From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 13 Dec 2010 19:47:58 +0000 (-0500)
Subject: Btrfs: EIO when we fail to read tree roots
X-Git-Tag: v2.6.37-rc6~30^2~2
X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=68433b73b104bff388aac376631d32abbbd872b0;p=~emulex%2Finfiniband.git

Btrfs: EIO when we fail to read tree roots

If we just get a plain IO error when we read tree roots, the code
wasn't properly sending that error up the chain.  This allowed mounts to
continue when they should failed, and allowed operations
on partially setup root structs.  The end result was usually oopsen
on spinlocks that hadn't been spun up correctly.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index b803c266767..a5d2249e6da 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1007,7 +1007,10 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
 	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     blocksize, generation);
-	BUG_ON(!root->node);
+	if (!root->node || !btrfs_buffer_uptodate(root->node, generation)) {
+		free_extent_buffer(root->node);
+		return -EIO;
+	}
 	root->commit_root = btrfs_root_node(root);
 	return 0;
 }