xfs: add CRC checks to block format directory blocks

Now that directory buffers are made from a single struct xfs_buf, we
can add CRC calculation and checking callbacks. While there, add all
the fields to the on disk structures for future functionality such
as d_type support, uuids, block numbers, owner inode, etc.

To distinguish between the different on disk formats, change the
magic numbers for the new format directory blocks.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Ben Myers <bpm@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

7 files changed, 344 insertions, 105 deletions

diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c
index 12afe07a91d7..79a3f1b2bc7d 100644
--- a/fs/xfs/xfs_dir2_block.c
+++ b/fs/xfs/xfs_dir2_block.c
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -28,11 +29,13 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_inode_item.h"
+#include "xfs_buf_item.h"
 #include "xfs_dir2.h"
 #include "xfs_dir2_format.h"
 #include "xfs_dir2_priv.h"
 #include "xfs_error.h"
 #include "xfs_trace.h"
+#include "xfs_cksum.h"
 
 /*
  * Local function prototypes.
@@ -56,44 +59,74 @@ xfs_dir_startup(void)
         xfs_dir_hash_dotdot = xfs_da_hashname((unsigned char *)"..", 2);
 }
 
-static void
-xfs_dir2_block_verify(
+static bool
+xfs_dir3_block_verify(
         struct xfs_buf          *bp)
 {
         struct xfs_mount        *mp = bp->b_target->bt_mount;
-        struct xfs_dir2_data_hdr *hdr = bp->b_addr;
-        int                     block_ok = 0;
-
-        block_ok = hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
-        block_ok = block_ok && __xfs_dir2_data_check(NULL, bp) == 0;
-
-        if (!block_ok) {
-                XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, hdr);
-                xfs_buf_ioerror(bp, EFSCORRUPTED);
+        struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+
+        if (xfs_sb_version_hascrc(&mp->m_sb)) {
+                if (hdr3->magic != cpu_to_be32(XFS_DIR3_BLOCK_MAGIC))
+                        return false;
+                if (!uuid_equal(&hdr3->uuid, &mp->m_sb.sb_uuid))
+                        return false;
+                if (be64_to_cpu(hdr3->blkno) != bp->b_bn)
+                        return false;
+        } else {
+                if (hdr3->magic != cpu_to_be32(XFS_DIR2_BLOCK_MAGIC))
+                        return false;
         }
+        if (__xfs_dir2_data_check(NULL, bp))
+                return false;
+        return true;
 }
 
 static void
-xfs_dir2_block_read_verify(
+xfs_dir3_block_read_verify(
         struct xfs_buf  *bp)
 {
-        xfs_dir2_block_verify(bp);
+        struct xfs_mount        *mp = bp->b_target->bt_mount;
+
+        if ((xfs_sb_version_hascrc(&mp->m_sb) &&
+             !xfs_verify_cksum(bp->b_addr, BBTOB(bp->b_length),
+                                          XFS_DIR3_DATA_CRC_OFF)) ||
+            !xfs_dir3_block_verify(bp)) {
+                XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr);
+                xfs_buf_ioerror(bp, EFSCORRUPTED);
+        }
 }
 
 static void
-xfs_dir2_block_write_verify(
+xfs_dir3_block_write_verify(
         struct xfs_buf  *bp)
 {
-        xfs_dir2_block_verify(bp);
+        struct xfs_mount        *mp = bp->b_target->bt_mount;
+        struct xfs_buf_log_item *bip = bp->b_fspriv;
+        struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+
+        if (!xfs_dir3_block_verify(bp)) {
+                XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, bp->b_addr);
+                xfs_buf_ioerror(bp, EFSCORRUPTED);
+                return;
+        }
+
+        if (!xfs_sb_version_hascrc(&mp->m_sb))
+                return;
+
+        if (bip)
+                hdr3->lsn = cpu_to_be64(bip->bli_item.li_lsn);
+
+        xfs_update_cksum(bp->b_addr, BBTOB(bp->b_length), XFS_DIR3_DATA_CRC_OFF);
 }
 
-const struct xfs_buf_ops xfs_dir2_block_buf_ops = {
-        .verify_read = xfs_dir2_block_read_verify,
-        .verify_write = xfs_dir2_block_write_verify,
+const struct xfs_buf_ops xfs_dir3_block_buf_ops = {
+        .verify_read = xfs_dir3_block_read_verify,
+        .verify_write = xfs_dir3_block_write_verify,
 };
 
 static int
-xfs_dir2_block_read(
+xfs_dir3_block_read(
         struct xfs_trans        *tp,
         struct xfs_inode        *dp,
         struct xfs_buf          **bpp)
@@ -101,7 +134,29 @@ xfs_dir2_block_read(
         struct xfs_mount        *mp = dp->i_mount;
 
         return xfs_da_read_buf(tp, dp, mp->m_dirdatablk, -1, bpp,
-                                XFS_DATA_FORK, &xfs_dir2_block_buf_ops);
+                                XFS_DATA_FORK, &xfs_dir3_block_buf_ops);
+}
+
+static void
+xfs_dir3_block_init(
+        struct xfs_mount        *mp,
+        struct xfs_buf          *bp,
+        struct xfs_inode        *dp)
+{
+        struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+
+        bp->b_ops = &xfs_dir3_block_buf_ops;
+
+        if (xfs_sb_version_hascrc(&mp->m_sb)) {
+                memset(hdr3, 0, sizeof(*hdr3));
+                hdr3->magic = cpu_to_be32(XFS_DIR3_BLOCK_MAGIC);
+                hdr3->blkno = cpu_to_be64(bp->b_bn);
+                hdr3->owner = cpu_to_be64(dp->i_ino);
+                uuid_copy(&hdr3->uuid, &mp->m_sb.sb_uuid);
+                return;
+
+        }
+        hdr3->magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
 }
 
 static void
@@ -121,7 +176,7 @@ xfs_dir2_block_need_space(
         struct xfs_dir2_data_unused     *enddup = NULL;
 
         *compact = 0;
-        bf = hdr->bestfree;
+        bf = xfs_dir3_data_bestfree_p(hdr);
 
         /*
          * If there are stale entries we'll use one for the leaf.
@@ -303,7 +358,7 @@ xfs_dir2_block_addname(
         mp = dp->i_mount;
 
         /* Read the (one and only) directory block into bp. */
-        error = xfs_dir2_block_read(tp, dp, &bp);
+        error = xfs_dir3_block_read(tp, dp, &bp);
         if (error)
                 return error;
 
@@ -531,7 +586,7 @@ xfs_dir2_block_getdents(
         if (xfs_dir2_dataptr_to_db(mp, *offset) > mp->m_dirdatablk)
                 return 0;
 
-        error = xfs_dir2_block_read(NULL, dp, &bp);
+        error = xfs_dir3_block_read(NULL, dp, &bp);
         if (error)
                 return error;
 
@@ -546,7 +601,7 @@ xfs_dir2_block_getdents(
          * Set up values for the loop.
          */
         btp = xfs_dir2_block_tail_p(mp, hdr);
-        ptr = (char *)(hdr + 1);
+        ptr = (char *)xfs_dir3_data_entry_p(hdr);
         endptr = (char *)xfs_dir2_block_leaf_p(btp);
 
         /*
@@ -711,7 +766,7 @@ xfs_dir2_block_lookup_int(
         tp = args->trans;
         mp = dp->i_mount;
 
-        error = xfs_dir2_block_read(tp, dp, &bp);
+        error = xfs_dir3_block_read(tp, dp, &bp);
         if (error)
                 return error;
 
@@ -974,9 +1029,12 @@ xfs_dir2_leaf_to_block(
          * These will show up in the leaf bests table.
          */
         while (dp->i_d.di_size > mp->m_dirblksize) {
+                int hdrsz;
+
+                hdrsz = xfs_dir3_data_hdr_size(xfs_sb_version_hascrc(&mp->m_sb));
                 bestsp = xfs_dir2_leaf_bests_p(ltp);
                 if (be16_to_cpu(bestsp[be32_to_cpu(ltp->bestcount) - 1]) ==
-                    mp->m_dirblksize - (uint)sizeof(*hdr)) {
+                                            mp->m_dirblksize - hdrsz) {
                         if ((error =
                             xfs_dir2_leaf_trim_data(args, lbp,
                                     (xfs_dir2_db_t)(be32_to_cpu(ltp->bestcount) - 1))))
@@ -1014,8 +1072,8 @@ xfs_dir2_leaf_to_block(
         /*
          * Start converting it to block form.
          */
-        dbp->b_ops = &xfs_dir2_block_buf_ops;
-        hdr->magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
+        xfs_dir3_block_init(mp, dbp, dp);
+
         needlog = 1;
         needscan = 0;
         /*
@@ -1137,16 +1195,16 @@ xfs_dir2_sf_to_block(
                 return error;
         }
         /*
-         * Initialize the data block.
+         * Initialize the data block, then convert it to block format.
          */
-        error = xfs_dir2_data_init(args, blkno, &bp);
+        error = xfs_dir3_data_init(args, blkno, &bp);
         if (error) {
                 kmem_free(sfp);
                 return error;
         }
-        bp->b_ops = &xfs_dir2_block_buf_ops;
+        xfs_dir3_block_init(mp, bp, dp);
         hdr = bp->b_addr;
-        hdr->magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
+
         /*
          * Compute size of block "tail" area.
          */
@@ -1156,7 +1214,7 @@ xfs_dir2_sf_to_block(
          * The whole thing is initialized to free by the init routine.
          * Say we're using the leaf and tail area.
          */
-        dup = (xfs_dir2_data_unused_t *)(hdr + 1);
+        dup = xfs_dir3_data_unused_p(hdr);
         needlog = needscan = 0;
         xfs_dir2_data_use_free(tp, bp, dup, mp->m_dirblksize - i, i, &needlog,
                 &needscan);
@@ -1178,8 +1236,7 @@ xfs_dir2_sf_to_block(
         /*
          * Create entry for .
          */
-        dep = (xfs_dir2_data_entry_t *)
-              ((char *)hdr + XFS_DIR2_DATA_DOT_OFFSET);
+        dep = xfs_dir3_data_dot_entry_p(hdr);
         dep->inumber = cpu_to_be64(dp->i_ino);
         dep->namelen = 1;
         dep->name[0] = '.';
@@ -1192,8 +1249,7 @@ xfs_dir2_sf_to_block(
         /*
          * Create entry for ..
          */
-        dep = (xfs_dir2_data_entry_t *)
-                ((char *)hdr + XFS_DIR2_DATA_DOTDOT_OFFSET);
+        dep = xfs_dir3_data_dotdot_entry_p(hdr);
         dep->inumber = cpu_to_be64(xfs_dir2_sf_get_parent_ino(sfp));
         dep->namelen = 2;
         dep->name[0] = dep->name[1] = '.';
@@ -1203,7 +1259,7 @@ xfs_dir2_sf_to_block(
         blp[1].hashval = cpu_to_be32(xfs_dir_hash_dotdot);
         blp[1].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(mp,
                                 (char *)dep - (char *)hdr));
-        offset = XFS_DIR2_DATA_FIRST_OFFSET;
+        offset = xfs_dir3_data_first_offset(hdr);
         /*
          * Loop over existing entries, stuff them in.
          */
diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c
index ffcf1774152e..8484ec4848ea 100644
--- a/fs/xfs/xfs_dir2_data.c
+++ b/fs/xfs/xfs_dir2_data.c
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -65,10 +66,11 @@ __xfs_dir2_data_check(
 
         mp = bp->b_target->bt_mount;
         hdr = bp->b_addr;
-        bf = hdr->bestfree;
-        p = (char *)(hdr + 1);
+        bf = xfs_dir3_data_bestfree_p(hdr);
+        p = (char *)xfs_dir3_data_entry_p(hdr);
 
         switch (hdr->magic) {
+        case cpu_to_be32(XFS_DIR3_BLOCK_MAGIC):
         case cpu_to_be32(XFS_DIR2_BLOCK_MAGIC):
                 btp = xfs_dir2_block_tail_p(mp, hdr);
                 lep = xfs_dir2_block_leaf_p(btp);
@@ -148,7 +150,8 @@ __xfs_dir2_data_check(
                                                (char *)dep - (char *)hdr);
                 count++;
                 lastfree = 0;
-                if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC)) {
+                if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+                    hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
                         addr = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
                                 (xfs_dir2_data_aoff_t)
                                 ((char *)dep - (char *)hdr));
@@ -168,7 +171,8 @@ __xfs_dir2_data_check(
          * Need to have seen all the entries and all the bestfree slots.
          */
         XFS_WANT_CORRUPTED_RETURN(freeseen == 7);
-        if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC)) {
+        if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+            hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
                 for (i = stale = 0; i < be32_to_cpu(btp->count); i++) {
                         if (lep[i].address ==
                             cpu_to_be32(XFS_DIR2_NULL_DATAPTR))
@@ -216,7 +220,8 @@ xfs_dir2_data_reada_verify(
 
         switch (hdr->magic) {
         case cpu_to_be32(XFS_DIR2_BLOCK_MAGIC):
-                bp->b_ops = &xfs_dir2_block_buf_ops;
+        case cpu_to_be32(XFS_DIR3_BLOCK_MAGIC):
+                bp->b_ops = &xfs_dir3_block_buf_ops;
                 bp->b_ops->verify_read(bp);
                 return;
         case cpu_to_be32(XFS_DIR2_DATA_MAGIC):
@@ -288,12 +293,15 @@ xfs_dir2_data_freefind(
 {
         xfs_dir2_data_free_t    *dfp;           /* bestfree entry */
         xfs_dir2_data_aoff_t    off;            /* offset value needed */
+        struct xfs_dir2_data_free *bf;
 #if defined(DEBUG) && defined(__KERNEL__)
         int                     matched;        /* matched the value */
         int                     seenzero;       /* saw a 0 bestfree entry */
 #endif
 
         off = (xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr);
+        bf = xfs_dir3_data_bestfree_p(hdr);
+
 #if defined(DEBUG) && defined(__KERNEL__)
         /*
          * Validate some consistency in the bestfree table.
@@ -301,9 +309,10 @@ xfs_dir2_data_freefind(
          * one we're looking for it has to be exact.
          */
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
-        for (dfp = &hdr->bestfree[0], seenzero = matched = 0;
-             dfp < &hdr->bestfree[XFS_DIR2_DATA_FD_COUNT];
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
+        for (dfp = &bf[0], seenzero = matched = 0;
+             dfp < &bf[XFS_DIR2_DATA_FD_COUNT];
              dfp++) {
                 if (!dfp->offset) {
                         ASSERT(!dfp->length);
@@ -319,7 +328,7 @@ xfs_dir2_data_freefind(
                 else
                         ASSERT(be16_to_cpu(dfp->offset) + be16_to_cpu(dfp->length) <= off);
                 ASSERT(matched || be16_to_cpu(dfp->length) >= be16_to_cpu(dup->length));
-                if (dfp > &hdr->bestfree[0])
+                if (dfp > &bf[0])
                         ASSERT(be16_to_cpu(dfp[-1].length) >= be16_to_cpu(dfp[0].length));
         }
 #endif
@@ -328,14 +337,12 @@ xfs_dir2_data_freefind(
          * it can't be there since they're sorted.
          */
         if (be16_to_cpu(dup->length) <
-            be16_to_cpu(hdr->bestfree[XFS_DIR2_DATA_FD_COUNT - 1].length))
+            be16_to_cpu(bf[XFS_DIR2_DATA_FD_COUNT - 1].length))
                 return NULL;
         /*
          * Look at the three bestfree entries for our guy.
          */
-        for (dfp = &hdr->bestfree[0];
-             dfp < &hdr->bestfree[XFS_DIR2_DATA_FD_COUNT];
-             dfp++) {
+        for (dfp = &bf[0]; dfp < &bf[XFS_DIR2_DATA_FD_COUNT]; dfp++) {
                 if (!dfp->offset)
                         return NULL;
                 if (be16_to_cpu(dfp->offset) == off)
@@ -359,11 +366,12 @@ xfs_dir2_data_freeinsert(
         xfs_dir2_data_free_t    *dfp;           /* bestfree table pointer */
         xfs_dir2_data_free_t    new;            /* new bestfree entry */
 
-#ifdef __KERNEL__
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
-#endif
-        dfp = hdr->bestfree;
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
+
+        dfp = xfs_dir3_data_bestfree_p(hdr);
         new.length = dup->length;
         new.offset = cpu_to_be16((char *)dup - (char *)hdr);
 
@@ -400,32 +408,36 @@ xfs_dir2_data_freeremove(
         xfs_dir2_data_free_t    *dfp,           /* bestfree entry pointer */
         int                     *loghead)       /* out: log data header */
 {
-#ifdef __KERNEL__
+        struct xfs_dir2_data_free *bf;
+
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
-#endif
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
+
         /*
          * It's the first entry, slide the next 2 up.
          */
-        if (dfp == &hdr->bestfree[0]) {
-                hdr->bestfree[0] = hdr->bestfree[1];
-                hdr->bestfree[1] = hdr->bestfree[2];
+        bf = xfs_dir3_data_bestfree_p(hdr);
+        if (dfp == &bf[0]) {
+                bf[0] = bf[1];
+                bf[1] = bf[2];
         }
         /*
          * It's the second entry, slide the 3rd entry up.
          */
-        else if (dfp == &hdr->bestfree[1])
-                hdr->bestfree[1] = hdr->bestfree[2];
+        else if (dfp == &bf[1])
+                bf[1] = bf[2];
         /*
          * Must be the last entry.
          */
         else
-                ASSERT(dfp == &hdr->bestfree[2]);
+                ASSERT(dfp == &bf[2]);
         /*
          * Clear the 3rd entry, must be zero now.
          */
-        hdr->bestfree[2].length = 0;
-        hdr->bestfree[2].offset = 0;
+        bf[2].length = 0;
+        bf[2].offset = 0;
         *loghead = 1;
 }
 
@@ -441,23 +453,26 @@ xfs_dir2_data_freescan(
         xfs_dir2_block_tail_t   *btp;           /* block tail */
         xfs_dir2_data_entry_t   *dep;           /* active data entry */
         xfs_dir2_data_unused_t  *dup;           /* unused data entry */
+        struct xfs_dir2_data_free *bf;
         char                    *endp;          /* end of block's data */
         char                    *p;             /* current entry pointer */
 
-#ifdef __KERNEL__
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
-#endif
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
+
         /*
          * Start by clearing the table.
          */
-        memset(hdr->bestfree, 0, sizeof(hdr->bestfree));
+        bf = xfs_dir3_data_bestfree_p(hdr);
+        memset(bf, 0, sizeof(*bf) * XFS_DIR2_DATA_FD_COUNT);
         *loghead = 1;
         /*
          * Set up pointers.
          */
-        p = (char *)(hdr + 1);
-        if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC)) {
+        p = (char *)xfs_dir3_data_entry_p(hdr);
+        if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+            hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
                 btp = xfs_dir2_block_tail_p(mp, hdr);
                 endp = (char *)xfs_dir2_block_leaf_p(btp);
         } else
@@ -493,7 +508,7 @@ xfs_dir2_data_freescan(
  * Give back the buffer for the created block.
  */
 int                                             /* error */
-xfs_dir2_data_init(
+xfs_dir3_data_init(
         xfs_da_args_t           *args,          /* directory operation args */
         xfs_dir2_db_t           blkno,          /* logical dir block number */
         struct xfs_buf          **bpp)          /* output block buffer */
@@ -502,6 +517,7 @@ xfs_dir2_data_init(
         xfs_dir2_data_hdr_t     *hdr;           /* data block header */
         xfs_inode_t             *dp;            /* incore directory inode */
         xfs_dir2_data_unused_t  *dup;           /* unused entry pointer */
+        struct xfs_dir2_data_free *bf;
         int                     error;          /* error return value */
         int                     i;              /* bestfree index */
         xfs_mount_t             *mp;            /* filesystem mount point */
@@ -524,21 +540,34 @@ xfs_dir2_data_init(
          * Initialize the header.
          */
         hdr = bp->b_addr;
-        hdr->magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
-        hdr->bestfree[0].offset = cpu_to_be16(sizeof(*hdr));
+
+        if (xfs_sb_version_hascrc(&mp->m_sb)) {
+                struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+
+                memset(hdr3, 0, sizeof(*hdr3));
+                hdr3->magic = cpu_to_be32(XFS_DIR3_DATA_MAGIC);
+                hdr3->blkno = cpu_to_be64(bp->b_bn);
+                hdr3->owner = cpu_to_be64(dp->i_ino);
+                uuid_copy(&hdr3->uuid, &mp->m_sb.sb_uuid);
+
+        } else
+                hdr->magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
+
+        bf = xfs_dir3_data_bestfree_p(hdr);
+        bf[0].offset = cpu_to_be16(xfs_dir3_data_entry_offset(hdr));
         for (i = 1; i < XFS_DIR2_DATA_FD_COUNT; i++) {
-                hdr->bestfree[i].length = 0;
-                hdr->bestfree[i].offset = 0;
+                bf[i].length = 0;
+                bf[i].offset = 0;
         }
 
         /*
          * Set up an unused entry for the block's body.
          */
-        dup = (xfs_dir2_data_unused_t *)(hdr + 1);
+        dup = xfs_dir3_data_unused_p(hdr);
         dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
 
-        t = mp->m_dirblksize - (uint)sizeof(*hdr);
-        hdr->bestfree[0].length = cpu_to_be16(t);
+        t = mp->m_dirblksize - (uint)xfs_dir3_data_entry_offset(hdr);
+        bf[0].length = cpu_to_be16(t);
         dup->length = cpu_to_be16(t);
         *xfs_dir2_data_unused_tag_p(dup) = cpu_to_be16((char *)dup - (char *)hdr);
         /*
@@ -562,7 +591,8 @@ xfs_dir2_data_log_entry(
         xfs_dir2_data_hdr_t     *hdr = bp->b_addr;
 
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
 
         xfs_trans_log_buf(tp, bp, (uint)((char *)dep - (char *)hdr),
                 (uint)((char *)(xfs_dir2_data_entry_tag_p(dep) + 1) -
@@ -580,9 +610,10 @@ xfs_dir2_data_log_header(
         xfs_dir2_data_hdr_t     *hdr = bp->b_addr;
 
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
 
-        xfs_trans_log_buf(tp, bp, 0, sizeof(*hdr) - 1);
+        xfs_trans_log_buf(tp, bp, 0, xfs_dir3_data_entry_offset(hdr) - 1);
 }
 
 /*
@@ -597,7 +628,8 @@ xfs_dir2_data_log_unused(
         xfs_dir2_data_hdr_t     *hdr = bp->b_addr;
 
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
 
         /*
          * Log the first part of the unused entry.
@@ -635,6 +667,7 @@ xfs_dir2_data_make_free(
         xfs_dir2_data_unused_t  *newdup;        /* new unused entry */
         xfs_dir2_data_unused_t  *postdup;       /* unused entry after us */
         xfs_dir2_data_unused_t  *prevdup;       /* unused entry before us */
+        struct xfs_dir2_data_free *bf;
 
         mp = tp->t_mountp;
         hdr = bp->b_addr;
@@ -647,7 +680,8 @@ xfs_dir2_data_make_free(
         else {
                 xfs_dir2_block_tail_t   *btp;   /* block tail */
 
-                ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
+                ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+                        hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
                 btp = xfs_dir2_block_tail_p(mp, hdr);
                 endptr = (char *)xfs_dir2_block_leaf_p(btp);
         }
@@ -655,7 +689,7 @@ xfs_dir2_data_make_free(
          * If this isn't the start of the block, then back up to
          * the previous entry and see if it's free.
          */
-        if (offset > sizeof(*hdr)) {
+        if (offset > xfs_dir3_data_entry_offset(hdr)) {
                 __be16                  *tagp;  /* tag just before us */
 
                 tagp = (__be16 *)((char *)hdr + offset) - 1;
@@ -681,6 +715,7 @@ xfs_dir2_data_make_free(
          * Previous and following entries are both free,
          * merge everything into a single free entry.
          */
+        bf = xfs_dir3_data_bestfree_p(hdr);
         if (prevdup && postdup) {
                 xfs_dir2_data_free_t    *dfp2;  /* another bestfree pointer */
 
@@ -695,7 +730,7 @@ xfs_dir2_data_make_free(
                  * since the third bestfree is there, there might be more
                  * entries.
                  */
-                needscan = (hdr->bestfree[2].length != 0);
+                needscan = (bf[2].length != 0);
                 /*
                  * Fix up the new big freespace.
                  */
@@ -711,10 +746,10 @@ xfs_dir2_data_make_free(
                          * Remove entry 1 first then entry 0.
                          */
                         ASSERT(dfp && dfp2);
-                        if (dfp == &hdr->bestfree[1]) {
-                                dfp = &hdr->bestfree[0];
+                        if (dfp == &bf[1]) {
+                                dfp = &bf[0];
                                 ASSERT(dfp2 == dfp);
-                                dfp2 = &hdr->bestfree[1];
+                                dfp2 = &bf[1];
                         }
                         xfs_dir2_data_freeremove(hdr, dfp2, needlogp);
                         xfs_dir2_data_freeremove(hdr, dfp, needlogp);
@@ -722,7 +757,7 @@ xfs_dir2_data_make_free(
                          * Now insert the new entry.
                          */
                         dfp = xfs_dir2_data_freeinsert(hdr, prevdup, needlogp);
-                        ASSERT(dfp == &hdr->bestfree[0]);
+                        ASSERT(dfp == &bf[0]);
                         ASSERT(dfp->length == prevdup->length);
                         ASSERT(!dfp[1].length);
                         ASSERT(!dfp[2].length);
@@ -751,7 +786,7 @@ xfs_dir2_data_make_free(
                  */
                 else {
                         needscan = be16_to_cpu(prevdup->length) >
-                                   be16_to_cpu(hdr->bestfree[2].length);
+                                   be16_to_cpu(bf[2].length);
                 }
         }
         /*
@@ -779,7 +814,7 @@ xfs_dir2_data_make_free(
                  */
                 else {
                         needscan = be16_to_cpu(newdup->length) >
-                                   be16_to_cpu(hdr->bestfree[2].length);
+                                   be16_to_cpu(bf[2].length);
                 }
         }
         /*
@@ -818,10 +853,12 @@ xfs_dir2_data_use_free(
         xfs_dir2_data_unused_t  *newdup;        /* new unused entry */
         xfs_dir2_data_unused_t  *newdup2;       /* another new unused entry */
         int                     oldlen;         /* old unused entry's length */
+        struct xfs_dir2_data_free *bf;
 
         hdr = bp->b_addr;
         ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC));
+               hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
+               hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
         ASSERT(be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG);
         ASSERT(offset >= (char *)dup - (char *)hdr);
         ASSERT(offset + len <= (char *)dup + be16_to_cpu(dup->length) - (char *)hdr);
@@ -831,7 +868,8 @@ xfs_dir2_data_use_free(
          */
         dfp = xfs_dir2_data_freefind(hdr, dup);
         oldlen = be16_to_cpu(dup->length);
-        ASSERT(dfp || oldlen <= be16_to_cpu(hdr->bestfree[2].length));
+        bf = xfs_dir3_data_bestfree_p(hdr);
+        ASSERT(dfp || oldlen <= be16_to_cpu(bf[2].length));
         /*
          * Check for alignment with front and back of the entry.
          */
@@ -845,7 +883,7 @@ xfs_dir2_data_use_free(
          */
         if (matchfront && matchback) {
                 if (dfp) {
-                        needscan = (hdr->bestfree[2].offset != 0);
+                        needscan = (bf[2].offset != 0);
                         if (!needscan)
                                 xfs_dir2_data_freeremove(hdr, dfp, needlogp);
                 }
@@ -875,7 +913,7 @@ xfs_dir2_data_use_free(
                          * that means we don't know if there was a better
                          * choice for the last slot, or not.  Rescan.
                          */
-                        needscan = dfp == &hdr->bestfree[2];
+                        needscan = dfp == &bf[2];
                 }
         }
         /*
@@ -902,7 +940,7 @@ xfs_dir2_data_use_free(
                          * that means we don't know if there was a better
                          * choice for the last slot, or not.  Rescan.
                          */
-                        needscan = dfp == &hdr->bestfree[2];
+                        needscan = dfp == &bf[2];
                 }
         }
         /*
@@ -930,7 +968,7 @@ xfs_dir2_data_use_free(
                  * the 2 new will work.
                  */
                 if (dfp) {
-                        needscan = (hdr->bestfree[2].length != 0);
+                        needscan = (bf[2].length != 0);
                         if (!needscan) {
                                 xfs_dir2_data_freeremove(hdr, dfp, needlogp);
                                 xfs_dir2_data_freeinsert(hdr, newdup, needlogp);
diff --git a/fs/xfs/xfs_dir2_format.h b/fs/xfs/xfs_dir2_format.h
index 07270981f48f..845979a4ee96 100644
--- a/fs/xfs/xfs_dir2_format.h
+++ b/fs/xfs/xfs_dir2_format.h
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -36,6 +37,37 @@
 #define XFS_DIR2_FREE_MAGIC     0x58443246      /* XD2F: free index blocks */
 
 /*
+ * Directory Version 3 With CRCs.
+ *
+ * The tree formats are the same as for version 2 directories.  The difference
+ * is in the block header and dirent formats. In many cases the v3 structures
+ * use v2 definitions as they are no different and this makes code sharing much
+ * easier.
+ *
+ * Also, the xfs_dir3_*() functions handle both v2 and v3 formats - if the
+ * format is v2 then they switch to the existing v2 code, or the format is v3
+ * they implement the v3 functionality. This means the existing dir2 is a mix of
+ * xfs_dir2/xfs_dir3 calls and functions. The xfs_dir3 functions are called
+ * where there is a difference in the formats, otherwise the code is unchanged.
+ *
+ * Where it is possible, the code decides what to do based on the magic numbers
+ * in the blocks rather than feature bits in the superblock. This means the code
+ * is as independent of the external XFS code as possible as doesn't require
+ * passing struct xfs_mount pointers into places where it isn't really
+ * necessary.
+ *
+ * Version 3 includes:
+ *
+ *      - a larger block header for CRC and identification purposes and so the
+ *      offsets of all the structures inside the blocks are different.
+ *
+ *      - new magic numbers to be able to detect the v2/v3 types on the fly.
+ */
+
+#define XFS_DIR3_BLOCK_MAGIC    0x58444233      /* XDB3: single block dirs */
+#define XFS_DIR3_DATA_MAGIC     0x58444433      /* XDD3: multiblock dirs */
+
+/*
  * Byte offset in data block and shortform entry.
  */
 typedef __uint16_t      xfs_dir2_data_off_t;
@@ -226,6 +258,38 @@ typedef struct xfs_dir2_data_hdr {
 } xfs_dir2_data_hdr_t;
 
 /*
+ * define a structure for all the verification fields we are adding to the
+ * directory block structures. This will be used in several structures.
+ * The magic number must be the first entry to align with all the dir2
+ * structures so we determine how to decode them just by the magic number.
+ */
+struct xfs_dir3_blk_hdr {
+        __be32                  magic;  /* magic number */
+        __be32                  crc;    /* CRC of block */
+        __be64                  blkno;  /* first block of the buffer */
+        __be64                  lsn;    /* sequence number of last write */
+        uuid_t                  uuid;   /* filesystem we belong to */
+        __be64                  owner;  /* inode that owns the block */
+};
+
+struct xfs_dir3_data_hdr {
+        struct xfs_dir3_blk_hdr hdr;
+        xfs_dir2_data_free_t    best_free[XFS_DIR2_DATA_FD_COUNT];
+};
+
+#define XFS_DIR3_DATA_CRC_OFF  offsetof(struct xfs_dir3_data_hdr, hdr.crc)
+
+static inline struct xfs_dir2_data_free *
+xfs_dir3_data_bestfree_p(struct xfs_dir2_data_hdr *hdr)
+{
+        if (hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
+                struct xfs_dir3_data_hdr *hdr3 = (struct xfs_dir3_data_hdr *)hdr;
+                return hdr3->best_free;
+        }
+        return hdr->bestfree;
+}
+
+/*
  * Active entry in a data block.
  *
  * Aligned to 8 bytes.  After the variable length name field there is a
@@ -280,6 +344,85 @@ xfs_dir2_data_unused_tag_p(struct xfs_dir2_data_unused *dup)
                         be16_to_cpu(dup->length) - sizeof(__be16));
 }
 
+static inline struct xfs_dir2_data_unused *
+xfs_dir3_data_unused_p(struct xfs_dir2_data_hdr *hdr)
+{
+        if (hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
+                return (struct xfs_dir2_data_unused *)
+                        ((char *)hdr + sizeof(struct xfs_dir3_data_hdr));
+        }
+        return (struct xfs_dir2_data_unused *)
+                ((char *)hdr + sizeof(struct xfs_dir2_data_hdr));
+}
+
+static inline size_t
+xfs_dir3_data_hdr_size(bool dir3)
+{
+        if (dir3)
+                return sizeof(struct xfs_dir3_data_hdr);
+        return sizeof(struct xfs_dir2_data_hdr);
+}
+
+static inline size_t
+xfs_dir3_data_entry_offset(struct xfs_dir2_data_hdr *hdr)
+{
+        bool dir3 = hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
+                    hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC);
+        return xfs_dir3_data_hdr_size(dir3);
+}
+
+static inline struct xfs_dir2_data_entry *
+xfs_dir3_data_entry_p(struct xfs_dir2_data_hdr *hdr)
+{
+        return (struct xfs_dir2_data_entry *)
+                ((char *)hdr + xfs_dir3_data_entry_offset(hdr));
+}
+
+/*
+ * Offsets of . and .. in data space (always block 0)
+ */
+static inline xfs_dir2_data_aoff_t
+xfs_dir3_data_dot_offset(struct xfs_dir2_data_hdr *hdr)
+{
+        return xfs_dir3_data_entry_offset(hdr);
+}
+
+static inline xfs_dir2_data_aoff_t
+xfs_dir3_data_dotdot_offset(struct xfs_dir2_data_hdr *hdr)
+{
+        return xfs_dir3_data_dot_offset(hdr) + xfs_dir2_data_entsize(1);
+}
+
+static inline xfs_dir2_data_aoff_t
+xfs_dir3_data_first_offset(struct xfs_dir2_data_hdr *hdr)
+{
+        return xfs_dir3_data_dotdot_offset(hdr) + xfs_dir2_data_entsize(2);
+}
+
+/*
+ * location of . and .. in data space (always block 0)
+ */
+static inline struct xfs_dir2_data_entry *
+xfs_dir3_data_dot_entry_p(struct xfs_dir2_data_hdr *hdr)
+{
+        return (struct xfs_dir2_data_entry *)
+                ((char *)hdr + xfs_dir3_data_dot_offset(hdr));
+}
+
+static inline struct xfs_dir2_data_entry *
+xfs_dir3_data_dotdot_entry_p(struct xfs_dir2_data_hdr *hdr)
+{
+        return (struct xfs_dir2_data_entry *)
+                ((char *)hdr + xfs_dir3_data_dotdot_offset(hdr));
+}
+
+static inline struct xfs_dir2_data_entry *
+xfs_dir3_data_first_entry_p(struct xfs_dir2_data_hdr *hdr)
+{
+        return (struct xfs_dir2_data_entry *)
+                ((char *)hdr + xfs_dir3_data_first_offset(hdr));
+}
+
 /*
  * Leaf block structures.
  *
diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c
index 60cd2fa4e047..ef163f06fd16 100644
--- a/fs/xfs/xfs_dir2_leaf.c
+++ b/fs/xfs/xfs_dir2_leaf.c
@@ -149,6 +149,7 @@ xfs_dir2_block_to_leaf(
         int                     needlog;        /* need to log block header */
         int                     needscan;       /* need to rescan bestfree */
         xfs_trans_t             *tp;            /* transaction pointer */
+        struct xfs_dir2_data_free       *bf;
 
         trace_xfs_dir2_block_to_leaf(args);
 
@@ -177,6 +178,7 @@ xfs_dir2_block_to_leaf(
         xfs_dir2_data_check(dp, dbp);
         btp = xfs_dir2_block_tail_p(mp, hdr);
         blp = xfs_dir2_block_leaf_p(btp);
+        bf = xfs_dir3_data_bestfree_p(hdr);
         /*
          * Set the counts in the leaf header.
          */
@@ -212,7 +214,7 @@ xfs_dir2_block_to_leaf(
         ltp = xfs_dir2_leaf_tail_p(mp, leaf);
         ltp->bestcount = cpu_to_be32(1);
         bestsp = xfs_dir2_leaf_bests_p(ltp);
-        bestsp[0] =  hdr->bestfree[0].length;
+        bestsp[0] =  bf[0].length;
         /*
          * Log the data header and leaf bests table.
          */
@@ -544,7 +546,7 @@ xfs_dir2_leaf_addname(
                 /*
                  * Initialize the block.
                  */
-                if ((error = xfs_dir2_data_init(args, use_block, &dbp))) {
+                if ((error = xfs_dir3_data_init(args, use_block, &dbp))) {
                         xfs_trans_brelse(tp, lbp);
                         return error;
                 }
diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c
index 5980f9b7fa9b..985d70bc391d 100644
--- a/fs/xfs/xfs_dir2_node.c
+++ b/fs/xfs/xfs_dir2_node.c
@@ -1588,7 +1588,7 @@ xfs_dir2_node_addname_int(
                 if (unlikely((error = xfs_dir2_grow_inode(args,
                                                          XFS_DIR2_DATA_SPACE,
                                                          &dbno)) ||
-                    (error = xfs_dir2_data_init(args, dbno, &dbp))))
+                    (error = xfs_dir3_data_init(args, dbno, &dbp))))
                         return error;
 
                 /*
diff --git a/fs/xfs/xfs_dir2_priv.h b/fs/xfs/xfs_dir2_priv.h
index 7da79f6515fd..e6f2e0a7bb65 100644
--- a/fs/xfs/xfs_dir2_priv.h
+++ b/fs/xfs/xfs_dir2_priv.h
@@ -30,7 +30,7 @@ extern int xfs_dir_cilookup_result(struct xfs_da_args *args,
                                 const unsigned char *name, int len);
 
 /* xfs_dir2_block.c */
-extern const struct xfs_buf_ops xfs_dir2_block_buf_ops;
+extern const struct xfs_buf_ops xfs_dir3_block_buf_ops;
 
 extern int xfs_dir2_block_addname(struct xfs_da_args *args);
 extern int xfs_dir2_block_getdents(struct xfs_inode *dp, void *dirent,
@@ -61,7 +61,7 @@ xfs_dir2_data_freeinsert(struct xfs_dir2_data_hdr *hdr,
                 struct xfs_dir2_data_unused *dup, int *loghead);
 extern void xfs_dir2_data_freescan(struct xfs_mount *mp,
                 struct xfs_dir2_data_hdr *hdr, int *loghead);
-extern int xfs_dir2_data_init(struct xfs_da_args *args, xfs_dir2_db_t blkno,
+extern int xfs_dir3_data_init(struct xfs_da_args *args, xfs_dir2_db_t blkno,
                 struct xfs_buf **bpp);
 extern void xfs_dir2_data_log_entry(struct xfs_trans *tp, struct xfs_buf *bp,
                 struct xfs_dir2_data_entry *dep);
diff --git a/fs/xfs/xfs_dir2_sf.c b/fs/xfs/xfs_dir2_sf.c
index 1b9fc3ec7e4b..b9756228e9c0 100644
--- a/fs/xfs/xfs_dir2_sf.c
+++ b/fs/xfs/xfs_dir2_sf.c
@@ -278,7 +278,7 @@ xfs_dir2_block_to_sf(
          * Set up to loop over the block's entries.
          */
         btp = xfs_dir2_block_tail_p(mp, hdr);
-        ptr = (char *)(hdr + 1);
+        ptr = (char *)xfs_dir3_data_entry_p(hdr);
         endptr = (char *)xfs_dir2_block_leaf_p(btp);
         sfep = xfs_dir2_sf_firstentry(sfp);
         /*