xfs: Add write support for dirent filetype field

Add support to propagate and add filetype values into the on-disk
directs. This involves passing the filetype into the xfs_da_args
structure along with the name and namelength for direct operations,
and encoding it into the dirent at the same time we write the inode
number into the dirent.

With write support, add the feature flag to the
XFS_SB_FEAT_INCOMPAT_ALL mask so we can now mount filesystems with
this feature set.

Performance of directory recursion is now much improved. Parallel
walk of ~50 million directory entries across hundreds of directories
improves significantly. Unpatched, no CRCs:

Walking via ls -R

real    3m19.886s
user    6m36.960s
sys     28m19.087s

THis is doing roughly 500 getdents() calls per second, and 250,000
inode lookups per second to determine the inode type at roughly
17,000 read IOPS. CPU usage is 90% kernel space.

With dtype support patched in and the fileset recreated with CRCs
enabled:

Walking via ls -R

real    0m31.316s
user    6m32.975s
sys     0m21.111s

This is doing roughly 3500 getdents() calls per second at 16,000
IOPS. There are no inode lookups at all. CPU usages is almost 100%
userspace.

This is a big win for recursive directory walks that only need to
find file names and file types.

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

7 files changed, 47 insertions, 9 deletions

diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c
index f9825b125fcf..d3ff96c321f9 100644
--- a/fs/xfs/xfs_dir2.c
+++ b/fs/xfs/xfs_dir2.c
@@ -209,6 +209,7 @@ xfs_dir_createname(
         memset(&args, 0, sizeof(xfs_da_args_t));
         args.name = name->name;
         args.namelen = name->len;
+        args.filetype = name->type;
         args.hashval = dp->i_mount->m_dirnameops->hashname(name);
         args.inumber = inum;
         args.dp = dp;
@@ -283,6 +284,7 @@ xfs_dir_lookup(
         memset(&args, 0, sizeof(xfs_da_args_t));
         args.name = name->name;
         args.namelen = name->len;
+        args.filetype = name->type;
         args.hashval = dp->i_mount->m_dirnameops->hashname(name);
         args.dp = dp;
         args.whichfork = XFS_DATA_FORK;
@@ -338,6 +340,7 @@ xfs_dir_removename(
         memset(&args, 0, sizeof(xfs_da_args_t));
         args.name = name->name;
         args.namelen = name->len;
+        args.filetype = name->type;
         args.hashval = dp->i_mount->m_dirnameops->hashname(name);
         args.inumber = ino;
         args.dp = dp;
diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c
index 1cd2f564e374..0957aa98b6c0 100644
--- a/fs/xfs/xfs_dir2_block.c
+++ b/fs/xfs/xfs_dir2_block.c
@@ -549,6 +549,7 @@ xfs_dir2_block_addname(
         dep->inumber = cpu_to_be64(args->inumber);
         dep->namelen = args->namelen;
         memcpy(dep->name, args->name, args->namelen);
+        xfs_dir3_dirent_put_ftype(mp, dep, args->filetype);
         tagp = xfs_dir3_data_entry_tag_p(mp, dep);
         *tagp = cpu_to_be16((char *)dep - (char *)hdr);
         /*
@@ -641,6 +642,7 @@ xfs_dir2_block_lookup(
          * Fill in inode number, CI name if appropriate, release the block.
          */
         args->inumber = be64_to_cpu(dep->inumber);
+        args->filetype = xfs_dir3_dirent_get_ftype(mp, dep);
         error = xfs_dir_cilookup_result(args, dep->name, dep->namelen);
         xfs_trans_brelse(args->trans, bp);
         return XFS_ERROR(error);
@@ -873,6 +875,7 @@ xfs_dir2_block_replace(
          * Change the inode number to the new value.
          */
         dep->inumber = cpu_to_be64(args->inumber);
+        xfs_dir3_dirent_put_ftype(mp, dep, args->filetype);
         xfs_dir2_data_log_entry(args->trans, bp, dep);
         xfs_dir3_data_check(dp, bp);
         return 0;
@@ -1159,6 +1162,7 @@ xfs_dir2_sf_to_block(
         dep->inumber = cpu_to_be64(dp->i_ino);
         dep->namelen = 1;
         dep->name[0] = '.';
+        xfs_dir3_dirent_put_ftype(mp, dep, XFS_DIR3_FT_DIR);
         tagp = xfs_dir3_data_entry_tag_p(mp, dep);
         *tagp = cpu_to_be16((char *)dep - (char *)hdr);
         xfs_dir2_data_log_entry(tp, bp, dep);
@@ -1172,6 +1176,7 @@ xfs_dir2_sf_to_block(
         dep->inumber = cpu_to_be64(xfs_dir2_sf_get_parent_ino(sfp));
         dep->namelen = 2;
         dep->name[0] = dep->name[1] = '.';
+        xfs_dir3_dirent_put_ftype(mp, dep, XFS_DIR3_FT_DIR);
         tagp = xfs_dir3_data_entry_tag_p(mp, dep);
         *tagp = cpu_to_be16((char *)dep - (char *)hdr);
         xfs_dir2_data_log_entry(tp, bp, dep);
@@ -1219,6 +1224,8 @@ xfs_dir2_sf_to_block(
                 dep = (xfs_dir2_data_entry_t *)((char *)hdr + newoffset);
                 dep->inumber = cpu_to_be64(xfs_dir3_sfe_get_ino(mp, sfp, sfep));
                 dep->namelen = sfep->namelen;
+                xfs_dir3_dirent_put_ftype(mp, dep,
+                                        xfs_dir3_sfe_get_ftype(mp, sfp, sfep));
                 memcpy(dep->name, sfep->name, dep->namelen);
                 tagp = xfs_dir3_data_entry_tag_p(mp, dep);
                 *tagp = cpu_to_be16((char *)dep - (char *)hdr);
diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c
index 1b59e43d97d9..47e1326c169a 100644
--- a/fs/xfs/xfs_dir2_data.c
+++ b/fs/xfs/xfs_dir2_data.c
@@ -149,6 +149,8 @@ __xfs_dir3_data_check(
                 XFS_WANT_CORRUPTED_RETURN(
                         be16_to_cpu(*xfs_dir3_data_entry_tag_p(mp, dep)) ==
                                                (char *)dep - (char *)hdr);
+                XFS_WANT_CORRUPTED_RETURN(
+                        xfs_dir3_dirent_get_ftype(mp, dep) < XFS_DIR3_FT_MAX);
                 count++;
                 lastfree = 0;
                 if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c
index 887b1bdec6dd..08984eeee159 100644
--- a/fs/xfs/xfs_dir2_leaf.c
+++ b/fs/xfs/xfs_dir2_leaf.c
@@ -897,6 +897,7 @@ xfs_dir2_leaf_addname(
         dep->inumber = cpu_to_be64(args->inumber);
         dep->namelen = args->namelen;
         memcpy(dep->name, args->name, dep->namelen);
+        xfs_dir3_dirent_put_ftype(mp, dep, args->filetype);
         tagp = xfs_dir3_data_entry_tag_p(mp, dep);
         *tagp = cpu_to_be16((char *)dep - (char *)hdr);
         /*
@@ -1225,6 +1226,7 @@ xfs_dir2_leaf_lookup(
          * Return the found inode number & CI name if appropriate
          */
         args->inumber = be64_to_cpu(dep->inumber);
+        args->filetype = xfs_dir3_dirent_get_ftype(dp->i_mount, dep);
         error = xfs_dir_cilookup_result(args, dep->name, dep->namelen);
         xfs_trans_brelse(tp, dbp);
         xfs_trans_brelse(tp, lbp);
@@ -1555,6 +1557,7 @@ xfs_dir2_leaf_replace(
          * Put the new inode number in, log it.
          */
         dep->inumber = cpu_to_be64(args->inumber);
+        xfs_dir3_dirent_put_ftype(dp->i_mount, dep, args->filetype);
         tp = args->trans;
         xfs_dir2_data_log_entry(tp, dbp, dep);
         xfs_dir3_leaf_check(dp->i_mount, lbp);
diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c
index 49f1e9ed492c..4c3dba7ffb74 100644
--- a/fs/xfs/xfs_dir2_node.c
+++ b/fs/xfs/xfs_dir2_node.c
@@ -816,6 +816,7 @@ xfs_dir2_leafn_lookup_for_entry(
                                 xfs_trans_brelse(tp, state->extrablk.bp);
                         args->cmpresult = cmp;
                         args->inumber = be64_to_cpu(dep->inumber);
+                        args->filetype = xfs_dir3_dirent_get_ftype(mp, dep);
                         *indexp = index;
                         state->extravalid = 1;
                         state->extrablk.bp = curbp;
@@ -2007,6 +2008,7 @@ xfs_dir2_node_addname_int(
         dep->inumber = cpu_to_be64(args->inumber);
         dep->namelen = args->namelen;
         memcpy(dep->name, args->name, dep->namelen);
+        xfs_dir3_dirent_put_ftype(mp, dep, args->filetype);
         tagp = xfs_dir3_data_entry_tag_p(mp, dep);
         *tagp = cpu_to_be16((char *)dep - (char *)hdr);
         xfs_dir2_data_log_entry(tp, dbp, dep);
@@ -2227,6 +2229,7 @@ xfs_dir2_node_replace(
                  * Fill in the new inode number and log the entry.