btrfs: Validate child tree block's level and first key

We have several reports about node pointer points to incorrect child tree blocks, which could have even wrong owner and level but still with valid generation and checksum.

Although btrfs check could handle it and print error message like:
leaf parent key incorrect 60670574592

Kernel doesn't have enough check on this type of corruption correctly. At least add such check to read_tree_block() and btrfs_read_buffer(), where we need two new parameters @first_key and @level to verify the child tree block.

There are call sites which can't use this new verfication by their nature:
1) Root node/leaf
It's referred by bytenr directly, so no @first_key to verify
2) Backref walker
FULL_BACKREF only has parent bytenr, no key to verify too.

Despite above cases, about 50% call sites can take advantage of this verification.

Another note of this verification is, it needs extra info from nodeptr, so it can't fit into current tree-checker framework, which is limited to node/leaf boundary.

Signed-off-by Qu Wenruo <[email protected]>

---
To David,

Since half of the new parameters are already filled, I'm not sure if it's a good idea to make a new function for this.

---
 fs/btrfs/backref.c     |  6 +++--
 fs/btrfs/ctree.c       | 25 +++++++++++++-----
 fs/btrfs/disk-io.c     | 69 ++++++++++++++++++++++++++++++++++++++++----------
 fs/btrfs/disk-io.h     |  8 +++---
 fs/btrfs/extent-tree.c |  6 ++++-
 fs/btrfs/print-tree.c  | 10 +++++---
 fs/btrfs/qgroup.c      |  7 +++--
 fs/btrfs/ref-verify.c  |  7 ++++-
 fs/btrfs/relocation.c  | 21 ++++++++++++---
 fs/btrfs/tree-log.c    | 12 ++++++---
 10 files changed, 132 insertions(+), 39 deletions(-)

fs/btrfs/backref.c +0 -0

index 26484648d090..3866b8ab20f1 100644

fs/btrfs/backref.c +0 -0

 				if (IS_ERR(eb)) {
 					ret = PTR_ERR(eb);
 					goto out;

fs/btrfs/ctree.c +19 -6

index b88a79e69ddf..e49ca6d529e8 100644
@@ -1436,7 +1436,7 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 	if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 		btrfs_tree_read_unlock(eb_root);
 		free_extent_buffer(eb_root);
-		old = read_tree_block(fs_info, logical, 0);
+		old = read_tree_block(fs_info, logical, 0, NULL, 0);
 		if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
 			if (!IS_ERR(old))
 				free_extent_buffer(old);

@@ -1656,6 +1656,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	btrfs_set_lock_blocking(parent);
 
 	for (i = start_slot; i <= end_slot; i++) {
+		struct btrfs_key first_key;
 		int close = 1;
 
 		btrfs_node_key(parent, &disk_key, i);

@@ -1665,6 +1666,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		progress_passed = 1;
 		blocknr = btrfs_node_blockptr(parent, i);
 		gen = btrfs_node_ptr_generation(parent, i);
+		btrfs_node_key_to_cpu(parent, &first_key, i);
 		if (last_block == 0)
 			last_block = blocknr;
 

@@ -1688,7 +1690,9 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 			uptodate = 0;
 		if (!cur || !uptodate) {
 			if (!cur) {
-				cur = read_tree_block(fs_info, blocknr, gen);
+				cur = read_tree_block(fs_info, blocknr, gen,
+						      &first_key,
+						      parent_level - 1);
 				if (IS_ERR(cur)) {
 					return PTR_ERR(cur);
 				} else if (!extent_buffer_uptodate(cur)) {

@@ -1696,7 +1700,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 					return -EIO;
 				}
 			} else if (!uptodate) {
-				err = btrfs_read_buffer(cur, gen);
+				err = btrfs_read_buffer(cur, gen, &first_key,
+							parent_level - 1);
 				if (err) {
 					free_extent_buffer(cur);
 					return err;

@@ -1849,14 +1854,17 @@ read_node_slot(struct btrfs_fs_info *fs_info, struct extent_buffer *parent,
 {
 	int level = btrfs_header_level(parent);
 	struct extent_buffer *eb;
+	struct btrfs_key first_key;
 
 	if (slot < 0 || slot >= btrfs_header_nritems(parent))
 		return ERR_PTR(-ENOENT);
 
 	BUG_ON(level == 0);
 
+	btrfs_node_key_to_cpu(parent, &first_key, slot);
 	eb = read_tree_block(fs_info, btrfs_node_blockptr(parent, slot),
-			     btrfs_node_ptr_generation(parent, slot));
+			     btrfs_node_ptr_generation(parent, slot),
+			     &first_key, level - 1);
 	if (!IS_ERR(eb) && !extent_buffer_uptodate(eb)) {
 		free_extent_buffer(eb);
 		eb = ERR_PTR(-EIO);

@@ -2445,10 +2453,14 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	u64 gen;
 	struct extent_buffer *b = *eb_ret;
 	struct extent_buffer *tmp;
+	struct btrfs_key first_key;
 	int ret;
+	int parent_level;
 
 	blocknr = btrfs_node_blockptr(b, slot);
 	gen = btrfs_node_ptr_generation(b, slot);
+	parent_level = btrfs_header_level(b);
+	btrfs_node_key_to_cpu(b, &first_key, slot);
 
 	tmp = find_extent_buffer(fs_info, blocknr);
 	if (tmp) {

@@ -2467,7 +2479,7 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 		btrfs_set_path_blocking(p);
 
 		/* now we're allowed to do a blocking uptodate check */
-		ret = btrfs_read_buffer(tmp, gen);
+		ret = btrfs_read_buffer(tmp, gen, &first_key, parent_level - 1);
 		if (!ret) {
 			*eb_ret = tmp;
 			return 0;

@@ -2494,7 +2506,8 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	btrfs_release_path(p);
 
 	ret = -EAGAIN;
-	tmp = read_tree_block(fs_info, blocknr, 0);
+	tmp = read_tree_block(fs_info, blocknr, 0, &first_key,
+			      parent_level - 1);
 	if (!IS_ERR(tmp)) {
 		/*
 		 * If the read above didn't mark this buffer up to date,

fs/btrfs/disk-io.c +56 -13

index 21f34ad0d411..46fbbbf34568 100644
@@ -428,13 +428,40 @@ static int btrfs_check_super_csum(struct btrfs_fs_info *fs_info,
 	return ret;
 }
 
+static int verify_parent_key(struct extent_buffer *eb,
+			     struct btrfs_key *first_key,
+			     int level)
+{
+	int found_level;
+	struct btrfs_key found_key;
+
+	if (!first_key)
+		return 0;
+
+	found_level = btrfs_header_level(eb);
+	if (found_level != level)
+		return -EIO;
+
+	if (found_level)
+		btrfs_node_key_to_cpu(eb, &found_key, 0);
+	else
+		btrfs_item_key_to_cpu(eb, &found_key, 0);
+	return btrfs_comp_cpu_keys(first_key, &found_key);
+}
+
 /*
  * helper to read a given tree block, doing retries as required when
  * the checksums don't match and we have alternate mirrors to try.
+ *
+ * @parent_transid:	expected transid, skip check if 0
+ * @first_key:		expected key of first slot, skip check if NULL
+ * @level:		expected level, skip check if @first_key is NULL
  */
 static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 					  struct extent_buffer *eb,
-					  u64 parent_transid)
+					  u64 parent_transid,
+					  struct btrfs_key *first_key,
+					  int level)
 {
 	struct extent_io_tree *io_tree;
 	int failed = 0;

@@ -449,11 +476,13 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		ret = read_extent_buffer_pages(io_tree, eb, WAIT_COMPLETE,
 					       mirror_num);
 		if (!ret) {
-			if (!verify_parent_transid(io_tree, eb,
+			if (verify_parent_transid(io_tree, eb,
 						   parent_transid, 0))
-				break;
-			else
 				ret = -EIO;
+			else if (verify_parent_key(eb, first_key, level))
+				ret = -EUCLEAN;
+			else
+				break;
 		}
 
 		/*

@@ -461,7 +490,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		 * there is no reason to read the other copies, they won't be
 		 * any less wrong.
 		 */
-		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
+		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags) ||
+		    ret == -EUCLEAN)
 			break;
 
 		num_copies = btrfs_num_copies(fs_info,

@@ -1062,8 +1092,17 @@ void btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
 			        buf->start, buf->start + buf->len - 1);
 }
 
+/*
+ * Read tree block at logical address @bytenr and do variant basic but critical
+ * verfication.
+ *
+ * @parent_transid:	expected transid of this tree block, skip check if 0
+ * @first_key:		expected key in slot 0, skip check if NULL
+ * @level:		eppected level, skip check if @first_key is NULL
+ */
 struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
-				      u64 parent_transid)
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level)
 {
 	struct extent_buffer *buf = NULL;
 	int ret;

@@ -1072,7 +1111,8 @@ struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
 	if (IS_ERR(buf))
 		return buf;
 
-	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					     first_key, level);
 	if (ret) {
 		free_extent_buffer(buf);
 		return ERR_PTR(ret);

@@ -1425,7 +1465,7 @@ static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
 	generation = btrfs_root_generation(&root->root_item);
 	root->node = read_tree_block(fs_info,
 				     btrfs_root_bytenr(&root->root_item),
-				     generation);
+				     generation, NULL, 0);
 	if (IS_ERR(root->node)) {
 		ret = PTR_ERR(root->node);
 		goto find_fail;

@@ -2289,7 +2329,8 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 	__setup_root(log_tree_root, fs_info, BTRFS_TREE_LOG_OBJECTID);
 
 	log_tree_root->node = read_tree_block(fs_info, bytenr,
-					      fs_info->generation + 1);
+					      fs_info->generation + 1,
+					      NULL, 0);
 	if (IS_ERR(log_tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read log tree");
 		ret = PTR_ERR(log_tree_root->node);

@@ -2746,7 +2787,7 @@ int open_ctree(struct super_block *sb,
 
 	chunk_root->node = read_tree_block(fs_info,
 					   btrfs_super_chunk_root(disk_super),
-					   generation);
+					   generation, NULL, 0);
 	if (IS_ERR(chunk_root->node) ||
 	    !extent_buffer_uptodate(chunk_root->node)) {
 		btrfs_err(fs_info, "failed to read chunk root");

@@ -2783,7 +2824,7 @@ int open_ctree(struct super_block *sb,
 
 	tree_root->node = read_tree_block(fs_info,
 					  btrfs_super_root(disk_super),
-					  generation);
+					  generation, NULL, 0);
 	if (IS_ERR(tree_root->node) ||
 	    !extent_buffer_uptodate(tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read tree root");

@@ -3890,12 +3931,14 @@ void btrfs_btree_balance_dirty_nodelay(struct btrfs_fs_info *fs_info)
 	__btrfs_btree_balance_dirty(fs_info, 0);
 }
 
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level)
 {
 	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					      first_key, level);
 }
 
 static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info)

fs/btrfs/disk-io.h +5 -3

index 301151a50ac1..3a1f4b2fa6ff 100644
@@ -52,8 +52,9 @@ static inline u64 btrfs_sb_offset(int mirror)
 struct btrfs_device;
 struct btrfs_fs_devices;
 
-struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info,
-				      u64 bytenr, u64 parent_transid);
+struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level);
 void readahead_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr);
 int reada_tree_block_flagged(struct btrfs_fs_info *fs_info, u64 bytenr,
 			 int mirror_num, struct extent_buffer **eb);

@@ -123,7 +124,8 @@ static inline void btrfs_put_fs_root(struct btrfs_root *root)
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
 int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
 			  int atomic);
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level);
 u32 btrfs_csum_data(const char *data, u32 seed, size_t len);
 void btrfs_csum_final(u32 crc, u8 *result);
 blk_status_t btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,

fs/btrfs/extent-tree.c +5 -1

index c1618ab9fecf..156ac8e4ff62 100644
@@ -8701,6 +8701,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	u64 parent;
 	u32 blocksize;
 	struct btrfs_key key;
+	struct btrfs_key first_key;
 	struct extent_buffer *next;
 	int level = wc->level;
 	int reada = 0;

@@ -8721,6 +8722,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	}
 
 	bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]);
+	btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+			      path->slots[level]);
 	blocksize = fs_info->nodesize;
 
 	next = find_extent_buffer(fs_info, bytenr);

@@ -8785,7 +8788,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	if (!next) {
 		if (reada && level == 1)
 			reada_walk_down(trans, root, wc, path);
-		next = read_tree_block(fs_info, bytenr, generation);
+		next = read_tree_block(fs_info, bytenr, generation, &first_key,
+				       level - 1);
 		if (IS_ERR(next)) {
 			return PTR_ERR(next);
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/print-tree.c +7 -3

index 569205e651c7..b00d88291085 100644
@@ -365,9 +365,13 @@ void btrfs_print_tree(struct extent_buffer *c)
 		       btrfs_node_blockptr(c, i));
 	}
 	for (i = 0; i < nr; i++) {
-		struct extent_buffer *next = read_tree_block(fs_info,
-					btrfs_node_blockptr(c, i),
-					btrfs_node_ptr_generation(c, i));
+		struct btrfs_key first_key;
+		struct extent_buffer *next;
+
+		btrfs_node_key_to_cpu(c, &first_key, i);
+		next = read_tree_block(fs_info, btrfs_node_blockptr(c, i),
+				       btrfs_node_ptr_generation(c, i),
+				       &first_key, level - 1);
 		if (IS_ERR(next)) {
 			continue;
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/qgroup.c +5 -2

index aa259d6986e1..86698702d08c 100644
@@ -1614,7 +1614,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 		return 0;
 
 	if (!extent_buffer_uptodate(root_eb)) {
-		ret = btrfs_read_buffer(root_eb, root_gen);
+		ret = btrfs_read_buffer(root_eb, root_gen, NULL, 0);
 		if (ret)
 			goto out;
 	}

@@ -1645,6 +1645,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 	level = root_level;
 	while (level >= 0) {
 		if (path->nodes[level] == NULL) {
+			struct btrfs_key first_key;
 			int parent_slot;
 			u64 child_gen;
 			u64 child_bytenr;

@@ -1657,8 +1658,10 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 			parent_slot = path->slots[level + 1];
 			child_bytenr = btrfs_node_blockptr(eb, parent_slot);
 			child_gen = btrfs_node_ptr_generation(eb, parent_slot);
+			btrfs_node_key_to_cpu(eb, &first_key, parent_slot);
 
-			eb = read_tree_block(fs_info, child_bytenr, child_gen);
+			eb = read_tree_block(fs_info, child_bytenr, child_gen,
+					     &first_key, level);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				goto out;

fs/btrfs/ref-verify.c +6 -1

index 171f3cce30e6..25c767635426 100644
@@ -579,11 +579,16 @@ static int walk_down_tree(struct btrfs_root *root, struct btrfs_path *path,
 
 	while (level >= 0) {
 		if (level) {
+			struct btrfs_key first_key;
+
 			block_bytenr = btrfs_node_blockptr(path->nodes[level],
 							   path->slots[level]);
 			gen = btrfs_node_ptr_generation(path->nodes[level],
 							path->slots[level]);
-			eb = read_tree_block(fs_info, block_bytenr, gen);
+			btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+					      path->slots[level]);
+			eb = read_tree_block(fs_info, block_bytenr, gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb))
 				return PTR_ERR(eb);
 			if (!extent_buffer_uptodate(eb)) {

fs/btrfs/relocation.c +17 -4

index cd2298d185dd..4a309f8c9535 100644
@@ -1839,6 +1839,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 
 	parent = eb;
 	while (1) {
+		struct btrfs_key first_key;
+
 		level = btrfs_header_level(parent);
 		BUG_ON(level < lowest_level);
 

@@ -1852,6 +1854,7 @@ int replace_path(struct btrfs_trans_handle *trans,
 		old_bytenr = btrfs_node_blockptr(parent, slot);
 		blocksize = fs_info->nodesize;
 		old_ptr_gen = btrfs_node_ptr_generation(parent, slot);
+		btrfs_node_key_to_cpu(parent, &key, slot);
 
 		if (level <= max_level) {
 			eb = path->nodes[level];

@@ -1876,7 +1879,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 				break;
 			}
 
-			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen);
+			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				break;

@@ -2036,6 +2040,8 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 	last_snapshot = btrfs_root_last_snapshot(&root->root_item);
 
 	for (i = *level; i > 0; i--) {
+		struct btrfs_key first_key;
+
 		eb = path->nodes[i];
 		nritems = btrfs_header_nritems(eb);
 		while (path->slots[i] < nritems) {

@@ -2056,7 +2062,9 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 		}
 
 		bytenr = btrfs_node_blockptr(eb, path->slots[i]);
-		eb = read_tree_block(fs_info, bytenr, ptr_gen);
+		btrfs_node_key_to_cpu(eb, &first_key, path->slots[i]);
+		eb = read_tree_block(fs_info, bytenr, ptr_gen, &first_key,
+				     i - 1);
 		if (IS_ERR(eb)) {
 			return PTR_ERR(eb);
 		} else if (!extent_buffer_uptodate(eb)) {

@@ -2714,6 +2722,8 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 	path->lowest_level = node->level + 1;
 	rc->backref_cache.path[node->level] = node;
 	list_for_each_entry(edge, &node->upper, list[LOWER]) {
+		struct btrfs_key first_key;
+
 		cond_resched();
 
 		upper = edge->node[UPPER];

@@ -2779,7 +2789,9 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 
 		blocksize = root->fs_info->nodesize;
 		generation = btrfs_node_ptr_generation(upper->eb, slot);
-		eb = read_tree_block(fs_info, bytenr, generation);
+		btrfs_node_key_to_cpu(upper->eb, &first_key, slot);
+		eb = read_tree_block(fs_info, bytenr, generation, &first_key,
+				     upper->level - 1);
 		if (IS_ERR(eb)) {
 			err = PTR_ERR(eb);
 			goto next;

@@ -2944,7 +2956,8 @@ static int get_tree_block_key(struct btrfs_fs_info *fs_info,
 	struct extent_buffer *eb;
 
 	BUG_ON(block->key_ready);
-	eb = read_tree_block(fs_info, block->bytenr, block->key.offset);
+	eb = read_tree_block(fs_info, block->bytenr, block->key.offset, NULL,
+			     0);
 	if (IS_ERR(eb)) {
 		return PTR_ERR(eb);
 	} else if (!extent_buffer_uptodate(eb)) {

fs/btrfs/tree-log.c +8 -4

index 434457794c27..b98a1801b406 100644
@@ -304,7 +304,7 @@ static int process_one_buffer(struct btrfs_root *log,
 	 * pin down any logged extents, so we have to read the block.
 	 */
 	if (btrfs_fs_incompat(fs_info, MIXED_GROUPS)) {
-		ret = btrfs_read_buffer(eb, gen);
+		ret = btrfs_read_buffer(eb, gen, NULL, 0);
 		if (ret)
 			return ret;
 	}

@@ -2420,7 +2420,7 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 	int i;
 	int ret;
 
-	ret = btrfs_read_buffer(eb, gen);
+	ret = btrfs_read_buffer(eb, gen, NULL, 0);
 	if (ret)
 		return ret;
 

@@ -2537,6 +2537,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 	WARN_ON(*level >= BTRFS_MAX_LEVEL);
 
 	while (*level > 0) {
+		struct btrfs_key first_key;
+
 		WARN_ON(*level < 0);
 		WARN_ON(*level >= BTRFS_MAX_LEVEL);
 		cur = path->nodes[*level];

@@ -2549,6 +2551,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
 		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+		btrfs_node_key_to_cpu(cur, &first_key, path->slots[*level]);
 		blocksize = fs_info->nodesize;
 
 		parent = path->nodes[*level];

@@ -2567,7 +2570,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 			path->slots[*level]++;
 			if (wc->free) {
-				ret = btrfs_read_buffer(next, ptr_gen);
+				ret = btrfs_read_buffer(next, ptr_gen,
+							&first_key, *level - 1);
 				if (ret) {
 					free_extent_buffer(next);
 					return ret;

@@ -2597,7 +2601,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 			free_extent_buffer(next);
 			continue;
 		}
-		ret = btrfs_read_buffer(next, ptr_gen);
+		ret = btrfs_read_buffer(next, ptr_gen, &first_key, *level - 1);
 		if (ret) {
 			free_extent_buffer(next);
 			return ret;

Quoted 52 lines · Qu Wenruo

We have several reports about node pointer points to incorrect child tree blocks, which could have even wrong owner and level but still with valid generation and checksum.

Although btrfs check could handle it and print error message like:
leaf parent key incorrect 60670574592

Kernel doesn't have enough check on this type of corruption correctly. At least add such check to read_tree_block() and btrfs_read_buffer(), where we need two new parameters @first_key and @level to verify the child tree block.

There are call sites which can't use this new verfication by their nature:
1) Root node/leaf
It's referred by bytenr directly, so no @first_key to verify
2) Backref walker
FULL_BACKREF only has parent bytenr, no key to verify too.

Despite above cases, about 50% call sites can take advantage of this verification.

Another note of this verification is, it needs extra info from nodeptr, so it can't fit into current tree-checker framework, which is limited to node/leaf boundary.

Signed-off-by Qu Wenruo <[email protected]>

---
To David,

Since half of the new parameters are already filled, I'm not sure if it's a good idea to make a new function for this.

---
 fs/btrfs/backref.c     |  6 +++--
 fs/btrfs/ctree.c       | 25 +++++++++++++-----
 fs/btrfs/disk-io.c     | 69 ++++++++++++++++++++++++++++++++++++++++----------
 fs/btrfs/disk-io.h     |  8 +++---
 fs/btrfs/extent-tree.c |  6 ++++-
 fs/btrfs/print-tree.c  | 10 +++++---
 fs/btrfs/qgroup.c      |  7 +++--
 fs/btrfs/ref-verify.c  |  7 ++++-
 fs/btrfs/relocation.c  | 21 ++++++++++++---
 fs/btrfs/tree-log.c    | 12 ++++++---
 10 files changed, 132 insertions(+), 39 deletions(-)

fs/btrfs/backref.c +0 -0

index 26484648d090..3866b8ab20f1 100644
@@ -738,7 +738,8 @@ static int add_missing_keys(struct btrfs_fs_info *fs_info,
 		BUG_ON(ref->key_for_search.type);
 		BUG_ON(!ref->wanted_disk_byte);
 

fs/btrfs/backref.c static int add_missing_keys(struct btrfs_fs_info *fs_info, · line 742

Quoted 3 lines

fs/btrfs/backref.c +2 -1

-		eb = read_tree_block(fs_info, ref->wanted_disk_byte, 0);
+		eb = read_tree_block(fs_info, ref->wanted_disk_byte, 0, NULL,
+				     0);

I had a call with Jeff and he suggested that for this call side you can actually derive the parameters. I guess the level could really be ref->level - 1 (based on quick scan of add_delayed_refs, though you'd have to check it). The first key could be a bit tricky though. I'm not too familiar with this code.

Quoted 8 lines

fs/btrfs/backref.c +0 -0

 		if (IS_ERR(eb)) {
 			free_pref(ref);
 			return PTR_ERR(eb);

@@ -1291,7 +1292,8 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
 			    ref->level == 0) {
 				struct extent_buffer *eb;
 

fs/btrfs/backref.c static int find_parent_nodes(struct btrfs_trans_handle *trans, · line 1296

Quoted 3 lines

fs/btrfs/backref.c +2 -1

-				eb = read_tree_block(fs_info, ref->parent, 0);
+				eb = read_tree_block(fs_info, ref->parent, 0,
+						     NULL, 0);

He also suggested something like that could be done for this.

Quoted 528 lines

fs/btrfs/backref.c +0 -0

 				if (IS_ERR(eb)) {
 					ret = PTR_ERR(eb);
 					goto out;

fs/btrfs/ctree.c +19 -6

index b88a79e69ddf..e49ca6d529e8 100644
@@ -1436,7 +1436,7 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 	if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 		btrfs_tree_read_unlock(eb_root);
 		free_extent_buffer(eb_root);
-		old = read_tree_block(fs_info, logical, 0);
+		old = read_tree_block(fs_info, logical, 0, NULL, 0);
 		if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
 			if (!IS_ERR(old))
 				free_extent_buffer(old);

@@ -1656,6 +1656,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	btrfs_set_lock_blocking(parent);
 
 	for (i = start_slot; i <= end_slot; i++) {
+		struct btrfs_key first_key;
 		int close = 1;
 
 		btrfs_node_key(parent, &disk_key, i);

@@ -1665,6 +1666,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		progress_passed = 1;
 		blocknr = btrfs_node_blockptr(parent, i);
 		gen = btrfs_node_ptr_generation(parent, i);
+		btrfs_node_key_to_cpu(parent, &first_key, i);
 		if (last_block == 0)
 			last_block = blocknr;
 

@@ -1688,7 +1690,9 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 			uptodate = 0;
 		if (!cur || !uptodate) {
 			if (!cur) {
-				cur = read_tree_block(fs_info, blocknr, gen);
+				cur = read_tree_block(fs_info, blocknr, gen,
+						      &first_key,
+						      parent_level - 1);
 				if (IS_ERR(cur)) {
 					return PTR_ERR(cur);
 				} else if (!extent_buffer_uptodate(cur)) {

@@ -1696,7 +1700,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 					return -EIO;
 				}
 			} else if (!uptodate) {
-				err = btrfs_read_buffer(cur, gen);
+				err = btrfs_read_buffer(cur, gen, &first_key,
+							parent_level - 1);
 				if (err) {
 					free_extent_buffer(cur);
 					return err;

@@ -1849,14 +1854,17 @@ read_node_slot(struct btrfs_fs_info *fs_info, struct extent_buffer *parent,
 {
 	int level = btrfs_header_level(parent);
 	struct extent_buffer *eb;
+	struct btrfs_key first_key;
 
 	if (slot < 0 || slot >= btrfs_header_nritems(parent))
 		return ERR_PTR(-ENOENT);
 
 	BUG_ON(level == 0);
 
+	btrfs_node_key_to_cpu(parent, &first_key, slot);
 	eb = read_tree_block(fs_info, btrfs_node_blockptr(parent, slot),
-			     btrfs_node_ptr_generation(parent, slot));
+			     btrfs_node_ptr_generation(parent, slot),
+			     &first_key, level - 1);
 	if (!IS_ERR(eb) && !extent_buffer_uptodate(eb)) {
 		free_extent_buffer(eb);
 		eb = ERR_PTR(-EIO);

@@ -2445,10 +2453,14 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	u64 gen;
 	struct extent_buffer *b = *eb_ret;
 	struct extent_buffer *tmp;
+	struct btrfs_key first_key;
 	int ret;
+	int parent_level;
 
 	blocknr = btrfs_node_blockptr(b, slot);
 	gen = btrfs_node_ptr_generation(b, slot);
+	parent_level = btrfs_header_level(b);
+	btrfs_node_key_to_cpu(b, &first_key, slot);
 
 	tmp = find_extent_buffer(fs_info, blocknr);
 	if (tmp) {

@@ -2467,7 +2479,7 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 		btrfs_set_path_blocking(p);
 
 		/* now we're allowed to do a blocking uptodate check */
-		ret = btrfs_read_buffer(tmp, gen);
+		ret = btrfs_read_buffer(tmp, gen, &first_key, parent_level - 1);
 		if (!ret) {
 			*eb_ret = tmp;
 			return 0;

@@ -2494,7 +2506,8 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	btrfs_release_path(p);
 
 	ret = -EAGAIN;
-	tmp = read_tree_block(fs_info, blocknr, 0);
+	tmp = read_tree_block(fs_info, blocknr, 0, &first_key,
+			      parent_level - 1);
 	if (!IS_ERR(tmp)) {
 		/*
 		 * If the read above didn't mark this buffer up to date,

fs/btrfs/disk-io.c +56 -13

index 21f34ad0d411..46fbbbf34568 100644
@@ -428,13 +428,40 @@ static int btrfs_check_super_csum(struct btrfs_fs_info *fs_info,
 	return ret;
 }
 
+static int verify_parent_key(struct extent_buffer *eb,
+			     struct btrfs_key *first_key,
+			     int level)
+{
+	int found_level;
+	struct btrfs_key found_key;
+
+	if (!first_key)
+		return 0;
+
+	found_level = btrfs_header_level(eb);
+	if (found_level != level)
+		return -EIO;
+
+	if (found_level)
+		btrfs_node_key_to_cpu(eb, &found_key, 0);
+	else
+		btrfs_item_key_to_cpu(eb, &found_key, 0);
+	return btrfs_comp_cpu_keys(first_key, &found_key);
+}
+
 /*
  * helper to read a given tree block, doing retries as required when
  * the checksums don't match and we have alternate mirrors to try.
+ *
+ * @parent_transid:	expected transid, skip check if 0
+ * @first_key:		expected key of first slot, skip check if NULL
+ * @level:		expected level, skip check if @first_key is NULL
  */
 static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 					  struct extent_buffer *eb,
-					  u64 parent_transid)
+					  u64 parent_transid,
+					  struct btrfs_key *first_key,
+					  int level)
 {
 	struct extent_io_tree *io_tree;
 	int failed = 0;

@@ -449,11 +476,13 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		ret = read_extent_buffer_pages(io_tree, eb, WAIT_COMPLETE,
 					       mirror_num);
 		if (!ret) {
-			if (!verify_parent_transid(io_tree, eb,
+			if (verify_parent_transid(io_tree, eb,
 						   parent_transid, 0))
-				break;
-			else
 				ret = -EIO;
+			else if (verify_parent_key(eb, first_key, level))
+				ret = -EUCLEAN;
+			else
+				break;
 		}
 
 		/*

@@ -461,7 +490,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		 * there is no reason to read the other copies, they won't be
 		 * any less wrong.
 		 */
-		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
+		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags) ||
+		    ret == -EUCLEAN)
 			break;
 
 		num_copies = btrfs_num_copies(fs_info,

@@ -1062,8 +1092,17 @@ void btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
 			        buf->start, buf->start + buf->len - 1);
 }
 
+/*
+ * Read tree block at logical address @bytenr and do variant basic but critical
+ * verfication.
+ *
+ * @parent_transid:	expected transid of this tree block, skip check if 0
+ * @first_key:		expected key in slot 0, skip check if NULL
+ * @level:		eppected level, skip check if @first_key is NULL
+ */
 struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
-				      u64 parent_transid)
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level)
 {
 	struct extent_buffer *buf = NULL;
 	int ret;

@@ -1072,7 +1111,8 @@ struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
 	if (IS_ERR(buf))
 		return buf;
 
-	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					     first_key, level);
 	if (ret) {
 		free_extent_buffer(buf);
 		return ERR_PTR(ret);

@@ -1425,7 +1465,7 @@ static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
 	generation = btrfs_root_generation(&root->root_item);
 	root->node = read_tree_block(fs_info,
 				     btrfs_root_bytenr(&root->root_item),
-				     generation);
+				     generation, NULL, 0);
 	if (IS_ERR(root->node)) {
 		ret = PTR_ERR(root->node);
 		goto find_fail;

@@ -2289,7 +2329,8 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 	__setup_root(log_tree_root, fs_info, BTRFS_TREE_LOG_OBJECTID);
 
 	log_tree_root->node = read_tree_block(fs_info, bytenr,
-					      fs_info->generation + 1);
+					      fs_info->generation + 1,
+					      NULL, 0);
 	if (IS_ERR(log_tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read log tree");
 		ret = PTR_ERR(log_tree_root->node);

@@ -2746,7 +2787,7 @@ int open_ctree(struct super_block *sb,
 
 	chunk_root->node = read_tree_block(fs_info,
 					   btrfs_super_chunk_root(disk_super),
-					   generation);
+					   generation, NULL, 0);
 	if (IS_ERR(chunk_root->node) ||
 	    !extent_buffer_uptodate(chunk_root->node)) {
 		btrfs_err(fs_info, "failed to read chunk root");

@@ -2783,7 +2824,7 @@ int open_ctree(struct super_block *sb,
 
 	tree_root->node = read_tree_block(fs_info,
 					  btrfs_super_root(disk_super),
-					  generation);
+					  generation, NULL, 0);
 	if (IS_ERR(tree_root->node) ||
 	    !extent_buffer_uptodate(tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read tree root");

@@ -3890,12 +3931,14 @@ void btrfs_btree_balance_dirty_nodelay(struct btrfs_fs_info *fs_info)
 	__btrfs_btree_balance_dirty(fs_info, 0);
 }
 
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level)
 {
 	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					      first_key, level);
 }
 
 static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info)

fs/btrfs/disk-io.h +5 -3

index 301151a50ac1..3a1f4b2fa6ff 100644
@@ -52,8 +52,9 @@ static inline u64 btrfs_sb_offset(int mirror)
 struct btrfs_device;
 struct btrfs_fs_devices;
 
-struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info,
-				      u64 bytenr, u64 parent_transid);
+struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level);
 void readahead_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr);
 int reada_tree_block_flagged(struct btrfs_fs_info *fs_info, u64 bytenr,
 			 int mirror_num, struct extent_buffer **eb);

@@ -123,7 +124,8 @@ static inline void btrfs_put_fs_root(struct btrfs_root *root)
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
 int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
 			  int atomic);
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level);
 u32 btrfs_csum_data(const char *data, u32 seed, size_t len);
 void btrfs_csum_final(u32 crc, u8 *result);
 blk_status_t btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,

fs/btrfs/extent-tree.c +5 -1

index c1618ab9fecf..156ac8e4ff62 100644
@@ -8701,6 +8701,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	u64 parent;
 	u32 blocksize;
 	struct btrfs_key key;
+	struct btrfs_key first_key;
 	struct extent_buffer *next;
 	int level = wc->level;
 	int reada = 0;

@@ -8721,6 +8722,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	}
 
 	bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]);
+	btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+			      path->slots[level]);
 	blocksize = fs_info->nodesize;
 
 	next = find_extent_buffer(fs_info, bytenr);

@@ -8785,7 +8788,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	if (!next) {
 		if (reada && level == 1)
 			reada_walk_down(trans, root, wc, path);
-		next = read_tree_block(fs_info, bytenr, generation);
+		next = read_tree_block(fs_info, bytenr, generation, &first_key,
+				       level - 1);
 		if (IS_ERR(next)) {
 			return PTR_ERR(next);
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/print-tree.c +7 -3

index 569205e651c7..b00d88291085 100644
@@ -365,9 +365,13 @@ void btrfs_print_tree(struct extent_buffer *c)
 		       btrfs_node_blockptr(c, i));
 	}
 	for (i = 0; i < nr; i++) {
-		struct extent_buffer *next = read_tree_block(fs_info,
-					btrfs_node_blockptr(c, i),
-					btrfs_node_ptr_generation(c, i));
+		struct btrfs_key first_key;
+		struct extent_buffer *next;
+
+		btrfs_node_key_to_cpu(c, &first_key, i);
+		next = read_tree_block(fs_info, btrfs_node_blockptr(c, i),
+				       btrfs_node_ptr_generation(c, i),
+				       &first_key, level - 1);
 		if (IS_ERR(next)) {
 			continue;
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/qgroup.c +5 -2

index aa259d6986e1..86698702d08c 100644
@@ -1614,7 +1614,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 		return 0;
 
 	if (!extent_buffer_uptodate(root_eb)) {
-		ret = btrfs_read_buffer(root_eb, root_gen);
+		ret = btrfs_read_buffer(root_eb, root_gen, NULL, 0);
 		if (ret)
 			goto out;
 	}

@@ -1645,6 +1645,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 	level = root_level;
 	while (level >= 0) {
 		if (path->nodes[level] == NULL) {
+			struct btrfs_key first_key;
 			int parent_slot;
 			u64 child_gen;
 			u64 child_bytenr;

@@ -1657,8 +1658,10 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 			parent_slot = path->slots[level + 1];
 			child_bytenr = btrfs_node_blockptr(eb, parent_slot);
 			child_gen = btrfs_node_ptr_generation(eb, parent_slot);
+			btrfs_node_key_to_cpu(eb, &first_key, parent_slot);
 
-			eb = read_tree_block(fs_info, child_bytenr, child_gen);
+			eb = read_tree_block(fs_info, child_bytenr, child_gen,
+					     &first_key, level);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				goto out;

fs/btrfs/ref-verify.c +6 -1

index 171f3cce30e6..25c767635426 100644
@@ -579,11 +579,16 @@ static int walk_down_tree(struct btrfs_root *root, struct btrfs_path *path,
 
 	while (level >= 0) {
 		if (level) {
+			struct btrfs_key first_key;
+
 			block_bytenr = btrfs_node_blockptr(path->nodes[level],
 							   path->slots[level]);
 			gen = btrfs_node_ptr_generation(path->nodes[level],
 							path->slots[level]);
-			eb = read_tree_block(fs_info, block_bytenr, gen);
+			btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+					      path->slots[level]);
+			eb = read_tree_block(fs_info, block_bytenr, gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb))
 				return PTR_ERR(eb);
 			if (!extent_buffer_uptodate(eb)) {

fs/btrfs/relocation.c +17 -4

index cd2298d185dd..4a309f8c9535 100644
@@ -1839,6 +1839,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 
 	parent = eb;
 	while (1) {
+		struct btrfs_key first_key;
+
 		level = btrfs_header_level(parent);
 		BUG_ON(level < lowest_level);
 

@@ -1852,6 +1854,7 @@ int replace_path(struct btrfs_trans_handle *trans,
 		old_bytenr = btrfs_node_blockptr(parent, slot);
 		blocksize = fs_info->nodesize;
 		old_ptr_gen = btrfs_node_ptr_generation(parent, slot);
+		btrfs_node_key_to_cpu(parent, &key, slot);
 
 		if (level <= max_level) {
 			eb = path->nodes[level];

@@ -1876,7 +1879,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 				break;
 			}
 
-			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen);
+			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				break;

@@ -2036,6 +2040,8 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 	last_snapshot = btrfs_root_last_snapshot(&root->root_item);
 
 	for (i = *level; i > 0; i--) {
+		struct btrfs_key first_key;
+
 		eb = path->nodes[i];
 		nritems = btrfs_header_nritems(eb);
 		while (path->slots[i] < nritems) {

@@ -2056,7 +2062,9 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 		}
 
 		bytenr = btrfs_node_blockptr(eb, path->slots[i]);
-		eb = read_tree_block(fs_info, bytenr, ptr_gen);
+		btrfs_node_key_to_cpu(eb, &first_key, path->slots[i]);
+		eb = read_tree_block(fs_info, bytenr, ptr_gen, &first_key,
+				     i - 1);
 		if (IS_ERR(eb)) {
 			return PTR_ERR(eb);
 		} else if (!extent_buffer_uptodate(eb)) {

@@ -2714,6 +2722,8 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 	path->lowest_level = node->level + 1;
 	rc->backref_cache.path[node->level] = node;
 	list_for_each_entry(edge, &node->upper, list[LOWER]) {
+		struct btrfs_key first_key;
+
 		cond_resched();
 
 		upper = edge->node[UPPER];

@@ -2779,7 +2789,9 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 
 		blocksize = root->fs_info->nodesize;
 		generation = btrfs_node_ptr_generation(upper->eb, slot);
-		eb = read_tree_block(fs_info, bytenr, generation);
+		btrfs_node_key_to_cpu(upper->eb, &first_key, slot);
+		eb = read_tree_block(fs_info, bytenr, generation, &first_key,
+				     upper->level - 1);
 		if (IS_ERR(eb)) {
 			err = PTR_ERR(eb);
 			goto next;

@@ -2944,7 +2956,8 @@ static int get_tree_block_key(struct btrfs_fs_info *fs_info,
 	struct extent_buffer *eb;
 
 	BUG_ON(block->key_ready);
-	eb = read_tree_block(fs_info, block->bytenr, block->key.offset);
+	eb = read_tree_block(fs_info, block->bytenr, block->key.offset, NULL,
+			     0);
 	if (IS_ERR(eb)) {
 		return PTR_ERR(eb);
 	} else if (!extent_buffer_uptodate(eb)) {

fs/btrfs/tree-log.c +7 -3

index 434457794c27..b98a1801b406 100644
@@ -304,7 +304,7 @@ static int process_one_buffer(struct btrfs_root *log,
 	 * pin down any logged extents, so we have to read the block.
 	 */
 	if (btrfs_fs_incompat(fs_info, MIXED_GROUPS)) {
-		ret = btrfs_read_buffer(eb, gen);
+		ret = btrfs_read_buffer(eb, gen, NULL, 0);
 		if (ret)
 			return ret;
 	}

@@ -2420,7 +2420,7 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 	int i;
 	int ret;
 
-	ret = btrfs_read_buffer(eb, gen);
+	ret = btrfs_read_buffer(eb, gen, NULL, 0);
 	if (ret)
 		return ret;
 

@@ -2537,6 +2537,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 	WARN_ON(*level >= BTRFS_MAX_LEVEL);
 
 	while (*level > 0) {
+		struct btrfs_key first_key;
+
 		WARN_ON(*level < 0);
 		WARN_ON(*level >= BTRFS_MAX_LEVEL);
 		cur = path->nodes[*level];

@@ -2549,6 +2551,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
 		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+		btrfs_node_key_to_cpu(cur, &first_key, path->slots[*level]);
 		blocksize = fs_info->nodesize;
 
 		parent = path->nodes[*level];

@@ -2567,7 +2570,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 			path->slots[*level]++;
 			if (wc->free) {
-				ret = btrfs_read_buffer(next, ptr_gen);
+				ret = btrfs_read_buffer(next, ptr_gen,
+							&first_key, *level - 1);
 				if (ret) {
 					free_extent_buffer(next);
 					return ret;

fs/btrfs/tree-log.c static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, · line 2607

Quoted 8 lines

fs/btrfs/tree-log.c +1 -1

@@ -2597,7 +2601,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 			free_extent_buffer(next);
 			continue;
 		}
-		ret = btrfs_read_buffer(next, ptr_gen);
+		ret = btrfs_read_buffer(next, ptr_gen, &first_key, *level - 1);
 		if (ret) {
 			free_extent_buffer(next);
 			return ret;

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 55 lines · Nikolay Borisov

Quoted 52 lines · Qu Wenruo

We have several reports about node pointer points to incorrect child tree blocks, which could have even wrong owner and level but still with valid generation and checksum.

Although btrfs check could handle it and print error message like:
leaf parent key incorrect 60670574592

Kernel doesn't have enough check on this type of corruption correctly. At least add such check to read_tree_block() and btrfs_read_buffer(), where we need two new parameters @first_key and @level to verify the child tree block.

There are call sites which can't use this new verfication by their nature:
1) Root node/leaf
It's referred by bytenr directly, so no @first_key to verify
2) Backref walker
FULL_BACKREF only has parent bytenr, no key to verify too.

Despite above cases, about 50% call sites can take advantage of this verification.

Another note of this verification is, it needs extra info from nodeptr, so it can't fit into current tree-checker framework, which is limited to node/leaf boundary.

Signed-off-by Qu Wenruo <[email protected]>

---
To David,

Since half of the new parameters are already filled, I'm not sure if it's a good idea to make a new function for this.

---
 fs/btrfs/backref.c     |  6 +++--
 fs/btrfs/ctree.c       | 25 +++++++++++++-----
 fs/btrfs/disk-io.c     | 69 ++++++++++++++++++++++++++++++++++++++++----------
 fs/btrfs/disk-io.h     |  8 +++---
 fs/btrfs/extent-tree.c |  6 ++++-
 fs/btrfs/print-tree.c  | 10 +++++---
 fs/btrfs/qgroup.c      |  7 +++--
 fs/btrfs/ref-verify.c  |  7 ++++-
 fs/btrfs/relocation.c  | 21 ++++++++++++---
 fs/btrfs/tree-log.c    | 12 ++++++---
 10 files changed, 132 insertions(+), 39 deletions(-)

fs/btrfs/backref.c +0 -0

index 26484648d090..3866b8ab20f1 100644
@@ -738,7 +738,8 @@ static int add_missing_keys(struct btrfs_fs_info *fs_info,
 		BUG_ON(ref->key_for_search.type);
 		BUG_ON(!ref->wanted_disk_byte);
 

Quoted 8 lines

fs/btrfs/backref.c static int add_missing_keys(struct btrfs_fs_info *fs_info, · line 742

Quoted 3 lines

fs/btrfs/backref.c +2 -1

-		eb = read_tree_block(fs_info, ref->wanted_disk_byte, 0);
+		eb = read_tree_block(fs_info, ref->wanted_disk_byte, 0, NULL,
+				     0);

I had a call with Jeff and he suggested that for this call side you can actually derive the parameters. I guess the level could really be ref->level - 1 (based on quick scan of add_delayed_refs, though you'd have to check it). The first key could be a bit tricky though. I'm not too familiar with this code.

For level it's true we could reach it without much problem. And in fact for root node/leaf, we can also get level info from ROOT_ITEM or superblock.
So for level we can enrich the check (and needs a new macro to skip level check)

For first key I think it's not possible as that's what we're resolving in this function, so @first_key check will still be optional IIRC.

Quoted 8 lines

Quoted 8 lines

fs/btrfs/backref.c +0 -0

 		if (IS_ERR(eb)) {
 			free_pref(ref);
 			return PTR_ERR(eb);

@@ -1291,7 +1292,8 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
 			    ref->level == 0) {
 				struct extent_buffer *eb;
 

Quoted 5 lines

fs/btrfs/backref.c static int find_parent_nodes(struct btrfs_trans_handle *trans, · line 1296

Quoted 3 lines

fs/btrfs/backref.c +2 -1

-				eb = read_tree_block(fs_info, ref->parent, 0);
+				eb = read_tree_block(fs_info, ref->parent, 0,
+						     NULL, 0);

He also suggested something like that could be done for this.

For level, sure.

But for first key I think it's pretty hard in this call site. We are looking for EXTENT_DATA here so in fact any item in fs/subvol trees can be possible.

Quoted 541 lines

Quoted 528 lines

fs/btrfs/backref.c +0 -0

 				if (IS_ERR(eb)) {
 					ret = PTR_ERR(eb);
 					goto out;

fs/btrfs/ctree.c +19 -6

index b88a79e69ddf..e49ca6d529e8 100644
@@ -1436,7 +1436,7 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 	if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 		btrfs_tree_read_unlock(eb_root);
 		free_extent_buffer(eb_root);
-		old = read_tree_block(fs_info, logical, 0);
+		old = read_tree_block(fs_info, logical, 0, NULL, 0);
 		if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
 			if (!IS_ERR(old))
 				free_extent_buffer(old);

@@ -1656,6 +1656,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	btrfs_set_lock_blocking(parent);
 
 	for (i = start_slot; i <= end_slot; i++) {
+		struct btrfs_key first_key;
 		int close = 1;
 
 		btrfs_node_key(parent, &disk_key, i);

@@ -1665,6 +1666,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		progress_passed = 1;
 		blocknr = btrfs_node_blockptr(parent, i);
 		gen = btrfs_node_ptr_generation(parent, i);
+		btrfs_node_key_to_cpu(parent, &first_key, i);
 		if (last_block == 0)
 			last_block = blocknr;
 

@@ -1688,7 +1690,9 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 			uptodate = 0;
 		if (!cur || !uptodate) {
 			if (!cur) {
-				cur = read_tree_block(fs_info, blocknr, gen);
+				cur = read_tree_block(fs_info, blocknr, gen,
+						      &first_key,
+						      parent_level - 1);
 				if (IS_ERR(cur)) {
 					return PTR_ERR(cur);
 				} else if (!extent_buffer_uptodate(cur)) {

@@ -1696,7 +1700,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 					return -EIO;
 				}
 			} else if (!uptodate) {
-				err = btrfs_read_buffer(cur, gen);
+				err = btrfs_read_buffer(cur, gen, &first_key,
+							parent_level - 1);
 				if (err) {
 					free_extent_buffer(cur);
 					return err;

@@ -1849,14 +1854,17 @@ read_node_slot(struct btrfs_fs_info *fs_info, struct extent_buffer *parent,
 {
 	int level = btrfs_header_level(parent);
 	struct extent_buffer *eb;
+	struct btrfs_key first_key;
 
 	if (slot < 0 || slot >= btrfs_header_nritems(parent))
 		return ERR_PTR(-ENOENT);
 
 	BUG_ON(level == 0);
 
+	btrfs_node_key_to_cpu(parent, &first_key, slot);
 	eb = read_tree_block(fs_info, btrfs_node_blockptr(parent, slot),
-			     btrfs_node_ptr_generation(parent, slot));
+			     btrfs_node_ptr_generation(parent, slot),
+			     &first_key, level - 1);
 	if (!IS_ERR(eb) && !extent_buffer_uptodate(eb)) {
 		free_extent_buffer(eb);
 		eb = ERR_PTR(-EIO);

@@ -2445,10 +2453,14 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	u64 gen;
 	struct extent_buffer *b = *eb_ret;
 	struct extent_buffer *tmp;
+	struct btrfs_key first_key;
 	int ret;
+	int parent_level;
 
 	blocknr = btrfs_node_blockptr(b, slot);
 	gen = btrfs_node_ptr_generation(b, slot);
+	parent_level = btrfs_header_level(b);
+	btrfs_node_key_to_cpu(b, &first_key, slot);
 
 	tmp = find_extent_buffer(fs_info, blocknr);
 	if (tmp) {

@@ -2467,7 +2479,7 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 		btrfs_set_path_blocking(p);
 
 		/* now we're allowed to do a blocking uptodate check */
-		ret = btrfs_read_buffer(tmp, gen);
+		ret = btrfs_read_buffer(tmp, gen, &first_key, parent_level - 1);
 		if (!ret) {
 			*eb_ret = tmp;
 			return 0;

@@ -2494,7 +2506,8 @@ read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
 	btrfs_release_path(p);
 
 	ret = -EAGAIN;
-	tmp = read_tree_block(fs_info, blocknr, 0);
+	tmp = read_tree_block(fs_info, blocknr, 0, &first_key,
+			      parent_level - 1);
 	if (!IS_ERR(tmp)) {
 		/*
 		 * If the read above didn't mark this buffer up to date,

fs/btrfs/disk-io.c +56 -13

index 21f34ad0d411..46fbbbf34568 100644
@@ -428,13 +428,40 @@ static int btrfs_check_super_csum(struct btrfs_fs_info *fs_info,
 	return ret;
 }
 
+static int verify_parent_key(struct extent_buffer *eb,
+			     struct btrfs_key *first_key,
+			     int level)
+{
+	int found_level;
+	struct btrfs_key found_key;
+
+	if (!first_key)
+		return 0;
+
+	found_level = btrfs_header_level(eb);
+	if (found_level != level)
+		return -EIO;
+
+	if (found_level)
+		btrfs_node_key_to_cpu(eb, &found_key, 0);
+	else
+		btrfs_item_key_to_cpu(eb, &found_key, 0);
+	return btrfs_comp_cpu_keys(first_key, &found_key);
+}
+
 /*
  * helper to read a given tree block, doing retries as required when
  * the checksums don't match and we have alternate mirrors to try.
+ *
+ * @parent_transid:	expected transid, skip check if 0
+ * @first_key:		expected key of first slot, skip check if NULL
+ * @level:		expected level, skip check if @first_key is NULL
  */
 static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 					  struct extent_buffer *eb,
-					  u64 parent_transid)
+					  u64 parent_transid,
+					  struct btrfs_key *first_key,
+					  int level)
 {
 	struct extent_io_tree *io_tree;
 	int failed = 0;

@@ -449,11 +476,13 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		ret = read_extent_buffer_pages(io_tree, eb, WAIT_COMPLETE,
 					       mirror_num);
 		if (!ret) {
-			if (!verify_parent_transid(io_tree, eb,
+			if (verify_parent_transid(io_tree, eb,
 						   parent_transid, 0))
-				break;
-			else
 				ret = -EIO;
+			else if (verify_parent_key(eb, first_key, level))
+				ret = -EUCLEAN;
+			else
+				break;
 		}
 
 		/*

@@ -461,7 +490,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_fs_info *fs_info,
 		 * there is no reason to read the other copies, they won't be
 		 * any less wrong.
 		 */
-		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
+		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags) ||
+		    ret == -EUCLEAN)
 			break;
 
 		num_copies = btrfs_num_copies(fs_info,

@@ -1062,8 +1092,17 @@ void btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
 			        buf->start, buf->start + buf->len - 1);
 }
 
+/*
+ * Read tree block at logical address @bytenr and do variant basic but critical
+ * verfication.
+ *
+ * @parent_transid:	expected transid of this tree block, skip check if 0
+ * @first_key:		expected key in slot 0, skip check if NULL
+ * @level:		eppected level, skip check if @first_key is NULL
+ */
 struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
-				      u64 parent_transid)
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level)
 {
 	struct extent_buffer *buf = NULL;
 	int ret;

@@ -1072,7 +1111,8 @@ struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
 	if (IS_ERR(buf))
 		return buf;
 
-	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	ret = btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					     first_key, level);
 	if (ret) {
 		free_extent_buffer(buf);
 		return ERR_PTR(ret);

@@ -1425,7 +1465,7 @@ static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
 	generation = btrfs_root_generation(&root->root_item);
 	root->node = read_tree_block(fs_info,
 				     btrfs_root_bytenr(&root->root_item),
-				     generation);
+				     generation, NULL, 0);
 	if (IS_ERR(root->node)) {
 		ret = PTR_ERR(root->node);
 		goto find_fail;

@@ -2289,7 +2329,8 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 	__setup_root(log_tree_root, fs_info, BTRFS_TREE_LOG_OBJECTID);
 
 	log_tree_root->node = read_tree_block(fs_info, bytenr,
-					      fs_info->generation + 1);
+					      fs_info->generation + 1,
+					      NULL, 0);
 	if (IS_ERR(log_tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read log tree");
 		ret = PTR_ERR(log_tree_root->node);

@@ -2746,7 +2787,7 @@ int open_ctree(struct super_block *sb,
 
 	chunk_root->node = read_tree_block(fs_info,
 					   btrfs_super_chunk_root(disk_super),
-					   generation);
+					   generation, NULL, 0);
 	if (IS_ERR(chunk_root->node) ||
 	    !extent_buffer_uptodate(chunk_root->node)) {
 		btrfs_err(fs_info, "failed to read chunk root");

@@ -2783,7 +2824,7 @@ int open_ctree(struct super_block *sb,
 
 	tree_root->node = read_tree_block(fs_info,
 					  btrfs_super_root(disk_super),
-					  generation);
+					  generation, NULL, 0);
 	if (IS_ERR(tree_root->node) ||
 	    !extent_buffer_uptodate(tree_root->node)) {
 		btrfs_warn(fs_info, "failed to read tree root");

@@ -3890,12 +3931,14 @@ void btrfs_btree_balance_dirty_nodelay(struct btrfs_fs_info *fs_info)
 	__btrfs_btree_balance_dirty(fs_info, 0);
 }
 
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level)
 {
 	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
-	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid);
+	return btree_read_extent_buffer_pages(fs_info, buf, parent_transid,
+					      first_key, level);
 }
 
 static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info)

fs/btrfs/disk-io.h +5 -3

index 301151a50ac1..3a1f4b2fa6ff 100644
@@ -52,8 +52,9 @@ static inline u64 btrfs_sb_offset(int mirror)
 struct btrfs_device;
 struct btrfs_fs_devices;
 
-struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info,
-				      u64 bytenr, u64 parent_transid);
+struct extent_buffer *read_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr,
+				      u64 parent_transid,
+				      struct btrfs_key *first_key, int level);
 void readahead_tree_block(struct btrfs_fs_info *fs_info, u64 bytenr);
 int reada_tree_block_flagged(struct btrfs_fs_info *fs_info, u64 bytenr,
 			 int mirror_num, struct extent_buffer **eb);

@@ -123,7 +124,8 @@ static inline void btrfs_put_fs_root(struct btrfs_root *root)
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
 int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
 			  int atomic);
-int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid,
+		      struct btrfs_key *first_key, int level);
 u32 btrfs_csum_data(const char *data, u32 seed, size_t len);
 void btrfs_csum_final(u32 crc, u8 *result);
 blk_status_t btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,

fs/btrfs/extent-tree.c +5 -1

index c1618ab9fecf..156ac8e4ff62 100644
@@ -8701,6 +8701,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	u64 parent;
 	u32 blocksize;
 	struct btrfs_key key;
+	struct btrfs_key first_key;
 	struct extent_buffer *next;
 	int level = wc->level;
 	int reada = 0;

@@ -8721,6 +8722,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	}
 
 	bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]);
+	btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+			      path->slots[level]);
 	blocksize = fs_info->nodesize;
 
 	next = find_extent_buffer(fs_info, bytenr);

@@ -8785,7 +8788,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	if (!next) {
 		if (reada && level == 1)
 			reada_walk_down(trans, root, wc, path);
-		next = read_tree_block(fs_info, bytenr, generation);
+		next = read_tree_block(fs_info, bytenr, generation, &first_key,
+				       level - 1);
 		if (IS_ERR(next)) {
 			return PTR_ERR(next);
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/print-tree.c +7 -3

index 569205e651c7..b00d88291085 100644
@@ -365,9 +365,13 @@ void btrfs_print_tree(struct extent_buffer *c)
 		       btrfs_node_blockptr(c, i));
 	}
 	for (i = 0; i < nr; i++) {
-		struct extent_buffer *next = read_tree_block(fs_info,
-					btrfs_node_blockptr(c, i),
-					btrfs_node_ptr_generation(c, i));
+		struct btrfs_key first_key;
+		struct extent_buffer *next;
+
+		btrfs_node_key_to_cpu(c, &first_key, i);
+		next = read_tree_block(fs_info, btrfs_node_blockptr(c, i),
+				       btrfs_node_ptr_generation(c, i),
+				       &first_key, level - 1);
 		if (IS_ERR(next)) {
 			continue;
 		} else if (!extent_buffer_uptodate(next)) {

fs/btrfs/qgroup.c +5 -2

index aa259d6986e1..86698702d08c 100644
@@ -1614,7 +1614,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 		return 0;
 
 	if (!extent_buffer_uptodate(root_eb)) {
-		ret = btrfs_read_buffer(root_eb, root_gen);
+		ret = btrfs_read_buffer(root_eb, root_gen, NULL, 0);
 		if (ret)
 			goto out;
 	}

@@ -1645,6 +1645,7 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 	level = root_level;
 	while (level >= 0) {
 		if (path->nodes[level] == NULL) {
+			struct btrfs_key first_key;
 			int parent_slot;
 			u64 child_gen;
 			u64 child_bytenr;

@@ -1657,8 +1658,10 @@ int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
 			parent_slot = path->slots[level + 1];
 			child_bytenr = btrfs_node_blockptr(eb, parent_slot);
 			child_gen = btrfs_node_ptr_generation(eb, parent_slot);
+			btrfs_node_key_to_cpu(eb, &first_key, parent_slot);
 
-			eb = read_tree_block(fs_info, child_bytenr, child_gen);
+			eb = read_tree_block(fs_info, child_bytenr, child_gen,
+					     &first_key, level);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				goto out;

fs/btrfs/ref-verify.c +6 -1

index 171f3cce30e6..25c767635426 100644
@@ -579,11 +579,16 @@ static int walk_down_tree(struct btrfs_root *root, struct btrfs_path *path,
 
 	while (level >= 0) {
 		if (level) {
+			struct btrfs_key first_key;
+
 			block_bytenr = btrfs_node_blockptr(path->nodes[level],
 							   path->slots[level]);
 			gen = btrfs_node_ptr_generation(path->nodes[level],
 							path->slots[level]);
-			eb = read_tree_block(fs_info, block_bytenr, gen);
+			btrfs_node_key_to_cpu(path->nodes[level], &first_key,
+					      path->slots[level]);
+			eb = read_tree_block(fs_info, block_bytenr, gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb))
 				return PTR_ERR(eb);
 			if (!extent_buffer_uptodate(eb)) {

fs/btrfs/relocation.c +17 -4

index cd2298d185dd..4a309f8c9535 100644
@@ -1839,6 +1839,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 
 	parent = eb;
 	while (1) {
+		struct btrfs_key first_key;
+
 		level = btrfs_header_level(parent);
 		BUG_ON(level < lowest_level);
 

@@ -1852,6 +1854,7 @@ int replace_path(struct btrfs_trans_handle *trans,
 		old_bytenr = btrfs_node_blockptr(parent, slot);
 		blocksize = fs_info->nodesize;
 		old_ptr_gen = btrfs_node_ptr_generation(parent, slot);
+		btrfs_node_key_to_cpu(parent, &key, slot);
 
 		if (level <= max_level) {
 			eb = path->nodes[level];

@@ -1876,7 +1879,8 @@ int replace_path(struct btrfs_trans_handle *trans,
 				break;
 			}
 
-			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen);
+			eb = read_tree_block(fs_info, old_bytenr, old_ptr_gen,
+					     &first_key, level - 1);
 			if (IS_ERR(eb)) {
 				ret = PTR_ERR(eb);
 				break;

@@ -2036,6 +2040,8 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 	last_snapshot = btrfs_root_last_snapshot(&root->root_item);
 
 	for (i = *level; i > 0; i--) {
+		struct btrfs_key first_key;
+
 		eb = path->nodes[i];
 		nritems = btrfs_header_nritems(eb);
 		while (path->slots[i] < nritems) {

@@ -2056,7 +2062,9 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 		}
 
 		bytenr = btrfs_node_blockptr(eb, path->slots[i]);
-		eb = read_tree_block(fs_info, bytenr, ptr_gen);
+		btrfs_node_key_to_cpu(eb, &first_key, path->slots[i]);
+		eb = read_tree_block(fs_info, bytenr, ptr_gen, &first_key,
+				     i - 1);
 		if (IS_ERR(eb)) {
 			return PTR_ERR(eb);
 		} else if (!extent_buffer_uptodate(eb)) {

@@ -2714,6 +2722,8 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 	path->lowest_level = node->level + 1;
 	rc->backref_cache.path[node->level] = node;
 	list_for_each_entry(edge, &node->upper, list[LOWER]) {
+		struct btrfs_key first_key;
+
 		cond_resched();
 
 		upper = edge->node[UPPER];

@@ -2779,7 +2789,9 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 
 		blocksize = root->fs_info->nodesize;
 		generation = btrfs_node_ptr_generation(upper->eb, slot);
-		eb = read_tree_block(fs_info, bytenr, generation);
+		btrfs_node_key_to_cpu(upper->eb, &first_key, slot);
+		eb = read_tree_block(fs_info, bytenr, generation, &first_key,
+				     upper->level - 1);
 		if (IS_ERR(eb)) {
 			err = PTR_ERR(eb);
 			goto next;

@@ -2944,7 +2956,8 @@ static int get_tree_block_key(struct btrfs_fs_info *fs_info,
 	struct extent_buffer *eb;
 
 	BUG_ON(block->key_ready);
-	eb = read_tree_block(fs_info, block->bytenr, block->key.offset);
+	eb = read_tree_block(fs_info, block->bytenr, block->key.offset, NULL,
+			     0);
 	if (IS_ERR(eb)) {
 		return PTR_ERR(eb);
 	} else if (!extent_buffer_uptodate(eb)) {

fs/btrfs/tree-log.c +7 -3

index 434457794c27..b98a1801b406 100644
@@ -304,7 +304,7 @@ static int process_one_buffer(struct btrfs_root *log,
 	 * pin down any logged extents, so we have to read the block.
 	 */
 	if (btrfs_fs_incompat(fs_info, MIXED_GROUPS)) {
-		ret = btrfs_read_buffer(eb, gen);
+		ret = btrfs_read_buffer(eb, gen, NULL, 0);
 		if (ret)
 			return ret;
 	}

@@ -2420,7 +2420,7 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
 	int i;
 	int ret;
 
-	ret = btrfs_read_buffer(eb, gen);
+	ret = btrfs_read_buffer(eb, gen, NULL, 0);
 	if (ret)
 		return ret;
 

@@ -2537,6 +2537,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 	WARN_ON(*level >= BTRFS_MAX_LEVEL);
 
 	while (*level > 0) {
+		struct btrfs_key first_key;
+
 		WARN_ON(*level < 0);
 		WARN_ON(*level >= BTRFS_MAX_LEVEL);
 		cur = path->nodes[*level];

@@ -2549,6 +2551,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
 		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+		btrfs_node_key_to_cpu(cur, &first_key, path->slots[*level]);
 		blocksize = fs_info->nodesize;
 
 		parent = path->nodes[*level];

@@ -2567,7 +2570,8 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 			path->slots[*level]++;
 			if (wc->free) {
-				ret = btrfs_read_buffer(next, ptr_gen);
+				ret = btrfs_read_buffer(next, ptr_gen,
+							&first_key, *level - 1);
 				if (ret) {
 					free_extent_buffer(next);
 					return ret;

fs/btrfs/tree-log.c static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, · line 2607

Quoted 8 lines

fs/btrfs/tree-log.c +1 -1

@@ -2597,7 +2601,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 			free_extent_buffer(next);
 			continue;
 		}
-		ret = btrfs_read_buffer(next, ptr_gen);
+		ret = btrfs_read_buffer(next, ptr_gen, &first_key, *level - 1);
 		if (ret) {
 			free_extent_buffer(next);
 			return ret;

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html