[PATCH v3 0/3] btrfs: REF_COWS bit rework

[PATCH v3 0/3] btrfs: REF_COWS bit rework

[Qu Wenruo]

This small patchset reworks the REF_COWS bit, by renaming it, and remove
that bit for data relocation root.


The basic idea of such rework is to reduce the confusion caused by the
name REF_COWS.

With the new bit called SHAREABLE, it should be clear that no user can
really create snapshot for data reloc tree, thus its tree blocks
shouldn't be shareable.

This would make data balance for reloc tree a little simpler.

Changelog:
v2:
- Add new patch to address the log tree check in
  btrfs_truncate_inode_items()
  Thanks for the advice from David, now it's much simpler than original
  check, and data reloc tree no longer needs extra hanlding

- Grab data reloc root in create_reloc_inode() and
  btrfs_recover_relocation()

- Comment update

v3:
- Remove ALIGN_DOWN() -> round_down() change

- Remove the confusing comment on the log tree inode

Qu Wenruo (3):
  btrfs: Rename BTRFS_ROOT_REF_COWS to BTRFS_ROOT_SHAREABLE
  btrfs: inode: Cleanup the log tree exceptions in
    btrfs_truncate_inode_items()
  btrfs: Don't set SHAREABLE flag for data reloc tree

 fs/btrfs/backref.c     |  4 ++--
 fs/btrfs/backref.h     |  2 +-
 fs/btrfs/block-rsv.c   |  2 +-
 fs/btrfs/ctree.c       | 26 +++++++++++++-------------
 fs/btrfs/ctree.h       | 25 +++++++++++++++++++++++--
 fs/btrfs/disk-io.c     | 27 ++++++++++++++++++++-------
 fs/btrfs/extent-tree.c |  2 +-
 fs/btrfs/file.c        |  2 +-
 fs/btrfs/inode.c       | 35 +++++++++++++++++------------------
 fs/btrfs/ioctl.c       |  2 +-
 fs/btrfs/relocation.c  | 41 +++++++++++++++++++----------------------
 fs/btrfs/transaction.c | 12 ++++++------
 fs/btrfs/tree-defrag.c |  2 +-
 13 files changed, 106 insertions(+), 76 deletions(-)

-- 
2.26.2

[PATCH v3 1/3] btrfs: Rename BTRFS_ROOT_REF_COWS to BTRFS_ROOT_SHAREABLE

[Qu Wenruo]

The name BTRFS_ROOT_REF_COWS is not helpful to show what it really
means.

In fact, that bit can only be set to those trees:
- Subvolume roots
- Data reloc root
- Reloc roots for above roots

All other trees won't get this bit set.
So just by the result, it is obvious that, roots with this bit set can
have tree blocks shared with other trees.
Either shared by snapshots, or by reloc roots (an special snapshot
created by relocation).

This patch will rename BTRFS_ROOT_REF_COWS to BTRFS_ROOT_SHAREABLE to
make it easier to understand, and update all comment mentioning
"reference counted" to follow the rename.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/backref.c     |  4 ++--
 fs/btrfs/backref.h     |  2 +-
 fs/btrfs/block-rsv.c   |  2 +-
 fs/btrfs/ctree.c       | 26 +++++++++++++-------------
 fs/btrfs/ctree.h       | 24 ++++++++++++++++++++++--
 fs/btrfs/disk-io.c     | 13 +++++++------
 fs/btrfs/extent-tree.c |  2 +-
 fs/btrfs/file.c        |  2 +-
 fs/btrfs/inode.c       | 18 ++++++++++--------
 fs/btrfs/ioctl.c       |  2 +-
 fs/btrfs/relocation.c  | 24 +++++++++++++-----------
 fs/btrfs/transaction.c | 12 ++++++------
 fs/btrfs/tree-defrag.c |  2 +-
 13 files changed, 79 insertions(+), 54 deletions(-)

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 60a69f7c0b36..f1ff92eb14f3 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -2702,7 +2702,7 @@ static int handle_indirect_tree_backref(struct btrfs_backref_cache *cache,
 	root = btrfs_get_fs_root(fs_info, &root_key, false);
 	if (IS_ERR(root))
 		return PTR_ERR(root);
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		cur->cowonly = 1;
 
 	if (btrfs_root_level(&root->root_item) == cur->level) {
@@ -2789,7 +2789,7 @@ static int handle_indirect_tree_backref(struct btrfs_backref_cache *cache,
 				goto out;
 			}
 			upper->owner = btrfs_header_owner(eb);
-			if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+			if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 				upper->cowonly = 1;
 
 			/*
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index 18393cc05ca2..ff705cc564a9 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -184,7 +184,7 @@ struct btrfs_backref_node {
 	struct extent_buffer *eb;
 	/* Level of the tree block */
 	unsigned int level:8;
-	/* Is the block in non-reference counted tree */
+	/* Is the block in a non-shareable tree */
 	unsigned int cowonly:1;
 	/* 1 if no child node is in the cache */
 	unsigned int lowest:1;
diff --git a/fs/btrfs/block-rsv.c b/fs/btrfs/block-rsv.c
index dbba53e712e6..7e1549a84fcc 100644
--- a/fs/btrfs/block-rsv.c
+++ b/fs/btrfs/block-rsv.c
@@ -458,7 +458,7 @@ static struct btrfs_block_rsv *get_block_rsv(
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_block_rsv *block_rsv = NULL;
 
-	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 	    (root == fs_info->csum_root && trans->adding_csums) ||
 	    (root == fs_info->uuid_root))
 		block_rsv = trans->block_rsv;
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 6c28efe5b14a..90a9e238ead8 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -144,9 +144,10 @@ struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
 	return eb;
 }
 
-/* cowonly root (everything not a reference counted cow subvolume), just get
- * put onto a simple dirty list.  transaction.c walks this to make sure they
- * get properly updated on disk.
+/*
+ * Cowonly root (not shareable trees, everything not subvolume roots or reloc
+ * roots), just get put onto a simple dirty list.  transaction.c walks this to
+ * make sure they get properly updated on disk.
  */
 static void add_root_to_dirty_list(struct btrfs_root *root)
 {
@@ -185,9 +186,9 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans,
 	int level;
 	struct btrfs_disk_key disk_key;
 
-	WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != fs_info->running_transaction->transid);
-	WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != root->last_trans);
 
 	level = btrfs_header_level(buf);
@@ -826,12 +827,11 @@ int btrfs_block_can_be_shared(struct btrfs_root *root,
 			      struct extent_buffer *buf)
 {
 	/*
-	 * Tree blocks not in reference counted trees and tree roots
-	 * are never shared. If a block was allocated after the last
-	 * snapshot and the block was not allocated by tree relocation,
-	 * we know the block is not shared.
+	 * Tree blocks not in shareable trees and tree roots are never shared.
+	 * If a block was allocated after the last snapshot and the block was
+	 * not allocated by tree relocation, we know the block is not shared.
 	 */
-	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 	    buf != root->node && buf != root->commit_root &&
 	    (btrfs_header_generation(buf) <=
 	     btrfs_root_last_snapshot(&root->root_item) ||
@@ -1024,9 +1024,9 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 
 	btrfs_assert_tree_locked(buf);
 
-	WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != fs_info->running_transaction->transid);
-	WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	WARN_ON(test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 		trans->transid != root->last_trans);
 
 	level = btrfs_header_level(buf);
@@ -1065,7 +1065,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) {
+	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state)) {
 		ret = btrfs_reloc_cow_block(trans, root, buf, cow);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 03ea7370aea7..65c09aea4cb9 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -969,7 +969,27 @@ enum {
 	 * is used to tell us when more checks are required
 	 */
 	BTRFS_ROOT_IN_TRANS_SETUP,
-	BTRFS_ROOT_REF_COWS,
+
+	/*
+	 * If tree blocks of this root can be shared between other roots.
+	 * Only subvolume trees and their reloc trees have this bit set.
+	 * Conflicts with TRACK_DIRTY bit.
+	 *
+	 * This affects two things:
+	 * - How balance works
+	 *   For shareable roots, btrfs needs to use reloc tree and do path
+	 *   replacement for balance, and needs various pre/post hooks for
+	 *   snapshot creation to handle them.
+	 *
+	 *   While for non shareable trees, we just simply do a tree search
+	 *   with COW.
+	 *
+	 * - How dirty roots are tracked
+	 *   For shareable roots, btrfs_record_root_in_trans() is needed to
+	 *   trace them, while non subvolume roots have TRACK_DIRTY bit, they
+	 *   don't need manual record.
+	 */
+	BTRFS_ROOT_SHAREABLE,
 	BTRFS_ROOT_TRACK_DIRTY,
 	BTRFS_ROOT_IN_RADIX,
 	BTRFS_ROOT_ORPHAN_ITEM_INSERTED,
@@ -1055,7 +1075,7 @@ struct btrfs_root {
 	struct btrfs_key defrag_progress;
 	struct btrfs_key defrag_max;
 
-	/* the dirty list is only used by non-reference counted roots */
+	/* the dirty list is only used by non-shareable roots */
 	struct list_head dirty_list;
 
 	struct list_head root_list;
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 8ad451695d49..76b165ad497f 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1275,12 +1275,13 @@ static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
 	root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
 
 	/*
-	 * DON'T set REF_COWS for log trees
+	 * DON'T set SHAREABLE bit for log trees.
 	 *
-	 * log trees do not get reference counted because they go away
-	 * before a real commit is actually done.  They do store pointers
-	 * to file data extents, and those reference counts still get
-	 * updated (along with back refs to the log tree).
+	 * Log trees are not exposed to user space thus can't be snapshotted,
+	 * and they go away before a real commit is actually done.
+	 *
+	 * They do store pointers to file data extents, and those reference
+	 * counts still get updated (along with back refs to the log tree).
 	 */
 
 	leaf = btrfs_alloc_tree_block(trans, root, 0, BTRFS_TREE_LOG_OBJECTID,
@@ -1419,7 +1420,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root)
 		goto fail;
 
 	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
-		set_bit(BTRFS_ROOT_REF_COWS, &root->state);
+		set_bit(BTRFS_ROOT_SHAREABLE, &root->state);
 		btrfs_check_and_init_root_item(&root->root_item);
 	}
 
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index faa585d54eb7..36d8014ce452 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -2442,7 +2442,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 	nritems = btrfs_header_nritems(buf);
 	level = btrfs_header_level(buf);
 
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state) && level == 0)
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state) && level == 0)
 		return 0;
 
 	if (full_backref)
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 719e68ab552c..606c2f3c1a38 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -775,7 +775,7 @@ int __btrfs_drop_extents(struct btrfs_trans_handle *trans,
 	if (start >= BTRFS_I(inode)->disk_i_size && !replace_extent)
 		modify_tree = 0;
 
-	update_refs = (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+	update_refs = (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 		       root == fs_info->tree_root);
 	while (1) {
 		recow = 0;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 5d567082f95a..a6c26c10ffc5 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -4107,11 +4107,13 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY);
 
 	/*
-	 * for non-free space inodes and ref cows, we want to back off from
-	 * time to time
+	 * for non-free space inodes and non-shareable roots, we want to back
+	 * off from time to time.
+	 * This means all inodes in subvolume roots, reloc roots, and data
+	 * reloc roots.
 	 */
 	if (!btrfs_is_free_space_inode(BTRFS_I(inode)) &&
-	    test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	    test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		be_nice = true;
 
 	path = btrfs_alloc_path();
@@ -4128,7 +4130,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	 * not block aligned since we will be keeping the last block of the
 	 * extent just the way it is.
 	 */
-	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 	    root == fs_info->tree_root)
 		btrfs_drop_extent_cache(BTRFS_I(inode), ALIGN(new_size,
 					fs_info->sectorsize),
@@ -4240,7 +4242,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 							 extent_num_bytes);
 				num_dec = (orig_num_bytes -
 					   extent_num_bytes);
-				if (test_bit(BTRFS_ROOT_REF_COWS,
+				if (test_bit(BTRFS_ROOT_SHAREABLE,
 					     &root->state) &&
 				    extent_start != 0)
 					inode_sub_bytes(inode, num_dec);
@@ -4256,7 +4258,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 				num_dec = btrfs_file_extent_num_bytes(leaf, fi);
 				if (extent_start != 0) {
 					found_extent = 1;
-					if (test_bit(BTRFS_ROOT_REF_COWS,
+					if (test_bit(BTRFS_ROOT_SHAREABLE,
 						     &root->state))
 						inode_sub_bytes(inode, num_dec);
 				}
@@ -4292,7 +4294,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 				clear_len = fs_info->sectorsize;
 			}
 
-			if (test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+			if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 				inode_sub_bytes(inode, item_end + 1 - new_size);
 		}
 delete:
@@ -4333,7 +4335,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 		should_throttle = false;
 
 		if (found_extent &&
-		    (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+		    (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 		     root == fs_info->tree_root)) {
 			struct btrfs_ref ref = { 0 };
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 40b729dce91c..709d9446896a 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -750,7 +750,7 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	int ret;
 	bool snapshot_force_cow = false;
 
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		return -EINVAL;
 
 	if (atomic_read(&root->nr_swapfiles)) {
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index f25deca18a5d..437b782c57e6 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -321,7 +321,7 @@ int btrfs_should_ignore_reloc_root(struct btrfs_root *root)
 {
 	struct btrfs_root *reloc_root;
 
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		return 0;
 
 	/* This root has been merged with its reloc tree, we can ignore it */
@@ -808,7 +808,7 @@ static struct btrfs_root *create_reloc_root(struct btrfs_trans_handle *trans,
 
 	reloc_root = btrfs_read_tree_root(fs_info->tree_root, &root_key);
 	BUG_ON(IS_ERR(reloc_root));
-	set_bit(BTRFS_ROOT_REF_COWS, &reloc_root->state);
+	set_bit(BTRFS_ROOT_SHAREABLE, &reloc_root->state);
 	reloc_root->last_trans = trans->transid;
 	return reloc_root;
 }
@@ -2018,7 +2018,7 @@ struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
 		next = walk_up_backref(next, edges, &index);
 		root = next->root;
 		BUG_ON(!root);
-		BUG_ON(!test_bit(BTRFS_ROOT_REF_COWS, &root->state));
+		BUG_ON(!test_bit(BTRFS_ROOT_SHAREABLE, &root->state));
 
 		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
 			record_reloc_root_in_trans(trans, root);
@@ -2062,10 +2062,12 @@ struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
 }
 
 /*
- * select a tree root for relocation. return NULL if the block
- * is reference counted. we should use do_relocation() in this
- * case. return a tree root pointer if the block isn't reference
- * counted. return -ENOENT if the block is root of reloc tree.
+ * Select a tree root for relocation.
+ *
+ * Return NULL if the block is not shareable. we should use do_relocation() in this
+ * case.
+ * Return a tree root pointer if the block is shareable.
+ * Return -ENOENT if the block is root of reloc tree.
  */
 static noinline_for_stack
 struct btrfs_root *select_one_root(struct btrfs_backref_node *node)
@@ -2083,8 +2085,8 @@ struct btrfs_root *select_one_root(struct btrfs_backref_node *node)
 		root = next->root;
 		BUG_ON(!root);
 
-		/* no other choice for non-references counted tree */
-		if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+		/* no other choice for non-shareable tree */
+		if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 			return root;
 
 		if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID)
@@ -2480,7 +2482,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans,
 	}
 
 	if (root) {
-		if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) {
+		if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state)) {
 			BUG_ON(node->new_bytenr);
 			BUG_ON(!list_empty(&node->list));
 			btrfs_record_root_in_trans(trans, root);
@@ -3765,7 +3767,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 			goto out;
 		}
 
-		set_bit(BTRFS_ROOT_REF_COWS, &reloc_root->state);
+		set_bit(BTRFS_ROOT_SHAREABLE, &reloc_root->state);
 		list_add(&reloc_root->root_list, &reloc_roots);
 
 		if (btrfs_root_refs(&reloc_root->root_item) > 0) {
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 96eb313a5080..a04b5442ac39 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -349,10 +349,10 @@ static noinline int join_transaction(struct btrfs_fs_info *fs_info,
 }
 
 /*
- * this does all the record keeping required to make sure that a reference
- * counted root is properly recorded in a given transaction.  This is required
+ * This does all the record keeping required to make sure that a shareable
+ * root is properly recorded in a given transaction.  This is required
  * to make sure the old root from before we joined the transaction is deleted
- * when the transaction commits
+ * when the transaction commits.
  */
 static int record_root_in_trans(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
@@ -360,7 +360,7 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	if ((test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
+	if ((test_bit(BTRFS_ROOT_SHAREABLE, &root->state) &&
 	    root->last_trans < trans->transid) || force) {
 		WARN_ON(root == fs_info->extent_root);
 		WARN_ON(!force && root->commit_root != root->node);
@@ -439,7 +439,7 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		return 0;
 
 	/*
@@ -504,7 +504,7 @@ static inline bool need_reserve_reloc_root(struct btrfs_root *root)
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	if (!fs_info->reloc_ctl ||
-	    !test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+	    !test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
 	    root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
 	    root->reloc_root)
 		return false;
diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c
index 5f9e2dd413af..16c3a6d2586d 100644
--- a/fs/btrfs/tree-defrag.c
+++ b/fs/btrfs/tree-defrag.c
@@ -35,7 +35,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 		goto out;
 	}
 
-	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
 		goto out;
 
 	path = btrfs_alloc_path();
-- 
2.26.2

[PATCH v3 2/3] btrfs: inode: Cleanup the log tree exceptions in btrfs_truncate_inode_items()

[Qu Wenruo]

There are a lot of root owner check in btrfs_truncate_inode_items()
like:

	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
	    root == fs_info->tree_root)

But considering that, there are only those trees can have INODE_ITEMs:
- tree root (For v1 space cache)
- subvolume trees
- tree reloc trees
- data reloc tree
- log trees

And since subvolume/tree reloc/data reloc trees all have SHAREABLE bit,
and we're checking tree root manually, so above check is just excluding
log trees.

This patch will replace two of such checks to a much simpler one:

	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)

This would merge btrfs_drop_extent_cache() and lock_extent_bits() call
into the same if branch.

Finally replace ALIGN() with round_up(), as I'm always bad at determing
the alignement direction.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/inode.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index a6c26c10ffc5..ae6b47edabc5 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -4121,20 +4121,18 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 	path->reada = READA_BACK;
 
-	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
+		/*
+		 * We want to drop from the next block forward in case this
+		 * new size is not block aligned since we will be keeping the
+		 * last block of the extent just the way it is.
+		 */
 		lock_extent_bits(&BTRFS_I(inode)->io_tree, lock_start, (u64)-1,
 				 &cached_state);
-
-	/*
-	 * We want to drop from the next block forward in case this new size is
-	 * not block aligned since we will be keeping the last block of the
-	 * extent just the way it is.
-	 */
-	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
-	    root == fs_info->tree_root)
-		btrfs_drop_extent_cache(BTRFS_I(inode), ALIGN(new_size,
+		btrfs_drop_extent_cache(BTRFS_I(inode), round_up(new_size,
 					fs_info->sectorsize),
 					(u64)-1, 0);
+	}
 
 	/*
 	 * This function is also used to drop the items in the log tree before
@@ -4335,8 +4333,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 		should_throttle = false;
 
 		if (found_extent &&
-		    (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
-		     root == fs_info->tree_root)) {
+		    root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
 			struct btrfs_ref ref = { 0 };
 
 			bytes_deleted += extent_num_bytes;
-- 
2.26.2

[PATCH v3 3/3] btrfs: Don't set SHAREABLE flag for data reloc tree

[Qu Wenruo]

SHAREABLE flag is set for subvolumes because users can create snapshot
for subvolumes, thus sharing tree blocks of them.

But data reloc tree are not exposed to user space, as it's only an
internal tree for data relocation, thus it doesn't need the full path
replacement treat at all.

This patch will make data reloc tree just a non-shareable tree, and add
btrfs_fs_info::data_reloc_root for data reloc tree, so relocation code can
grab it from fs_info directly.

This would slightly improve tree relocation, as now data reloc tree
can go through regular COW routine to get relocated, without bothering
the complex tree reloc tree routine.

Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/ctree.h      |  1 +
 fs/btrfs/disk-io.c    | 14 +++++++++++++-
 fs/btrfs/relocation.c | 17 ++++++-----------
 3 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 65c09aea4cb9..d3be34316677 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -582,6 +582,7 @@ struct btrfs_fs_info {
 	struct btrfs_root *quota_root;
 	struct btrfs_root *uuid_root;
 	struct btrfs_root *free_space_root;
+	struct btrfs_root *data_reloc_root;
 
 	/* the log root tree is a directory of all the other log roots */
 	struct btrfs_root *log_root_tree;
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 76b165ad497f..fced949b150c 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1419,7 +1419,8 @@ static int btrfs_init_fs_root(struct btrfs_root *root)
 	if (ret)
 		goto fail;
 
-	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID &&
+	    root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) {
 		set_bit(BTRFS_ROOT_SHAREABLE, &root->state);
 		btrfs_check_and_init_root_item(&root->root_item);
 	}
@@ -1525,6 +1526,7 @@ void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
 	btrfs_put_root(fs_info->uuid_root);
 	btrfs_put_root(fs_info->free_space_root);
 	btrfs_put_root(fs_info->fs_root);
+	btrfs_put_root(fs_info->data_reloc_root);
 	btrfs_check_leaked_roots(fs_info);
 	btrfs_extent_buffer_leak_debug_check(fs_info);
 	kfree(fs_info->super_copy);
@@ -1981,6 +1983,7 @@ static void free_root_pointers(struct btrfs_fs_info *info, bool free_chunk_root)
 	free_root_extent_buffers(info->quota_root);
 	free_root_extent_buffers(info->uuid_root);
 	free_root_extent_buffers(info->fs_root);
+	free_root_extent_buffers(info->data_reloc_root);
 	if (free_chunk_root)
 		free_root_extent_buffers(info->chunk_root);
 	free_root_extent_buffers(info->free_space_root);
@@ -2287,6 +2290,15 @@ static int btrfs_read_roots(struct btrfs_fs_info *fs_info)
 	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
 	fs_info->csum_root = root;
 
+	location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
+	root = btrfs_get_fs_root(fs_info, &location, true);
+	if (IS_ERR(root)) {
+		ret = PTR_ERR(root);
+		goto out;
+	}
+	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
+	fs_info->data_reloc_root = root;
+
 	location.objectid = BTRFS_QUOTA_TREE_OBJECTID;
 	root = btrfs_read_tree_root(tree_root, &location);
 	if (!IS_ERR(root)) {
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 437b782c57e6..9afc1a6928cf 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3470,14 +3470,12 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 {
 	struct inode *inode = NULL;
 	struct btrfs_trans_handle *trans;
-	struct btrfs_root *root;
+	struct btrfs_root *root = btrfs_grab_root(fs_info->data_reloc_root);
 	struct btrfs_key key;
 	u64 objectid;
 	int err = 0;
 
-	root = read_fs_root(fs_info, BTRFS_DATA_RELOC_TREE_OBJECTID);
-	if (IS_ERR(root))
-		return ERR_CAST(root);
+	ASSERT(root);
 
 	trans = btrfs_start_transaction(root, 6);
 	if (IS_ERR(trans)) {
@@ -3870,13 +3868,10 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 	if (err == 0) {
 		/* cleanup orphan inode in data relocation tree */
-		fs_root = read_fs_root(fs_info, BTRFS_DATA_RELOC_TREE_OBJECTID);
-		if (IS_ERR(fs_root)) {
-			err = PTR_ERR(fs_root);
-		} else {
-			err = btrfs_orphan_cleanup(fs_root);
-			btrfs_put_root(fs_root);
-		}
+		fs_root = btrfs_grab_root(fs_info->data_reloc_root);
+		ASSERT(fs_root);
+		err = btrfs_orphan_cleanup(fs_root);
+		btrfs_put_root(fs_root);
 	}
 	return err;
 }
-- 
2.26.2

Re: [PATCH v3 3/3] btrfs: Don't set SHAREABLE flag for data reloc tree

[David Sterba]

On Fri, May 15, 2020 at 02:01:42PM +0800, Qu Wenruo wrote:
> @@ -1525,6 +1526,7 @@ void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
>  	btrfs_put_root(fs_info->uuid_root);
>  	btrfs_put_root(fs_info->free_space_root);
>  	btrfs_put_root(fs_info->fs_root);
> +	btrfs_put_root(fs_info->data_reloc_root);
>  	btrfs_check_leaked_roots(fs_info);
>  	btrfs_extent_buffer_leak_debug_check(fs_info);
>  	kfree(fs_info->super_copy);

> +	location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
> +	root = btrfs_get_fs_root(fs_info, &location, true);
> +	if (IS_ERR(root)) {
> +		ret = PTR_ERR(root);
> +		goto out;
> +	}
> +	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
> +	fs_info->data_reloc_root = root;

I've read the code more carefully, the data reloc tree needs to be read
as the others, like fs_tree as you said before. A new tree has the
reference count set to 1, which is what I missed before, so calling
btrfs_get_fs_root would set it to 2 and then btrfs_free_fs_info won't
free it as it expects refs == 1. Sorry for misleading you.

> @@ -3470,14 +3470,12 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
>  {
>  	struct inode *inode = NULL;
>  	struct btrfs_trans_handle *trans;
> -	struct btrfs_root *root;
> +	struct btrfs_root *root = btrfs_grab_root(fs_info->data_reloc_root);

Which means you can use the pointer directly as you had in previous
versions.

>  	struct btrfs_key key;
>  	u64 objectid;
>  	int err = 0;
>  
> -	root = read_fs_root(fs_info, BTRFS_DATA_RELOC_TREE_OBJECTID);
> -	if (IS_ERR(root))
> -		return ERR_CAST(root);
> +	ASSERT(root);
>  
>  	trans = btrfs_start_transaction(root, 6);
>  	if (IS_ERR(trans)) {
> @@ -3870,13 +3868,10 @@ int btrfs_recover_relocation(struct btrfs_root *root)
>  
>  	if (err == 0) {
>  		/* cleanup orphan inode in data relocation tree */
> -		fs_root = read_fs_root(fs_info, BTRFS_DATA_RELOC_TREE_OBJECTID);
> -		if (IS_ERR(fs_root)) {
> -			err = PTR_ERR(fs_root);
> -		} else {
> -			err = btrfs_orphan_cleanup(fs_root);
> -			btrfs_put_root(fs_root);
> -		}
> +		fs_root = btrfs_grab_root(fs_info->data_reloc_root);
> +		ASSERT(fs_root);
> +		err = btrfs_orphan_cleanup(fs_root);
> +		btrfs_put_root(fs_root);

Here it's fine to grab/put so it's clear the tree is in use but it's
merely for clarity.

Re: [PATCH v3 2/3] btrfs: inode: Cleanup the log tree exceptions in btrfs_truncate_inode_items()

[David Sterba]

On Fri, May 15, 2020 at 02:01:41PM +0800, Qu Wenruo wrote:
> There are a lot of root owner check in btrfs_truncate_inode_items()
> like:
> 
> 	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
> 	    root == fs_info->tree_root)
> 
> But considering that, there are only those trees can have INODE_ITEMs:
> - tree root (For v1 space cache)
> - subvolume trees
> - tree reloc trees
> - data reloc tree
> - log trees
> 
> And since subvolume/tree reloc/data reloc trees all have SHAREABLE bit,
> and we're checking tree root manually, so above check is just excluding
> log trees.
> 
> This patch will replace two of such checks to a much simpler one:
> 
> 	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
> 
> This would merge btrfs_drop_extent_cache() and lock_extent_bits() call
> into the same if branch.
> 
> Finally replace ALIGN() with round_up(), as I'm always bad at determing
> the alignement direction.

Alert

> 
> Signed-off-by: Qu Wenruo <wqu@suse.com>
> ---
>  fs/btrfs/inode.c | 21 +++++++++------------
>  1 file changed, 9 insertions(+), 12 deletions(-)
> 
> diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
> index a6c26c10ffc5..ae6b47edabc5 100644
> --- a/fs/btrfs/inode.c
> +++ b/fs/btrfs/inode.c
> @@ -4121,20 +4121,18 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
>  		return -ENOMEM;
>  	path->reada = READA_BACK;
>  
> -	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
> +	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
> +		/*
> +		 * We want to drop from the next block forward in case this
> +		 * new size is not block aligned since we will be keeping the
> +		 * last block of the extent just the way it is.
> +		 */

The comment is misplaced, it belongs to the 'drop extent cache' part

>  		lock_extent_bits(&BTRFS_I(inode)->io_tree, lock_start, (u64)-1,
>  				 &cached_state);
> -
> -	/*
> -	 * We want to drop from the next block forward in case this new size is
> -	 * not block aligned since we will be keeping the last block of the
> -	 * extent just the way it is.
> -	 */
> -	if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state) ||
> -	    root == fs_info->tree_root)
> -		btrfs_drop_extent_cache(BTRFS_I(inode), ALIGN(new_size,

here.

Regarding ALIGN -> round_up, please don't put such unrelated changes to
one patch.

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[David Sterba]

On Fri, May 15, 2020 at 02:01:39PM +0800, Qu Wenruo wrote:
> This small patchset reworks the REF_COWS bit, by renaming it, and remove
> that bit for data relocation root.
> 
> 
> The basic idea of such rework is to reduce the confusion caused by the
> name REF_COWS.
> 
> With the new bit called SHAREABLE, it should be clear that no user can
> really create snapshot for data reloc tree, thus its tree blocks
> shouldn't be shareable.
> 
> This would make data balance for reloc tree a little simpler.
> 
> Changelog:
> v2:
> - Add new patch to address the log tree check in
>   btrfs_truncate_inode_items()
>   Thanks for the advice from David, now it's much simpler than original
>   check, and data reloc tree no longer needs extra hanlding
> 
> - Grab data reloc root in create_reloc_inode() and
>   btrfs_recover_relocation()
> 
> - Comment update
> 
> v3:
> - Remove ALIGN_DOWN() -> round_down() change
> 
> - Remove the confusing comment on the log tree inode

I've added the patches to misc-next, with some fixups. I'll let it also
go through fstests, but a quick run has hit this write-time corruption
very early. I haven't analyzed it and it's possible that it's caused by
my fixups.

btrfs/003		[19:39:32][   79.959953] run fstests btrfs/003 at 2020-05-15 19:39:32
[   80.527008] BTRFS info (device vda): disk space caching is enabled
[   80.530862] BTRFS info (device vda): has skinny extents
[   81.588702] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (11353)
[   81.598929] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (11353)
[   81.607316] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (11353)
[   81.616304] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (11353)
[   81.625220] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (11353)
[   81.633532] BTRFS: device fsid 168518c8-d58c-40bb-b921-1853b5f43e13 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (11353)
[   81.677123] BTRFS info (device vdb): disk space caching is enabled
[   81.680643] BTRFS info (device vdb): has skinny extents
[   81.682275] BTRFS info (device vdb): flagging fs with big metadata feature
[   81.701692] BTRFS info (device vdb): checking UUID tree
[   89.421089] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (12984)
[   89.426854] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (12984)
[   89.433740] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (12984)
[   89.442686] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (12984)
[   89.451243] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (12984)
[   89.459020] BTRFS: device fsid 62194f9a-4bb9-4f11-9d72-34bb627770a8 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (12984)
[   89.499238] BTRFS info (device vdb): disk space caching is enabled
[   89.501868] BTRFS info (device vdb): has skinny extents
[   89.503909] BTRFS info (device vdb): flagging fs with big metadata feature
[   89.519384] BTRFS info (device vdb): checking UUID tree
[   97.219663] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (14616)
[   97.228409] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (14616)
[   97.233077] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (14616)
[   97.238235] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (14616)
[   97.246465] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (14616)
[   97.251105] BTRFS: device fsid 18c8f856-5d32-44e3-8a97-f898c01c4806 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (14616)
[   97.291084] BTRFS info (device vdb): disk space caching is enabled
[   97.293793] BTRFS info (device vdb): has skinny extents
[   97.295700] BTRFS info (device vdb): flagging fs with big metadata feature
[   97.317883] BTRFS info (device vdb): checking UUID tree
[  104.896913] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (16207)
[  104.904143] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (16207)
[  104.912371] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (16207)
[  104.916558] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (16207)
[  104.922823] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (16207)
[  104.927640] BTRFS: device fsid 0860e370-740c-4f73-a631-8155b52a2a9c devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (16207)
[  104.968049] BTRFS info (device vdb): disk space caching is enabled
[  104.972725] BTRFS info (device vdb): has skinny extents
[  104.975199] BTRFS info (device vdb): flagging fs with big metadata feature
[  104.994401] BTRFS info (device vdb): checking UUID tree
[  112.217591] BTRFS: device fsid 25db6680-c72d-494b-b8db-e9719d8dfd36 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (17799)
[  112.293518] BTRFS info (device vdb): disk space caching is enabled
[  112.296310] BTRFS info (device vdb): has skinny extents
[  112.298225] BTRFS info (device vdb): flagging fs with big metadata feature
[  112.320796] BTRFS info (device vdb): checking UUID tree
[  119.314197] BTRFS info (device vdb): disk added /dev/vdd
[  119.411823] BTRFS info (device vdb): disk added /dev/vde
[  119.491205] BTRFS info (device vdb): disk added /dev/vdf
[  119.585185] BTRFS info (device vdb): disk added /dev/vdg
[  119.624572] BTRFS info (device vdb): balance: start -d -m -s
[  119.630843] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
[  119.640113] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=298909696 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
[  119.647511] BTRFS info (device vdb): leaf 298909696 gen 11 total ptrs 4 free space 15851 owner 18446744073709551607
[  119.652214] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 19404
[  119.656275] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
[  119.658436] 		inode generation 1 size 0 mode 100600
[  119.660812] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
[  119.663645] 		inode generation 4 size 0 mode 40755
[  119.665768] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
[  119.668146] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
[  119.671438] BTRFS error (device vdb): block=298909696 write time tree block corruption detected
[  119.675319] ------------[ cut here ]------------
[  119.677789] WARNING: CPU: 1 PID: 19404 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
[  119.683116] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
[  119.690407] CPU: 1 PID: 19404 Comm: btrfs Not tainted 5.7.0-rc5-default+ #1108
[  119.693724] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
[  119.698411] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
[  119.700653] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
[  119.706847] RSP: 0018:ffff9ec884397778 EFLAGS: 00010292
[  119.707997] RAX: 0000000000000000 RBX: ffff902c7faf0dc0 RCX: 0000000000000006
[  119.709428] RDX: 0000000000000000 RSI: ffff902c4fcd5dd0 RDI: ffff902c4fcd5500
[  119.711051] RBP: 0000000000000001 R08: 0000001bdd341101 R09: 0000000000000000
[  119.712587] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c74b3c000
[  119.714957] R13: ffff902c428a16b0 R14: 0000000000000000 R15: 00000000ffffff8b
[  119.716474] FS:  00007fdc12d658c0(0000) GS:ffff902c7b800000(0000) knlGS:0000000000000000
[  119.719045] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  119.720724] CR2: 000055a44869a2d8 CR3: 00000001367c2003 CR4: 0000000000160ee0
[  119.725194] Call Trace:
[  119.727333]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
[  119.730365]  submit_one_bio+0x2b/0x40 [btrfs]
[  119.732253]  btree_write_cache_pages+0x373/0x440 [btrfs]
[  119.734909]  do_writepages+0x40/0xe0
[  119.736667]  ? do_raw_spin_unlock+0x4b/0xc0
[  119.738667]  ? _raw_spin_unlock+0x1f/0x30
[  119.740122]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
[  119.743009]  __filemap_fdatawrite_range+0xce/0x110
[  119.744446]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
[  119.746166]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
[  119.748184]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
[  119.750072]  ? start_transaction+0xd2/0x5e0 [btrfs]
[  119.751896]  prepare_to_relocate+0x107/0x130 [btrfs]
[  119.753666]  relocate_block_group+0x5b/0x600 [btrfs]
[  119.755321]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
[  119.757111]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
[  119.758833]  __btrfs_balance+0x41c/0xcc0 [btrfs]
[  119.760526]  btrfs_balance+0x65b/0xbd0 [btrfs]
[  119.762050]  ? kmem_cache_alloc_trace+0x1a7/0x320
[  119.763777]  ? btrfs_ioctl_balance+0x21c/0x350 [btrfs]
[  119.765593]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
[  119.767444]  ? __handle_mm_fault+0x499/0x740
[  119.769007]  btrfs_ioctl+0x304/0x2590 [btrfs]
[  119.770537]  ? do_raw_spin_unlock+0x4b/0xc0
[  119.772015]  ? _raw_spin_unlock+0x1f/0x30
[  119.774671]  ? __handle_mm_fault+0x499/0x740
[  119.777106]  ? do_user_addr_fault+0x1d8/0x3f0
[  119.779366]  ? kvm_sched_clock_read+0x14/0x30
[  119.781451]  ? sched_clock+0x5/0x10
[  119.783584]  ? sched_clock_cpu+0x15/0x130
[  119.785507]  ? do_user_addr_fault+0x1d8/0x3f0
[  119.787587]  ? ksys_ioctl+0x68/0xa0
[  119.789577]  ksys_ioctl+0x68/0xa0
[  119.790909]  __x64_sys_ioctl+0x16/0x20
[  119.792361]  do_syscall_64+0x50/0x210
[  119.793749]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
[  119.795547] RIP: 0033:0x7fdc12e5e227
[  119.796967] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
[  119.802957] RSP: 002b:00007ffc3c0bf7b8 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
[  119.805509] RAX: ffffffffffffffda RBX: 00007ffc3c0bf860 RCX: 00007fdc12e5e227
[  119.807832] RDX: 00007ffc3c0bf860 RSI: 00000000c4009420 RDI: 0000000000000003
[  119.810088] RBP: 0000000000000003 R08: 000055a4486922a0 R09: 0000000000000231
[  119.812378] R10: 00007fdc13088cf0 R11: 0000000000000206 R12: 0000000000000001
[  119.814654] R13: 00007ffc3c0c2127 R14: 0000000000000002 R15: 0000000000000000
[  119.816912] irq event stamp: 19394
[  119.818225] hardirqs last  enabled at (19393): [<ffffffff8a10a166>] console_unlock+0x436/0x590
[  119.821106] hardirqs last disabled at (19394): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
[  119.824311] softirqs last  enabled at (19390): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
[  119.827257] softirqs last disabled at (19383): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
[  119.830080] ---[ end trace 247639532e5b557e ]---
[  119.832171] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
[  119.836020] BTRFS info (device vdb): forced readonly
[  119.837888] BTRFS warning (device vdb): Skipping commit of aborted transaction.
[  119.841381] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
[  119.845600] BTRFS info (device vdb): balance: ended with status: -30
[  120.282656] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (19409)
[  120.287896] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (19409)
[  120.298518] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (19409)
[  120.303545] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (19409)
[  120.311101] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (19409)
[  120.314834] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (19409)
[  120.349393] BTRFS info (device vdb): disk space caching is enabled
[  120.353340] BTRFS info (device vdb): has skinny extents
[  120.355944] BTRFS info (device vdb): flagging fs with big metadata feature
[  120.377592] BTRFS info (device vdb): checking UUID tree
[  127.395164] BTRFS info (device vdb): relocating block group 1104150528 flags data|raid0
[  127.448678] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=30769152 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
[  127.458117] BTRFS info (device vdb): leaf 30769152 gen 7 total ptrs 4 free space 15851 owner 18446744073709551607
[  127.462400] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 21002
[  127.466146] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
[  127.468412] 		inode generation 1 size 0 mode 100600
[  127.470284] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
[  127.472282] 		inode generation 4 size 0 mode 40755
[  127.473848] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
[  127.475601] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
[  127.477585] BTRFS error (device vdb): block=30769152 write time tree block corruption detected
[  127.480577] ------------[ cut here ]------------
[  127.482417] WARNING: CPU: 3 PID: 21002 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
[  127.486001] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
[  127.491491] CPU: 3 PID: 21002 Comm: btrfs Tainted: G        W         5.7.0-rc5-default+ #1108
[  127.494777] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
[  127.498791] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
[  127.500882] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
[  127.507102] RSP: 0018:ffff9ec8852ff700 EFLAGS: 00010282
[  127.509172] RAX: 0000000000000000 RBX: ffff902c7f20c680 RCX: 0000000000000006
[  127.511742] RDX: 0000000000000000 RSI: ffff902c4fd28910 RDI: ffff902c4fd28040
[  127.514884] RBP: 0000000000000005 R08: 0000001dae6ee2ba R09: 0000000000000000
[  127.517380] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c4a994000
[  127.519327] R13: ffff902c428a1ef0 R14: 0000000000000000 R15: 00000000ffffff8b
[  127.523221] FS:  00007f69ecb3b8c0(0000) GS:ffff902c7bc00000(0000) knlGS:0000000000000000
[  127.527813] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  127.529913] CR2: 00007f69ecca0af0 CR3: 000000010579c002 CR4: 0000000000160ee0
[  127.531942] Call Trace:
[  127.533156]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
[  127.535744]  submit_one_bio+0x2b/0x40 [btrfs]
[  127.537542]  submit_extent_page+0x104/0x210 [btrfs]
[  127.540137]  write_one_eb+0x1b1/0x390 [btrfs]
[  127.542825]  ? find_first_extent_bit_state+0x90/0x90 [btrfs]
[  127.559976]  btree_write_cache_pages+0x1af/0x440 [btrfs]
[  127.562025]  do_writepages+0x40/0xe0
[  127.563813]  ? do_raw_spin_unlock+0x4b/0xc0
[  127.565541]  ? _raw_spin_unlock+0x1f/0x30
[  127.567363]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
[  127.569545]  __filemap_fdatawrite_range+0xce/0x110
[  127.571519]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
[  127.574396]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
[  127.576877]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
[  127.579328]  ? start_transaction+0xd2/0x5e0 [btrfs]
[  127.581449]  prepare_to_relocate+0x107/0x130 [btrfs]
[  127.583958]  relocate_block_group+0x5b/0x600 [btrfs]
[  127.586867]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
[  127.588999]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
[  127.590844]  btrfs_shrink_device+0x214/0x530 [btrfs]
[  127.592345]  btrfs_rm_device+0x22e/0x7f0 [btrfs]
[  127.593913]  ? _copy_from_user+0x6a/0xa0
[  127.595367]  btrfs_ioctl+0x218f/0x2590 [btrfs]
[  127.596993]  ? __handle_mm_fault+0x1c1/0x740
[  127.598537]  ? do_user_addr_fault+0x1d8/0x3f0
[  127.600139]  ? kvm_sched_clock_read+0x14/0x30
[  127.601713]  ? sched_clock+0x5/0x10
[  127.603121]  ? sched_clock_cpu+0x15/0x130
[  127.604840]  ? do_user_addr_fault+0x1d8/0x3f0
[  127.606740]  ? ksys_ioctl+0x68/0xa0
[  127.608092]  ksys_ioctl+0x68/0xa0
[  127.609434]  __x64_sys_ioctl+0x16/0x20
[  127.610860]  do_syscall_64+0x50/0x210
[  127.612294]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
[  127.614042] RIP: 0033:0x7f69ecc34227
[  127.615464] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
[  127.621798] RSP: 002b:00007ffc4c0a4e08 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[  127.624742] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f69ecc34227
[  127.627020] RDX: 00007ffc4c0a4e30 RSI: 000000005000943a RDI: 0000000000000003
[  127.629398] RBP: 00007ffc4c0a6fd0 R08: 00007ffc4c0a4e68 R09: 006764762f766564
[  127.631559] R10: 00007f69ece5ecf0 R11: 0000000000000202 R12: 0000000000000000
[  127.634138] R13: 00007ffc4c0a4e30 R14: 000056046c20be8c R15: 0000000000000003
[  127.635759] irq event stamp: 108944
[  127.637314] hardirqs last  enabled at (108943): [<ffffffff8a10a166>] console_unlock+0x436/0x590
[  127.640821] hardirqs last disabled at (108944): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
[  127.644549] softirqs last  enabled at (108940): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
[  127.647486] softirqs last disabled at (108933): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
[  127.650608] ---[ end trace 247639532e5b557f ]---
[  127.653165] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
[  127.657655] BTRFS info (device vdb): forced readonly
[  127.659739] BTRFS warning (device vdb): Skipping commit of aborted transaction.
[  127.663588] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
[failed, exit status 1] [19:40:22]- output mismatch (see /tmp/fstests/results//btrfs/003.out.bad)
    --- tests/btrfs/003.out	2018-04-12 16:57:00.608225550 +0000
    +++ /tmp/fstests/results//btrfs/003.out.bad	2020-05-15 19:40:22.176000000 +0000
    @@ -1,2 +1,6 @@
     QA output created by 003
    -Silence is golden
    +ERROR: error during balancing '/tmp/scratch': Read-only file system
    +There may be more info in syslog - try dmesg | tail
    +ERROR: error removing device '/dev/vdg': Read-only file system
    +btrfs device delete failed
    +(see /tmp/fstests/results//btrfs/003.full for details)
    ...
    (Run 'diff -u /tmp/fstests/tests/btrfs/003.out /tmp/fstests/results//btrfs/003.out.bad'  to see the entire diff)

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[Qu Wenruo]

[-- Attachment #1.1: Type: text/plain, Size: 16088 bytes --]



On 2020/5/16 上午3:45, David Sterba wrote:
> On Fri, May 15, 2020 at 02:01:39PM +0800, Qu Wenruo wrote:
>> This small patchset reworks the REF_COWS bit, by renaming it, and remove
>> that bit for data relocation root.
>>
>>
>> The basic idea of such rework is to reduce the confusion caused by the
>> name REF_COWS.
>>
>> With the new bit called SHAREABLE, it should be clear that no user can
>> really create snapshot for data reloc tree, thus its tree blocks
>> shouldn't be shareable.
>>
>> This would make data balance for reloc tree a little simpler.
>>
>> Changelog:
>> v2:
>> - Add new patch to address the log tree check in
>>   btrfs_truncate_inode_items()
>>   Thanks for the advice from David, now it's much simpler than original
>>   check, and data reloc tree no longer needs extra hanlding
>>
>> - Grab data reloc root in create_reloc_inode() and
>>   btrfs_recover_relocation()
>>
>> - Comment update
>>
>> v3:
>> - Remove ALIGN_DOWN() -> round_down() change
>>
>> - Remove the confusing comment on the log tree inode
> 
> I've added the patches to misc-next, with some fixups. I'll let it also
> go through fstests, but a quick run has hit this write-time corruption
> very early. I haven't analyzed it and it's possible that it's caused by
> my fixups.

Passed my local balance group.

And furthermore, the error pattern is exactly what I saw during my
development.

> 
[...]
> [  119.624572] BTRFS info (device vdb): balance: start -d -m -s
> [  119.630843] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
> [  119.640113] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=298909696 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
> [  119.647511] BTRFS info (device vdb): leaf 298909696 gen 11 total ptrs 4 free space 15851 owner 18446744073709551607
> [  119.652214] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 19404
> [  119.656275] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
> [  119.658436] 		inode generation 1 size 0 mode 100600

This is using 1 as ino number, which means root::highest_objectid is not
properly initialized.

This happened when I'm using btrfs_read_tree_root() other than
btrfs_read_fs_root(), which initializes root::highest_objectid.

So I guess there is something wrong happened during the fixup.

Thanks,
Qu

> [  119.660812] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
> [  119.663645] 		inode generation 4 size 0 mode 40755
> [  119.665768] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
> [  119.668146] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
> [  119.671438] BTRFS error (device vdb): block=298909696 write time tree block corruption detected
> [  119.675319] ------------[ cut here ]------------
> [  119.677789] WARNING: CPU: 1 PID: 19404 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
> [  119.683116] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
> [  119.690407] CPU: 1 PID: 19404 Comm: btrfs Not tainted 5.7.0-rc5-default+ #1108
> [  119.693724] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
> [  119.698411] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
> [  119.700653] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
> [  119.706847] RSP: 0018:ffff9ec884397778 EFLAGS: 00010292
> [  119.707997] RAX: 0000000000000000 RBX: ffff902c7faf0dc0 RCX: 0000000000000006
> [  119.709428] RDX: 0000000000000000 RSI: ffff902c4fcd5dd0 RDI: ffff902c4fcd5500
> [  119.711051] RBP: 0000000000000001 R08: 0000001bdd341101 R09: 0000000000000000
> [  119.712587] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c74b3c000
> [  119.714957] R13: ffff902c428a16b0 R14: 0000000000000000 R15: 00000000ffffff8b
> [  119.716474] FS:  00007fdc12d658c0(0000) GS:ffff902c7b800000(0000) knlGS:0000000000000000
> [  119.719045] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  119.720724] CR2: 000055a44869a2d8 CR3: 00000001367c2003 CR4: 0000000000160ee0
> [  119.725194] Call Trace:
> [  119.727333]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
> [  119.730365]  submit_one_bio+0x2b/0x40 [btrfs]
> [  119.732253]  btree_write_cache_pages+0x373/0x440 [btrfs]
> [  119.734909]  do_writepages+0x40/0xe0
> [  119.736667]  ? do_raw_spin_unlock+0x4b/0xc0
> [  119.738667]  ? _raw_spin_unlock+0x1f/0x30
> [  119.740122]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
> [  119.743009]  __filemap_fdatawrite_range+0xce/0x110
> [  119.744446]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
> [  119.746166]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
> [  119.748184]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
> [  119.750072]  ? start_transaction+0xd2/0x5e0 [btrfs]
> [  119.751896]  prepare_to_relocate+0x107/0x130 [btrfs]
> [  119.753666]  relocate_block_group+0x5b/0x600 [btrfs]
> [  119.755321]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
> [  119.757111]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
> [  119.758833]  __btrfs_balance+0x41c/0xcc0 [btrfs]
> [  119.760526]  btrfs_balance+0x65b/0xbd0 [btrfs]
> [  119.762050]  ? kmem_cache_alloc_trace+0x1a7/0x320
> [  119.763777]  ? btrfs_ioctl_balance+0x21c/0x350 [btrfs]
> [  119.765593]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
> [  119.767444]  ? __handle_mm_fault+0x499/0x740
> [  119.769007]  btrfs_ioctl+0x304/0x2590 [btrfs]
> [  119.770537]  ? do_raw_spin_unlock+0x4b/0xc0
> [  119.772015]  ? _raw_spin_unlock+0x1f/0x30
> [  119.774671]  ? __handle_mm_fault+0x499/0x740
> [  119.777106]  ? do_user_addr_fault+0x1d8/0x3f0
> [  119.779366]  ? kvm_sched_clock_read+0x14/0x30
> [  119.781451]  ? sched_clock+0x5/0x10
> [  119.783584]  ? sched_clock_cpu+0x15/0x130
> [  119.785507]  ? do_user_addr_fault+0x1d8/0x3f0
> [  119.787587]  ? ksys_ioctl+0x68/0xa0
> [  119.789577]  ksys_ioctl+0x68/0xa0
> [  119.790909]  __x64_sys_ioctl+0x16/0x20
> [  119.792361]  do_syscall_64+0x50/0x210
> [  119.793749]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
> [  119.795547] RIP: 0033:0x7fdc12e5e227
> [  119.796967] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
> [  119.802957] RSP: 002b:00007ffc3c0bf7b8 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
> [  119.805509] RAX: ffffffffffffffda RBX: 00007ffc3c0bf860 RCX: 00007fdc12e5e227
> [  119.807832] RDX: 00007ffc3c0bf860 RSI: 00000000c4009420 RDI: 0000000000000003
> [  119.810088] RBP: 0000000000000003 R08: 000055a4486922a0 R09: 0000000000000231
> [  119.812378] R10: 00007fdc13088cf0 R11: 0000000000000206 R12: 0000000000000001
> [  119.814654] R13: 00007ffc3c0c2127 R14: 0000000000000002 R15: 0000000000000000
> [  119.816912] irq event stamp: 19394
> [  119.818225] hardirqs last  enabled at (19393): [<ffffffff8a10a166>] console_unlock+0x436/0x590
> [  119.821106] hardirqs last disabled at (19394): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
> [  119.824311] softirqs last  enabled at (19390): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
> [  119.827257] softirqs last disabled at (19383): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
> [  119.830080] ---[ end trace 247639532e5b557e ]---
> [  119.832171] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
> [  119.836020] BTRFS info (device vdb): forced readonly
> [  119.837888] BTRFS warning (device vdb): Skipping commit of aborted transaction.
> [  119.841381] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
> [  119.845600] BTRFS info (device vdb): balance: ended with status: -30
> [  120.282656] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (19409)
> [  120.287896] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (19409)
> [  120.298518] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (19409)
> [  120.303545] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (19409)
> [  120.311101] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (19409)
> [  120.314834] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (19409)
> [  120.349393] BTRFS info (device vdb): disk space caching is enabled
> [  120.353340] BTRFS info (device vdb): has skinny extents
> [  120.355944] BTRFS info (device vdb): flagging fs with big metadata feature
> [  120.377592] BTRFS info (device vdb): checking UUID tree
> [  127.395164] BTRFS info (device vdb): relocating block group 1104150528 flags data|raid0
> [  127.448678] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=30769152 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
> [  127.458117] BTRFS info (device vdb): leaf 30769152 gen 7 total ptrs 4 free space 15851 owner 18446744073709551607
> [  127.462400] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 21002
> [  127.466146] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
> [  127.468412] 		inode generation 1 size 0 mode 100600
> [  127.470284] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
> [  127.472282] 		inode generation 4 size 0 mode 40755
> [  127.473848] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
> [  127.475601] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
> [  127.477585] BTRFS error (device vdb): block=30769152 write time tree block corruption detected
> [  127.480577] ------------[ cut here ]------------
> [  127.482417] WARNING: CPU: 3 PID: 21002 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
> [  127.486001] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
> [  127.491491] CPU: 3 PID: 21002 Comm: btrfs Tainted: G        W         5.7.0-rc5-default+ #1108
> [  127.494777] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
> [  127.498791] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
> [  127.500882] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
> [  127.507102] RSP: 0018:ffff9ec8852ff700 EFLAGS: 00010282
> [  127.509172] RAX: 0000000000000000 RBX: ffff902c7f20c680 RCX: 0000000000000006
> [  127.511742] RDX: 0000000000000000 RSI: ffff902c4fd28910 RDI: ffff902c4fd28040
> [  127.514884] RBP: 0000000000000005 R08: 0000001dae6ee2ba R09: 0000000000000000
> [  127.517380] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c4a994000
> [  127.519327] R13: ffff902c428a1ef0 R14: 0000000000000000 R15: 00000000ffffff8b
> [  127.523221] FS:  00007f69ecb3b8c0(0000) GS:ffff902c7bc00000(0000) knlGS:0000000000000000
> [  127.527813] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  127.529913] CR2: 00007f69ecca0af0 CR3: 000000010579c002 CR4: 0000000000160ee0
> [  127.531942] Call Trace:
> [  127.533156]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
> [  127.535744]  submit_one_bio+0x2b/0x40 [btrfs]
> [  127.537542]  submit_extent_page+0x104/0x210 [btrfs]
> [  127.540137]  write_one_eb+0x1b1/0x390 [btrfs]
> [  127.542825]  ? find_first_extent_bit_state+0x90/0x90 [btrfs]
> [  127.559976]  btree_write_cache_pages+0x1af/0x440 [btrfs]
> [  127.562025]  do_writepages+0x40/0xe0
> [  127.563813]  ? do_raw_spin_unlock+0x4b/0xc0
> [  127.565541]  ? _raw_spin_unlock+0x1f/0x30
> [  127.567363]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
> [  127.569545]  __filemap_fdatawrite_range+0xce/0x110
> [  127.571519]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
> [  127.574396]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
> [  127.576877]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
> [  127.579328]  ? start_transaction+0xd2/0x5e0 [btrfs]
> [  127.581449]  prepare_to_relocate+0x107/0x130 [btrfs]
> [  127.583958]  relocate_block_group+0x5b/0x600 [btrfs]
> [  127.586867]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
> [  127.588999]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
> [  127.590844]  btrfs_shrink_device+0x214/0x530 [btrfs]
> [  127.592345]  btrfs_rm_device+0x22e/0x7f0 [btrfs]
> [  127.593913]  ? _copy_from_user+0x6a/0xa0
> [  127.595367]  btrfs_ioctl+0x218f/0x2590 [btrfs]
> [  127.596993]  ? __handle_mm_fault+0x1c1/0x740
> [  127.598537]  ? do_user_addr_fault+0x1d8/0x3f0
> [  127.600139]  ? kvm_sched_clock_read+0x14/0x30
> [  127.601713]  ? sched_clock+0x5/0x10
> [  127.603121]  ? sched_clock_cpu+0x15/0x130
> [  127.604840]  ? do_user_addr_fault+0x1d8/0x3f0
> [  127.606740]  ? ksys_ioctl+0x68/0xa0
> [  127.608092]  ksys_ioctl+0x68/0xa0
> [  127.609434]  __x64_sys_ioctl+0x16/0x20
> [  127.610860]  do_syscall_64+0x50/0x210
> [  127.612294]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
> [  127.614042] RIP: 0033:0x7f69ecc34227
> [  127.615464] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
> [  127.621798] RSP: 002b:00007ffc4c0a4e08 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
> [  127.624742] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f69ecc34227
> [  127.627020] RDX: 00007ffc4c0a4e30 RSI: 000000005000943a RDI: 0000000000000003
> [  127.629398] RBP: 00007ffc4c0a6fd0 R08: 00007ffc4c0a4e68 R09: 006764762f766564
> [  127.631559] R10: 00007f69ece5ecf0 R11: 0000000000000202 R12: 0000000000000000
> [  127.634138] R13: 00007ffc4c0a4e30 R14: 000056046c20be8c R15: 0000000000000003
> [  127.635759] irq event stamp: 108944
> [  127.637314] hardirqs last  enabled at (108943): [<ffffffff8a10a166>] console_unlock+0x436/0x590
> [  127.640821] hardirqs last disabled at (108944): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
> [  127.644549] softirqs last  enabled at (108940): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
> [  127.647486] softirqs last disabled at (108933): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
> [  127.650608] ---[ end trace 247639532e5b557f ]---
> [  127.653165] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
> [  127.657655] BTRFS info (device vdb): forced readonly
> [  127.659739] BTRFS warning (device vdb): Skipping commit of aborted transaction.
> [  127.663588] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
> [failed, exit status 1] [19:40:22]- output mismatch (see /tmp/fstests/results//btrfs/003.out.bad)
>     --- tests/btrfs/003.out	2018-04-12 16:57:00.608225550 +0000
>     +++ /tmp/fstests/results//btrfs/003.out.bad	2020-05-15 19:40:22.176000000 +0000
>     @@ -1,2 +1,6 @@
>      QA output created by 003
>     -Silence is golden
>     +ERROR: error during balancing '/tmp/scratch': Read-only file system
>     +There may be more info in syslog - try dmesg | tail
>     +ERROR: error removing device '/dev/vdg': Read-only file system
>     +btrfs device delete failed
>     +(see /tmp/fstests/results//btrfs/003.full for details)
>     ...
>     (Run 'diff -u /tmp/fstests/tests/btrfs/003.out /tmp/fstests/results//btrfs/003.out.bad'  to see the entire diff)
> 


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[Qu Wenruo]

[-- Attachment #1.1: Type: text/plain, Size: 16579 bytes --]



On 2020/5/16 下午3:01, Qu Wenruo wrote:
> 
> 
> On 2020/5/16 上午3:45, David Sterba wrote:
>> On Fri, May 15, 2020 at 02:01:39PM +0800, Qu Wenruo wrote:
>>> This small patchset reworks the REF_COWS bit, by renaming it, and remove
>>> that bit for data relocation root.
>>>
>>>
>>> The basic idea of such rework is to reduce the confusion caused by the
>>> name REF_COWS.
>>>
>>> With the new bit called SHAREABLE, it should be clear that no user can
>>> really create snapshot for data reloc tree, thus its tree blocks
>>> shouldn't be shareable.
>>>
>>> This would make data balance for reloc tree a little simpler.
>>>
>>> Changelog:
>>> v2:
>>> - Add new patch to address the log tree check in
>>>   btrfs_truncate_inode_items()
>>>   Thanks for the advice from David, now it's much simpler than original
>>>   check, and data reloc tree no longer needs extra hanlding
>>>
>>> - Grab data reloc root in create_reloc_inode() and
>>>   btrfs_recover_relocation()
>>>
>>> - Comment update
>>>
>>> v3:
>>> - Remove ALIGN_DOWN() -> round_down() change
>>>
>>> - Remove the confusing comment on the log tree inode
>>
>> I've added the patches to misc-next, with some fixups. I'll let it also
>> go through fstests, but a quick run has hit this write-time corruption
>> very early. I haven't analyzed it and it's possible that it's caused by
>> my fixups.
> 
> Passed my local balance group.
> 
> And furthermore, the error pattern is exactly what I saw during my
> development.
> 
>>
> [...]
>> [  119.624572] BTRFS info (device vdb): balance: start -d -m -s
>> [  119.630843] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
>> [  119.640113] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=298909696 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
>> [  119.647511] BTRFS info (device vdb): leaf 298909696 gen 11 total ptrs 4 free space 15851 owner 18446744073709551607
>> [  119.652214] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 19404
>> [  119.656275] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
>> [  119.658436] 		inode generation 1 size 0 mode 100600
> 
> This is using 1 as ino number, which means root::highest_objectid is not
> properly initialized.
> 
> This happened when I'm using btrfs_read_tree_root() other than
> btrfs_read_fs_root(), which initializes root::highest_objectid.

After fetching the misc-next branch, that's exactly the problem.

The 3rd patch is using the correct btrfs_get_fs_root() which won't
trigger the problem.

Thanks,
Qu
> 
> So I guess there is something wrong happened during the fixup.
> 
> Thanks,
> Qu
> 
>> [  119.660812] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
>> [  119.663645] 		inode generation 4 size 0 mode 40755
>> [  119.665768] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
>> [  119.668146] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
>> [  119.671438] BTRFS error (device vdb): block=298909696 write time tree block corruption detected
>> [  119.675319] ------------[ cut here ]------------
>> [  119.677789] WARNING: CPU: 1 PID: 19404 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
>> [  119.683116] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
>> [  119.690407] CPU: 1 PID: 19404 Comm: btrfs Not tainted 5.7.0-rc5-default+ #1108
>> [  119.693724] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
>> [  119.698411] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
>> [  119.700653] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
>> [  119.706847] RSP: 0018:ffff9ec884397778 EFLAGS: 00010292
>> [  119.707997] RAX: 0000000000000000 RBX: ffff902c7faf0dc0 RCX: 0000000000000006
>> [  119.709428] RDX: 0000000000000000 RSI: ffff902c4fcd5dd0 RDI: ffff902c4fcd5500
>> [  119.711051] RBP: 0000000000000001 R08: 0000001bdd341101 R09: 0000000000000000
>> [  119.712587] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c74b3c000
>> [  119.714957] R13: ffff902c428a16b0 R14: 0000000000000000 R15: 00000000ffffff8b
>> [  119.716474] FS:  00007fdc12d658c0(0000) GS:ffff902c7b800000(0000) knlGS:0000000000000000
>> [  119.719045] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [  119.720724] CR2: 000055a44869a2d8 CR3: 00000001367c2003 CR4: 0000000000160ee0
>> [  119.725194] Call Trace:
>> [  119.727333]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
>> [  119.730365]  submit_one_bio+0x2b/0x40 [btrfs]
>> [  119.732253]  btree_write_cache_pages+0x373/0x440 [btrfs]
>> [  119.734909]  do_writepages+0x40/0xe0
>> [  119.736667]  ? do_raw_spin_unlock+0x4b/0xc0
>> [  119.738667]  ? _raw_spin_unlock+0x1f/0x30
>> [  119.740122]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
>> [  119.743009]  __filemap_fdatawrite_range+0xce/0x110
>> [  119.744446]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
>> [  119.746166]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
>> [  119.748184]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
>> [  119.750072]  ? start_transaction+0xd2/0x5e0 [btrfs]
>> [  119.751896]  prepare_to_relocate+0x107/0x130 [btrfs]
>> [  119.753666]  relocate_block_group+0x5b/0x600 [btrfs]
>> [  119.755321]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
>> [  119.757111]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
>> [  119.758833]  __btrfs_balance+0x41c/0xcc0 [btrfs]
>> [  119.760526]  btrfs_balance+0x65b/0xbd0 [btrfs]
>> [  119.762050]  ? kmem_cache_alloc_trace+0x1a7/0x320
>> [  119.763777]  ? btrfs_ioctl_balance+0x21c/0x350 [btrfs]
>> [  119.765593]  btrfs_ioctl_balance+0x298/0x350 [btrfs]
>> [  119.767444]  ? __handle_mm_fault+0x499/0x740
>> [  119.769007]  btrfs_ioctl+0x304/0x2590 [btrfs]
>> [  119.770537]  ? do_raw_spin_unlock+0x4b/0xc0
>> [  119.772015]  ? _raw_spin_unlock+0x1f/0x30
>> [  119.774671]  ? __handle_mm_fault+0x499/0x740
>> [  119.777106]  ? do_user_addr_fault+0x1d8/0x3f0
>> [  119.779366]  ? kvm_sched_clock_read+0x14/0x30
>> [  119.781451]  ? sched_clock+0x5/0x10
>> [  119.783584]  ? sched_clock_cpu+0x15/0x130
>> [  119.785507]  ? do_user_addr_fault+0x1d8/0x3f0
>> [  119.787587]  ? ksys_ioctl+0x68/0xa0
>> [  119.789577]  ksys_ioctl+0x68/0xa0
>> [  119.790909]  __x64_sys_ioctl+0x16/0x20
>> [  119.792361]  do_syscall_64+0x50/0x210
>> [  119.793749]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
>> [  119.795547] RIP: 0033:0x7fdc12e5e227
>> [  119.796967] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
>> [  119.802957] RSP: 002b:00007ffc3c0bf7b8 EFLAGS: 00000206 ORIG_RAX: 0000000000000010
>> [  119.805509] RAX: ffffffffffffffda RBX: 00007ffc3c0bf860 RCX: 00007fdc12e5e227
>> [  119.807832] RDX: 00007ffc3c0bf860 RSI: 00000000c4009420 RDI: 0000000000000003
>> [  119.810088] RBP: 0000000000000003 R08: 000055a4486922a0 R09: 0000000000000231
>> [  119.812378] R10: 00007fdc13088cf0 R11: 0000000000000206 R12: 0000000000000001
>> [  119.814654] R13: 00007ffc3c0c2127 R14: 0000000000000002 R15: 0000000000000000
>> [  119.816912] irq event stamp: 19394
>> [  119.818225] hardirqs last  enabled at (19393): [<ffffffff8a10a166>] console_unlock+0x436/0x590
>> [  119.821106] hardirqs last disabled at (19394): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
>> [  119.824311] softirqs last  enabled at (19390): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
>> [  119.827257] softirqs last disabled at (19383): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
>> [  119.830080] ---[ end trace 247639532e5b557e ]---
>> [  119.832171] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
>> [  119.836020] BTRFS info (device vdb): forced readonly
>> [  119.837888] BTRFS warning (device vdb): Skipping commit of aborted transaction.
>> [  119.841381] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
>> [  119.845600] BTRFS info (device vdb): balance: ended with status: -30
>> [  120.282656] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 1 transid 5 /dev/vdb scanned by mkfs.btrfs (19409)
>> [  120.287896] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 2 transid 5 /dev/vdc scanned by mkfs.btrfs (19409)
>> [  120.298518] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 3 transid 5 /dev/vdd scanned by mkfs.btrfs (19409)
>> [  120.303545] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 4 transid 5 /dev/vde scanned by mkfs.btrfs (19409)
>> [  120.311101] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 5 transid 5 /dev/vdf scanned by mkfs.btrfs (19409)
>> [  120.314834] BTRFS: device fsid b28016df-bc50-4ac6-a337-926c19fb18f3 devid 6 transid 5 /dev/vdg scanned by mkfs.btrfs (19409)
>> [  120.349393] BTRFS info (device vdb): disk space caching is enabled
>> [  120.353340] BTRFS info (device vdb): has skinny extents
>> [  120.355944] BTRFS info (device vdb): flagging fs with big metadata feature
>> [  120.377592] BTRFS info (device vdb): checking UUID tree
>> [  127.395164] BTRFS info (device vdb): relocating block group 1104150528 flags data|raid0
>> [  127.448678] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=30769152 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
>> [  127.458117] BTRFS info (device vdb): leaf 30769152 gen 7 total ptrs 4 free space 15851 owner 18446744073709551607
>> [  127.462400] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 21002
>> [  127.466146] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
>> [  127.468412] 		inode generation 1 size 0 mode 100600
>> [  127.470284] 	item 1 key (256 1 0) itemoff 15963 itemsize 160
>> [  127.472282] 		inode generation 4 size 0 mode 40755
>> [  127.473848] 	item 2 key (256 12 256) itemoff 15951 itemsize 12
>> [  127.475601] 	item 3 key (18446744073709551611 48 1) itemoff 15951 itemsize 0
>> [  127.477585] BTRFS error (device vdb): block=30769152 write time tree block corruption detected
>> [  127.480577] ------------[ cut here ]------------
>> [  127.482417] WARNING: CPU: 3 PID: 21002 at fs/btrfs/disk-io.c:537 btree_csum_one_bio+0x297/0x2a0 [btrfs]
>> [  127.486001] Modules linked in: btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
>> [  127.491491] CPU: 3 PID: 21002 Comm: btrfs Tainted: G        W         5.7.0-rc5-default+ #1108
>> [  127.494777] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
>> [  127.498791] RIP: 0010:btree_csum_one_bio+0x297/0x2a0 [btrfs]
>> [  127.500882] Code: bc fd ff ff e8 9a 24 c1 c9 31 f6 48 89 3c 24 e8 ef 7b ff ff 48 8b 3c 24 48 c7 c6 f0 45 55 c0 48 8b 17 4c 89 e7 e8 94 cc 0b 00 <0f> 0b e9 8f fd ff ff 66 90 0f 1f 44 00 00 48 89 f7 e9 53 fd ff ff
>> [  127.507102] RSP: 0018:ffff9ec8852ff700 EFLAGS: 00010282
>> [  127.509172] RAX: 0000000000000000 RBX: ffff902c7f20c680 RCX: 0000000000000006
>> [  127.511742] RDX: 0000000000000000 RSI: ffff902c4fd28910 RDI: ffff902c4fd28040
>> [  127.514884] RBP: 0000000000000005 R08: 0000001dae6ee2ba R09: 0000000000000000
>> [  127.517380] R10: 0000000000000000 R11: 0000000000000000 R12: ffff902c4a994000
>> [  127.519327] R13: ffff902c428a1ef0 R14: 0000000000000000 R15: 00000000ffffff8b
>> [  127.523221] FS:  00007f69ecb3b8c0(0000) GS:ffff902c7bc00000(0000) knlGS:0000000000000000
>> [  127.527813] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [  127.529913] CR2: 00007f69ecca0af0 CR3: 000000010579c002 CR4: 0000000000160ee0
>> [  127.531942] Call Trace:
>> [  127.533156]  btree_submit_bio_hook+0x74/0xc0 [btrfs]
>> [  127.535744]  submit_one_bio+0x2b/0x40 [btrfs]
>> [  127.537542]  submit_extent_page+0x104/0x210 [btrfs]
>> [  127.540137]  write_one_eb+0x1b1/0x390 [btrfs]
>> [  127.542825]  ? find_first_extent_bit_state+0x90/0x90 [btrfs]
>> [  127.559976]  btree_write_cache_pages+0x1af/0x440 [btrfs]
>> [  127.562025]  do_writepages+0x40/0xe0
>> [  127.563813]  ? do_raw_spin_unlock+0x4b/0xc0
>> [  127.565541]  ? _raw_spin_unlock+0x1f/0x30
>> [  127.567363]  ? wbc_attach_and_unlock_inode+0x194/0x2a0
>> [  127.569545]  __filemap_fdatawrite_range+0xce/0x110
>> [  127.571519]  btrfs_write_marked_extents+0x68/0x160 [btrfs]
>> [  127.574396]  btrfs_write_and_wait_transaction+0x4f/0xd0 [btrfs]
>> [  127.576877]  btrfs_commit_transaction+0x76a/0xae0 [btrfs]
>> [  127.579328]  ? start_transaction+0xd2/0x5e0 [btrfs]
>> [  127.581449]  prepare_to_relocate+0x107/0x130 [btrfs]
>> [  127.583958]  relocate_block_group+0x5b/0x600 [btrfs]
>> [  127.586867]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
>> [  127.588999]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
>> [  127.590844]  btrfs_shrink_device+0x214/0x530 [btrfs]
>> [  127.592345]  btrfs_rm_device+0x22e/0x7f0 [btrfs]
>> [  127.593913]  ? _copy_from_user+0x6a/0xa0
>> [  127.595367]  btrfs_ioctl+0x218f/0x2590 [btrfs]
>> [  127.596993]  ? __handle_mm_fault+0x1c1/0x740
>> [  127.598537]  ? do_user_addr_fault+0x1d8/0x3f0
>> [  127.600139]  ? kvm_sched_clock_read+0x14/0x30
>> [  127.601713]  ? sched_clock+0x5/0x10
>> [  127.603121]  ? sched_clock_cpu+0x15/0x130
>> [  127.604840]  ? do_user_addr_fault+0x1d8/0x3f0
>> [  127.606740]  ? ksys_ioctl+0x68/0xa0
>> [  127.608092]  ksys_ioctl+0x68/0xa0
>> [  127.609434]  __x64_sys_ioctl+0x16/0x20
>> [  127.610860]  do_syscall_64+0x50/0x210
>> [  127.612294]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
>> [  127.614042] RIP: 0033:0x7f69ecc34227
>> [  127.615464] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
>> [  127.621798] RSP: 002b:00007ffc4c0a4e08 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
>> [  127.624742] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f69ecc34227
>> [  127.627020] RDX: 00007ffc4c0a4e30 RSI: 000000005000943a RDI: 0000000000000003
>> [  127.629398] RBP: 00007ffc4c0a6fd0 R08: 00007ffc4c0a4e68 R09: 006764762f766564
>> [  127.631559] R10: 00007f69ece5ecf0 R11: 0000000000000202 R12: 0000000000000000
>> [  127.634138] R13: 00007ffc4c0a4e30 R14: 000056046c20be8c R15: 0000000000000003
>> [  127.635759] irq event stamp: 108944
>> [  127.637314] hardirqs last  enabled at (108943): [<ffffffff8a10a166>] console_unlock+0x436/0x590
>> [  127.640821] hardirqs last disabled at (108944): [<ffffffff8a002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
>> [  127.644549] softirqs last  enabled at (108940): [<ffffffff8aa0031e>] __do_softirq+0x31e/0x55d
>> [  127.647486] softirqs last disabled at (108933): [<ffffffff8a08d91d>] irq_exit+0x9d/0xb0
>> [  127.650608] ---[ end trace 247639532e5b557f ]---
>> [  127.653165] BTRFS: error (device vdb) in btrfs_commit_transaction:2323: errno=-5 IO failure (Error while writing out transaction)
>> [  127.657655] BTRFS info (device vdb): forced readonly
>> [  127.659739] BTRFS warning (device vdb): Skipping commit of aborted transaction.
>> [  127.663588] BTRFS: error (device vdb) in cleanup_transaction:1894: errno=-5 IO failure
>> [failed, exit status 1] [19:40:22]- output mismatch (see /tmp/fstests/results//btrfs/003.out.bad)
>>     --- tests/btrfs/003.out	2018-04-12 16:57:00.608225550 +0000
>>     +++ /tmp/fstests/results//btrfs/003.out.bad	2020-05-15 19:40:22.176000000 +0000
>>     @@ -1,2 +1,6 @@
>>      QA output created by 003
>>     -Silence is golden
>>     +ERROR: error during balancing '/tmp/scratch': Read-only file system
>>     +There may be more info in syslog - try dmesg | tail
>>     +ERROR: error removing device '/dev/vdg': Read-only file system
>>     +btrfs device delete failed
>>     +(see /tmp/fstests/results//btrfs/003.full for details)
>>     ...
>>     (Run 'diff -u /tmp/fstests/tests/btrfs/003.out /tmp/fstests/results//btrfs/003.out.bad'  to see the entire diff)
>>
> 


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[David Sterba]

On Mon, May 18, 2020 at 01:13:34PM +0800, Qu Wenruo wrote:
> >> [  119.624572] BTRFS info (device vdb): balance: start -d -m -s
> >> [  119.630843] BTRFS info (device vdb): relocating block group 30408704 flags metadata|dup
> >> [  119.640113] BTRFS critical (device vdb): corrupt leaf: root=18446744073709551607 block=298909696 slot=0, invalid key objectid: has 1 expect 6 or [256, 18446744073709551360] or 18446744073709551604
> >> [  119.647511] BTRFS info (device vdb): leaf 298909696 gen 11 total ptrs 4 free space 15851 owner 18446744073709551607
> >> [  119.652214] BTRFS info (device vdb): refs 3 lock (w:0 r:0 bw:0 br:0 sw:0 sr:0) lock_owner 0 current 19404
> >> [  119.656275] 	item 0 key (1 1 0) itemoff 16123 itemsize 160
> >> [  119.658436] 		inode generation 1 size 0 mode 100600
> > 
> > This is using 1 as ino number, which means root::highest_objectid is not
> > properly initialized.
> > 
> > This happened when I'm using btrfs_read_tree_root() other than
> > btrfs_read_fs_root(), which initializes root::highest_objectid.
> 
> After fetching the misc-next branch, that's exactly the problem.
> 
> The 3rd patch is using the correct btrfs_get_fs_root() which won't
> trigger the problem.

I see, thanks. That the data reloc tree does not fit into the pattern of
other trees being initialized in the function with btrfs_read_tree_root
needs to be documented then.

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[David Sterba]

On Mon, May 18, 2020 at 01:13:34PM +0800, Qu Wenruo wrote:
> > This is using 1 as ino number, which means root::highest_objectid is not
> > properly initialized.
> > 
> > This happened when I'm using btrfs_read_tree_root() other than
> > btrfs_read_fs_root(), which initializes root::highest_objectid.
> 
> After fetching the misc-next branch, that's exactly the problem.
> 
> The 3rd patch is using the correct btrfs_get_fs_root() which won't
> trigger the problem.

Looking at it futher, btrfs_get_fs_root does a few more things that does
not seem to be necessary for the data reloc tree.

The following two should be sufficient:

- btrfs_read_tree_root
- btrfs_init_fs_root

but with btrfs_get_fs_root there's an orphan item check and the data
reloc tree is added to fs_roots_radix.

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[David Sterba]

On Mon, May 18, 2020 at 05:03:31PM +0200, David Sterba wrote:
> On Mon, May 18, 2020 at 01:13:34PM +0800, Qu Wenruo wrote:
> - btrfs_read_tree_root
> - btrfs_init_fs_root

With

2291         location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
2292         root = btrfs_read_tree_root(tree_root, &location);
2293         if (IS_ERR(root)) {
2294                 ret = PTR_ERR(root);
2295                 goto out;
2296         }
2297         ret = btrfs_init_fs_root(root);
2298         if (ret)
2299                 goto out;
2300         set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
2301         fs_info->data_reloc_root = root;

still not fine, btrfs/003

[  104.188077] BTRFS error (device vdb): unable to find ref byte nr 31883264 parent 0 root 18446744073709551607  owner 0 offset 0

[  103.476564] BTRFS info (device vdb): relocating block group 22020096 flags system|raid1
[  103.481260] ------------[ cut here ]------------
[  103.483638] WARNING: CPU: 0 PID: 21011 at fs/btrfs/extent-tree.c:3055 __btrfs_free_extent+0x66c/0x900 [btrfs]
[  103.488388] Modules linked in: xxhash_generic btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
[  103.492733] CPU: 0 PID: 21011 Comm: btrfs Not tainted 5.7.0-rc6-default+ #1109
[  103.495057] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
[  103.498331] RIP: 0010:__btrfs_free_extent+0x66c/0x900 [btrfs]
[  103.500037] Code: 48 c7 44 24 48 00 00 00 00 48 89 ea 4c 89 f6 e8 da a3 ff ff 41 89 c7 e9 c8 fb ff ff 0f 0b 48 c7 c7 60 fa 2d c0 e8 2a 8f 2e f1 <0f> 0b 49 8b 3e e8 8a 76 00 00 ff 74 24 18 49 89 d9 4d 89 e8 48 8b
[  103.505108] RSP: 0018:ffffac68455678b0 EFLAGS: 00010246
[  103.506711] RAX: 00000000fffffffe RBX: 0000000000000000 RCX: 0000000000000002
[  103.508748] RDX: 00000000fffffffe RSI: 0000000000000000 RDI: 000151b4a32a2068
[  103.510624] RBP: 0000000001e68000 R08: 000313fad20cf648 R09: ffff8f6e74474a88
[  103.512589] R10: 0000000000000000 R11: ffff8f6e781ffb88 R12: 0000000000000000
[  103.514486] R13: fffffffffffffff7 R14: ffff8f6e74474a88 R15: 00000000fffffffe
[  103.516691] FS:  00007f9ab400b8c0(0000) GS:ffff8f6e7d600000(0000) knlGS:0000000000000000
[  103.519541] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  103.521430] CR2: 000055f4d1e28e18 CR3: 00000000747d3006 CR4: 0000000000160ef0
[  103.523684] Call Trace:
[  103.524743]  ? native_sched_clock_from_tsc+0x41/0xc0
[  103.526212]  ? btrfs_run_delayed_refs_for_head+0x197/0xcd0 [btrfs]
[  103.527983]  btrfs_run_delayed_refs_for_head+0x248/0xcd0 [btrfs]
[  103.529704]  ? _raw_read_unlock+0x1f/0x30
[  103.530973]  ? btrfs_merge_delayed_refs+0x3d3/0x480 [btrfs]
[  103.532566]  __btrfs_run_delayed_refs+0x9d/0x680 [btrfs]
[  103.534327]  ? join_transaction+0x15d/0x4c0 [btrfs]
[  103.536146]  ? kvm_sched_clock_read+0x14/0x30
[  103.537535]  ? sched_clock+0x5/0x10
[  103.538720]  ? sched_clock_cpu+0x15/0x130
[  103.540023]  btrfs_run_delayed_refs+0x86/0x1e0 [btrfs]
[  103.541486]  btrfs_commit_transaction+0x57/0xae0 [btrfs]
[  103.543049]  ? start_transaction+0xd2/0x5e0 [btrfs]
[  103.544578]  prepare_to_relocate+0x107/0x130 [btrfs]
[  103.546117]  relocate_block_group+0x5b/0x600 [btrfs]
[  103.547677]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
[  103.549589]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
[  103.551293]  btrfs_shrink_device+0x214/0x530 [btrfs]
[  103.553023]  btrfs_rm_device+0x22e/0x7f0 [btrfs]
[  103.554635]  ? _copy_from_user+0x6a/0xa0
[  103.556118]  btrfs_ioctl+0x218f/0x2590 [btrfs]
[  103.557709]  ? __handle_mm_fault+0x1c1/0x740
[  103.559204]  ? do_user_addr_fault+0x1d8/0x3f0
[  103.560590]  ? kvm_sched_clock_read+0x14/0x30
[  103.561927]  ? sched_clock+0x5/0x10
[  103.563126]  ? sched_clock_cpu+0x15/0x130
[  103.564458]  ? do_user_addr_fault+0x1d8/0x3f0
[  103.565817]  ? ksys_ioctl+0x68/0xa0
[  103.566965]  ksys_ioctl+0x68/0xa0
[  103.568135]  __x64_sys_ioctl+0x16/0x20
[  103.569426]  do_syscall_64+0x50/0x210
[  103.570718]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
[  103.572304] RIP: 0033:0x7f9ab4104227
[  103.573566] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
[  103.578827] RSP: 002b:00007ffc5dd62938 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[  103.581443] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f9ab4104227
[  103.583643] RDX: 00007ffc5dd62960 RSI: 000000005000943a RDI: 0000000000000003
[  103.585796] RBP: 00007ffc5dd64b00 R08: 00007ffc5dd62998 R09: 006764762f766564
[  103.587920] R10: 00007f9ab432ecf0 R11: 0000000000000202 R12: 0000000000000000
[  103.590108] R13: 00007ffc5dd62960 R14: 000055ee38909e8c R15: 0000000000000003
[  103.592356] irq event stamp: 1219702
[  103.593688] hardirqs last  enabled at (1219701): [<ffffffffb17b090e>] _raw_spin_unlock_irqrestore+0x3e/0x50
[  103.596941] hardirqs last disabled at (1219702): [<ffffffffb1002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
[  103.600110] softirqs last  enabled at (1219564): [<ffffffffb1a0031e>] __do_softirq+0x31e/0x55d
[  103.602911] softirqs last disabled at (1219503): [<ffffffffb108d9ed>] irq_exit+0x9d/0xb0
[  103.605756] ---[ end trace 0cea18996e5ca624 ]---
[  103.607373] BTRFS info (device vdb): leaf 12916031488 gen 22 total ptrs 111 free space 9383 owner 2
[  103.610495] BTRFS info (device vdb): refs 2 lock (w:1 r:0 bw:0 br:0 sw:1 sr:0) lock_owner 21011 current 21011
[  103.613684] 	item 0 key (22020096 169 0) itemoff 16250 itemsize 33
[  103.615597] 		extent refs 1 gen 22 flags 2
[  103.617040] 		ref#0: tree block backref root 3
[  103.618523] 	item 1 key (22020096 192 8388608) itemoff 16226 itemsize 24
[  103.620533] 		block group used 16384 chunk_objectid 256 flags 18
[  103.622401] 	item 2 key (7546601472 168 262144) itemoff 16173 itemsize 53
[  103.624531] 		extent refs 1 gen 19 flags 1
[  103.625954] 		ref#0: extent data backref root 1 objectid 256 offset 0 count 1
[  103.628154] 	item 3 key (7546601472 192 5368709120) itemoff 16149 itemsize 24
[  103.630330] 		block group used 4538368 chunk_objectid 256 flags 9
[  103.632218] 	item 4 key (7547125760 168 262144) itemoff 16096 itemsize 53
[  103.634240] 		extent refs 1 gen 21 flags 1
[  103.635634] 		ref#0: extent data backref root 1 objectid 259 offset 0 count 1
[  103.637862] 	item 5 key (7547387904 168 1310720) itemoff 16043 itemsize 53
[  103.639966] 		extent refs 1 gen 21 flags 1
[  103.641218] 		ref#0: extent data backref root 1 objectid 258 offset 0 count 1
[  103.643165] 	item 6 key (7549747200 168 12288) itemoff 16006 itemsize 37
[  103.645120] 		extent refs 1 gen 8 flags 1
[  103.646443] 		ref#0: shared data backref parent 12915392512 count 1
[  103.648219] 	item 7 key (7549759488 168 12288) itemoff 15969 itemsize 37
[  103.649972] 		extent refs 1 gen 8 flags 1
[  103.651361] 		ref#0: shared data backref parent 12915392512 count 1
[  103.653234] 	item 8 key (7549771776 168 12288) itemoff 15932 itemsize 37
[  103.654963] 		extent refs 1 gen 8 flags 1
[  103.656235] 		ref#0: shared data backref parent 12915392512 count 1
[  103.657918] 	item 9 key (7549784064 168 12288) itemoff 15895 itemsize 37
[  103.659702] 		extent refs 1 gen 8 flags 1
[  103.661039] 		ref#0: shared data backref parent 12915392512 count 1
[  103.662705] 	item 10 key (7549796352 168 12288) itemoff 15858 itemsize 37
[  103.664576] 		extent refs 1 gen 8 flags 1
[  103.665833] 		ref#0: shared data backref parent 12915392512 count 1
[  103.667512] 	item 11 key (7549808640 168 12288) itemoff 15821 itemsize 37
[  103.669324] 		extent refs 1 gen 8 flags 1
[  103.670581] 		ref#0: shared data backref parent 12915392512 count 1
[  103.672305] 	item 12 key (7549820928 168 12288) itemoff 15784 itemsize 37
[  103.682923] 		extent refs 1 gen 8 flags 1
[  103.684201] 		ref#0: shared data backref parent 12915392512 count 1
[  103.685913] 	item 13 key (7549833216 168 12288) itemoff 15747 itemsize 37
[  103.687798] 		extent refs 1 gen 8 flags 1
[  103.689116] 		ref#0: shared data backref parent 12915392512 count 1
[  103.690910] 	item 14 key (7549845504 168 12288) itemoff 15710 itemsize 37
[  103.692850] 		extent refs 1 gen 8 flags 1
[  103.694070] 		ref#0: shared data backref parent 12915392512 count 1
[  103.695744] 	item 15 key (7549857792 168 12288) itemoff 15673 itemsize 37
[  103.697593] 		extent refs 1 gen 8 flags 1
[  103.698824] 		ref#0: shared data backref parent 12915392512 count 1
[  103.700589] 	item 16 key (7549870080 168 12288) itemoff 15636 itemsize 37
[  103.702508] 		extent refs 1 gen 8 flags 1
[  103.703873] 		ref#0: shared data backref parent 12915392512 count 1
[  103.705573] 	item 17 key (7549882368 168 12288) itemoff 15599 itemsize 37
[  103.707363] 		extent refs 1 gen 8 flags 1
[  103.708644] 		ref#0: shared data backref parent 12915392512 count 1
[  103.710300] 	item 18 key (7549894656 168 12288) itemoff 15562 itemsize 37
[  103.712431] 		extent refs 1 gen 8 flags 1
[  103.713877] 		ref#0: shared data backref parent 12915392512 count 1
[  103.715866] 	item 19 key (7549906944 168 12288) itemoff 15525 itemsize 37
[  103.717995] 		extent refs 1 gen 8 flags 1
[  103.719438] 		ref#0: shared data backref parent 12915392512 count 1
[  103.721452] 	item 20 key (7549919232 168 12288) itemoff 15488 itemsize 37
[  103.723574] 		extent refs 1 gen 8 flags 1
[  103.725085] 		ref#0: shared data backref parent 12915392512 count 1
[  103.727163] 	item 21 key (7549931520 168 12288) itemoff 15451 itemsize 37
[  103.729031] 		extent refs 1 gen 8 flags 1
[  103.730305] 		ref#0: shared data backref parent 12915392512 count 1
[  103.732026] 	item 22 key (7549943808 168 12288) itemoff 15414 itemsize 37
[  103.733761] 		extent refs 1 gen 8 flags 1
[  103.734983] 		ref#0: shared data backref parent 12915392512 count 1
[  103.736710] 	item 23 key (7549956096 168 12288) itemoff 15377 itemsize 37
[  103.738589] 		extent refs 1 gen 8 flags 1
[  103.739962] 		ref#0: shared data backref parent 12915392512 count 1
[  103.741925] 	item 24 key (7549968384 168 12288) itemoff 15340 itemsize 37
[  103.744059] 		extent refs 1 gen 8 flags 1
[  103.745476] 		ref#0: shared data backref parent 12915392512 count 1
[  103.747416] 	item 25 key (7549980672 168 12288) itemoff 15303 itemsize 37
[  103.749495] 		extent refs 1 gen 8 flags 1
[  103.750908] 		ref#0: shared data backref parent 12915392512 count 1
[  103.752878] 	item 26 key (7549992960 168 12288) itemoff 15266 itemsize 37
[  103.754966] 		extent refs 1 gen 8 flags 1
[  103.756378] 		ref#0: shared data backref parent 12915408896 count 1
[  103.758320] 	item 27 key (7550005248 168 12288) itemoff 15229 itemsize 37
[  103.760225] 		extent refs 1 gen 8 flags 1
[  103.761495] 		ref#0: shared data backref parent 12915408896 count 1
[  103.763406] 	item 28 key (7550017536 168 12288) itemoff 15192 itemsize 37
[  103.765476] 		extent refs 1 gen 8 flags 1
[  103.766879] 		ref#0: shared data backref parent 12915408896 count 1
[  103.768840] 	item 29 key (7550029824 168 12288) itemoff 15155 itemsize 37
[  103.770881] 		extent refs 1 gen 8 flags 1
[  103.772294] 		ref#0: shared data backref parent 12915408896 count 1
[  103.774244] 	item 30 key (7550042112 168 12288) itemoff 15118 itemsize 37
[  103.776246] 		extent refs 1 gen 8 flags 1
[  103.777593] 		ref#0: shared data backref parent 12915408896 count 1
[  103.779486] 	item 31 key (7550054400 168 12288) itemoff 15081 itemsize 37
[  103.781516] 		extent refs 1 gen 8 flags 1
[  103.782920] 		ref#0: shared data backref parent 12915408896 count 1
[  103.784833] 	item 32 key (7550066688 168 12288) itemoff 15044 itemsize 37
[  103.786906] 		extent refs 1 gen 8 flags 1
[  103.788360] 		ref#0: shared data backref parent 12915408896 count 1
[  103.790355] 	item 33 key (7550078976 168 12288) itemoff 15007 itemsize 37
[  103.792499] 		extent refs 1 gen 8 flags 1
[  103.793941] 		ref#0: shared data backref parent 12915408896 count 1
[  103.795927] 	item 34 key (7550091264 168 12288) itemoff 14970 itemsize 37
[  103.798011] 		extent refs 1 gen 8 flags 1
[  103.799220] 		ref#0: shared data backref parent 12915408896 count 1
[  103.800900] 	item 35 key (7550103552 168 12288) itemoff 14933 itemsize 37
[  103.802949] 		extent refs 1 gen 8 flags 1
[  103.804399] 		ref#0: shared data backref parent 12915408896 count 1
[  103.806333] 	item 36 key (7550115840 168 12288) itemoff 14896 itemsize 37
[  103.808405] 		extent refs 1 gen 8 flags 1
[  103.809805] 		ref#0: shared data backref parent 12915408896 count 1
[  103.811731] 	item 37 key (7550128128 168 12288) itemoff 14859 itemsize 37
[  103.813803] 		extent refs 1 gen 8 flags 1
[  103.815196] 		ref#0: shared data backref parent 12915408896 count 1
[  103.817131] 	item 38 key (7550140416 168 12288) itemoff 14822 itemsize 37
[  103.819033] 		extent refs 1 gen 8 flags 1
[  103.820379] 		ref#0: shared data backref parent 12915408896 count 1
[  103.822148] 	item 39 key (7550152704 168 12288) itemoff 14785 itemsize 37
[  103.823891] 		extent refs 1 gen 8 flags 1
[  103.825117] 		ref#0: shared data backref parent 12915408896 count 1
[  103.826770] 	item 40 key (7550164992 168 12288) itemoff 14748 itemsize 37
[  103.828634] 		extent refs 1 gen 8 flags 1
[  103.829825] 		ref#0: shared data backref parent 12915408896 count 1
[  103.831482] 	item 41 key (7550177280 168 12288) itemoff 14711 itemsize 37
[  103.833271] 		extent refs 1 gen 8 flags 1
[  103.834512] 		ref#0: shared data backref parent 12915408896 count 1
[  103.836181] 	item 42 key (7550189568 168 12288) itemoff 14674 itemsize 37
[  103.837971] 		extent refs 1 gen 8 flags 1
[  103.839255] 		ref#0: shared data backref parent 12915408896 count 1
[  103.841066] 	item 43 key (7550201856 168 12288) itemoff 14637 itemsize 37
[  103.843080] 		extent refs 1 gen 8 flags 1
[  103.844450] 		ref#0: shared data backref parent 12915408896 count 1
[  103.846317] 	item 44 key (7550214144 168 12288) itemoff 14600 itemsize 37
[  103.848286] 		extent refs 1 gen 8 flags 1
[  103.849647] 		ref#0: shared data backref parent 12915408896 count 1
[  103.851373] 	item 45 key (7550226432 168 12288) itemoff 14563 itemsize 37
[  103.853123] 		extent refs 1 gen 8 flags 1
[  103.854370] 		ref#0: shared data backref parent 12915408896 count 1
[  103.855922] 	item 46 key (7550238720 168 12288) itemoff 14526 itemsize 37
[  103.857613] 		extent refs 1 gen 8 flags 1
[  103.858815] 		ref#0: shared data backref parent 12915408896 count 1
[  103.860629] 	item 47 key (7550251008 168 12288) itemoff 14489 itemsize 37
[  103.862379] 		extent refs 1 gen 8 flags 1
[  103.863572] 		ref#0: shared data backref parent 12915408896 count 1
[  103.865192] 	item 48 key (7550263296 168 12288) itemoff 14452 itemsize 37
[  103.867156] 		extent refs 1 gen 8 flags 1
[  103.868559] 		ref#0: shared data backref parent 12915408896 count 1
[  103.870353] 	item 49 key (7550275584 168 12288) itemoff 14415 itemsize 37
[  103.872282] 		extent refs 1 gen 8 flags 1
[  103.873621] 		ref#0: shared data backref parent 12915408896 count 1
[  103.875483] 	item 50 key (7550287872 168 12288) itemoff 14378 itemsize 37
[  103.877456] 		extent refs 1 gen 8 flags 1
[  103.878808] 		ref#0: shared data backref parent 12915408896 count 1
[  103.880700] 	item 51 key (7550300160 168 12288) itemoff 14341 itemsize 37
[  103.882595] 		extent refs 1 gen 8 flags 1
[  103.883928] 		ref#0: shared data backref parent 12915408896 count 1
[  103.885784] 	item 52 key (7550312448 168 12288) itemoff 14304 itemsize 37
[  103.887773] 		extent refs 1 gen 8 flags 1
[  103.889119] 		ref#0: shared data backref parent 12915408896 count 1
[  103.890982] 	item 53 key (7550324736 168 12288) itemoff 14267 itemsize 37
[  103.892949] 		extent refs 1 gen 8 flags 1
[  103.894317] 		ref#0: shared data backref parent 12915408896 count 1
[  103.896196] 	item 54 key (7550337024 168 12288) itemoff 14230 itemsize 37
[  103.898119] 		extent refs 1 gen 8 flags 1
[  103.899457] 		ref#0: shared data backref parent 12915408896 count 1
[  103.901260] 	item 55 key (7550349312 168 12288) itemoff 14193 itemsize 37
[  103.903243] 		extent refs 1 gen 8 flags 1
[  103.904590] 		ref#0: shared data backref parent 12915408896 count 1
[  103.906360] 	item 56 key (7550361600 168 12288) itemoff 14156 itemsize 37
[  103.908352] 		extent refs 1 gen 8 flags 1
[  103.909696] 		ref#0: shared data backref parent 12915408896 count 1
[  103.911544] 	item 57 key (7550373888 168 12288) itemoff 14119 itemsize 37
[  103.913484] 		extent refs 1 gen 8 flags 1
[  103.914840] 		ref#0: shared data backref parent 12915408896 count 1
[  103.916722] 	item 58 key (7550386176 168 12288) itemoff 14082 itemsize 37
[  103.918610] 		extent refs 1 gen 8 flags 1
[  103.919917] 		ref#0: shared data backref parent 12915408896 count 1
[  103.921737] 	item 59 key (7550398464 168 12288) itemoff 14045 itemsize 37
[  103.923633] 		extent refs 1 gen 8 flags 1
[  103.924949] 		ref#0: shared data backref parent 12915408896 count 1
[  103.926814] 	item 60 key (7550410752 168 12288) itemoff 14008 itemsize 37
[  103.928791] 		extent refs 1 gen 8 flags 1
[  103.930148] 		ref#0: shared data backref parent 12915408896 count 1
[  103.932015] 	item 61 key (7550423040 168 12288) itemoff 13971 itemsize 37
[  103.933978] 		extent refs 1 gen 8 flags 1
[  103.935300] 		ref#0: shared data backref parent 12915408896 count 1
[  103.937034] 	item 62 key (7550435328 168 12288) itemoff 13934 itemsize 37
[  103.938804] 		extent refs 1 gen 8 flags 1
[  103.940190] 		ref#0: shared data backref parent 12915408896 count 1
[  103.941935] 	item 63 key (7550447616 168 12288) itemoff 13897 itemsize 37
[  103.943663] 		extent refs 1 gen 8 flags 1
[  103.944856] 		ref#0: shared data backref parent 12915408896 count 1
[  103.946601] 	item 64 key (7550459904 168 12288) itemoff 13860 itemsize 37
[  103.948478] 		extent refs 1 gen 8 flags 1
[  103.949773] 		ref#0: shared data backref parent 12915408896 count 1
[  103.951587] 	item 65 key (7550472192 168 12288) itemoff 13823 itemsize 37
[  103.953509] 		extent refs 1 gen 8 flags 1
[  103.954809] 		ref#0: shared data backref parent 12915408896 count 1
[  103.956624] 	item 66 key (7550484480 168 12288) itemoff 13786 itemsize 37
[  103.958499] 		extent refs 1 gen 8 flags 1
[  103.959789] 		ref#0: shared data backref parent 12915425280 count 1
[  103.961575] 	item 67 key (7550496768 168 12288) itemoff 13749 itemsize 37
[  103.963453] 		extent refs 1 gen 8 flags 1
[  103.964763] 		ref#0: shared data backref parent 12915425280 count 1
[  103.966566] 	item 68 key (7550509056 168 12288) itemoff 13712 itemsize 37
[  103.968528] 		extent refs 1 gen 8 flags 1
[  103.969886] 		ref#0: shared data backref parent 12915425280 count 1
[  103.971797] 	item 69 key (7550521344 168 12288) itemoff 13675 itemsize 37
[  103.973722] 		extent refs 1 gen 8 flags 1
[  103.975026] 		ref#0: shared data backref parent 12915425280 count 1
[  103.976821] 	item 70 key (7550533632 168 12288) itemoff 13638 itemsize 37
[  103.978724] 		extent refs 1 gen 8 flags 1
[  103.980091] 		ref#0: shared data backref parent 12915425280 count 1
[  103.981963] 	item 71 key (7550545920 168 12288) itemoff 13601 itemsize 37
[  103.983946] 		extent refs 1 gen 8 flags 1
[  103.985294] 		ref#0: shared data backref parent 12915425280 count 1
[  103.987105] 	item 72 key (7550558208 168 12288) itemoff 13564 itemsize 37
[  103.997580] 		extent refs 1 gen 8 flags 1
[  103.998898] 		ref#0: shared data backref parent 12915425280 count 1
[  104.000716] 	item 73 key (7550570496 168 12288) itemoff 13527 itemsize 37
[  104.002621] 		extent refs 1 gen 8 flags 1
[  104.003925] 		ref#0: shared data backref parent 12915425280 count 1
[  104.005513] 	item 74 key (7550582784 168 12288) itemoff 13490 itemsize 37
[  104.007451] 		extent refs 1 gen 8 flags 1
[  104.008770] 		ref#0: shared data backref parent 12915425280 count 1
[  104.010582] 	item 75 key (7550595072 168 12288) itemoff 13453 itemsize 37
[  104.012568] 		extent refs 1 gen 8 flags 1
[  104.013918] 		ref#0: shared data backref parent 12915425280 count 1
[  104.015782] 	item 76 key (7550607360 168 12288) itemoff 13416 itemsize 37
[  104.017765] 		extent refs 1 gen 8 flags 1
[  104.019090] 		ref#0: shared data backref parent 12915425280 count 1
[  104.020901] 	item 77 key (7550619648 168 12288) itemoff 13379 itemsize 37
[  104.022809] 		extent refs 1 gen 8 flags 1
[  104.024160] 		ref#0: shared data backref parent 12915425280 count 1
[  104.025952] 	item 78 key (7550631936 168 12288) itemoff 13342 itemsize 37
[  104.027850] 		extent refs 1 gen 8 flags 1
[  104.029148] 		ref#0: shared data backref parent 12915425280 count 1
[  104.030891] 	item 79 key (7550644224 168 12288) itemoff 13305 itemsize 37
[  104.032760] 		extent refs 1 gen 8 flags 1
[  104.034032] 		ref#0: shared data backref parent 12915425280 count 1
[  104.035771] 	item 80 key (7550656512 168 12288) itemoff 13268 itemsize 37
[  104.037636] 		extent refs 1 gen 8 flags 1
[  104.038905] 		ref#0: shared data backref parent 12915425280 count 1
[  104.040709] 	item 81 key (7550668800 168 12288) itemoff 13231 itemsize 37
[  104.042632] 		extent refs 1 gen 8 flags 1
[  104.043927] 		ref#0: shared data backref parent 12915425280 count 1
[  104.045772] 	item 82 key (7550681088 168 12288) itemoff 13194 itemsize 37
[  104.047678] 		extent refs 1 gen 8 flags 1
[  104.048986] 		ref#0: shared data backref parent 12915425280 count 1
[  104.050793] 	item 83 key (7550693376 168 12288) itemoff 13157 itemsize 37
[  104.052722] 		extent refs 1 gen 8 flags 1
[  104.054036] 		ref#0: shared data backref parent 12915425280 count 1
[  104.055867] 	item 84 key (7550705664 168 12288) itemoff 13120 itemsize 37
[  104.057848] 		extent refs 1 gen 8 flags 1
[  104.059145] 		ref#0: shared data backref parent 12915425280 count 1
[  104.060941] 	item 85 key (7550717952 168 12288) itemoff 13083 itemsize 37
[  104.062845] 		extent refs 1 gen 8 flags 1
[  104.064128] 		ref#0: shared data backref parent 12915425280 count 1
[  104.065880] 	item 86 key (7550730240 168 12288) itemoff 13046 itemsize 37
[  104.067738] 		extent refs 1 gen 8 flags 1
[  104.069039] 		ref#0: shared data backref parent 12915425280 count 1
[  104.070841] 	item 87 key (7550742528 168 12288) itemoff 13009 itemsize 37
[  104.072842] 		extent refs 1 gen 8 flags 1
[  104.074149] 		ref#0: shared data backref parent 12915425280 count 1
[  104.075938] 	item 88 key (7550754816 168 12288) itemoff 12972 itemsize 37
[  104.077870] 		extent refs 1 gen 8 flags 1
[  104.079177] 		ref#0: shared data backref parent 12915425280 count 1
[  104.080995] 	item 89 key (7550767104 168 12288) itemoff 12935 itemsize 37
[  104.082924] 		extent refs 1 gen 8 flags 1
[  104.084297] 		ref#0: shared data backref parent 12915425280 count 1
[  104.086150] 	item 90 key (7550779392 168 12288) itemoff 12898 itemsize 37
[  104.088063] 		extent refs 1 gen 8 flags 1
[  104.089377] 		ref#0: shared data backref parent 12915425280 count 1
[  104.091150] 	item 91 key (7550791680 168 12288) itemoff 12861 itemsize 37
[  104.093056] 		extent refs 1 gen 8 flags 1
[  104.094350] 		ref#0: shared data backref parent 12915425280 count 1
[  104.096173] 	item 92 key (7550803968 168 12288) itemoff 12824 itemsize 37
[  104.098077] 		extent refs 1 gen 8 flags 1
[  104.099361] 		ref#0: shared data backref parent 12915425280 count 1
[  104.101121] 	item 93 key (7550816256 168 12288) itemoff 12787 itemsize 37
[  104.103008] 		extent refs 1 gen 8 flags 1
[  104.104297] 		ref#0: shared data backref parent 12915425280 count 1
[  104.106081] 	item 94 key (7550828544 168 12288) itemoff 12750 itemsize 37
[  104.107974] 		extent refs 1 gen 8 flags 1
[  104.109295] 		ref#0: shared data backref parent 12915425280 count 1
[  104.111066] 	item 95 key (7550840832 168 12288) itemoff 12713 itemsize 37
[  104.112967] 		extent refs 1 gen 8 flags 1
[  104.114261] 		ref#0: shared data backref parent 12915425280 count 1
[  104.115916] 	item 96 key (7550853120 168 12288) itemoff 12676 itemsize 37
[  104.117640] 		extent refs 1 gen 8 flags 1
[  104.118875] 		ref#0: shared data backref parent 12915425280 count 1
[  104.120509] 	item 97 key (7550865408 168 12288) itemoff 12639 itemsize 37
[  104.122156] 		extent refs 1 gen 8 flags 1
[  104.123308] 		ref#0: shared data backref parent 12915425280 count 1
[  104.124884] 	item 98 key (7550877696 168 12288) itemoff 12602 itemsize 37
[  104.126612] 		extent refs 1 gen 8 flags 1
[  104.127829] 		ref#0: shared data backref parent 12915425280 count 1
[  104.129434] 	item 99 key (7550889984 168 12288) itemoff 12565 itemsize 37
[  104.131155] 		extent refs 1 gen 8 flags 1
[  104.132501] 		ref#0: shared data backref parent 12915425280 count 1
[  104.134353] 	item 100 key (7550902272 168 12288) itemoff 12528 itemsize 37
[  104.136348] 		extent refs 1 gen 8 flags 1
[  104.137680] 		ref#0: shared data backref parent 12915425280 count 1
[  104.139515] 	item 101 key (7550914560 168 12288) itemoff 12491 itemsize 37
[  104.141450] 		extent refs 1 gen 8 flags 1
[  104.142747] 		ref#0: shared data backref parent 12915425280 count 1
[  104.144628] 	item 102 key (7550926848 168 12288) itemoff 12454 itemsize 37
[  104.146583] 		extent refs 1 gen 8 flags 1
[  104.147918] 		ref#0: shared data backref parent 12915425280 count 1
[  104.149800] 	item 103 key (7550939136 168 12288) itemoff 12417 itemsize 37
[  104.151820] 		extent refs 1 gen 8 flags 1
[  104.153185] 		ref#0: shared data backref parent 12915425280 count 1
[  104.155038] 	item 104 key (7550951424 168 12288) itemoff 12380 itemsize 37
[  104.156994] 		extent refs 1 gen 8 flags 1
[  104.158305] 		ref#0: shared data backref parent 12915425280 count 1
[  104.160059] 	item 105 key (7550963712 168 12288) itemoff 12343 itemsize 37
[  104.162041] 		extent refs 1 gen 8 flags 1
[  104.163343] 		ref#0: shared data backref parent 12915425280 count 1
[  104.165122] 	item 106 key (7550976000 168 12288) itemoff 12306 itemsize 37
[  104.166748] 		extent refs 1 gen 8 flags 1
[  104.167929] 		ref#0: shared data backref parent 12915441664 count 1
[  104.169624] 	item 107 key (7550988288 168 12288) itemoff 12269 itemsize 37
[  104.171436] 		extent refs 1 gen 8 flags 1
[  104.172703] 		ref#0: shared data backref parent 12915441664 count 1
[  104.174373] 	item 108 key (7551000576 168 12288) itemoff 12232 itemsize 37
[  104.176201] 		extent refs 1 gen 8 flags 1
[  104.177348] 		ref#0: shared data backref parent 12915441664 count 1
[  104.178988] 	item 109 key (7551012864 168 12288) itemoff 12195 itemsize 37
[  104.180779] 		extent refs 1 gen 8 flags 1
[  104.181980] 		ref#0: shared data backref parent 12915441664 count 1
[  104.183531] 	item 110 key (7551025152 168 12288) itemoff 12158 itemsize 37
[  104.185268] 		extent refs 1 gen 8 flags 1
[  104.186464] 		ref#0: shared data backref parent 12915441664 count 1
[  104.188077] BTRFS error (device vdb): unable to find ref byte nr 31883264 parent 0 root 18446744073709551607  owner 0 offset 0
[  104.190938] ------------[ cut here ]------------
[  104.192233] BTRFS: Transaction aborted (error -2)
[  104.193563] WARNING: CPU: 0 PID: 21011 at fs/btrfs/extent-tree.c:3061 __btrfs_free_extent+0x6c7/0x900 [btrfs]
[  104.196276] Modules linked in: xxhash_generic btrfs blake2b_generic libcrc32c crc32c_intel xor zstd_decompress zstd_compress xxhash lzo_compress lzo_decompress raid6_pq loop
[  104.200443] CPU: 0 PID: 21011 Comm: btrfs Tainted: G        W         5.7.0-rc6-default+ #1109
[  104.202857] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba527-rebuilt.opensuse.org 04/01/2014
[  104.206257] RIP: 0010:__btrfs_free_extent+0x6c7/0x900 [btrfs]
[  104.208018] Code: 8b 40 50 f0 48 0f ba a8 f8 1b 00 00 02 0f 92 c0 5e 84 c0 0f 85 44 d9 0c 00 be fe ff ff ff 48 c7 c7 e8 c4 27 c0 e8 19 92 ef f0 <0f> 0b e9 2c d9 0c 00 83 e8 01 49 8b 3e b9 11 00 00 00 48 8d 74 24
[  104.213484] RSP: 0018:ffffac68455678b0 EFLAGS: 00010282
[  104.215108] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000006
[  104.217146] RDX: ffff8f6e74a2d500 RSI: ffff8f6e74a2de08 RDI: ffff8f6e74a2d500
[  104.219176] RBP: 0000000001e68000 R08: 00000018426b0f4e R09: 0000000000000000
[  104.221126] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
[  104.223127] R13: fffffffffffffff7 R14: ffff8f6e74474a88 R15: 00000000fffffffe
[  104.225090] FS:  00007f9ab400b8c0(0000) GS:ffff8f6e7d600000(0000) knlGS:0000000000000000
[  104.227602] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  104.229265] CR2: 000055f4d1e28e18 CR3: 00000000747d3006 CR4: 0000000000160ef0
[  104.231230] Call Trace:
[  104.232217]  ? native_sched_clock_from_tsc+0x41/0xc0
[  104.233711]  ? btrfs_run_delayed_refs_for_head+0x197/0xcd0 [btrfs]
[  104.235452]  btrfs_run_delayed_refs_for_head+0x248/0xcd0 [btrfs]
[  104.237153]  ? _raw_read_unlock+0x1f/0x30
[  104.238443]  ? btrfs_merge_delayed_refs+0x3d3/0x480 [btrfs]
[  104.240071]  __btrfs_run_delayed_refs+0x9d/0x680 [btrfs]
[  104.241662]  ? join_transaction+0x15d/0x4c0 [btrfs]
[  104.243119]  ? kvm_sched_clock_read+0x14/0x30
[  104.244486]  ? sched_clock+0x5/0x10
[  104.245671]  ? sched_clock_cpu+0x15/0x130
[  104.246987]  btrfs_run_delayed_refs+0x86/0x1e0 [btrfs]
[  104.248558]  btrfs_commit_transaction+0x57/0xae0 [btrfs]
[  104.250129]  ? start_transaction+0xd2/0x5e0 [btrfs]
[  104.251577]  prepare_to_relocate+0x107/0x130 [btrfs]
[  104.253145]  relocate_block_group+0x5b/0x600 [btrfs]
[  104.254491]  btrfs_relocate_block_group+0x15e/0x340 [btrfs]
[  104.255998]  btrfs_relocate_chunk+0x38/0x110 [btrfs]
[  104.257371]  btrfs_shrink_device+0x214/0x530 [btrfs]
[  104.258701]  btrfs_rm_device+0x22e/0x7f0 [btrfs]
[  104.260057]  ? _copy_from_user+0x6a/0xa0
[  104.261241]  btrfs_ioctl+0x218f/0x2590 [btrfs]
[  104.262524]  ? __handle_mm_fault+0x1c1/0x740
[  104.263783]  ? do_user_addr_fault+0x1d8/0x3f0
[  104.265202]  ? kvm_sched_clock_read+0x14/0x30
[  104.266596]  ? sched_clock+0x5/0x10
[  104.267780]  ? sched_clock_cpu+0x15/0x130
[  104.269056]  ? do_user_addr_fault+0x1d8/0x3f0
[  104.270437]  ? ksys_ioctl+0x68/0xa0
[  104.271634]  ksys_ioctl+0x68/0xa0
[  104.272705]  __x64_sys_ioctl+0x16/0x20
[  104.273851]  do_syscall_64+0x50/0x210
[  104.275171]  entry_SYSCALL_64_after_hwframe+0x49/0xb3
[  104.276793] RIP: 0033:0x7f9ab4104227
[  104.278087] Code: 00 00 90 48 8b 05 69 8c 0c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 39 8c 0c 00 f7 d8 64 89 01 48
[  104.291704] RSP: 002b:00007ffc5dd62938 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[  104.294058] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f9ab4104227
[  104.295819] RDX: 00007ffc5dd62960 RSI: 000000005000943a RDI: 0000000000000003
[  104.297535] RBP: 00007ffc5dd64b00 R08: 00007ffc5dd62998 R09: 006764762f766564
[  104.299491] R10: 00007f9ab432ecf0 R11: 0000000000000202 R12: 0000000000000000
[  104.301225] R13: 00007ffc5dd62960 R14: 000055ee38909e8c R15: 0000000000000003
[  104.303016] irq event stamp: 1224070
[  104.304101] hardirqs last  enabled at (1224069): [<ffffffffb110a256>] console_unlock+0x436/0x590
[  104.306736] hardirqs last disabled at (1224070): [<ffffffffb1002b5b>] trace_hardirqs_off_thunk+0x1a/0x1c
[  104.309723] softirqs last  enabled at (1224058): [<ffffffffb1a0031e>] __do_softirq+0x31e/0x55d
[  104.312528] softirqs last disabled at (1224051): [<ffffffffb108d9ed>] irq_exit+0x9d/0xb0
[  104.315200] ---[ end trace 0cea18996e5ca625 ]---
[  104.316770] BTRFS: error (device vdb) in __btrfs_free_extent:3061: errno=-2 No such entry
[  104.319480] BTRFS info (device vdb): forced readonly
[  104.321109] BTRFS: error (device vdb) in btrfs_run_delayed_refs:2173: errno=-2 No such entry
[failed, exit status 1] [15:10:54]- output mismatch (see /tmp/fstests/results//btrfs/003.out.bad)
    --- tests/btrfs/003.out	2018-04-12 16:57:00.608225550 +0000
    +++ /tmp/fstests/results//btrfs/003.out.bad	2020-05-18 15:10:54.312000000 +0000
    @@ -1,2 +1,4 @@
     QA output created by 003
    -Silence is golden
    +ERROR: error removing device '/dev/vdg': Read-only file system
    +btrfs device delete failed
    +(see /tmp/fstests/results//btrfs/003.full for details)
    ...
    (Run 'diff -u /tmp/fstests/tests/btrfs/003.out /tmp/fstests/results//btrfs/003.out.bad'  to see the entire diff)

Re: [PATCH v3 0/3] btrfs: REF_COWS bit rework

[David Sterba]

On Mon, May 18, 2020 at 05:38:59PM +0200, David Sterba wrote:
> On Mon, May 18, 2020 at 05:03:31PM +0200, David Sterba wrote:
> > On Mon, May 18, 2020 at 01:13:34PM +0800, Qu Wenruo wrote:
> > - btrfs_read_tree_root
> > - btrfs_init_fs_root
> 
> With
> 
> 2291         location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
> 2292         root = btrfs_read_tree_root(tree_root, &location);
> 2293         if (IS_ERR(root)) {
> 2294                 ret = PTR_ERR(root);
> 2295                 goto out;
> 2296         }
> 2297         ret = btrfs_init_fs_root(root);
> 2298         if (ret)
> 2299                 goto out;
> 2300         set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
> 2301         fs_info->data_reloc_root = root;

Enough experimenting, I don't want to leave misc-next in a broken state
so will add get_fs_root back.