xfs: add version 3 inode format with CRCs

Add a new inode version with a larger core.  The primary objective is
to allow for a crc of the inode, and location information (uuid and ino)
to verify it was written in the right place.  We also extend it by:

	a creation time (for Samba);
	a changecount (for NFSv4);
	a flush sequence (in LSN format for recovery);
	an additional inode flags field; and
	some additional padding.

These additional fields are not implemented yet, but already laid
out in the structure.

[dchinner@redhat.com] Added LSN and flags field, some factoring and rework to
capture all the necessary information in the crc calculation.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Ben Myers <bpm@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

fs/xfs/xfs_buf_item.h

@@ -48,6 +48,7 @@ extern kmem_zone_t	*xfs_buf_item_zone;
 #define XFS_BLF_AGF_BUF		(1<<6)
 #define XFS_BLF_AGFL_BUF	(1<<7)
 #define XFS_BLF_AGI_BUF		(1<<8)
+#define XFS_BLF_DINO_BUF	(1<<9)
 
 #define XFS_BLF_TYPE_MASK	\
 		(XFS_BLF_UDQUOT_BUF | \
@@ -56,7 +57,8 @@ extern kmem_zone_t	*xfs_buf_item_zone;
 		 XFS_BLF_BTREE_BUF | \
 		 XFS_BLF_AGF_BUF | \
 		 XFS_BLF_AGFL_BUF | \
-		 XFS_BLF_AGI_BUF)
+		 XFS_BLF_AGI_BUF | \
+		 XFS_BLF_DINO_BUF)
 
 #define	XFS_BLF_CHUNK		128
 #define	XFS_BLF_SHIFT		7

fs/xfs/xfs_dinode.h

@@ -19,7 +19,7 @@
 #define	__XFS_DINODE_H__
 
 #define	XFS_DINODE_MAGIC		0x494e	/* 'IN' */
-#define XFS_DINODE_GOOD_VERSION(v)	(((v) == 1 || (v) == 2))
+#define XFS_DINODE_GOOD_VERSION(v)	((v) >= 1 && (v) <= 3)
 
 typedef struct xfs_timestamp {
 	__be32		t_sec;		/* timestamp seconds */
@@ -70,10 +70,35 @@ typedef struct xfs_dinode {
 
 	/* di_next_unlinked is the only non-core field in the old dinode */
 	__be32		di_next_unlinked;/* agi unlinked list ptr */
-} __attribute__((packed)) xfs_dinode_t;
+
+	/* start of the extended dinode, writable fields */
+	__le32		di_crc;		/* CRC of the inode */
+	__be64		di_changecount;	/* number of attribute changes */
+	__be64		di_lsn;		/* flush sequence */
+	__be64		di_flags2;	/* more random flags */
+	__u8		di_pad2[16];	/* more padding for future expansion */
+
+	/* fields only written to during inode creation */
+	xfs_timestamp_t	di_crtime;	/* time created */
+	__be64		di_ino;		/* inode number */
+	uuid_t		di_uuid;	/* UUID of the filesystem */
+
+	/* structure must be padded to 64 bit alignment */
+} xfs_dinode_t;
 
 #define DI_MAX_FLUSH 0xffff
 
+/*
+ * Size of the core inode on disk.  Version 1 and 2 inodes have
+ * the same size, but version 3 has grown a few additional fields.
+ */
+static inline uint xfs_dinode_size(int version)
+{
+	if (version == 3)
+		return sizeof(struct xfs_dinode);
+	return offsetof(struct xfs_dinode, di_crc);
+}
+
 /*
  * The 32 bit link count in the inode theoretically maxes out at UINT_MAX.
  * Since the pathconf interface is signed, we use 2^31 - 1 instead.
@@ -105,7 +130,7 @@ typedef enum xfs_dinode_fmt {
  * Inode size for given fs.
  */
 #define XFS_LITINO(mp, version) \
-	((int)(((mp)->m_sb.sb_inodesize) - sizeof(struct xfs_dinode)))
+	((int)(((mp)->m_sb.sb_inodesize) - xfs_dinode_size(version)))
 
 #define XFS_BROOT_SIZE_ADJ(ip) \
 	(XFS_BMBT_BLOCK_LEN((ip)->i_mount) - sizeof(xfs_bmdr_block_t))
@@ -133,7 +158,7 @@ typedef enum xfs_dinode_fmt {
  * Return pointers to the data or attribute forks.
  */
 #define XFS_DFORK_DPTR(dip) \
-	((char *)(dip) + sizeof(struct xfs_dinode))
+	((char *)dip + xfs_dinode_size(dip->di_version))
 #define XFS_DFORK_APTR(dip)	\
 	(XFS_DFORK_DPTR(dip) + XFS_DFORK_BOFF(dip))
 #define XFS_DFORK_PTR(dip,w)	\

fs/xfs/xfs_ialloc.c

@@ -167,6 +167,7 @@ xfs_ialloc_inode_init(
 	int			version;
 	int			i, j;
 	xfs_daddr_t		d;
+	xfs_ino_t		ino = 0;
 
 	/*
 	 * Loop over the new block(s), filling in the inodes.
@@ -185,13 +186,29 @@ xfs_ialloc_inode_init(
 	}
 
 	/*
-	 * Figure out what version number to use in the inodes we create.
-	 * If the superblock version has caught up to the one that supports
-	 * the new inode format, then use the new inode version.  Otherwise
-	 * use the old version so that old kernels will continue to be
-	 * able to use the file system.
+	 * Figure out what version number to use in the inodes we create.  If
+	 * the superblock version has caught up to the one that supports the new
+	 * inode format, then use the new inode version.  Otherwise use the old
+	 * version so that old kernels will continue to be able to use the file
+	 * system.
+	 *
+	 * For v3 inodes, we also need to write the inode number into the inode,
+	 * so calculate the first inode number of the chunk here as
+	 * XFS_OFFBNO_TO_AGINO() only works within a filesystem block, not
+	 * across multiple filesystem blocks (such as a cluster) and so cannot
+	 * be used in the cluster buffer loop below.
+	 *
+	 * Further, because we are writing the inode directly into the buffer
+	 * and calculating a CRC on the entire inode, we have ot log the entire
+	 * inode so that the entire range the CRC covers is present in the log.
+	 * That means for v3 inode we log the entire buffer rather than just the
+	 * inode cores.
 	 */
-	if (xfs_sb_version_hasnlink(&mp->m_sb))
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		version = 3;
+		ino = XFS_AGINO_TO_INO(mp, agno,
+				       XFS_OFFBNO_TO_AGINO(mp, agbno, 0));
+	} else if (xfs_sb_version_hasnlink(&mp->m_sb))
 		version = 2;
 	else
 		version = 1;
@@ -214,17 +231,32 @@ xfs_ialloc_inode_init(
 		 *	individual transactions causing a lot of log traffic.
 		 */
 		fbuf->b_ops = &xfs_inode_buf_ops;
-		xfs_buf_zero(fbuf, 0, ninodes << mp->m_sb.sb_inodelog);
+		xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
 		for (i = 0; i < ninodes; i++) {
 			int	ioffset = i << mp->m_sb.sb_inodelog;
-			uint	isize = sizeof(struct xfs_dinode);
+			uint	isize = xfs_dinode_size(version);
 
 			free = xfs_make_iptr(mp, fbuf, i);
 			free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
 			free->di_version = version;
 			free->di_gen = cpu_to_be32(gen);
 			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
-			xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
+
+			if (version == 3) {
+				free->di_ino = cpu_to_be64(ino);
+				ino++;
+				uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid);
+				xfs_dinode_calc_crc(mp, free);
+			} else {
+				/* just log the inode core */
+				xfs_trans_log_buf(tp, fbuf, ioffset,
+						  ioffset + isize - 1);
+			}
+		}
+		if (version == 3) {
+			/* need to log the entire buffer */
+			xfs_trans_log_buf(tp, fbuf, 0,
+					  BBTOB(fbuf->b_length) - 1);
 		}
 		xfs_trans_inode_alloc_buf(tp, fbuf);
 	}