Merge tag 'for-linus-v3.11-rc1' of git://oss.sgi.com/xfs/xfs

Pull xfs update from Ben Myers:
 "This includes several bugfixes, part of the work for project quotas
  and group quotas to be used together, performance improvements for
  inode creation/deletion, buffer readahead, and bulkstat,
  implementation of the inode change count, an inode create transaction,
  and the removal of a bunch of dead code.

  There are also some duplicate commits that you already have from the
  3.10-rc series.

   - part of the work to allow project quotas and group quotas to be
     used together
   - inode change count
   - inode create transaction
   - block queue plugging in buffer readahead and bulkstat
   - ordered log vector support
   - removal of dead code in and around xfs_sync_inode_grab,
     xfs_ialloc_get_rec, XFS_MOUNT_RETERR, XFS_ALLOCFREE_LOG_RES,
     XFS_DIROP_LOG_RES, xfs_chash, ctl_table, and
     xfs_growfs_data_private
   - don't keep silent if sunit/swidth can not be changed via mount
   - fix a leak of remote symlink blocks into the filesystem when xattrs
     are used on symlinks
   - fix for fiemap to return FIEMAP_EXTENT_UNKOWN flag on delay extents
   - part of a fix for xfs_fsr
   - disable speculative preallocation with small files
   - performance improvements for inode creates and deletes"

* tag 'for-linus-v3.11-rc1' of git://oss.sgi.com/xfs/xfs: (61 commits)
  xfs: Remove incore use of XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD
  xfs: Change xfs_dquot_acct to be a 2-dimensional array
  xfs: Code cleanup and removal of some typedef usage
  xfs: Replace macro XFS_DQ_TO_QIP with a function
  xfs: Replace macro XFS_DQUOT_TREE with a function
  xfs: Define a new function xfs_is_quota_inode()
  xfs: implement inode change count
  xfs: Use inode create transaction
  xfs: Inode create item recovery
  xfs: Inode create transaction reservations
  xfs: Inode create log items
  xfs: Introduce an ordered buffer item
  xfs: Introduce ordered log vector support
  xfs: xfs_ifree doesn't need to modify the inode buffer
  xfs: don't do IO when creating an new inode
  xfs: don't use speculative prealloc for small files
  xfs: plug directory buffer readahead
  xfs: add pluging for bulkstat readahead
  xfs: Remove dead function prototype xfs_sync_inode_grab()
  xfs: Remove the left function variable from xfs_ialloc_get_rec()
  ...

1 files changed, 32 insertions, 36 deletions

diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 7f7be5f98f52..9ecfe1e559fc 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -1028,6 +1028,11 @@ xfs_dinode_calc_crc(
 
 /*
  * Read the disk inode attributes into the in-core inode structure.
+ *
+ * If we are initialising a new inode and we are not utilising the
+ * XFS_MOUNT_IKEEP inode cluster mode, we can simple build the new inode core
+ * with a random generation number. If we are keeping inodes around, we need to
+ * read the inode cluster to get the existing generation number off disk.
  */
 int
 xfs_iread(
@@ -1047,6 +1052,22 @@ xfs_iread(
 	if (error)
 		return error;
 
+	/* shortcut IO on inode allocation if possible */
+	if ((iget_flags & XFS_IGET_CREATE) &&
+	    !(mp->m_flags & XFS_MOUNT_IKEEP)) {
+		/* initialise the on-disk inode core */
+		memset(&ip->i_d, 0, sizeof(ip->i_d));
+		ip->i_d.di_magic = XFS_DINODE_MAGIC;
+		ip->i_d.di_gen = prandom_u32();
+		if (xfs_sb_version_hascrc(&mp->m_sb)) {
+			ip->i_d.di_version = 3;
+			ip->i_d.di_ino = ip->i_ino;
+			uuid_copy(&ip->i_d.di_uuid, &mp->m_sb.sb_uuid);
+		} else
+			ip->i_d.di_version = 2;
+		return 0;
+	}
+
 	/*
 	 * Get pointers to the on-disk inode and the buffer containing it.
 	 */
@@ -1133,17 +1154,16 @@ xfs_iread(
 	xfs_buf_set_ref(bp, XFS_INO_REF);
 
 	/*
-	 * Use xfs_trans_brelse() to release the buffer containing the
-	 * on-disk inode, because it was acquired with xfs_trans_read_buf()
-	 * in xfs_imap_to_bp() above.  If tp is NULL, this is just a normal
+	 * Use xfs_trans_brelse() to release the buffer containing the on-disk
+	 * inode, because it was acquired with xfs_trans_read_buf() in
+	 * xfs_imap_to_bp() above.  If tp is NULL, this is just a normal
 	 * brelse().  If we're within a transaction, then xfs_trans_brelse()
 	 * will only release the buffer if it is not dirty within the
 	 * transaction.  It will be OK to release the buffer in this case,
-	 * because inodes on disk are never destroyed and we will be
-	 * locking the new in-core inode before putting it in the hash
-	 * table where other processes can find it.  Thus we don't have
-	 * to worry about the inode being changed just because we released
-	 * the buffer.
+	 * because inodes on disk are never destroyed and we will be locking the
+	 * new in-core inode before putting it in the cache where other
+	 * processes can find it.  Thus we don't have to worry about the inode
+	 * being changed just because we released the buffer.
 	 */
  out_brelse:
 	xfs_trans_brelse(tp, bp);
@@ -2028,8 +2048,6 @@ xfs_ifree(
 	int			error;
 	int			delete;
 	xfs_ino_t		first_ino;
-	xfs_dinode_t    	*dip;
-	xfs_buf_t       	*ibp;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(ip->i_d.di_nlink == 0);
@@ -2042,14 +2060,13 @@ xfs_ifree(
 	 * Pull the on-disk inode from the AGI unlinked list.
 	 */
 	error = xfs_iunlink_remove(tp, ip);
-	if (error != 0) {
+	if (error)
 		return error;
-	}
 
 	error = xfs_difree(tp, ip->i_ino, flist, &delete, &first_ino);
-	if (error != 0) {
+	if (error)
 		return error;
-	}
+
 	ip->i_d.di_mode = 0;		/* mark incore inode as free */
 	ip->i_d.di_flags = 0;
 	ip->i_d.di_dmevmask = 0;
@@ -2061,31 +2078,10 @@ xfs_ifree(
 	 * by reincarnations of this inode.
 	 */
 	ip->i_d.di_gen++;
-
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
-	error = xfs_imap_to_bp(ip->i_mount, tp, &ip->i_imap, &dip, &ibp,
-			       0, 0);
-	if (error)
-		return error;
-
-        /*
-	* Clear the on-disk di_mode. This is to prevent xfs_bulkstat
-	* from picking up this inode when it is reclaimed (its incore state
-	* initialzed but not flushed to disk yet). The in-core di_mode is
-	* already cleared  and a corresponding transaction logged.
-	* The hack here just synchronizes the in-core to on-disk
-	* di_mode value in advance before the actual inode sync to disk.
-	* This is OK because the inode is already unlinked and would never
-	* change its di_mode again for this inode generation.
-	* This is a temporary hack that would require a proper fix
-	* in the future.
-	*/
-	dip->di_mode = 0;
-
-	if (delete) {
+	if (delete)
 		error = xfs_ifree_cluster(ip, tp, first_ino);
-	}
 
 	return error;
 }