Signed-off-by: Chris Mason <chris.mason@oracle.com>
first_index = pos >> PAGE_CACHE_SHIFT;
last_index = (pos + count) >> PAGE_CACHE_SHIFT;
+ /*
+ * if this is a nodatasum mount, force summing off for the inode
+ * all the time. That way a later mount with summing on won't
+ * get confused
+ */
+ if (btrfs_test_opt(root, NODATASUM))
+ btrfs_set_flag(inode, NODATASUM);
+
/*
* there are lots of better ways to do this, but this code
* makes sure the first and last page in the file range are