[PATCH RFC 2/9] btrfs: zoned: get rid of relocation_bg_lock

[Johannes Thumshirn <jth@kernel.org>]

From: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Lockstat analysis of benchmark workloads shows a very high contention of
the relocation_bg_lock. But the relocation_bg_lock only protects a single
field in 'struct btrfs_fs_info', namely 'u64 data_reloc_bg'.

Use READ_ONCE()/WRITE_ONCE() to access 'btrfs_fs_info::data_reloc_bg'.

This is safe in the allocator path, as relocation I/O is only going to
block groups in the relocation sub-space_info and at the moment, there is
only one relocation block group in this space info.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 fs/btrfs/disk-io.c     |  1 -
 fs/btrfs/extent-tree.c | 28 +++++++++++-----------------
 fs/btrfs/fs.h          |  6 +-----
 fs/btrfs/zoned.c       | 11 +++++------
 4 files changed, 17 insertions(+), 29 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 6ac5be02dce7..9a13f5b1ed43 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -2791,7 +2791,6 @@ void btrfs_init_fs_info(struct btrfs_fs_info *fs_info)
 	spin_lock_init(&fs_info->unused_bgs_lock);
 	spin_lock_init(&fs_info->treelog_bg_lock);
 	spin_lock_init(&fs_info->zone_active_bgs_lock);
-	spin_lock_init(&fs_info->relocation_bg_lock);
 	rwlock_init(&fs_info->tree_mod_log_lock);
 	rwlock_init(&fs_info->global_root_lock);
 	mutex_init(&fs_info->unused_bg_unpin_mutex);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 10f50c725313..a9bda68a1883 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3865,14 +3865,10 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	 * Do not allow non-relocation blocks in the dedicated relocation block
 	 * group, and vice versa.
 	 */
-	spin_lock(&fs_info->relocation_bg_lock);
-	data_reloc_bytenr = fs_info->data_reloc_bg;
+	data_reloc_bytenr = READ_ONCE(fs_info->data_reloc_bg);
 	if (data_reloc_bytenr &&
 	    ((ffe_ctl->for_data_reloc && bytenr != data_reloc_bytenr) ||
 	     (!ffe_ctl->for_data_reloc && bytenr == data_reloc_bytenr)))
-		skip = true;
-	spin_unlock(&fs_info->relocation_bg_lock);
-	if (skip)
 		return 1;
 
 	/* Check RO and no space case before trying to activate it */
@@ -3899,7 +3895,6 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	spin_lock(&space_info->lock);
 	spin_lock(&block_group->lock);
 	spin_lock(&fs_info->treelog_bg_lock);
-	spin_lock(&fs_info->relocation_bg_lock);
 
 	if (ret)
 		goto out;
@@ -3908,8 +3903,8 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	       block_group->start == fs_info->treelog_bg ||
 	       fs_info->treelog_bg == 0);
 	ASSERT(!ffe_ctl->for_data_reloc ||
-	       block_group->start == fs_info->data_reloc_bg ||
-	       fs_info->data_reloc_bg == 0);
+	       block_group->start == data_reloc_bytenr ||
+	       data_reloc_bytenr == 0);
 
 	if (block_group->ro ||
 	    (!ffe_ctl->for_data_reloc &&
@@ -3932,7 +3927,7 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	 * Do not allow currently used block group to be the data relocation
 	 * dedicated block group.
 	 */
-	if (ffe_ctl->for_data_reloc && !fs_info->data_reloc_bg &&
+	if (ffe_ctl->for_data_reloc && data_reloc_bytenr == 0 &&
 	    (block_group->used || block_group->reserved)) {
 		ret = 1;
 		goto out;
@@ -3957,8 +3952,8 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 		fs_info->treelog_bg = block_group->start;
 
 	if (ffe_ctl->for_data_reloc) {
-		if (!fs_info->data_reloc_bg)
-			fs_info->data_reloc_bg = block_group->start;
+		if (READ_ONCE(fs_info->data_reloc_bg) == 0)
+			WRITE_ONCE(fs_info->data_reloc_bg, block_group->start);
 		/*
 		 * Do not allow allocations from this block group, unless it is
 		 * for data relocation. Compared to increasing the ->ro, setting
@@ -3994,8 +3989,7 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	if (ret && ffe_ctl->for_treelog)
 		fs_info->treelog_bg = 0;
 	if (ret && ffe_ctl->for_data_reloc)
-		fs_info->data_reloc_bg = 0;
-	spin_unlock(&fs_info->relocation_bg_lock);
+		WRITE_ONCE(fs_info->data_reloc_bg, 0);
 	spin_unlock(&fs_info->treelog_bg_lock);
 	spin_unlock(&block_group->lock);
 	spin_unlock(&space_info->lock);
@@ -4304,10 +4298,10 @@ static int prepare_allocation_zoned(struct btrfs_fs_info *fs_info,
 			ffe_ctl->hint_byte = fs_info->treelog_bg;
 		spin_unlock(&fs_info->treelog_bg_lock);
 	} else if (ffe_ctl->for_data_reloc) {
-		spin_lock(&fs_info->relocation_bg_lock);
-		if (fs_info->data_reloc_bg)
-			ffe_ctl->hint_byte = fs_info->data_reloc_bg;
-		spin_unlock(&fs_info->relocation_bg_lock);
+		u64 data_reloc_bg = READ_ONCE(fs_info->data_reloc_bg);
+
+		if (data_reloc_bg)
+			ffe_ctl->hint_byte = data_reloc_bg;
 	} else if (ffe_ctl->flags & BTRFS_BLOCK_GROUP_DATA) {
 		struct btrfs_block_group *block_group;
 
diff --git a/fs/btrfs/fs.h b/fs/btrfs/fs.h
index b239e4b8421c..570f4b85096c 100644
--- a/fs/btrfs/fs.h
+++ b/fs/btrfs/fs.h
@@ -849,11 +849,7 @@ struct btrfs_fs_info {
 	spinlock_t treelog_bg_lock;
 	u64 treelog_bg;
 
-	/*
-	 * Start of the dedicated data relocation block group, protected by
-	 * relocation_bg_lock.
-	 */
-	spinlock_t relocation_bg_lock;
+	/* Start of the dedicated data relocation block group */
 	u64 data_reloc_bg;
 	struct mutex zoned_data_reloc_io_lock;
 
diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index 0d5d6db72b62..388c277a84d3 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -2495,11 +2495,10 @@ void btrfs_schedule_zone_finish_bg(struct btrfs_block_group *bg,
 void btrfs_clear_data_reloc_bg(struct btrfs_block_group *bg)
 {
 	struct btrfs_fs_info *fs_info = bg->fs_info;
+	u64 data_reloc_bg = READ_ONCE(fs_info->data_reloc_bg);
 
-	spin_lock(&fs_info->relocation_bg_lock);
-	if (fs_info->data_reloc_bg == bg->start)
-		fs_info->data_reloc_bg = 0;
-	spin_unlock(&fs_info->relocation_bg_lock);
+	if (data_reloc_bg == bg->start)
+		WRITE_ONCE(fs_info->data_reloc_bg, 0);
 }
 
 void btrfs_zoned_reserve_data_reloc_bg(struct btrfs_fs_info *fs_info)
@@ -2518,7 +2517,7 @@ void btrfs_zoned_reserve_data_reloc_bg(struct btrfs_fs_info *fs_info)
 	if (!btrfs_is_zoned(fs_info))
 		return;
 
-	if (fs_info->data_reloc_bg)
+	if (READ_ONCE(fs_info->data_reloc_bg))
 		return;
 
 	if (sb_rdonly(fs_info->sb))
@@ -2539,7 +2538,7 @@ void btrfs_zoned_reserve_data_reloc_bg(struct btrfs_fs_info *fs_info)
 			continue;
 		}
 
-		fs_info->data_reloc_bg = bg->start;
+		WRITE_ONCE(fs_info->data_reloc_bg, bg->start);
 		set_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &bg->runtime_flags);
 		btrfs_zone_activate(bg);
 
-- 
2.49.0