1 files changed, 53 insertions, 27 deletions

diff --git a/fs/xfs/xfs_inode_buf.c b/fs/xfs/xfs_inode_buf.c
index 38fe509827d..cb35ae41d4a 100644
--- a/fs/xfs/xfs_inode_buf.c
+++ b/fs/xfs/xfs_inode_buf.c
@@ -17,20 +17,20 @@
  */
 #include "xfs.h"
 #include "xfs_fs.h"
+#include "xfs_shared.h"
 #include "xfs_format.h"
-#include "xfs_log.h"
-#include "xfs_trans.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
 #include "xfs_mount.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_error.h"
 #include "xfs_cksum.h"
 #include "xfs_icache.h"
+#include "xfs_trans.h"
 #include "xfs_ialloc.h"
+#include "xfs_dinode.h"
 
 /*
  * Check that none of the inode's in the buffer have a next
@@ -53,17 +53,29 @@ xfs_inobp_check(
 					i * mp->m_sb.sb_inodesize);
 		if (!dip->di_next_unlinked)  {
 			xfs_alert(mp,
-	"Detected bogus zero next_unlinked field in incore inode buffer 0x%p.",
-				bp);
-			ASSERT(dip->di_next_unlinked);
+	"Detected bogus zero next_unlinked field in inode %d buffer 0x%llx.",
+				i, (long long)bp->b_bn);
 		}
 	}
 }
 #endif
 
+/*
+ * If we are doing readahead on an inode buffer, we might be in log recovery
+ * reading an inode allocation buffer that hasn't yet been replayed, and hence
+ * has not had the inode cores stamped into it. Hence for readahead, the buffer
+ * may be potentially invalid.
+ *
+ * If the readahead buffer is invalid, we don't want to mark it with an error,
+ * but we do want to clear the DONE status of the buffer so that a followup read
+ * will re-read it from disk. This will ensure that we don't get an unnecessary
+ * warnings during log recovery and we don't get unnecssary panics on debug
+ * kernels.
+ */
 static void
 xfs_inode_buf_verify(
-	struct xfs_buf	*bp)
+	struct xfs_buf	*bp,
+	bool		readahead)
 {
 	struct xfs_mount *mp = bp->b_target->bt_mount;
 	int		i;
@@ -84,15 +96,18 @@ xfs_inode_buf_verify(
 		if (unlikely(XFS_TEST_ERROR(!di_ok, mp,
 						XFS_ERRTAG_ITOBP_INOTOBP,
 						XFS_RANDOM_ITOBP_INOTOBP))) {
+			if (readahead) {
+				bp->b_flags &= ~XBF_DONE;
+				return;
+			}
+
 			xfs_buf_ioerror(bp, EFSCORRUPTED);
-			XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_HIGH,
-					     mp, dip);
+			xfs_verifier_error(bp);
 #ifdef DEBUG
-			xfs_emerg(mp,
+			xfs_alert(mp,
 				"bad inode magic/vsn daddr %lld #%d (magic=%x)",
 				(unsigned long long)bp->b_bn, i,
 				be16_to_cpu(dip->di_magic));
-			ASSERT(0);
 #endif
 		}
 	}
@@ -104,14 +119,21 @@ static void
 xfs_inode_buf_read_verify(
 	struct xfs_buf	*bp)
 {
-	xfs_inode_buf_verify(bp);
+	xfs_inode_buf_verify(bp, false);
+}
+
+static void
+xfs_inode_buf_readahead_verify(
+	struct xfs_buf	*bp)
+{
+	xfs_inode_buf_verify(bp, true);
 }
 
 static void
 xfs_inode_buf_write_verify(
 	struct xfs_buf	*bp)
 {
-	xfs_inode_buf_verify(bp);
+	xfs_inode_buf_verify(bp, false);
 }
 
 const struct xfs_buf_ops xfs_inode_buf_ops = {
@@ -119,6 +141,11 @@ const struct xfs_buf_ops xfs_inode_buf_ops = {
 	.verify_write = xfs_inode_buf_write_verify,
 };
 
+const struct xfs_buf_ops xfs_inode_buf_ra_ops = {
+	.verify_read = xfs_inode_buf_readahead_verify,
+	.verify_write = xfs_inode_buf_write_verify,
+};
+
 
 /*
  * This routine is called to map an inode to the buffer containing the on-disk
@@ -166,7 +193,7 @@ xfs_imap_to_bp(
 	return 0;
 }
 
-STATIC void
+void
 xfs_dinode_from_disk(
 	xfs_icdinode_t		*to,
 	xfs_dinode_t		*from)
@@ -278,7 +305,7 @@ xfs_dinode_verify(
 	if (!xfs_sb_version_hascrc(&mp->m_sb))
 		return false;
 	if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize,
-			      offsetof(struct xfs_dinode, di_crc)))
+			      XFS_DINODE_CRC_OFF))
 		return false;
 	if (be64_to_cpu(dip->di_ino) != ip->i_ino)
 		return false;
@@ -299,7 +326,7 @@ xfs_dinode_calc_crc(
 
 	ASSERT(xfs_sb_version_hascrc(&mp->m_sb));
 	crc = xfs_start_cksum((char *)dip, mp->m_sb.sb_inodesize,
-			      offsetof(struct xfs_dinode, di_crc));
+			      XFS_DINODE_CRC_OFF);
 	dip->di_crc = xfs_end_cksum(crc);
 }
 
@@ -410,17 +437,16 @@ xfs_iread(
 	}
 
 	/*
-	 * The inode format changed when we moved the link count and
-	 * made it 32 bits long.  If this is an old format inode,
-	 * convert it in memory to look like a new one.  If it gets
-	 * flushed to disk we will convert back before flushing or
-	 * logging it.  We zero out the new projid field and the old link
-	 * count field.  We'll handle clearing the pad field (the remains
-	 * of the old uuid field) when we actually convert the inode to
-	 * the new format. We don't change the version number so that we
-	 * can distinguish this from a real new format inode.
+	 * Automatically convert version 1 inode formats in memory to version 2
+	 * inode format. If the inode is modified, it will get logged and
+	 * rewritten as a version 2 inode. We can do this because we set the
+	 * superblock feature bit for v2 inodes unconditionally during mount
+	 * and it means the reast of the code can assume the inode version is 2
+	 * or higher.
 	 */
 	if (ip->i_d.di_version == 1) {
+		ip->i_d.di_version = 2;
+		memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
 		ip->i_d.di_nlink = ip->i_d.di_onlink;
 		ip->i_d.di_onlink = 0;
 		xfs_set_projid(ip, 0);