1 files changed, 60 insertions, 25 deletions

diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c
index 3e39b7d40f25..de9096b8a47c 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.c
+++ b/fs/xfs/libxfs/xfs_attr_remote.c
@@ -25,6 +25,23 @@
 #define ATTR_RMTVALUE_MAPSIZE	1	/* # of map entries at once */
 
 /*
+ * Remote Attribute Values
+ * =======================
+ *
+ * Remote extended attribute values are conceptually simple -- they're written
+ * to data blocks mapped by an inode's attribute fork, and they have an upper
+ * size limit of 64k.  Setting a value does not involve the XFS log.
+ *
+ * However, on a v5 filesystem, maximally sized remote attr values require one
+ * block more than 64k worth of space to hold both the remote attribute value
+ * header (64 bytes).  On a 4k block filesystem this results in a 68k buffer;
+ * on a 64k block filesystem, this would be a 128k buffer.  Note that the log
+ * format can only handle a dirty buffer of XFS_MAX_BLOCKSIZE length (64k).
+ * Therefore, we /must/ ensure that remote attribute value buffers never touch
+ * the logging system and therefore never have a log item.
+ */
+
+/*
  * Each contiguous block has a header, so it is not just a simple attribute
  * length to FSB conversion.
  */
@@ -400,17 +417,25 @@ xfs_attr_rmtval_get(
 			       (map[i].br_startblock != HOLESTARTBLOCK));
 			dblkno = XFS_FSB_TO_DADDR(mp, map[i].br_startblock);
 			dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
-			error = xfs_trans_read_buf(mp, args->trans,
-						   mp->m_ddev_targp,
-						   dblkno, dblkcnt, 0, &bp,
-						   &xfs_attr3_rmt_buf_ops);
-			if (error)
+			bp = xfs_buf_read(mp->m_ddev_targp, dblkno, dblkcnt, 0,
+					&xfs_attr3_rmt_buf_ops);
+			if (!bp)
+				return -ENOMEM;
+			error = bp->b_error;
+			if (error) {
+				xfs_buf_ioerror_alert(bp, __func__);
+				xfs_buf_relse(bp);
+
+				/* bad CRC means corrupted metadata */
+				if (error == -EFSBADCRC)
+					error = -EFSCORRUPTED;
 				return error;
+			}
 
 			error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino,
 							&offset, &valuelen,
 							&dst);
-			xfs_trans_brelse(args->trans, bp);
+			xfs_buf_relse(bp);
 			if (error)
 				return error;
 
@@ -551,6 +576,32 @@ xfs_attr_rmtval_set(
 	return 0;
 }
 
+/* Mark stale any incore buffers for the remote value. */
+int
+xfs_attr_rmtval_stale(
+	struct xfs_inode	*ip,
+	struct xfs_bmbt_irec	*map,
+	xfs_buf_flags_t		incore_flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_buf		*bp;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	ASSERT((map->br_startblock != DELAYSTARTBLOCK) &&
+	       (map->br_startblock != HOLESTARTBLOCK));
+
+	bp = xfs_buf_incore(mp->m_ddev_targp,
+			XFS_FSB_TO_DADDR(mp, map->br_startblock),
+			XFS_FSB_TO_BB(mp, map->br_blockcount), incore_flags);
+	if (bp) {
+		xfs_buf_stale(bp);
+		xfs_buf_relse(bp);
+	}
+
+	return 0;
+}
+
 /*
  * Remove the value associated with an attribute by deleting the
  * out-of-line buffer that it is stored on.
@@ -559,7 +610,6 @@ int
 xfs_attr_rmtval_remove(
 	struct xfs_da_args	*args)
 {
-	struct xfs_mount	*mp = args->dp->i_mount;
 	xfs_dablk_t		lblkno;
 	int			blkcnt;
 	int			error;
@@ -574,9 +624,6 @@ xfs_attr_rmtval_remove(
 	blkcnt = args->rmtblkcnt;
 	while (blkcnt > 0) {
 		struct xfs_bmbt_irec	map;
-		struct xfs_buf		*bp;
-		xfs_daddr_t		dblkno;
-		int			dblkcnt;
 		int			nmap;
 
 		/*
@@ -588,21 +635,9 @@ xfs_attr_rmtval_remove(
 		if (error)
 			return error;
 		ASSERT(nmap == 1);
-		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-		       (map.br_startblock != HOLESTARTBLOCK));
-
-		dblkno = XFS_FSB_TO_DADDR(mp, map.br_startblock),
-		dblkcnt = XFS_FSB_TO_BB(mp, map.br_blockcount);
-
-		/*
-		 * If the "remote" value is in the cache, remove it.
-		 */
-		bp = xfs_buf_incore(mp->m_ddev_targp, dblkno, dblkcnt, XBF_TRYLOCK);
-		if (bp) {
-			xfs_buf_stale(bp);
-			xfs_buf_relse(bp);
-			bp = NULL;
-		}
+		error = xfs_attr_rmtval_stale(args->dp, &map, XBF_TRYLOCK);
+		if (error)
+			return error;
 
 		lblkno += map.br_blockcount;
 		blkcnt -= map.br_blockcount;