[PATCH v4 0/5] btrfs: defrag: don't waste CPU time on non-target extent

[PATCH v4 0/5] btrfs: defrag: don't waste CPU time on non-target extent

[Qu Wenruo]

Changelog:
v2:
- Rebased to lastest misc-next
  Just one small conflict with static_assert() update.
  And this time only those patches are rebased to misc-next, thus it may
  cause conflicts with fixes for defrag_check_next_extent() in the
  future.

- Several grammar fixes

- Report accurate btrfs_defrag_ctrl::sectors_defragged
  This is inspired by a comment from Filipe that the skip check
  should be done in the defrag_collect_targets() call inside
  defrag_one_range().

  This results a new patch in v2.

- Change the timing of btrfs_defrag_ctrl::last_scanned update
  Now it's updated inside defrag_one_range(), which will give
  us an accurate view, unlike the previous call site in
  defrag_one_cluster().

- Don't change the timing of extent threshold.

- Rename @last_target to @last_is_target in defrag_collect_targets()

v3:
- Add Reviewed-by tags

- Fix a wrong value in commit message of the 1st patch

- Make @orig_start const for the 3rd patch

- Fix a missing word "skip" in the 5th patch

- Remove one unnecessary assignment in the 5th patch
  As we don't return the defragged sectors to user space.

v4:
- Move the skip behavior before the btrfs_defrag_ctrl refactor
  So it can be backported to v5.16

- Slightly change how we use btrfs_defrag_ctrl::last_scanned
  Previously we use that value almost as range start, but that's not correct,
  as defrag_collect_targets() can update it before we do any defrag.

  Now we pass explicit range start/len to defrag_one_cluster(), without
  re-using ctrl::last_scanned.

- Remove one incorrect range skip in defrag_one_cluster()
  There used to be a check against last_scanned to skip certain ranges.
  But that is not correct, as the last_scanned can already be larger
  than the first target after the 1st defrag_collect_targets() call.

  Now last_scanned is only really utilized after one cluster is
  defragged.

  This also removes the reviewed-by tag for the refactor patch 

Qu Wenruo (5):
  btrfs: defrag: allow defrag_one_cluster() to skip large extent which
    is not a target
  btrfs: uapi: introduce BTRFS_DEFRAG_RANGE_MASK for later sanity check
  btrfs: defrag: introduce btrfs_defrag_ctrl structure for later usage
  btrfs: defrag: use btrfs_defrag_ctrl to replace
    btrfs_ioctl_defrag_range_args for btrfs_defrag_file()
  btrfs: defrag: make btrfs_defrag_file() to report accurate number of
    defragged sectors

 fs/btrfs/ctree.h           |  22 +++-
 fs/btrfs/file.c            |  21 ++--
 fs/btrfs/ioctl.c           | 200 ++++++++++++++++++++++---------------
 include/uapi/linux/btrfs.h |   6 +-
 4 files changed, 155 insertions(+), 94 deletions(-)

-- 
2.35.0

[PATCH v4 1/5] btrfs: defrag: allow defrag_one_cluster() to skip large extent which is not a target

[Qu Wenruo]

In the rework of btrfs_defrag_file(), we always call
defrag_one_cluster() and increase the offset by cluster size, which is
only 256K.

But there are cases where we have a large extent (e.g. 128M) which
doesn't need to be defragged at all.

Before the refactor, we can directly skip the range, but now we have to
scan that extent map again and again until the cluster moves after the
non-target extent.

Fix the problem by allow defrag_one_cluster() to increase
btrfs_defrag_ctrl::last_scanned to the end of an extent, if and only if
the last extent of the cluster is not a target.

The test script looks like this:

	mkfs.btrfs -f $dev > /dev/null

	mount $dev $mnt

	# As btrfs ioctl uses 32M as extent_threshold
	xfs_io -f -c "pwrite 0 64M" $mnt/file1
	sync
	# Some fragemented range to defrag
	xfs_io -s -c "pwrite 65548k 4k" \
		  -c "pwrite 65544k 4k" \
		  -c "pwrite 65540k 4k" \
		  -c "pwrite 65536k 4k" \
		  $mnt/file1
	sync

	echo "=== before ==="
	xfs_io -c "fiemap -v" $mnt/file1
	echo "=== after ==="
	btrfs fi defrag $mnt/file1
	sync
	xfs_io -c "fiemap -v" $mnt/file1
	umount $mnt

With extra ftrace put into defrag_one_cluster(), before the patch it
would result tons of loops:

(As defrag_one_cluster() is inlined, the function name is its caller)

  btrfs-126062  [005] .....  4682.816026: btrfs_defrag_file: r/i=5/257 start=0 len=262144
  btrfs-126062  [005] .....  4682.816027: btrfs_defrag_file: r/i=5/257 start=262144 len=262144
  btrfs-126062  [005] .....  4682.816028: btrfs_defrag_file: r/i=5/257 start=524288 len=262144
  btrfs-126062  [005] .....  4682.816028: btrfs_defrag_file: r/i=5/257 start=786432 len=262144
  btrfs-126062  [005] .....  4682.816028: btrfs_defrag_file: r/i=5/257 start=1048576 len=262144
  ...
  btrfs-126062  [005] .....  4682.816043: btrfs_defrag_file: r/i=5/257 start=67108864 len=262144

But with this patch there will be just one loop, then directly to the
end of the extent:

  btrfs-130471  [014] .....  5434.029558: defrag_one_cluster: r/i=5/257 start=0 len=262144
  btrfs-130471  [014] .....  5434.029559: defrag_one_cluster: r/i=5/257 start=67108864 len=16384

Cc: stable@vger.kernel.org # 5.16
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
---
 fs/btrfs/ioctl.c | 50 ++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 40 insertions(+), 10 deletions(-)

diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index e7a284239393..fa4d29026275 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1204,9 +1204,11 @@ struct defrag_target_range {
  */
 static int defrag_collect_targets(struct btrfs_inode *inode,
 				  u64 start, u64 len, u32 extent_thresh,
-				  u64 newer_than, bool do_compress,
-				  bool locked, struct list_head *target_list)
+				  u64 newer_than, bool do_compress, bool locked,
+				  struct list_head *target_list,
+				  u64 *last_scanned_ret)
 {
+	bool last_is_target = false;
 	u64 cur = start;
 	int ret = 0;
 
@@ -1216,6 +1218,7 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 		bool next_mergeable = true;
 		u64 range_len;
 
+		last_is_target = false;
 		em = defrag_lookup_extent(&inode->vfs_inode, cur, locked);
 		if (!em)
 			break;
@@ -1298,6 +1301,7 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 		}
 
 add:
+		last_is_target = true;
 		range_len = min(extent_map_end(em), start + len) - cur;
 		/*
 		 * This one is a good target, check if it can be merged into
@@ -1341,6 +1345,17 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 			kfree(entry);
 		}
 	}
+	if (!ret && last_scanned_ret) {
+		/*
+		 * If the last extent is not a target, the caller can skip to
+		 * the end of that extent.
+		 * Otherwise, we can only go the end of the spcified range.
+		 */
+		if (!last_is_target)
+			*last_scanned_ret = max(cur, *last_scanned_ret);
+		else
+			*last_scanned_ret = max(start + len, *last_scanned_ret);
+	}
 	return ret;
 }
 
@@ -1400,7 +1415,8 @@ static int defrag_one_locked_target(struct btrfs_inode *inode,
 }
 
 static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
-			    u32 extent_thresh, u64 newer_than, bool do_compress)
+			    u32 extent_thresh, u64 newer_than, bool do_compress,
+			    u64 *last_scanned_ret)
 {
 	struct extent_state *cached_state = NULL;
 	struct defrag_target_range *entry;
@@ -1446,7 +1462,7 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 	 */
 	ret = defrag_collect_targets(inode, start, len, extent_thresh,
 				     newer_than, do_compress, true,
-				     &target_list);
+				     &target_list, last_scanned_ret);
 	if (ret < 0)
 		goto unlock_extent;
 
@@ -1481,7 +1497,8 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 			      u64 start, u32 len, u32 extent_thresh,
 			      u64 newer_than, bool do_compress,
 			      unsigned long *sectors_defragged,
-			      unsigned long max_sectors)
+			      unsigned long max_sectors,
+			      u64 *last_scanned_ret)
 {
 	const u32 sectorsize = inode->root->fs_info->sectorsize;
 	struct defrag_target_range *entry;
@@ -1491,7 +1508,7 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 
 	ret = defrag_collect_targets(inode, start, len, extent_thresh,
 				     newer_than, do_compress, false,
-				     &target_list);
+				     &target_list, NULL);
 	if (ret < 0)
 		goto out;
 
@@ -1508,6 +1525,15 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 			range_len = min_t(u32, range_len,
 				(max_sectors - *sectors_defragged) * sectorsize);
 
+		/*
+		 * If defrag_one_range() has updated last_scanned_ret,
+		 * our range may already be invalid (e.g. hole punched).
+		 * Skip if our range is before last_scanned_ret, as there is
+		 * no need to defrag the range anymore.
+		 */
+		if (entry->start + range_len <= *last_scanned_ret)
+			continue;
+
 		if (ra)
 			page_cache_sync_readahead(inode->vfs_inode.i_mapping,
 				ra, NULL, entry->start >> PAGE_SHIFT,
@@ -1520,7 +1546,8 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 		 * accounting.
 		 */
 		ret = defrag_one_range(inode, entry->start, range_len,
-				       extent_thresh, newer_than, do_compress);
+				       extent_thresh, newer_than, do_compress,
+				       last_scanned_ret);
 		if (ret < 0)
 			break;
 		*sectors_defragged += range_len >>
@@ -1531,6 +1558,8 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 		list_del_init(&entry->list);
 		kfree(entry);
 	}
+	if (ret >= 0)
+		*last_scanned_ret = max(*last_scanned_ret, start + len);
 	return ret;
 }
 
@@ -1616,6 +1645,7 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 
 	while (cur < last_byte) {
 		const unsigned long prev_sectors_defragged = sectors_defragged;
+		u64 last_scanned = cur;
 		u64 cluster_end;
 
 		if (btrfs_defrag_cancelled(fs_info)) {
@@ -1642,8 +1672,8 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 			BTRFS_I(inode)->defrag_compress = compress_type;
 		ret = defrag_one_cluster(BTRFS_I(inode), ra, cur,
 				cluster_end + 1 - cur, extent_thresh,
-				newer_than, do_compress,
-				&sectors_defragged, max_to_defrag);
+				newer_than, do_compress, &sectors_defragged,
+				max_to_defrag, &last_scanned);
 
 		if (sectors_defragged > prev_sectors_defragged)
 			balance_dirty_pages_ratelimited(inode->i_mapping);
@@ -1651,7 +1681,7 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 		btrfs_inode_unlock(inode, 0);
 		if (ret < 0)
 			break;
-		cur = cluster_end + 1;
+		cur = max(cluster_end + 1, last_scanned);
 		if (ret > 0) {
 			ret = 0;
 			break;
-- 
2.35.0

[PATCH v4 2/5] btrfs: uapi: introduce BTRFS_DEFRAG_RANGE_MASK for later sanity check

[Qu Wenruo]

And since we're here, replace the hardcoded bit flags (1, 2) with
(1UL << 0) and (1UL << 1), respectively.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
---
 include/uapi/linux/btrfs.h | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h
index 1cb1a3860f1d..57332a529c4e 100644
--- a/include/uapi/linux/btrfs.h
+++ b/include/uapi/linux/btrfs.h
@@ -576,8 +576,10 @@ struct btrfs_ioctl_clone_range_args {
  * Used by:
  * struct btrfs_ioctl_defrag_range_args.flags
  */
-#define BTRFS_DEFRAG_RANGE_COMPRESS 1
-#define BTRFS_DEFRAG_RANGE_START_IO 2
+#define BTRFS_DEFRAG_RANGE_COMPRESS	(1UL << 0)
+#define BTRFS_DEFRAG_RANGE_START_IO	(1UL << 1)
+#define BTRFS_DEFRAG_RANGE_FLAGS_MASK	(BTRFS_DEFRAG_RANGE_COMPRESS |\
+					 BTRFS_DEFRAG_RANGE_START_IO)
 struct btrfs_ioctl_defrag_range_args {
 	/* start of the defrag operation */
 	__u64 start;
-- 
2.35.0

[PATCH v4 3/5] btrfs: defrag: introduce btrfs_defrag_ctrl structure for later usage

[Qu Wenruo]

Currently btrfs_defrag_file() accepts not only
btrfs_ioctl_defrag_range_args but also other parameters like @newer_than
and @max_sectors_to_defrag for extra policies.

Those extra values are hidden from defrag ioctl and even caused bugs in
the past due to different behaviors based on those extra values.

Here we introduce a new structure, btrfs_defrag_ctrl, to include:

- all members in btrfs_ioctl_defrag_range_args

- @max_sectors_to_defrag and @newer_than

- Extra values which callers of btrfs_defrag_file() may care
  Like @sectors_defragged and @last_scanned.

With the new structure, also introduce a new helper,
btrfs_defrag_ioctl_args_to_ctrl() to:

- Do extra sanity check on @compress and @flags

- Do range alignment when possible

- Set default values.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
---
 fs/btrfs/ctree.h | 19 +++++++++++++++++++
 fs/btrfs/ioctl.c | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 60 insertions(+)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 9f7a950b8a69..670622fddd62 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -3318,6 +3318,25 @@ int btrfs_fileattr_set(struct user_namespace *mnt_userns,
 int btrfs_ioctl_get_supported_features(void __user *arg);
 void btrfs_sync_inode_flags_to_i_flags(struct inode *inode);
 int __pure btrfs_is_empty_uuid(u8 *uuid);
+
+struct btrfs_defrag_ctrl {
+	/* Input, read-only fields */
+	u64	start;
+	u64	len;
+	u32	extent_thresh;
+	u64	newer_than;
+	u64	max_sectors_to_defrag;
+	u8	compress;
+	u8	flags;
+
+	/* Output fields */
+	u64	sectors_defragged;
+	u64	last_scanned;	/* Exclusive bytenr */
+};
+int btrfs_defrag_ioctl_args_to_ctrl(struct btrfs_fs_info *fs_info,
+				    struct btrfs_ioctl_defrag_range_args *args,
+				    struct btrfs_defrag_ctrl *ctrl,
+				    u64 max_sectors_to_defrag, u64 newer_than);
 int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 		      struct btrfs_ioctl_defrag_range_args *range,
 		      u64 newer_than, unsigned long max_to_defrag);
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index fa4d29026275..c6a971033336 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1563,6 +1563,47 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 	return ret;
 }
 
+/*
+ * Convert the old ioctl format to the new btrfs_defrag_ctrl structure.
+ *
+ * Will also do basic tasks like setting default values and sanity checks.
+ */
+int btrfs_defrag_ioctl_args_to_ctrl(struct btrfs_fs_info *fs_info,
+				    struct btrfs_ioctl_defrag_range_args *args,
+				    struct btrfs_defrag_ctrl *ctrl,
+				    u64 max_sectors_to_defrag, u64 newer_than)
+{
+	u64 range_end;
+
+	if (args->flags & ~BTRFS_DEFRAG_RANGE_FLAGS_MASK)
+		return -EOPNOTSUPP;
+	if (args->compress_type >= BTRFS_NR_COMPRESS_TYPES)
+		return -EOPNOTSUPP;
+
+	ctrl->start = round_down(args->start, fs_info->sectorsize);
+	/*
+	 * If @len does not overflow with @start nor is -1, align the length.
+	 * Otherwise set it to (u64)-1 so later btrfs_defrag_file() will
+	 * determine the length using isize.
+	 */
+	if (!check_add_overflow(args->start, args->len, &range_end) &&
+	    args->len != (u64)-1)
+		ctrl->len = round_up(range_end, fs_info->sectorsize) -
+			    ctrl->start;
+	else
+		ctrl->len = -1;
+	ctrl->flags = args->flags;
+	ctrl->compress = args->compress_type;
+	if (args->extent_thresh == 0)
+		ctrl->extent_thresh = SZ_256K;
+	else
+		ctrl->extent_thresh = args->extent_thresh;
+	ctrl->newer_than = newer_than;
+	ctrl->last_scanned = 0;
+	ctrl->sectors_defragged = 0;
+	return 0;
+}
+
 /*
  * Entry point to file defragmentation.
  *
-- 
2.35.0

[PATCH v4 4/5] btrfs: defrag: use btrfs_defrag_ctrl to replace btrfs_ioctl_defrag_range_args for btrfs_defrag_file()

[Qu Wenruo]

This brings the following benefits:

- No more strange range->start update to indicate last scanned bytenr
  We have btrfs_defrag_ctrl::last_scanned (exclusive) for it directly.

- No more return value to indicate defragged sectors
  Now btrfs_defrag_file() will just return 0 if no error happened.
  And btrfs_defrag_ctrl::sectors_defragged will show that value.

- Less parameters to carry around
  Now most defrag_* functions only need to fetch its policy parameters
  from btrfs_defrag_ctrl directly.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/ctree.h |   3 +-
 fs/btrfs/file.c  |  21 +++---
 fs/btrfs/ioctl.c | 165 ++++++++++++++++++++---------------------------
 3 files changed, 82 insertions(+), 107 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 670622fddd62..ac222a9ce166 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -3338,8 +3338,7 @@ int btrfs_defrag_ioctl_args_to_ctrl(struct btrfs_fs_info *fs_info,
 				    struct btrfs_defrag_ctrl *ctrl,
 				    u64 max_sectors_to_defrag, u64 newer_than);
 int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
-		      struct btrfs_ioctl_defrag_range_args *range,
-		      u64 newer_than, unsigned long max_to_defrag);
+		      struct btrfs_defrag_ctrl *ctrl);
 void btrfs_get_block_group_info(struct list_head *groups_list,
 				struct btrfs_ioctl_space_info *space);
 void btrfs_update_ioctl_balance_args(struct btrfs_fs_info *fs_info,
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index c462896122dc..12e63be6a35b 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -277,8 +277,7 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_root *inode_root;
 	struct inode *inode;
-	struct btrfs_ioctl_defrag_range_args range;
-	int num_defrag;
+	struct btrfs_defrag_ctrl ctrl = {0};
 	int ret;
 
 	/* get the inode */
@@ -297,21 +296,23 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
 
 	/* do a chunk of defrag */
 	clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags);
-	memset(&range, 0, sizeof(range));
-	range.len = (u64)-1;
-	range.start = defrag->last_offset;
+	ctrl.len = (u64)-1;
+	ctrl.start = defrag->last_offset;
+	ctrl.newer_than = defrag->transid;
+	ctrl.max_sectors_to_defrag = BTRFS_DEFRAG_BATCH;
 
 	sb_start_write(fs_info->sb);
-	num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid,
-				       BTRFS_DEFRAG_BATCH);
+	ret = btrfs_defrag_file(inode, NULL, &ctrl);
 	sb_end_write(fs_info->sb);
+	if (ret < 0)
+		goto out;
 	/*
 	 * if we filled the whole defrag batch, there
 	 * must be more work to do.  Queue this defrag
 	 * again
 	 */
-	if (num_defrag == BTRFS_DEFRAG_BATCH) {
-		defrag->last_offset = range.start;
+	if (ctrl.sectors_defragged == BTRFS_DEFRAG_BATCH) {
+		defrag->last_offset = ctrl.last_scanned;
 		btrfs_requeue_inode_defrag(BTRFS_I(inode), defrag);
 	} else if (defrag->last_offset && !defrag->cycled) {
 		/*
@@ -325,7 +326,7 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info,
 	} else {
 		kmem_cache_free(btrfs_inode_defrag_cachep, defrag);
 	}
-
+out:
 	iput(inode);
 	return 0;
 cleanup:
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index c6a971033336..28c81ac7b837 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1191,23 +1191,20 @@ struct defrag_target_range {
 /*
  * Collect all valid target extents.
  *
+ * @ctrl:	   extra defrag policy control
  * @start:	   file offset to lookup
  * @len:	   length to lookup
- * @extent_thresh: file extent size threshold, any extent size >= this value
- *		   will be ignored
- * @newer_than:    only defrag extents newer than this value
- * @do_compress:   whether the defrag is doing compression
- *		   if true, @extent_thresh will be ignored and all regular
- *		   file extents meeting @newer_than will be targets.
  * @locked:	   if the range has already held extent lock
  * @target_list:   list of targets file extents
+ *
+ * Will update ctrl::last_scanned.
  */
 static int defrag_collect_targets(struct btrfs_inode *inode,
-				  u64 start, u64 len, u32 extent_thresh,
-				  u64 newer_than, bool do_compress, bool locked,
-				  struct list_head *target_list,
-				  u64 *last_scanned_ret)
+				  struct btrfs_defrag_ctrl *ctrl,
+				  u64 start, u32 len, bool locked,
+				  struct list_head *target_list)
 {
+	bool do_compress = ctrl->flags & BTRFS_DEFRAG_RANGE_COMPRESS;
 	bool last_is_target = false;
 	u64 cur = start;
 	int ret = 0;
@@ -1229,7 +1226,7 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 			goto next;
 
 		/* Skip older extent */
-		if (em->generation < newer_than)
+		if (em->generation < ctrl->newer_than)
 			goto next;
 
 		/* This em is under writeback, no need to defrag */
@@ -1273,7 +1270,7 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 			goto add;
 
 		/* Skip too large extent */
-		if (range_len >= extent_thresh)
+		if (range_len >= ctrl->extent_thresh)
 			goto next;
 
 		/*
@@ -1345,16 +1342,21 @@ static int defrag_collect_targets(struct btrfs_inode *inode,
 			kfree(entry);
 		}
 	}
-	if (!ret && last_scanned_ret) {
+	if (!ret) {
 		/*
 		 * If the last extent is not a target, the caller can skip to
 		 * the end of that extent.
 		 * Otherwise, we can only go the end of the spcified range.
+		 *
+		 * And we may got a range smaller than current
+		 * ctrl->last_scanned (e.g. executed in the defrag_one_range
+		 * call), so we have to ensure we didn't decrease
+		 * ctrl->last_scanned.
 		 */
 		if (!last_is_target)
-			*last_scanned_ret = max(cur, *last_scanned_ret);
+			ctrl->last_scanned = max(cur, ctrl->last_scanned);
 		else
-			*last_scanned_ret = max(start + len, *last_scanned_ret);
+			ctrl->last_scanned = max(start + len, ctrl->last_scanned);
 	}
 	return ret;
 }
@@ -1414,9 +1416,8 @@ static int defrag_one_locked_target(struct btrfs_inode *inode,
 	return ret;
 }
 
-static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
-			    u32 extent_thresh, u64 newer_than, bool do_compress,
-			    u64 *last_scanned_ret)
+static int defrag_one_range(struct btrfs_inode *inode,
+			    struct btrfs_defrag_ctrl *ctrl, u64 start, u32 len)
 {
 	struct extent_state *cached_state = NULL;
 	struct defrag_target_range *entry;
@@ -1460,9 +1461,8 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 	 * And this time we have extent locked already, pass @locked = true
 	 * so that we won't relock the extent range and cause deadlock.
 	 */
-	ret = defrag_collect_targets(inode, start, len, extent_thresh,
-				     newer_than, do_compress, true,
-				     &target_list, last_scanned_ret);
+	ret = defrag_collect_targets(inode, ctrl, start, len, true,
+				     &target_list);
 	if (ret < 0)
 		goto unlock_extent;
 
@@ -1492,13 +1492,17 @@ static int defrag_one_range(struct btrfs_inode *inode, u64 start, u32 len,
 	return ret;
 }
 
+/*
+ * Return <0 for error.
+ * Return >0 if we hit the ctrl->max_sectors_to_defrag limit
+ * Return 0 if we finished the range without error.
+ *
+ * For >= 0 case, ctrl->last_scanned and ctrl->sectors_defragged will be updated.
+ */
 static int defrag_one_cluster(struct btrfs_inode *inode,
 			      struct file_ra_state *ra,
-			      u64 start, u32 len, u32 extent_thresh,
-			      u64 newer_than, bool do_compress,
-			      unsigned long *sectors_defragged,
-			      unsigned long max_sectors,
-			      u64 *last_scanned_ret)
+			      struct btrfs_defrag_ctrl *ctrl,
+			      u64 start, u32 len)
 {
 	const u32 sectorsize = inode->root->fs_info->sectorsize;
 	struct defrag_target_range *entry;
@@ -1506,9 +1510,8 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 	LIST_HEAD(target_list);
 	int ret;
 
-	ret = defrag_collect_targets(inode, start, len, extent_thresh,
-				     newer_than, do_compress, false,
-				     &target_list, NULL);
+	ret = defrag_collect_targets(inode, ctrl, start, len, false,
+				     &target_list);
 	if (ret < 0)
 		goto out;
 
@@ -1516,23 +1519,16 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 		u32 range_len = entry->len;
 
 		/* Reached or beyond the limit */
-		if (max_sectors && *sectors_defragged >= max_sectors) {
+		if (ctrl->max_sectors_to_defrag &&
+		    ctrl->sectors_defragged >= ctrl->max_sectors_to_defrag) {
 			ret = 1;
 			break;
 		}
 
-		if (max_sectors)
+		if (ctrl->max_sectors_to_defrag)
 			range_len = min_t(u32, range_len,
-				(max_sectors - *sectors_defragged) * sectorsize);
-
-		/*
-		 * If defrag_one_range() has updated last_scanned_ret,
-		 * our range may already be invalid (e.g. hole punched).
-		 * Skip if our range is before last_scanned_ret, as there is
-		 * no need to defrag the range anymore.
-		 */
-		if (entry->start + range_len <= *last_scanned_ret)
-			continue;
+				(ctrl->max_sectors_to_defrag -
+				 ctrl->sectors_defragged) * sectorsize);
 
 		if (ra)
 			page_cache_sync_readahead(inode->vfs_inode.i_mapping,
@@ -1545,21 +1541,17 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 		 * But that's fine, it only affects the @sectors_defragged
 		 * accounting.
 		 */
-		ret = defrag_one_range(inode, entry->start, range_len,
-				       extent_thresh, newer_than, do_compress,
-				       last_scanned_ret);
+		ret = defrag_one_range(inode, ctrl, entry->start, range_len);
 		if (ret < 0)
 			break;
-		*sectors_defragged += range_len >>
-				      inode->root->fs_info->sectorsize_bits;
+		ctrl->sectors_defragged += range_len >>
+					   inode->root->fs_info->sectorsize_bits;
 	}
 out:
 	list_for_each_entry_safe(entry, tmp, &target_list, list) {
 		list_del_init(&entry->list);
 		kfree(entry);
 	}
-	if (ret >= 0)
-		*last_scanned_ret = max(*last_scanned_ret, start + len);
 	return ret;
 }
 
@@ -1609,59 +1601,47 @@ int btrfs_defrag_ioctl_args_to_ctrl(struct btrfs_fs_info *fs_info,
  *
  * @inode:	   inode to be defragged
  * @ra:		   readahead state (can be NUL)
- * @range:	   defrag options including range and flags
- * @newer_than:	   minimum transid to defrag
- * @max_to_defrag: max number of sectors to be defragged, if 0, the whole inode
- *		   will be defragged.
+ * @ctrl:	   defrag options including range and various policy parameters
  *
  * Return <0 for error.
- * Return >=0 for the number of sectors defragged, and range->start will be updated
- * to indicate the file offset where next defrag should be started at.
- * (Mostly for autodefrag, which sets @max_to_defrag thus we may exit early without
- *  defragging all the range).
+ * Return 0 if the defrag is done without error, ctrl->last_scanned and
+ * ctrl->sectors_defragged will be updated.
  */
 int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
-		      struct btrfs_ioctl_defrag_range_args *range,
-		      u64 newer_than, unsigned long max_to_defrag)
+		      struct btrfs_defrag_ctrl *ctrl)
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
-	unsigned long sectors_defragged = 0;
 	u64 isize = i_size_read(inode);
 	u64 cur;
 	u64 last_byte;
-	bool do_compress = range->flags & BTRFS_DEFRAG_RANGE_COMPRESS;
+	bool do_compress = ctrl->flags & BTRFS_DEFRAG_RANGE_COMPRESS;
 	bool ra_allocated = false;
-	int compress_type = BTRFS_COMPRESS_ZLIB;
 	int ret = 0;
-	u32 extent_thresh = range->extent_thresh;
 	pgoff_t start_index;
 
 	if (isize == 0)
 		return 0;
 
-	if (range->start >= isize)
+	if (ctrl->start >= isize)
 		return -EINVAL;
 
-	if (do_compress) {
-		if (range->compress_type >= BTRFS_NR_COMPRESS_TYPES)
-			return -EINVAL;
-		if (range->compress_type)
-			compress_type = range->compress_type;
-	}
+	if (do_compress)
+		ASSERT(ctrl->compress < BTRFS_NR_COMPRESS_TYPES);
 
-	if (extent_thresh == 0)
-		extent_thresh = SZ_256K;
+	if (ctrl->extent_thresh == 0)
+		ctrl->extent_thresh = SZ_256K;
 
-	if (range->start + range->len > range->start) {
+	if (ctrl->start + ctrl->len > ctrl->start) {
 		/* Got a specific range */
-		last_byte = min(isize, range->start + range->len);
+		last_byte = min(isize, ctrl->start + ctrl->len);
 	} else {
 		/* Defrag until file end */
 		last_byte = isize;
 	}
 
 	/* Align the range */
-	cur = round_down(range->start, fs_info->sectorsize);
+	cur = round_down(ctrl->start, fs_info->sectorsize);
+	ctrl->last_scanned = cur;
 	last_byte = round_up(last_byte, fs_info->sectorsize) - 1;
 
 	/*
@@ -1685,8 +1665,7 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 		inode->i_mapping->writeback_index = start_index;
 
 	while (cur < last_byte) {
-		const unsigned long prev_sectors_defragged = sectors_defragged;
-		u64 last_scanned = cur;
+		const unsigned long prev_sectors_defragged = ctrl->sectors_defragged;
 		u64 cluster_end;
 
 		if (btrfs_defrag_cancelled(fs_info)) {
@@ -1710,19 +1689,17 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 			break;
 		}
 		if (do_compress)
-			BTRFS_I(inode)->defrag_compress = compress_type;
-		ret = defrag_one_cluster(BTRFS_I(inode), ra, cur,
-				cluster_end + 1 - cur, extent_thresh,
-				newer_than, do_compress, &sectors_defragged,
-				max_to_defrag, &last_scanned);
+			BTRFS_I(inode)->defrag_compress = ctrl->compress;
+		ret = defrag_one_cluster(BTRFS_I(inode), ra, ctrl, cur,
+				cluster_end + 1 - cur);
 
-		if (sectors_defragged > prev_sectors_defragged)
+		if (ctrl->sectors_defragged > prev_sectors_defragged)
 			balance_dirty_pages_ratelimited(inode->i_mapping);
 
 		btrfs_inode_unlock(inode, 0);
 		if (ret < 0)
 			break;
-		cur = max(cluster_end + 1, last_scanned);
+		cur = max(cluster_end + 1, ctrl->last_scanned);
 		if (ret > 0) {
 			ret = 0;
 			break;
@@ -1732,27 +1709,21 @@ int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
 
 	if (ra_allocated)
 		kfree(ra);
-	/*
-	 * Update range.start for autodefrag, this will indicate where to start
-	 * in next run.
-	 */
-	range->start = cur;
-	if (sectors_defragged) {
+	if (ctrl->sectors_defragged) {
 		/*
 		 * We have defragged some sectors, for compression case they
 		 * need to be written back immediately.
 		 */
-		if (range->flags & BTRFS_DEFRAG_RANGE_START_IO) {
+		if (ctrl->flags & BTRFS_DEFRAG_RANGE_START_IO) {
 			filemap_flush(inode->i_mapping);
 			if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
 				     &BTRFS_I(inode)->runtime_flags))
 				filemap_flush(inode->i_mapping);
 		}
-		if (range->compress_type == BTRFS_COMPRESS_LZO)
+		if (ctrl->compress == BTRFS_COMPRESS_LZO)
 			btrfs_set_fs_incompat(fs_info, COMPRESS_LZO);
-		else if (range->compress_type == BTRFS_COMPRESS_ZSTD)
+		else if (ctrl->compress == BTRFS_COMPRESS_ZSTD)
 			btrfs_set_fs_incompat(fs_info, COMPRESS_ZSTD);
-		ret = sectors_defragged;
 	}
 	if (do_compress) {
 		btrfs_inode_lock(inode, 0);
@@ -3267,6 +3238,7 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
 	struct inode *inode = file_inode(file);
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_ioctl_defrag_range_args range = {0};
+	struct btrfs_defrag_ctrl ctrl = {0};
 	int ret;
 
 	ret = mnt_want_write_file(file);
@@ -3312,8 +3284,11 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
 			/* the rest are all set to zero by kzalloc */
 			range.len = (u64)-1;
 		}
-		ret = btrfs_defrag_file(file_inode(file), &file->f_ra,
-					&range, BTRFS_OLDEST_GENERATION, 0);
+		ret = btrfs_defrag_ioctl_args_to_ctrl(root->fs_info, &range,
+						      &ctrl, 0, BTRFS_OLDEST_GENERATION);
+		if (ret < 0)
+			break;
+		ret = btrfs_defrag_file(file_inode(file), &file->f_ra, &ctrl);
 		if (ret > 0)
 			ret = 0;
 		break;
-- 
2.35.0

[PATCH v4 5/5] btrfs: defrag: make btrfs_defrag_file() to report accurate number of defragged sectors

[Qu Wenruo]

Previously rework btrfs_defrag_file() can only report the number of
sectors from the first run of defrag_collect_targets().

This number is not accurate as if holes are punched after the first
defrag_collect_targets() call, we will not choose to defrag the holes.

Originally this is to avoid passing @sectors_defragged to every involved
functions.

But now since we have btrfs_defrag_ctrl, there is no need to do such
inaccurate accounting, just update btrfs_defrag_ctrl::sectors_defragged
after a successful defrag_one_locked_target() call.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
---
 fs/btrfs/ioctl.c | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 28c81ac7b837..e49f745af942 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1471,6 +1471,8 @@ static int defrag_one_range(struct btrfs_inode *inode,
 					       &cached_state);
 		if (ret < 0)
 			break;
+		ctrl->sectors_defragged += entry->len >>
+					   inode->root->fs_info->sectorsize_bits;
 	}
 
 	list_for_each_entry_safe(entry, tmp, &target_list, list) {
@@ -1535,17 +1537,9 @@ static int defrag_one_cluster(struct btrfs_inode *inode,
 				ra, NULL, entry->start >> PAGE_SHIFT,
 				((entry->start + range_len - 1) >> PAGE_SHIFT) -
 				(entry->start >> PAGE_SHIFT) + 1);
-		/*
-		 * Here we may not defrag any range if holes are punched before
-		 * we locked the pages.
-		 * But that's fine, it only affects the @sectors_defragged
-		 * accounting.
-		 */
 		ret = defrag_one_range(inode, ctrl, entry->start, range_len);
 		if (ret < 0)
 			break;
-		ctrl->sectors_defragged += range_len >>
-					   inode->root->fs_info->sectorsize_bits;
 	}
 out:
 	list_for_each_entry_safe(entry, tmp, &target_list, list) {
-- 
2.35.0

Re: [PATCH v4 0/5] btrfs: defrag: don't waste CPU time on non-target extent

[David Sterba]

On Fri, Feb 11, 2022 at 02:41:38PM +0800, Qu Wenruo wrote:
> Changelog:
> v2:
> - Rebased to lastest misc-next
>   Just one small conflict with static_assert() update.
>   And this time only those patches are rebased to misc-next, thus it may
>   cause conflicts with fixes for defrag_check_next_extent() in the
>   future.
> 
> - Several grammar fixes
> 
> - Report accurate btrfs_defrag_ctrl::sectors_defragged
>   This is inspired by a comment from Filipe that the skip check
>   should be done in the defrag_collect_targets() call inside
>   defrag_one_range().
> 
>   This results a new patch in v2.
> 
> - Change the timing of btrfs_defrag_ctrl::last_scanned update
>   Now it's updated inside defrag_one_range(), which will give
>   us an accurate view, unlike the previous call site in
>   defrag_one_cluster().
> 
> - Don't change the timing of extent threshold.
> 
> - Rename @last_target to @last_is_target in defrag_collect_targets()
> 
> v3:
> - Add Reviewed-by tags
> 
> - Fix a wrong value in commit message of the 1st patch
> 
> - Make @orig_start const for the 3rd patch
> 
> - Fix a missing word "skip" in the 5th patch
> 
> - Remove one unnecessary assignment in the 5th patch
>   As we don't return the defragged sectors to user space.
> 
> v4:
> - Move the skip behavior before the btrfs_defrag_ctrl refactor
>   So it can be backported to v5.16

Thanks, I've picked the patch to misc-next. I'm not sure about the rest
because of the ctrl structure cleanup and there are still some pending
defrag fixes. So I'll add the remainig patches as topic branch for
testing.

Re: [PATCH v4 0/5] btrfs: defrag: don't waste CPU time on non-target extent

[David Sterba]

On Mon, Feb 14, 2022 at 05:40:11PM +0100, David Sterba wrote:
> On Fri, Feb 11, 2022 at 02:41:38PM +0800, Qu Wenruo wrote:
> > Changelog:
> > v2:
> > - Rebased to lastest misc-next
> >   Just one small conflict with static_assert() update.
> >   And this time only those patches are rebased to misc-next, thus it may
> >   cause conflicts with fixes for defrag_check_next_extent() in the
> >   future.
> > 
> > - Several grammar fixes
> > 
> > - Report accurate btrfs_defrag_ctrl::sectors_defragged
> >   This is inspired by a comment from Filipe that the skip check
> >   should be done in the defrag_collect_targets() call inside
> >   defrag_one_range().
> > 
> >   This results a new patch in v2.
> > 
> > - Change the timing of btrfs_defrag_ctrl::last_scanned update
> >   Now it's updated inside defrag_one_range(), which will give
> >   us an accurate view, unlike the previous call site in
> >   defrag_one_cluster().
> > 
> > - Don't change the timing of extent threshold.
> > 
> > - Rename @last_target to @last_is_target in defrag_collect_targets()
> > 
> > v3:
> > - Add Reviewed-by tags
> > 
> > - Fix a wrong value in commit message of the 1st patch
> > 
> > - Make @orig_start const for the 3rd patch
> > 
> > - Fix a missing word "skip" in the 5th patch
> > 
> > - Remove one unnecessary assignment in the 5th patch
> >   As we don't return the defragged sectors to user space.
> > 
> > v4:
> > - Move the skip behavior before the btrfs_defrag_ctrl refactor
> >   So it can be backported to v5.16
> 
> Thanks, I've picked the patch to misc-next. I'm not sure about the rest
> because of the ctrl structure cleanup and there are still some pending
> defrag fixes. So I'll add the remainig patches as topic branch for
> testing.

Rest of the patchset added to misc-next, as the remaining fixups ported
to older stable trees will be easy enough.