* recovery and we don't get unnecssary panics on debug kernels. We use EIO here
* because all we want to do is say readahead failed; there is no-one to report
* the error to, so this will distinguish it from a non-ra verifier failure.
+ * Changes to this readahead error behavour also need to be reflected in
+ * xfs_dquot_buf_readahead_verify().
*/
static void
xfs_inode_buf_verify(