btrfs: fix memory leak of path on error return path

Currently if the allocation of roots or tmp_ulist fails the error handling
does not free up the allocation of path causing a memory leak. Fix this and
other similar leaks by moving the call of btrfs_free_path from label out
to label out_free_ulist.

Kudos to David Sterba for spotting the issue in my original fix and suggesting
the correct way to fix the leak and Anand Jain for spotting a double free
issue.

Addresses-Coverity: ("Resource leak")
Fixes: 5911c8fe05c5 ("btrfs: fiemap: preallocate ulists for btrfs_check_shared")
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 6b154bce56872f6215c0cf1b82d77b7470fb6d17..aea990473392d81474b19516c97ffc2ad2fcef16 100644 (file)
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -4611,7 +4611,6 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
        ret = btrfs_lookup_file_extent(NULL, root, path,
                        btrfs_ino(BTRFS_I(inode)), -1, 0);
        if (ret < 0) {
-               btrfs_free_path(path);
                goto out_free_ulist;
        } else {
                WARN_ON(!ret);
@@ -4764,11 +4763,11 @@ out_free:
                ret = emit_last_fiemap_cache(fieinfo, &cache);
        free_extent_map(em);
 out:
-       btrfs_free_path(path);
        unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len - 1,
                             &cached_state);
 
 out_free_ulist:
+       btrfs_free_path(path);
        ulist_free(roots);
        ulist_free(tmp_ulist);
        return ret;