Merge tag 'for-5.19-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
 "Features:

   - subpage:
      - support for PAGE_SIZE > 4K (previously only 64K)
      - make it work with raid56

   - repair super block num_devices automatically if it does not match
     the number of device items

   - defrag can convert inline extents to regular extents, up to now
     inline files were skipped but the setting of mount option
     max_inline could affect the decision logic

   - zoned:
      - minimal accepted zone size is explicitly set to 4MiB
      - make zone reclaim less aggressive and don't reclaim if there are
        enough free zones
      - add per-profile sysfs tunable of the reclaim threshold

   - allow automatic block group reclaim for non-zoned filesystems, with
     sysfs tunables

   - tree-checker: new check, compare extent buffer owner against owner
     rootid

  Performance:

   - avoid blocking on space reservation when doing nowait direct io
     writes (+7% throughput for reads and writes)

   - NOCOW write throughput improvement due to refined locking (+3%)

   - send: reduce pressure to page cache by dropping extent pages right
     after they're processed

  Core:

   - convert all radix trees to xarray

   - add iterators for b-tree node items

   - support printk message index

   - user bulk page allocation for extent buffers

   - switch to bio_alloc API, use on-stack bios where convenient, other
     bio cleanups

   - use rw lock for block groups to favor concurrent reads

   - simplify workques, don't allocate high priority threads for all
     normal queues as we need only one

   - refactor scrub, process chunks based on their constraints and
     similarity

   - allocate direct io structures on stack and pass around only
     pointers, avoids allocation and reduces potential error handling

  Fixes:

   - fix count of reserved transaction items for various inode
     operations

   - fix deadlock between concurrent dio writes when low on free data
     space

   - fix a few cases when zones need to be finished

  VFS, iomap:

   - add helper to check if sb write has started (usable for assertions)

   - new helper iomap_dio_alloc_bio, export iomap_dio_bio_end_io"

* tag 'for-5.19-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (173 commits)
  btrfs: zoned: introduce a minimal zone size 4M and reject mount
  btrfs: allow defrag to convert inline extents to regular extents
  btrfs: add "0x" prefix for unsupported optional features
  btrfs: do not account twice for inode ref when reserving metadata units
  btrfs: zoned: fix comparison of alloc_offset vs meta_write_pointer
  btrfs: send: avoid trashing the page cache
  btrfs: send: keep the current inode open while processing it
  btrfs: allocate the btrfs_dio_private as part of the iomap dio bio
  btrfs: move struct btrfs_dio_private to inode.c
  btrfs: remove the disk_bytenr in struct btrfs_dio_private
  btrfs: allocate dio_data on stack
  iomap: add per-iomap_iter private data
  iomap: allow the file system to provide a bio_set for direct I/O
  btrfs: add a btrfs_dio_rw wrapper
  btrfs: zoned: zone finish unused block group
  btrfs: zoned: properly finish block group on metadata write
  btrfs: zoned: finish block group when there are no more allocatable bytes left
  btrfs: zoned: consolidate zone finish functions
  btrfs: zoned: introduce btrfs_zoned_bg_is_full
  btrfs: improve error reporting in lookup_inline_extent_backref
  ...

1 files changed, 125 insertions, 80 deletions

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 0dd6de994199..ede389f2602d 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -168,11 +168,12 @@ static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
 	struct rb_node **p;
 	struct rb_node *parent = NULL;
 	struct btrfs_block_group *cache;
+	bool leftmost = true;
 
 	ASSERT(block_group->length != 0);
 
-	spin_lock(&info->block_group_cache_lock);
-	p = &info->block_group_cache_tree.rb_node;
+	write_lock(&info->block_group_cache_lock);
+	p = &info->block_group_cache_tree.rb_root.rb_node;
 
 	while (*p) {
 		parent = *p;
@@ -181,20 +182,18 @@ static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
 			p = &(*p)->rb_left;
 		} else if (block_group->start > cache->start) {
 			p = &(*p)->rb_right;
+			leftmost = false;
 		} else {
-			spin_unlock(&info->block_group_cache_lock);
+			write_unlock(&info->block_group_cache_lock);
 			return -EEXIST;
 		}
 	}
 
 	rb_link_node(&block_group->cache_node, parent, p);
-	rb_insert_color(&block_group->cache_node,
-			&info->block_group_cache_tree);
+	rb_insert_color_cached(&block_group->cache_node,
+			       &info->block_group_cache_tree, leftmost);
 
-	if (info->first_logical_byte > block_group->start)
-		info->first_logical_byte = block_group->start;
-
-	spin_unlock(&info->block_group_cache_lock);
+	write_unlock(&info->block_group_cache_lock);
 
 	return 0;
 }
@@ -210,8 +209,8 @@ static struct btrfs_block_group *block_group_cache_tree_search(
 	struct rb_node *n;
 	u64 end, start;
 
-	spin_lock(&info->block_group_cache_lock);
-	n = info->block_group_cache_tree.rb_node;
+	read_lock(&info->block_group_cache_lock);
+	n = info->block_group_cache_tree.rb_root.rb_node;
 
 	while (n) {
 		cache = rb_entry(n, struct btrfs_block_group, cache_node);
@@ -233,12 +232,9 @@ static struct btrfs_block_group *block_group_cache_tree_search(
 			break;
 		}
 	}
-	if (ret) {
+	if (ret)
 		btrfs_get_block_group(ret);
-		if (bytenr == 0 && info->first_logical_byte > ret->start)
-			info->first_logical_byte = ret->start;
-	}
-	spin_unlock(&info->block_group_cache_lock);
+	read_unlock(&info->block_group_cache_lock);
 
 	return ret;
 }
@@ -267,15 +263,15 @@ struct btrfs_block_group *btrfs_next_block_group(
 	struct btrfs_fs_info *fs_info = cache->fs_info;
 	struct rb_node *node;
 
-	spin_lock(&fs_info->block_group_cache_lock);
+	read_lock(&fs_info->block_group_cache_lock);
 
 	/* If our block group was removed, we need a full search. */
 	if (RB_EMPTY_NODE(&cache->cache_node)) {
 		const u64 next_bytenr = cache->start + cache->length;
 
-		spin_unlock(&fs_info->block_group_cache_lock);
+		read_unlock(&fs_info->block_group_cache_lock);
 		btrfs_put_block_group(cache);
-		cache = btrfs_lookup_first_block_group(fs_info, next_bytenr); return cache;
+		return btrfs_lookup_first_block_group(fs_info, next_bytenr);
 	}
 	node = rb_next(&cache->cache_node);
 	btrfs_put_block_group(cache);
@@ -284,46 +280,70 @@ struct btrfs_block_group *btrfs_next_block_group(
 		btrfs_get_block_group(cache);
 	} else
 		cache = NULL;
-	spin_unlock(&fs_info->block_group_cache_lock);
+	read_unlock(&fs_info->block_group_cache_lock);
 	return cache;
 }
 
-bool btrfs_inc_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr)
+/**
+ * Check if we can do a NOCOW write for a given extent.
+ *
+ * @fs_info:       The filesystem information object.
+ * @bytenr:        Logical start address of the extent.
+ *
+ * Check if we can do a NOCOW write for the given extent, and increments the
+ * number of NOCOW writers in the block group that contains the extent, as long
+ * as the block group exists and it's currently not in read-only mode.
+ *
+ * Returns: A non-NULL block group pointer if we can do a NOCOW write, the caller
+ *          is responsible for calling btrfs_dec_nocow_writers() later.
+ *
+ *          Or NULL if we can not do a NOCOW write
+ */
+struct btrfs_block_group *btrfs_inc_nocow_writers(struct btrfs_fs_info *fs_info,
+						  u64 bytenr)
 {
 	struct btrfs_block_group *bg;
-	bool ret = true;
+	bool can_nocow = true;
 
 	bg = btrfs_lookup_block_group(fs_info, bytenr);
 	if (!bg)
-		return false;
+		return NULL;
 
 	spin_lock(&bg->lock);
 	if (bg->ro)
-		ret = false;
+		can_nocow = false;
 	else
 		atomic_inc(&bg->nocow_writers);
 	spin_unlock(&bg->lock);
 
-	/* No put on block group, done by btrfs_dec_nocow_writers */
-	if (!ret)
+	if (!can_nocow) {
 		btrfs_put_block_group(bg);
+		return NULL;
+	}
 
-	return ret;
+	/* No put on block group, done by btrfs_dec_nocow_writers(). */
+	return bg;
 }
 
-void btrfs_dec_nocow_writers(struct btrfs_fs_info *fs_info, u64 bytenr)
+/**
+ * Decrement the number of NOCOW writers in a block group.
+ *
+ * @bg:       The block group.
+ *
+ * This is meant to be called after a previous call to btrfs_inc_nocow_writers(),
+ * and on the block group returned by that call. Typically this is called after
+ * creating an ordered extent for a NOCOW write, to prevent races with scrub and
+ * relocation.
+ *
+ * After this call, the caller should not use the block group anymore. It it wants
+ * to use it, then it should get a reference on it before calling this function.
+ */
+void btrfs_dec_nocow_writers(struct btrfs_block_group *bg)
 {
-	struct btrfs_block_group *bg;
-
-	bg = btrfs_lookup_block_group(fs_info, bytenr);
-	ASSERT(bg);
 	if (atomic_dec_and_test(&bg->nocow_writers))
 		wake_up_var(&bg->nocow_writers);
-	/*
-	 * Once for our lookup and once for the lookup done by a previous call
-	 * to btrfs_inc_nocow_writers()
-	 */
-	btrfs_put_block_group(bg);
+
+	/* For the lookup done by a previous call to btrfs_inc_nocow_writers(). */
 	btrfs_put_block_group(bg);
 }
 
@@ -772,10 +792,10 @@ int btrfs_cache_block_group(struct btrfs_block_group *cache, int load_cache_only
 	cache->has_caching_ctl = 1;
 	spin_unlock(&cache->lock);
 
-	spin_lock(&fs_info->block_group_cache_lock);
+	write_lock(&fs_info->block_group_cache_lock);
 	refcount_inc(&caching_ctl->count);
 	list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups);
-	spin_unlock(&fs_info->block_group_cache_lock);
+	write_unlock(&fs_info->block_group_cache_lock);
 
 	btrfs_get_block_group(cache);
 
@@ -957,17 +977,15 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	if (ret)
 		goto out;
 
-	spin_lock(&fs_info->block_group_cache_lock);
-	rb_erase(&block_group->cache_node,
-		 &fs_info->block_group_cache_tree);
+	write_lock(&fs_info->block_group_cache_lock);
+	rb_erase_cached(&block_group->cache_node,
+			&fs_info->block_group_cache_tree);
 	RB_CLEAR_NODE(&block_group->cache_node);
 
 	/* Once for the block groups rbtree */
 	btrfs_put_block_group(block_group);
 
-	if (fs_info->first_logical_byte == block_group->start)
-		fs_info->first_logical_byte = (u64)-1;
-	spin_unlock(&fs_info->block_group_cache_lock);
+	write_unlock(&fs_info->block_group_cache_lock);
 
 	down_write(&block_group->space_info->groups_sem);
 	/*
@@ -992,7 +1010,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	if (block_group->cached == BTRFS_CACHE_STARTED)
 		btrfs_wait_block_group_cache_done(block_group);
 	if (block_group->has_caching_ctl) {
-		spin_lock(&fs_info->block_group_cache_lock);
+		write_lock(&fs_info->block_group_cache_lock);
 		if (!caching_ctl) {
 			struct btrfs_caching_control *ctl;
 
@@ -1006,7 +1024,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 		}
 		if (caching_ctl)
 			list_del_init(&caching_ctl->list);
-		spin_unlock(&fs_info->block_group_cache_lock);
+		write_unlock(&fs_info->block_group_cache_lock);
 		if (caching_ctl) {
 			/* Once for the caching bgs list and once for us. */
 			btrfs_put_caching_control(caching_ctl);
@@ -1367,6 +1385,14 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 			goto next;
 		}
 
+		ret = btrfs_zone_finish(block_group);
+		if (ret < 0) {
+			btrfs_dec_block_group_ro(block_group);
+			if (ret == -EAGAIN)
+				ret = 0;
+			goto next;
+		}
+
 		/*
 		 * Want to do this before we do anything else so we can recover
 		 * properly if we fail to join the transaction.
@@ -1512,6 +1538,13 @@ static int reclaim_bgs_cmp(void *unused, const struct list_head *a,
 	return bg1->used > bg2->used;
 }
 
+static inline bool btrfs_should_reclaim(struct btrfs_fs_info *fs_info)
+{
+	if (btrfs_is_zoned(fs_info))
+		return btrfs_zoned_should_reclaim(fs_info);
+	return true;
+}
+
 void btrfs_reclaim_bgs_work(struct work_struct *work)
 {
 	struct btrfs_fs_info *fs_info =
@@ -1522,6 +1555,9 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 	if (!test_bit(BTRFS_FS_OPEN, &fs_info->flags))
 		return;
 
+	if (!btrfs_should_reclaim(fs_info))
+		return;
+
 	sb_start_write(fs_info->sb);
 
 	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_BALANCE)) {
@@ -1692,35 +1728,13 @@ static int find_first_block_group(struct btrfs_fs_info *fs_info,
 	struct btrfs_root *root = btrfs_block_group_root(fs_info);
 	int ret;
 	struct btrfs_key found_key;
-	struct extent_buffer *leaf;
-	int slot;
-
-	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
-	if (ret < 0)
-		return ret;
-
-	while (1) {
-		slot = path->slots[0];
-		leaf = path->nodes[0];
-		if (slot >= btrfs_header_nritems(leaf)) {
-			ret = btrfs_next_leaf(root, path);
-			if (ret == 0)
-				continue;
-			if (ret < 0)
-				goto out;
-			break;
-		}
-		btrfs_item_key_to_cpu(leaf, &found_key, slot);
 
+	btrfs_for_each_slot(root, key, &found_key, path, ret) {
 		if (found_key.objectid >= key->objectid &&
 		    found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
-			ret = read_bg_from_eb(fs_info, &found_key, path);
-			break;
+			return read_bg_from_eb(fs_info, &found_key, path);
 		}
-
-		path->slots[0]++;
 	}
-out:
 	return ret;
 }
 
@@ -3220,6 +3234,31 @@ int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans)
 	return ret;
 }
 
+static inline bool should_reclaim_block_group(struct btrfs_block_group *bg,
+					      u64 bytes_freed)
+{
+	const struct btrfs_space_info *space_info = bg->space_info;
+	const int reclaim_thresh = READ_ONCE(space_info->bg_reclaim_threshold);
+	const u64 new_val = bg->used;
+	const u64 old_val = new_val + bytes_freed;
+	u64 thresh;
+
+	if (reclaim_thresh == 0)
+		return false;
+
+	thresh = div_factor_fine(bg->length, reclaim_thresh);
+
+	/*
+	 * If we were below the threshold before don't reclaim, we are likely a
+	 * brand new block group and we don't want to relocate new block groups.
+	 */
+	if (old_val < thresh)
+		return false;
+	if (new_val >= thresh)
+		return false;
+	return true;
+}
+
 int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 			     u64 bytenr, u64 num_bytes, bool alloc)
 {
@@ -3242,6 +3281,8 @@ int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 	spin_unlock(&info->delalloc_root_lock);
 
 	while (total) {
+		bool reclaim;
+
 		cache = btrfs_lookup_block_group(info, bytenr);
 		if (!cache) {
 			ret = -ENOENT;
@@ -3287,6 +3328,8 @@ int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 					cache->space_info, num_bytes);
 			cache->space_info->bytes_used -= num_bytes;
 			cache->space_info->disk_used -= num_bytes * factor;
+
+			reclaim = should_reclaim_block_group(cache, num_bytes);
 			spin_unlock(&cache->lock);
 			spin_unlock(&cache->space_info->lock);
 
@@ -3313,6 +3356,8 @@ int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 		if (!alloc && old_val == 0) {
 			if (!btrfs_test_opt(info, DISCARD_ASYNC))
 				btrfs_mark_bg_unused(cache);
+		} else if (!alloc && reclaim) {
+			btrfs_mark_bg_to_reclaim(cache);
 		}
 
 		btrfs_put_block_group(cache);
@@ -3957,14 +4002,14 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	struct btrfs_caching_control *caching_ctl;
 	struct rb_node *n;
 
-	spin_lock(&info->block_group_cache_lock);
+	write_lock(&info->block_group_cache_lock);
 	while (!list_empty(&info->caching_block_groups)) {
 		caching_ctl = list_entry(info->caching_block_groups.next,
 					 struct btrfs_caching_control, list);
 		list_del(&caching_ctl->list);
 		btrfs_put_caching_control(caching_ctl);
 	}
-	spin_unlock(&info->block_group_cache_lock);
+	write_unlock(&info->block_group_cache_lock);
 
 	spin_lock(&info->unused_bgs_lock);
 	while (!list_empty(&info->unused_bgs)) {
@@ -3994,14 +4039,14 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	}
 	spin_unlock(&info->zone_active_bgs_lock);
 
-	spin_lock(&info->block_group_cache_lock);
-	while ((n = rb_last(&info->block_group_cache_tree)) != NULL) {
+	write_lock(&info->block_group_cache_lock);
+	while ((n = rb_last(&info->block_group_cache_tree.rb_root)) != NULL) {
 		block_group = rb_entry(n, struct btrfs_block_group,
 				       cache_node);
-		rb_erase(&block_group->cache_node,
-			 &info->block_group_cache_tree);
+		rb_erase_cached(&block_group->cache_node,
+				&info->block_group_cache_tree);
 		RB_CLEAR_NODE(&block_group->cache_node);
-		spin_unlock(&info->block_group_cache_lock);
+		write_unlock(&info->block_group_cache_lock);
 
 		down_write(&block_group->space_info->groups_sem);
 		list_del(&block_group->list);
@@ -4024,9 +4069,9 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 		ASSERT(block_group->swap_extents == 0);
 		btrfs_put_block_group(block_group);
 
-		spin_lock(&info->block_group_cache_lock);
+		write_lock(&info->block_group_cache_lock);
 	}
-	spin_unlock(&info->block_group_cache_lock);
+	write_unlock(&info->block_group_cache_lock);
 
 	btrfs_release_global_block_rsv(info);