4.14-stable review patch. If anyone has any objections, please let me know.
------------------
commit c3267bbaa9cae09b62960eafe33ad19196803285 upstream.
Current check_leaf() function does a good job checking key order and item offset/size.
However it only checks from slot 0 to the last but one slot, this is good but makes later expansion hard.
So this refactoring iterates from slot 0 to the last slot. For key comparison, it uses a key with all 0 as initial key, so all valid keys should be larger than that.
And for item size/offset checks, it compares current item end with previous item offset. For slot 0, use leaf end as a special case.
This makes later item/key offset checks and item size checks easier to be implemented.
Also, makes check_leaf() to return -EUCLEAN other than -EIO to indicate error.
Signed-off-by: Qu Wenruo quwenruo.btrfs@gmx.com Reviewed-by: Nikolay Borisov nborisov@suse.com Reviewed-by: David Sterba dsterba@suse.com Signed-off-by: David Sterba dsterba@suse.com Signed-off-by: Ben Hutchings ben.hutchings@codethink.co.uk Signed-off-by: Sasha Levin sashal@kernel.org --- fs/btrfs/disk-io.c | 50 +++++++++++++++++++++++++--------------------- 1 file changed, 27 insertions(+), 23 deletions(-)
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 0e67cee73c53..4a1e63df1183 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -554,8 +554,9 @@ static noinline int check_leaf(struct btrfs_root *root, struct extent_buffer *leaf) { struct btrfs_fs_info *fs_info = root->fs_info; + /* No valid key type is 0, so all key should be larger than this key */ + struct btrfs_key prev_key = {0, 0, 0}; struct btrfs_key key; - struct btrfs_key leaf_key; u32 nritems = btrfs_header_nritems(leaf); int slot;
@@ -588,7 +589,7 @@ static noinline int check_leaf(struct btrfs_root *root, CORRUPT("non-root leaf's nritems is 0", leaf, check_root, 0); free_extent_buffer(eb); - return -EIO; + return -EUCLEAN; } free_extent_buffer(eb); } @@ -598,28 +599,23 @@ static noinline int check_leaf(struct btrfs_root *root, if (nritems == 0) return 0;
- /* Check the 0 item */ - if (btrfs_item_offset_nr(leaf, 0) + btrfs_item_size_nr(leaf, 0) != - BTRFS_LEAF_DATA_SIZE(fs_info)) { - CORRUPT("invalid item offset size pair", leaf, root, 0); - return -EIO; - } - /* - * Check to make sure each items keys are in the correct order and their - * offsets make sense. We only have to loop through nritems-1 because - * we check the current slot against the next slot, which verifies the - * next slot's offset+size makes sense and that the current's slot - * offset is correct. + * Check the following things to make sure this is a good leaf, and + * leaf users won't need to bother with similar sanity checks: + * + * 1) key order + * 2) item offset and size + * No overlap, no hole, all inside the leaf. */ - for (slot = 0; slot < nritems - 1; slot++) { - btrfs_item_key_to_cpu(leaf, &leaf_key, slot); - btrfs_item_key_to_cpu(leaf, &key, slot + 1); + for (slot = 0; slot < nritems; slot++) { + u32 item_end_expected; + + btrfs_item_key_to_cpu(leaf, &key, slot);
/* Make sure the keys are in the right order */ - if (btrfs_comp_cpu_keys(&leaf_key, &key) >= 0) { + if (btrfs_comp_cpu_keys(&prev_key, &key) >= 0) { CORRUPT("bad key order", leaf, root, slot); - return -EIO; + return -EUCLEAN; }
/* @@ -627,10 +623,14 @@ static noinline int check_leaf(struct btrfs_root *root, * item data starts at the end of the leaf and grows towards the * front. */ - if (btrfs_item_offset_nr(leaf, slot) != - btrfs_item_end_nr(leaf, slot + 1)) { + if (slot == 0) + item_end_expected = BTRFS_LEAF_DATA_SIZE(fs_info); + else + item_end_expected = btrfs_item_offset_nr(leaf, + slot - 1); + if (btrfs_item_end_nr(leaf, slot) != item_end_expected) { CORRUPT("slot offset bad", leaf, root, slot); - return -EIO; + return -EUCLEAN; }
/* @@ -641,8 +641,12 @@ static noinline int check_leaf(struct btrfs_root *root, if (btrfs_item_end_nr(leaf, slot) > BTRFS_LEAF_DATA_SIZE(fs_info)) { CORRUPT("slot end outside of leaf", leaf, root, slot); - return -EIO; + return -EUCLEAN; } + + prev_key.objectid = key.objectid; + prev_key.type = key.type; + prev_key.offset = key.offset; }
return 0;