[PATCH v4 0/2] add FALLOC_FL_WRITE_ZEROES support to xfs

[PATCH v4 0/2] add FALLOC_FL_WRITE_ZEROES support to xfs

[Pankaj Raghav]

The benefits of FALLOC_FL_WRITE_ZEROES was already discussed as a part
of Zhang Yi's initial patches[1]. Postgres developer Andres also
mentioned they would like to use this feature in Postgres [2].

This series has the changes proposed by Dave [3].

I tested the changes with fsstress and fsx based on the xfstests patch I
sent recently to test this flag[4]. generic/363 helped me debug the
crash I noticed when I did the initial implementation[3].

Changes since v3:
- Introduce xfs_bmap_alloc_or_convert_range() in xfs_iomap.c for easy
  review experience (christoph)
- Add extsz hint and rt support in xfs_bmap_alloc_or_convert_range()

Changes since v2:
- Add allow_write_zeroes to xfs_global so that we can enable this
  feature independent of the HW underneath.

Changes since v1 [5.1 5.2]:
- Added a new function xfs_bmap_alloc_or_convert_range() based on Dave's
  feedback.
- Changed the xfs_falloc_write_zeroes to use
  xfs_bmap_alloc_or_convert_range() instead of doing prealloc and
  convert approach.

[1] https://lore.kernel.org/linux-fsdevel/20250619111806.3546162-1-yi.zhang@huaweicloud.com/
[2] https://lore.kernel.org/linux-fsdevel/20260217055103.GA6174@lst.de/T/#m7935b9bab32bb5ff372507f84803b8753ad1c814
[3] https://lore.kernel.org/linux-xfs/6i2jvzn3lyugjlbgmjzpped3gogzyqv5mpe2uqaifz4vjpaega@pomzoq7ley77/
[4] https://lore.kernel.org/linux-xfs/20260312195308.738189-1-p.raghav@samsung.com/
[5.1] https://lore.kernel.org/linux-xfs/20260309180708.427553-2-lukas@herbolt.com/
[5.2] https://lore.kernel.org/linux-xfs/abC1LvRElctaHPe5@dread/

Pankaj Raghav (2):
  xfs: add xfs_bmap_alloc_or_convert_range function
  xfs: add support for FALLOC_FL_WRITE_ZEROES

 fs/xfs/xfs_file.c  |  56 +++++++++++-
 fs/xfs/xfs_iomap.c | 215 ++++++++++++++++++++++++++++++++-------------
 fs/xfs/xfs_iomap.h |   2 +
 3 files changed, 210 insertions(+), 63 deletions(-)


base-commit: ad345c96913e1a5b507cea4e49d22e91d5947f56
-- 
2.51.2

[PATCH v4 1/2] xfs: add xfs_bmap_alloc_or_convert_range function

[Pankaj Raghav]

Add xfs_bmap_alloc_or_convert_range() that can either allocate a range
and/or convert unwritten extents to written extents.

This function is based on xfs_iomap_write_unwritten() but we add an
extra flag parameter. Only XFS_BMAPI_CONVERT and/or XFS_BMAPI_ZERO is
accepted as flags to this function. This function also additionally
accounts while starting the transaction for the blocks that might
be created because of XFS_BMAPI_ZERO.

This is done as a preparation to add FALLOC_FL_WRITE_ZEROES flag.

xfs_iomap_write_unwritten() function will now just call
xfs_bmap_alloc_or_convert_range() with flag set to XFS_BMAPI_CONVERT.

Suggested-by: Dave Chinner <dgc@kernel.org>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
I feel like xfs_bmap_alloc_or_convert_range() has become a bit big and
logic repeated, should we rethink of going back to the original solution? [1]

[1] https://lore.kernel.org/linux-xfs/20260310194245.848034-2-lukas@herbolt.com/

 fs/xfs/xfs_iomap.c | 215 ++++++++++++++++++++++++++++++++-------------
 fs/xfs/xfs_iomap.h |   2 +
 2 files changed, 155 insertions(+), 62 deletions(-)

diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index f20a02f49ed9..40715589814a 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -609,12 +609,30 @@ xfs_iomap_prealloc_size(
 	return alloc_blocks;
 }
 
+/*
+ * This function is used to allocate written extents over holes and/or convert
+ * unwritten extents to written extents based on the @flags passed to it.
+ *
+ * If @flags is zero, it will allocate written extents for holes and delalloc
+ * extents across the range.
+ *
+ * If XFS_BMAPI_CONVERT is specified in @flags, then it will also do conversion
+ * of unwritten extents in the range to written extents.
+ *
+ * If XFS_BMAPI_ZERO is specified in @flags, then both newly allocated extents
+ * and converted unwritten extents will be initialised to contain zeroes.
+ *
+ * If @update_isize is true, then if the range we are operating on extends
+ * beyond the current EOF, extend i_size to offset+len incrementally as extents
+ * in the range are allocated/converted.
+ */
 int
-xfs_iomap_write_unwritten(
-	xfs_inode_t	*ip,
-	xfs_off_t	offset,
-	xfs_off_t	count,
-	bool		update_isize)
+xfs_bmap_alloc_or_convert_range(
+	struct xfs_inode	*ip,
+	xfs_off_t		offset,
+	xfs_off_t		count,
+	uint32_t		flags,
+	bool			update_isize)
 {
 	xfs_mount_t	*mp = ip->i_mount;
 	xfs_fileoff_t	offset_fsb;
@@ -625,98 +643,159 @@ xfs_iomap_write_unwritten(
 	xfs_bmbt_irec_t imap;
 	struct inode	*inode = VFS_I(ip);
 	xfs_fsize_t	i_size;
-	uint		resblks;
 	int		error;
+	int rt = XFS_IS_REALTIME_INODE(ip);
+	xfs_extlen_t	extsz, temp;
 
-	trace_xfs_unwritten_convert(ip, offset, count);
+	ASSERT((flags & ~(XFS_BMAPI_ZERO | XFS_BMAPI_CONVERT)) == 0);
 
+	extsz = xfs_get_extsz_hint(ip);
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
 	count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
 	count_fsb = (xfs_filblks_t)(count_fsb - offset_fsb);
 
-	/*
-	 * Reserve enough blocks in this transaction for two complete extent
-	 * btree splits.  We may be converting the middle part of an unwritten
-	 * extent and in this case we will insert two new extents in the btree
-	 * each of which could cause a full split.
-	 *
-	 * This reservation amount will be used in the first call to
-	 * xfs_bmbt_split() to select an AG with enough space to satisfy the
-	 * rest of the operation.
-	 */
-	resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1;
-
 	/* Attach dquots so that bmbt splits are accounted correctly. */
 	error = xfs_qm_dqattach(ip);
 	if (error)
 		return error;
 
 	do {
+
+		uint	resblks, dblocks, bmapi_total, rblocks;
+		int	bmapi_error;
+
+		if (flags == XFS_BMAPI_CONVERT) {
+			/*
+			 * Reserve enough blocks in this transaction for two
+			 * complete extent btree splits.  We may be converting
+			 * the middle part of an unwritten extent and in this
+			 * case we will insert two new extents in the btree each
+			 * of which could cause a full split.
+			 *
+			 * This reservation amount will be used in the first call
+			 * to xfs_bmbt_split() to select an AG with enough space
+			 * to satisfy the rest of the operation.
+			 */
+			resblks = 0;
+			dblocks = XFS_DIOSTRAT_SPACE_RES(mp, resblks) << 1;
+			rblocks = 0;
+			bmapi_total = dblocks;
+		} else {
+			xfs_fileoff_t	s, e;
+			if (unlikely(extsz)) {
+				s = offset_fsb;
+				do_div(s, extsz);
+				s *= extsz;
+				e = offset_fsb + count_fsb;
+				div_u64_rem(offset_fsb, extsz, &temp);
+				if (temp)
+					e += temp;
+				div_u64_rem(e, extsz, &temp);
+				if (temp)
+					e += extsz - temp;
+			} else {
+				s = 0;
+				e = count_fsb;
+			}
+
+			/*
+			 * We might allocate data blocks (needs resblks + 1 split)
+			 * or convert an unwritten extent (needs 0 data blocks +
+			 * 2 splits). Ensure we have enough block reservation for
+			 * the worst case.
+			 */
+			resblks = XFS_FILBLKS_MIN((e - s), XFS_MAX_BMBT_EXTLEN);
+			bmapi_total = 0;
+			if (unlikely(rt)) {
+				rblocks = XFS_DIOSTRAT_SPACE_RES(mp, resblks);
+				dblocks = XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1;
+			} else {
+				rblocks = 0;
+				dblocks = XFS_DIOSTRAT_SPACE_RES(mp, resblks);
+				dblocks = max(dblocks,
+					XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1);
+			}
+		}
+
 		/*
-		 * Set up a transaction to convert the range of extents
-		 * from unwritten to real. Do allocations in a loop until
-		 * we have covered the range passed in.
+		 * Set up a transaction to convert the range of extents based on
+		 * the flags. Do allocations in a loop until we have covered the
+		 * range passed in.
 		 *
 		 * Note that we can't risk to recursing back into the filesystem
 		 * here as we might be asked to write out the same inode that we
 		 * complete here and might deadlock on the iolock.
 		 */
-		error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write, resblks,
-				0, true, &tp);
+		error = xfs_trans_alloc_inode(ip, &M_RES(mp)->tr_write, dblocks,
+				rblocks, true, &tp);
 		if (error)
 			return error;
 
-		error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
-				XFS_IEXT_WRITE_UNWRITTEN_CNT);
+		if (flags & XFS_BMAPI_CONVERT)
+			error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
+					XFS_IEXT_WRITE_UNWRITTEN_CNT);
+		else
+			error = xfs_iext_count_extend(tp, ip, XFS_DATA_FORK,
+					XFS_IEXT_ADD_NOSPLIT_CNT);
+
 		if (error)
 			goto error_on_bmapi_transaction;
 
-		/*
-		 * Modify the unwritten extent state of the buffer.
-		 */
 		nimaps = 1;
 		error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb,
-					XFS_BMAPI_CONVERT, resblks, &imap,
-					&nimaps);
-		if (error)
-			goto error_on_bmapi_transaction;
+					flags, bmapi_total, &imap, &nimaps);
+		bmapi_error = error;
+
+		if (error) {
+			if (error != -ENOSR)
+				goto error_on_bmapi_transaction;
+			/*
+			 * Keep searching until we get one contiguous
+			 * extent if we get ENOSR
+			 */
+			error = 0;
+		} else {
+			/*
+			 * Log the updated inode size as we go. We have to be
+			 * careful to only log it up to the actual write offset
+			 * if it is halfway into a block.
+			 */
+			i_size = XFS_FSB_TO_B(mp, offset_fsb + count_fsb);
+			if (i_size > offset + count)
+				i_size = offset + count;
+			if (update_isize && i_size > i_size_read(inode))
+				i_size_write(inode, i_size);
+			i_size = xfs_new_eof(ip, i_size);
+			if (i_size) {
+				ip->i_disk_size = i_size;
+				xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+			}
 
-		/*
-		 * Log the updated inode size as we go.  We have to be careful
-		 * to only log it up to the actual write offset if it is
-		 * halfway into a block.
-		 */
-		i_size = XFS_FSB_TO_B(mp, offset_fsb + count_fsb);
-		if (i_size > offset + count)
-			i_size = offset + count;
-		if (update_isize && i_size > i_size_read(inode))
-			i_size_write(inode, i_size);
-		i_size = xfs_new_eof(ip, i_size);
-		if (i_size) {
-			ip->i_disk_size = i_size;
-			xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 		}
 
 		error = xfs_trans_commit(tp);
 		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		if (error)
-			return error;
 
-		if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock))) {
-			xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
-			return xfs_alert_fsblock_zero(ip, &imap);
+		if (!bmapi_error) {
+			if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock))) {
+				xfs_bmap_mark_sick(ip, XFS_DATA_FORK);
+				return xfs_alert_fsblock_zero(ip, &imap);
+			}
+			if ((numblks_fsb = imap.br_blockcount) == 0) {
+				/*
+				 * The numblks_fsb value should always get smaller,
+				 * otherwise the loop is stuck.
+				 */
+				ASSERT(imap.br_blockcount);
+				break;
+			}
+			offset_fsb += numblks_fsb;
+			count_fsb -= numblks_fsb;
 		}
 
-		if ((numblks_fsb = imap.br_blockcount) == 0) {
-			/*
-			 * The numblks_fsb value should always get
-			 * smaller, otherwise the loop is stuck.
-			 */
-			ASSERT(imap.br_blockcount);
-			break;
-		}
-		offset_fsb += numblks_fsb;
-		count_fsb -= numblks_fsb;
+		if (error)
+			return error;
+
 	} while (count_fsb > 0);
 
 	return 0;
@@ -727,6 +806,18 @@ xfs_iomap_write_unwritten(
 	return error;
 }
 
+int
+xfs_iomap_write_unwritten(
+	xfs_inode_t	*ip,
+	xfs_off_t	offset,
+	xfs_off_t	count,
+	bool		update_isize)
+{
+	trace_xfs_unwritten_convert(ip, offset, count);
+	return xfs_bmap_alloc_or_convert_range(ip, offset, count,
+					XFS_BMAPI_CONVERT, update_isize);
+}
+
 static inline bool
 imap_needs_alloc(
 	struct inode		*inode,
diff --git a/fs/xfs/xfs_iomap.h b/fs/xfs/xfs_iomap.h
index ebcce7d49446..ba7b06074539 100644
--- a/fs/xfs/xfs_iomap.h
+++ b/fs/xfs/xfs_iomap.h
@@ -16,6 +16,8 @@ int xfs_iomap_write_direct(struct xfs_inode *ip, xfs_fileoff_t offset_fsb,
 		xfs_fileoff_t count_fsb, unsigned int flags,
 		struct xfs_bmbt_irec *imap, u64 *sequence);
 int xfs_iomap_write_unwritten(struct xfs_inode *, xfs_off_t, xfs_off_t, bool);
+int xfs_bmap_alloc_or_convert_range(struct xfs_inode *, xfs_off_t, xfs_off_t,
+		uint32_t, bool);
 xfs_fileoff_t xfs_iomap_eof_align_last_fsb(struct xfs_inode *ip,
 		xfs_fileoff_t end_fsb);
 
-- 
2.51.2

[PATCH v4 2/2] xfs: add support for FALLOC_FL_WRITE_ZEROES

[Pankaj Raghav]

If the underlying block device supports the unmap write zeroes
operation, this flag allows users to quickly preallocate a file with
written extents that contain zeroes. This is beneficial for subsequent
overwrites as it prevents the need for unwritten-to-written extent
conversions, thereby significantly reducing metadata updates and journal
I/O overhead, improving overwrite performance.

Co-developed-by: Lukas Herbolt <lukas@herbolt.com>
Signed-off-by: Lukas Herbolt <lukas@herbolt.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/xfs/xfs_file.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 55 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 845a97c9b063..99a02982154a 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1368,6 +1368,57 @@ xfs_falloc_force_zero(
 	return XFS_TEST_ERROR(ip->i_mount, XFS_ERRTAG_FORCE_ZERO_RANGE);
 }
 
+static int
+xfs_falloc_write_zeroes(
+	struct file		*file,
+	int			mode,
+	loff_t			offset,
+	loff_t			len,
+	struct xfs_zone_alloc_ctx *ac)
+{
+	struct inode		*inode = file_inode(file);
+	struct xfs_inode	*ip = XFS_I(inode);
+	loff_t			new_size = 0;
+	loff_t			old_size = XFS_ISIZE(ip);
+	int			error;
+	unsigned int		blksize = i_blocksize(inode);
+	loff_t			offset_aligned = round_down(offset, blksize);
+	bool			did_zero;
+
+	if (xfs_is_always_cow_inode(ip) ||
+	    !bdev_write_zeroes_unmap_sectors(
+		    xfs_inode_buftarg(XFS_I(inode))->bt_bdev))
+		return -EOPNOTSUPP;
+
+	error = xfs_falloc_newsize(file, mode, offset, len, &new_size);
+	if (error)
+		return error;
+
+	error = xfs_free_file_space(ip, offset, len, ac);
+	if (error)
+		return error;
+
+	/*
+	 * Zero the tail of the old EOF block and any space up to the new
+	 * offset.
+	 * In the usual truncate path, xfs_falloc_setsize takes care of
+	 * zeroing those blocks.
+	 */
+	if (offset_aligned > old_size)
+		error = xfs_zero_range(ip, old_size, offset_aligned - old_size,
+				NULL, &did_zero);
+	if (error)
+		return error;
+
+	error = xfs_bmap_alloc_or_convert_range(ip, offset, len,
+			XFS_BMAPI_CONVERT | XFS_BMAPI_ZERO,
+			new_size ? true : false);
+	if (error)
+		return error;
+
+	return error;
+}
+
 /*
  * Punch a hole and prealloc the range.  We use a hole punch rather than
  * unwritten extent conversion for two reasons:
@@ -1470,7 +1521,7 @@ xfs_falloc_allocate_range(
 		(FALLOC_FL_ALLOCATE_RANGE | FALLOC_FL_KEEP_SIZE |	\
 		 FALLOC_FL_PUNCH_HOLE |	FALLOC_FL_COLLAPSE_RANGE |	\
 		 FALLOC_FL_ZERO_RANGE |	FALLOC_FL_INSERT_RANGE |	\
-		 FALLOC_FL_UNSHARE_RANGE)
+		 FALLOC_FL_UNSHARE_RANGE | FALLOC_FL_WRITE_ZEROES)
 
 STATIC long
 __xfs_file_fallocate(
@@ -1522,6 +1573,9 @@ __xfs_file_fallocate(
 	case FALLOC_FL_ALLOCATE_RANGE:
 		error = xfs_falloc_allocate_range(file, mode, offset, len);
 		break;
+	case FALLOC_FL_WRITE_ZEROES:
+		error = xfs_falloc_write_zeroes(file, mode, offset, len, ac);
+		break;
 	default:
 		error = -EOPNOTSUPP;
 		break;
-- 
2.51.2

Re: [PATCH v4 1/2] xfs: add xfs_bmap_alloc_or_convert_range function

[Christoph Hellwig]

> I feel like xfs_bmap_alloc_or_convert_range() has become a bit big and
> logic repeated, should we rethink of going back to the original solution? [1]
> 
> [1] https://lore.kernel.org/linux-xfs/20260310194245.848034-2-lukas@herbolt.com/

Looking at the changes here there is almost no shared logic between
the two uses, which doesn't make core "sharing" look useful when there
is nothing left to share.