[PATCH] Btrfs-progs: run_next_block: Return error code when extent buffer read fails

[Praveen K Pandey <praveen@linux.vnet.ibm.com>]

run_next_block() ignores the return value of read_tree_block() and
always returns success status. This might cause deal_root_from_list()
to loop infinitely when reading metadata block fails. This bug fixes
the issue by extracting and returning the error code from the return
value of read_tree_block().

Bug ID : https://bugzilla.kernel.org/show_bug.cgi?id=155551

Signed-off-by: Praveen K Pandey <praveen@linux.vnet.ibm.com>
---
 cmds-check.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/cmds-check.c b/cmds-check.c
index 17b7efb..40e3c99 100644
--- a/cmds-check.c
+++ b/cmds-check.c
@@ -7625,6 +7625,7 @@ static int run_next_block(struct btrfs_root *root,
 	if (!extent_buffer_uptodate(buf)) {
 		record_bad_block_io(root->fs_info,
 				    extent_cache, bytenr, size);
+		ret = PTR_ERR(buf);
 		goto out;
 	}
 
-- 
2.5.5