap_sector(&wpc->iomap, pos) !=
+	    bio_end_sector(&wpc->ioend->io_bio))
 		return false;
 	/*
 	 * Limit ioend bio chain lengths to minimise IO completion latency. This
 	 * also prevents long tight loops ending page writeback on all the
 	 * folios in the ioend.
 	 */
-	if (wpc->ioend->io_folios >= IOEND_BATCH_SIZE)
+	if (wpc->nr_folios >= IOEND_BATCH_SIZE)
 		return false;
 	return true;
 }
@@ -1730,255 +1730,238 @@ iomap_can_add_to_ioend(struct iomap_writepage_ctx *wpc, loff_t offset,
 /*
  * Test to see if we have an existing ioend structure that we could append to
  * first; otherwise finish off the current ioend and start another.
+ *
+ * If a new ioend is created and cached, the old ioend is submitted to the block
+ * layer instantly.  Batching optimisations are provided by higher level block
+ * plugging.
+ *
+ * At the end of a writeback pass, there will be a cached ioend remaining on the
+ * writepage context that the caller will need to submit.
  */
-static void
-iomap_add_to_ioend(struct inode *inode, loff_t pos, struct folio *folio,
-		struct iomap_folio_state *ifs, struct iomap_writepage_ctx *wpc,
-		struct writeback_control *wbc, struct list_head *iolist)
+static int iomap_add_to_ioend(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio,
+		struct inode *inode, loff_t pos, unsigned len)
 {
-	sector_t sector = iomap_sector(&wpc->iomap, pos);
-	unsigned len = i_blocksize(inode);
+	struct iomap_folio_state *ifs = folio->private;
 	size_t poff = offset_in_folio(folio, pos);
+	int error;
 
-	if (!wpc->ioend || !iomap_can_add_to_ioend(wpc, pos, sector)) {
-		if (wpc->ioend)
-			list_add(&wpc->ioend->io_list, iolist);
-		wpc->ioend = iomap_alloc_ioend(inode, wpc, pos, sector, wbc);
+	if (!wpc->ioend || !iomap_can_add_to_ioend(wpc, pos)) {
+new_ioend:
+		error = iomap_submit_ioend(wpc, 0);
+		if (error)
+			return error;
+		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);
 	wpc->ioend->io_size += len;
 	wbc_account_cgroup_owner(wbc, &folio->page, len);
+	return 0;
 }
 
-/*
- * We implement an immediate ioend submission policy here to avoid needing to
- * chain multiple ioends and hence nest mempool allocations which can violate
- * the forward progress guarantees we need to provide. The current ioend we're
- * adding blocks to is cached in the writepage context, and if the new block
- * doesn't append to the cached ioend, it will create a new ioend and cache that
- * instead.
- *
- * If a new ioend is created and cached, the old ioend is returned and queued
- * locally for submission once the entire page is processed or an error has been
- * detected.  While ioends are submitted immediately after they are completed,
- * batching optimisations are provided by higher level block plugging.
- *
- * At the end of a writeback pass, there will be a cached ioend remaining on the
- * writepage context that the caller will need to submit.
- */
-static int
-iomap_writepage_map(struct iomap_writepage_ctx *wpc,
-		struct writeback_control *wbc, struct inode *inode,
-		struct folio *folio, u64 end_pos)
+static int iomap_writepage_map_blocks(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio,
+		struct inode *inode, u64 pos, unsigned dirty_len,
+		unsigned *count)
 {
-	struct iomap_folio_state *ifs = folio->private;
-	struct iomap_ioend *ioend, *next;
-	unsigned len = i_blocksize(inode);
-	unsigned nblocks = i_blocks_per_folio(inode, folio);
-	u64 pos = folio_pos(folio);
-	int error = 0, count = 0, i;
-	LIST_HEAD(submit_list);
-
-	WARN_ON_ONCE(end_pos <= pos);
-
-	if (!ifs && nblocks > 1) {
-		ifs = ifs_alloc(inode, folio, 0);
-		iomap_set_range_dirty(folio, 0, end_pos - pos);
-	}
+	int error;
 
-	WARN_ON_ONCE(ifs && atomic_read(&ifs->write_bytes_pending) != 0);
-
-	/*
-	 * Walk through the folio to find areas to write back. If we
-	 * run off the end of the current map or find the current map
-	 * invalid, grab a new one.
-	 */
-	for (i = 0; i < nblocks && pos < end_pos; i++, pos += len) {
-		if (ifs && !ifs_block_is_dirty(folio, ifs, i))
-			continue;
+	do {
+		unsigned map_len;
 
-		error = wpc->ops->map_blocks(wpc, inode, pos);
+		error = wpc->ops->map_blocks(wpc, inode, pos, dirty_len);
 		if (error)
 			break;
-		trace_iomap_writepage_map(inode, &wpc->iomap);
-		if (WARN_ON_ONCE(wpc->iomap.type == IOMAP_INLINE))
-			continue;
-		if (wpc->iomap.type == IOMAP_HOLE)
-			continue;
-		iomap_add_to_ioend(inode, pos, folio, ifs, wpc, wbc,
-				 &submit_list);
-		count++;
-	}
-	if (count)
-		wpc->ioend->io_folios++;
+		trace_iomap_writepage_map(inode, pos, dirty_len, &wpc->iomap);
 
-	WARN_ON_ONCE(!wpc->ioend && !list_empty(&submit_list));
-	WARN_ON_ONCE(!folio_test_locked(folio));
-	WARN_ON_ONCE(folio_test_writeback(folio));
-	WARN_ON_ONCE(folio_test_dirty(folio));
+		map_len = min_t(u64, dirty_len,
+			wpc->iomap.offset + wpc->iomap.length - pos);
+		WARN_ON_ONCE(!folio->private && map_len < dirty_len);
+
+		switch (wpc->iomap.type) {
+		case IOMAP_INLINE:
+			WARN_ON_ONCE(1);
+			error = -EIO;
+			break;
+		case IOMAP_HOLE:
+			break;
+		default:
+			error = iomap_add_to_ioend(wpc, wbc, folio, inode, pos,
+					map_len);
+			if (!error)
+				(*count)++;
+			break;
+		}
+		dirty_len -= map_len;
+		pos += map_len;
+	} while (dirty_len && !error);
 
 	/*
 	 * We cannot cancel the ioend directly here on error.  We may have
 	 * already set other pages under writeback and hence we have to run I/O
 	 * completion to mark the error state of the pages under writeback
 	 * appropriately.
+	 *
+	 * Just let the file system know what portion of the folio failed to
+	 * map.
 	 */
-	if (unlikely(error)) {
-		/*
-		 * Let the filesystem know what portion of the current page
-		 * failed to map. If the page hasn't been added to ioend, it
-		 * won't be affected by I/O completion and we must unlock it
-		 * now.
-		 */
-		if (wpc->ops->discard_folio)
-			wpc->ops->discard_folio(folio, pos);
-		if (!count) {
-			folio_unlock(folio);
-			goto done;
-		}
-	}
-
-	/*
-	 * We can have dirty bits set past end of file in page_mkwrite path
-	 * while mapping the last partial folio. Hence it's better to clear
-	 * all the dirty bits in the folio here.
-	 */
-	iomap_clear_range_dirty(folio, 0, folio_size(folio));
-	folio_start_writeback(folio);
-	folio_unlock(folio);
-
-	/*
-	 * Preserve the original error if there was one; catch
-	 * submission errors here and propagate into subsequent ioend
-	 * submissions.
-	 */
-	list_for_each_entry_safe(ioend, next, &submit_list, io_list) {
-		int error2;
-
-		list_del_init(&ioend->io_list);
-		error2 = iomap_submit_ioend(wpc, ioend, error);
-		if (error2 && !error)
-			error = error2;
-	}
-
-	/*
-	 * We can end up here with no error and nothing to write only if we race
-	 * with a partial page truncate on a sub-page block sized filesystem.
-	 */
-	if (!count)
-		folio_end_writeback(folio);
-done:
-	mapping_set_error(inode->i_mapping, error);
+	if (error && wpc->ops->discard_folio)
+		wpc->ops->discard_folio(folio, pos);
 	return error;
 }
 
 /*
- * Write out a dirty page.
+ * Check interaction of the folio with the file end.
  *
- * For delalloc space on the page, we need to allocate space and flush it.
- * For unwritten space on the page, we need to start the conversion to
- * regular allocated space.
+ * If the folio is entirely beyond i_size, return false.  If it straddles
+ * i_size, adjust end_pos and zero all data beyond i_size.
  */
-static int iomap_do_writepage(struct folio *folio,
-		struct writeback_control *wbc, void *data)
+static bool iomap_writepage_handle_eof(struct folio *folio, struct inode *inode,
+		u64 *end_pos)
 {
-	struct iomap_writepage_ctx *wpc = data;
-	struct inode *inode = folio->mapping->host;
-	u64 end_pos, isize;
-
-	trace_iomap_writepage(inode, folio_pos(folio), folio_size(folio));
+	u64 isize = i_size_read(inode);
 
-	/*
-	 * Refuse to write the folio out if we're called from reclaim context.
-	 *
-	 * This avoids stack overflows when called from deeply used stacks in
-	 * random callers for direct reclaim or memcg reclaim.  We explicitly
-	 * allow reclaim from kswapd as the stack usage there is relatively low.
-	 *
-	 * This should never happen except in the case of a VM regression so
-	 * warn about it.
-	 */
-	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
-			PF_MEMALLOC))
-		goto redirty;
-
-	/*
-	 * Is this folio beyond the end of the file?
-	 *
-	 * The folio index is less than the end_index, adjust the end_pos
-	 * to the highest offset that this folio should represent.
-	 * -----------------------------------------------------
-	 * |			file mapping	       | <EOF> |
-	 * -----------------------------------------------------
-	 * | Page ... | Page N-2 | Page N-1 |  Page N  |       |
-	 * ^--------------------------------^----------|--------
-	 * |     desired writeback range    |      see else    |
-	 * ---------------------------------^------------------|
-	 */
-	isize = i_size_read(inode);
-	end_pos = folio_pos(folio) + folio_size(folio);
-	if (end_pos > isize) {
-		/*
-		 * Check whether the page to write out is beyond or straddles
-		 * i_size or not.
-		 * -------------------------------------------------------
-		 * |		file mapping		        | <EOF>  |
-		 * -------------------------------------------------------
-		 * | Page ... | Page N-2 | Page N-1 |  Page N   | Beyond |
-		 * ^--------------------------------^-----------|---------
-		 * |				    |      Straddles     |
-		 * ---------------------------------^-----------|--------|
-		 */
+	if (*end_pos > isize) {
 		size_t poff = offset_in_folio(folio, isize);
 		pgoff_t end_index = isize >> PAGE_SHIFT;
 
 		/*
-		 * Skip the page if it's fully outside i_size, e.g.
-		 * due to a truncate operation that's in progress.  We've
-		 * cleaned this page and truncate will finish things off for
-		 * us.
+		 * If the folio is entirely ouside of i_size, skip it.
+		 *
+		 * This can happen due to a truncate operation that is in
+		 * progress and in that case truncate will finish it off once
+		 * we've dropped the folio lock.
 		 *
-		 * Note that the end_index is unsigned long.  If the given
-		 * offset is greater than 16TB on a 32-bit system then if we
-		 * checked if the page is fully outside i_size with
-		 * "if (page->index >= end_index + 1)", "end_index + 1" would
-		 * overflow and evaluate to 0.  Hence this page would be
+		 * Note that the pgoff_t used for end_index is an unsigned long.
+		 * If the given offset is greater than 16TB on a 32-bit system,
+		 * then if we checked if the folio is fully outside i_size with
+		 * "if (folio->index >= end_index + 1)", "end_index + 1" would
+		 * overflow and evaluate to 0.  Hence this folio would be
 		 * redirtied and written out repeatedly, which would result in
 		 * an infinite loop; the user program performing this operation
 		 * would hang.  Instead, we can detect this situation by
-		 * checking if the page is totally beyond i_size or if its
+		 * checking if the folio is totally beyond i_size or if its
 		 * offset is just equal to the EOF.
 		 */
 		if (folio->index > end_index ||
 		    (folio->index == end_index && poff == 0))
-			goto unlock;
+			return false;
 
 		/*
-		 * The page straddles i_size.  It must be zeroed out on each
-		 * and every writepage invocation because it may be mmapped.
-		 * "A file is mapped in multiples of the page size.  For a file
-		 * that is not a multiple of the page size, the remaining
-		 * memory is zeroed when mapped, and writes to that region are
-		 * not written out to the file."
+		 * The folio straddles i_size.
+		 *
+		 * It must be zeroed out on each and every writepage invocation
+		 * because it may be mmapped:
+		 *
+		 *    A file is mapped in multiples of the page size.  For a
+		 *    file that is not a multiple of the page size, the
+		 *    remaining memory is zeroed when mapped, and writes to that
+		 *    region are not written out to the file.
+		 *
+		 * Also adjust the writeback range to skip all blocks entirely
+		 * beyond i_size.
 		 */
 		folio_zero_segment(folio, poff, folio_size(folio));
-		end_pos = isize;
+		*end_pos = round_up(isize, i_blocksize(inode));
+	}
+
+	return true;
+}
+
+static int iomap_writepage_map(struct iomap_writepage_ctx *wpc,
+		struct writeback_control *wbc, struct folio *folio)
+{
+	struct iomap_folio_state *ifs = folio->private;
+	struct inode *inode = folio->mapping->host;
+	u64 pos = folio_pos(folio);
+	u64 end_pos = pos + folio_size(folio);
+	unsigned count = 0;
+	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);
+
+	if (i_blocks_per_folio(inode, folio) > 1) {
+		if (!ifs) {
+			ifs = ifs_alloc(inode, folio, 0);
+			iomap_set_range_dirty(folio, 0, end_pos - pos);
+		}
+
+		/*
+		 * Keep the I/O completion handler from clearing the writeback
+		 * bit until we have submitted all blocks by adding a bias to
+		 * ifs->write_bytes_pending, which is dropped after submitting
+		 * all blocks.
+		 */
+		WARN_ON_ONCE(atomic_read(&ifs->write_bytes_pending) != 0);
+		atomic_inc(&ifs->write_bytes_pending);
 	}
 
-	return iomap_writepage_map(wpc, wbc, inode, folio, end_pos);
+	/*
+	 * Set the writeback bit ASAP, as the I/O completion for the single
+	 * block per folio case happen hit as soon as we're submitting the bio.
+	 */
+	folio_start_writeback(folio);
 
-redirty:
-	folio_redirty_for_writepage(wbc, folio);
-unlock:
+	/*
+	 * Walk through the folio to find dirty areas to write back.
+	 */
+	while ((rlen = iomap_find_dirty_range(folio, &pos, end_pos))) {
+		error = iomap_writepage_map_blocks(wpc, wbc, folio, inode,
+				pos, rlen, &count);
+		if (error)
+			break;
+		pos += rlen;
+	}
+
+	if (count)
+		wpc->nr_folios++;
+
+	/*
+	 * We can have dirty bits set past end of file in page_mkwrite path
+	 * while mapping the last partial folio. Hence it's better to clear
+	 * all the dirty bits in the folio here.
+	 */
+	iomap_clear_range_dirty(folio, 0, folio_size(folio));
+
+	/*
+	 * Usually the writeback bit is cleared by the I/O completion handler.
+	 * But we may end up either not actually writing any blocks, or (when
+	 * there are multiple blocks in a folio) all I/O might have finished
+	 * already at this point.  In that case we need to clear the writeback
+	 * bit ourselves right after unlocking the page.
+	 */
 	folio_unlock(folio);
-	return 0;
+	if (ifs) {
+		if (atomic_dec_and_test(&ifs->write_bytes_pending))
+			folio_end_writeback(folio);
+	} else {
+		if (!count)
+			folio_end_writeback(folio);
+	}
+	mapping_set_error(inode->i_mapping, error);
+	return error;
+}
+
+static int iomap_do_writepage(struct folio *folio,
+		struct writeback_control *wbc, void *data)
+{
+	return iomap_writepage_map(data, wbc, folio);
 }
 
 int
@@ -1988,18 +1971,24 @@ iomap_writepages(struct address_space *mapping, struct writeback_control *wbc,
 {
 	int			ret;
 
+	/*
+	 * Writeback from reclaim context should never happen except in the case
+	 * of a VM regression so warn about it and refuse to write the data.
+	 */
+	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC | PF_KSWAPD)) ==
+			PF_MEMALLOC))
+		return -EIO;
+
 	wpc->ops = ops;
 	ret = write_cache_pages(mapping, wbc, iomap_do_writepage, wpc);
-	if (!wpc->ioend)
-		return ret;
-	return iomap_submit_ioend(wpc, wpc->ioend, ret);
+	return iomap_submit_ioend(wpc, ret);
 }
 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);
diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
index bcd3f8cf5ea4..f3b43d223a46 100644
--- a/fs/iomap/direct-io.c
+++ b/fs/iomap/direct-io.c
@@ -380,6 +380,7 @@ static loff_t iomap_dio_bio_iter(const struct iomap_iter *iter,
 		fscrypt_set_bio_crypt_ctx(bio, inode, pos >> inode->i_blkbits,
 					  GFP_KERNEL);
 		bio->bi_iter.bi_sector = iomap_sector(iomap, pos);
+		bio->bi_write_hint = inode->i_write_hint;
 		bio->bi_ioprio = dio->iocb->ki_ioprio;
 		bio->bi_private = dio;
 		bio->bi_end_io = iomap_dio_bio_end_io;
diff --git a/fs/iomap/trace.h b/fs/iomap/trace.h
index c16fd55f5595..0a991c4ce87d 100644
--- a/fs/iomap/trace.h
+++ b/fs/iomap/trace.h
@@ -154,7 +154,48 @@ DEFINE_EVENT(iomap_class, name,	\
 	TP_ARGS(inode, iomap))
 DEFINE_IOMAP_EVENT(iomap_iter_dstmap);
 DEFINE_IOMAP_EVENT(iomap_iter_srcmap);
-DEFINE_IOMAP_EVENT(iomap_writepage_map);
+
+TRACE_EVENT(iomap_writepage_map,
+	TP_PROTO(struct inode *inode, u64 pos, unsigned int dirty_len,
+		 struct iomap *iomap),
+	TP_ARGS(inode, pos, dirty_len, iomap),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(u64, ino)
+		__field(u64, pos)
+		__field(u64, dirty_len)
+		__field(u64, addr)
+		__field(loff_t, offset)
+		__field(u64, length)
+		__field(u16, type)
+		__field(u16, flags)
+		__field(dev_t, bdev)
+	),
+	TP_fast_assign(
+		__entry->dev = inode->i_sb->s_dev;
+		__entry->ino = inode->i_ino;
+		__entry->pos = pos;
+		__entry->dirty_len = dirty_len;
+		__entry->addr = iomap->addr;
+		__entry->offset = iomap->offset;
+		__entry->length = iomap->length;
+		__entry->type = iomap->type;
+		__entry->flags = iomap->flags;
+		__entry->bdev = iomap->bdev ? iomap->bdev->bd_dev : 0;
+	),
+	TP_printk("dev %d:%d ino 0x%llx bdev %d:%d pos 0x%llx dirty len 0x%llx "
+		  "addr 0x%llx offset 0x%llx length 0x%llx type %s flags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  MAJOR(__entry->bdev), MINOR(__entry->bdev),
+		  __entry->pos,
+		  __entry->dirty_len,
+		  __entry->addr,
+		  __entry->offset,
+		  __entry->length,
+		  __print_symbolic(__entry->type, IOMAP_TYPE_STRINGS),
+		  __print_flags(__entry->flags, "|", IOMAP_F_FLAGS_STRINGS))
+);
 
 TRACE_EVENT(iomap_iter,
 	TP_PROTO(struct iomap_iter *iter, const void *ops,
@@ -165,6 +206,7 @@ TRACE_EVENT(iomap_iter,
 		__field(u64, ino)
 		__field(loff_t, pos)
 		__field(u64, length)
+		__field(s64, processed)
 		__field(unsigned int, flags)
 		__field(const void *, ops)
 		__field(unsigned long, caller)
@@ -174,15 +216,17 @@ TRACE_EVENT(iomap_iter,
 		__entry->ino = iter->inode->i_ino;
 		__entry->pos = iter->pos;
 		__entry->length = iomap_length(iter);
+		__entry->processed = iter->processed;
 		__entry->flags = iter->flags;
 		__entry->ops = ops;
 		__entry->caller = caller;
 	),
-	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%llx flags %s (0x%x) ops %ps caller %pS",
+	TP_printk("dev %d:%d ino 0x%llx pos 0x%llx length 0x%llx processed %lld flags %s (0x%x) ops %ps caller %pS",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		   __entry->ino,
 		   __entry->pos,
 		   __entry->length,
+		   __entry->processed,
 		   __print_flags(__entry->flags, "|", IOMAP_FLAGS_STRINGS),
 		   __entry->flags,
 		   __entry->ops,
diff --git a/fs/mpage.c b/fs/mpage.c
index 738882e0766d..fa8b99a199fa 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -605,6 +605,7 @@ alloc_new:
 				GFP_NOFS);
 		bio->bi_iter.bi_sector = first_block << (blkbits - 9);
 		wbc_init_bio(wbc, bio);
+		bio->bi_write_hint = inode->i_write_hint;
 	}
 
 	/*
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 813f85156b0c..1698507d1ac7 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -112,7 +112,7 @@ xfs_end_ioend(
 	 * longer dirty. If we don't remove delalloc blocks here, they become
 	 * stale and can corrupt free space accounting on unmount.
 	 */
-	error = blk_status_to_errno(ioend->io_bio->bi_status);
+	error = blk_status_to_errno(ioend->io_bio.bi_status);
 	if (unlikely(error)) {
 		if (ioend->io_flags & IOMAP_F_SHARED) {
 			xfs_reflink_cancel_cow_range(ip, offset, size, true);
@@ -179,7 +179,7 @@ STATIC void
 xfs_end_bio(
 	struct bio		*bio)
 {
-	struct iomap_ioend	*ioend = bio->bi_private;
+	struct iomap_ioend	*ioend = iomap_ioend_from_bio(bio);
 	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
 	unsigned long		flags;
 
@@ -276,7 +276,8 @@ static int
 xfs_map_blocks(
 	struct iomap_writepage_ctx *wpc,
 	struct inode		*inode,
-	loff_t			offset)
+	loff_t			offset,
+	unsigned int		len)
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
@@ -444,7 +445,7 @@ xfs_prepare_ioend(
 	/* send ioends that might require a transaction to the completion wq */
 	if (xfs_ioend_is_append(ioend) || ioend->io_type == IOMAP_UNWRITTEN ||
 	    (ioend->io_flags & IOMAP_F_SHARED))
-		ioend->io_bio->bi_end_io = xfs_end_bio;
+		ioend->io_bio.bi_end_io = xfs_end_bio;
 	return status;
 }
 
diff --git a/fs/zonefs/file.c b/fs/zonefs/file.c
index dba5dcb62bef..3b103715acc9 100644
--- a/fs/zonefs/file.c
+++ b/fs/zonefs/file.c
@@ -125,7 +125,8 @@ static void zonefs_readahead(struct readahead_control *rac)
  * which implies that the page range can only be within the fixed inode size.
  */
 static int zonefs_write_map_blocks(struct iomap_writepage_ctx *wpc,
-				   struct inode *inode, loff_t offset)
+				   struct inode *inode, loff_t offset,
+				   unsigned int len)
 {
 	struct zonefs_zone *z = zonefs_inode_zone(inode);