Merge tag 'xfs-for-linus-v3.14-rc1' of git://oss.sgi.com/xfs/xfs

Pull xfs update from Ben Myers:
 "This is primarily bug fixes, many of which you already have.  New
  stuff includes a series to decouple the in-memory and on-disk log
  format, helpers in the area of inode clusters, and i_version handling.

  We decided to try to use more topic branches this release, so there
  are some merge commits in there on account of that.  I'm afraid I
  didn't do a good job of putting meaningful comments in the first
  couple of merges.  Sorry about that.  I think I have the hang of it
  now.

  For 3.14-rc1 there are fixes in the areas of remote attributes,
  discard, growfs, memory leaks in recovery, directory v2, quotas, the
  MAINTAINERS file, allocation alignment, extent list locking, and in
  xfs_bmapi_allocate.  There are cleanups in xfs_setsize_buftarg,
  removing unused macros, quotas, setattr, and freeing of inode
  clusters.  The in-memory and on-disk log format have been decoupled, a
  common helper to calculate the number of blocks in an inode cluster
  has been added, and handling of i_version has been pulled into the
  filesystems that use it.

   - cleanup in xfs_setsize_buftarg
   - removal of remaining unused flags for vop toss/flush/flushinval
   - fix for memory corruption in xfs_attrlist_by_handle
   - fix for out-of-date comment in xfs_trans_dqlockedjoin
   - fix for discard if range length is less than one block
   - fix for overrun of agfl buffer using growfs on v4 superblock
     filesystems
   - pull i_version handling out into the filesystems that use it
   - don't leak recovery items on error
   - fix for memory leak in xfs_dir2_node_removename
   - several cleanups for quotas
   - fix bad assertion in xfs_qm_vop_create_dqattach
   - cleanup for xfs_setattr_mode, and add xfs_setattr_time
   - fix quota assert in xfs_setattr_nonsize
   - fix an infinite loop when turning off group/project quota before
     user quota
   - fix for temporary buffer allocation failure in xfs_dir2_block_to_sf
     with large directory block sizes
   - fix Dave's email address in MAINTAINERS
   - cleanup calculation of freed inode cluster blocks
   - fix alignment of initial file allocations to match filesystem
     geometry
   - decouple in-memory and on-disk log format
   - introduce a common helper to calculate the number of filesystem
     blocks in an inode cluster
   - fixes for extent list locking
   - fix for off-by-one in xfs_attr3_rmt_verify
   - fix for missing destroy_work_on_stack in xfs_bmapi_allocate"

* tag 'xfs-for-linus-v3.14-rc1' of git://oss.sgi.com/xfs/xfs: (51 commits)
  xfs: Calling destroy_work_on_stack() to pair with INIT_WORK_ONSTACK()
  xfs: fix off-by-one error in xfs_attr3_rmt_verify
  xfs: assert that we hold the ilock for extent map access
  xfs: use xfs_ilock_attr_map_shared in xfs_attr_list_int
  xfs: use xfs_ilock_attr_map_shared in xfs_attr_get
  xfs: use xfs_ilock_data_map_shared in xfs_qm_dqiterate
  xfs: use xfs_ilock_data_map_shared in xfs_qm_dqtobp
  xfs: take the ilock around xfs_bmapi_read in xfs_zero_remaining_bytes
  xfs: reinstate the ilock in xfs_readdir
  xfs: add xfs_ilock_attr_map_shared
  xfs: rename xfs_ilock_map_shared
  xfs: remove xfs_iunlock_map_shared
  xfs: no need to lock the inode in xfs_find_handle
  xfs: use xfs_icluster_size_fsb in xfs_imap
  xfs: use xfs_icluster_size_fsb in xfs_ifree_cluster
  xfs: use xfs_icluster_size_fsb in xfs_ialloc_inode_init
  xfs: use xfs_icluster_size_fsb in xfs_bulkstat
  xfs: introduce a common helper xfs_icluster_size_fsb
  xfs: get rid of XFS_IALLOC_BLOCKS macros
  xfs: get rid of XFS_INODE_CLUSTER_SIZE macros
  ...

fs/attr.c

@@ -202,11 +202,6 @@ int notify_change(struct dentry * dentry, struct iattr * attr, struct inode **de
 			return -EPERM;
 	}
 
-	if ((ia_valid & ATTR_SIZE) && IS_I_VERSION(inode)) {
-		if (attr->ia_size != inode->i_size)
-			inode_inc_iversion(inode);
-	}
-
 	if ((ia_valid & ATTR_MODE)) {
 		umode_t amode = attr->ia_mode;
 		/* Flag setting protected by i_mutex */

fs/btrfs/inode.c

@@ -4354,8 +4354,12 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
 	 * these flags set.  For all other operations the VFS set these flags
 	 * explicitly if it wants a timestamp update.
 	 */
-	if (newsize != oldsize && (!(mask & (ATTR_CTIME | ATTR_MTIME))))
-		inode->i_ctime = inode->i_mtime = current_fs_time(inode->i_sb);
+	if (newsize != oldsize) {
+		inode_inc_iversion(inode);
+		if (!(mask & (ATTR_CTIME | ATTR_MTIME)))
+			inode->i_ctime = inode->i_mtime =
+				current_fs_time(inode->i_sb);
+	}
 
 	if (newsize > oldsize) {
 		truncate_pagecache(inode, newsize);

fs/ext4/inode.c

@@ -4586,6 +4586,10 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
 			if (attr->ia_size > sbi->s_bitmap_maxbytes)
 				return -EFBIG;
 		}
+
+		if (IS_I_VERSION(inode) && attr->ia_size != inode->i_size)
+			inode_inc_iversion(inode);
+
 		if (S_ISREG(inode->i_mode) &&
 		    (attr->ia_size < inode->i_size)) {
 			if (ext4_should_order_data(inode)) {

fs/xfs/xfs_aops.c

@@ -1217,7 +1217,7 @@ __xfs_get_blocks(
 		lockmode = XFS_ILOCK_EXCL;
 		xfs_ilock(ip, lockmode);
 	} else {
-		lockmode = xfs_ilock_map_shared(ip);
+		lockmode = xfs_ilock_data_map_shared(ip);
 	}
 
 	ASSERT(offset <= mp->m_super->s_maxbytes);

fs/xfs/xfs_attr.c

@@ -164,6 +164,7 @@ xfs_attr_get(
 {
 	int		error;
 	struct xfs_name	xname;
+	uint		lock_mode;
 
 	XFS_STATS_INC(xs_attr_get);
 
@@ -174,9 +175,9 @@ xfs_attr_get(
 	if (error)
 		return error;
 
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	lock_mode = xfs_ilock_attr_map_shared(ip);
 	error = xfs_attr_get_int(ip, &xname, value, valuelenp, flags);
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	xfs_iunlock(ip, lock_mode);
 	return(error);
 }
 

fs/xfs/xfs_attr_list.c

@@ -507,17 +507,17 @@ xfs_attr_list_int(
 {
 	int error;
 	xfs_inode_t *dp = context->dp;
+	uint		lock_mode;
 
 	XFS_STATS_INC(xs_attr_list);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return EIO;
 
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-
 	/*
 	 * Decide on what work routines to call based on the inode size.
 	 */
+	lock_mode = xfs_ilock_attr_map_shared(dp);
 	if (!xfs_inode_hasattr(dp)) {
 		error = 0;
 	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
@@ -527,9 +527,7 @@ xfs_attr_list_int(
 	} else {
 		error = xfs_attr_node_list(context);
 	}
-
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-
+	xfs_iunlock(dp, lock_mode);
 	return error;
 }
 

fs/xfs/xfs_bmap.c

@@ -4013,6 +4013,7 @@ xfs_bmapi_read(
 	ASSERT(*nmap >= 1);
 	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
 			   XFS_BMAPI_IGSTATE)));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
@@ -4207,6 +4208,7 @@ xfs_bmapi_delay(
 	ASSERT(*nmap >= 1);
 	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
 	ASSERT(!(flags & ~XFS_BMAPI_ENTIRE));
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS &&
@@ -4500,6 +4502,7 @@ xfs_bmapi_write(
 	ASSERT(tp != NULL);
 	ASSERT(len > 0);
 	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	if (unlikely(XFS_TEST_ERROR(
 	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
@@ -5051,6 +5054,7 @@ xfs_bunmapi(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
 
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(len > 0);
 	ASSERT(nexts >= 0);
 

fs/xfs/xfs_bmap_util.c

@@ -618,22 +618,27 @@ xfs_getbmap(
 		return XFS_ERROR(ENOMEM);
 
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
-	if (whichfork == XFS_DATA_FORK && !(iflags & BMV_IF_DELALLOC)) {
-		if (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size) {
+	if (whichfork == XFS_DATA_FORK) {
+		if (!(iflags & BMV_IF_DELALLOC) &&
+		    (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size)) {
 			error = -filemap_write_and_wait(VFS_I(ip)->i_mapping);
 			if (error)
 				goto out_unlock_iolock;
+
+			/*
+			 * Even after flushing the inode, there can still be
+			 * delalloc blocks on the inode beyond EOF due to
+			 * speculative preallocation.  These are not removed
+			 * until the release function is called or the inode
+			 * is inactivated.  Hence we cannot assert here that
+			 * ip->i_delayed_blks == 0.
+			 */
 		}
-		/*
-		 * even after flushing the inode, there can still be delalloc
-		 * blocks on the inode beyond EOF due to speculative
-		 * preallocation. These are not removed until the release
-		 * function is called or the inode is inactivated. Hence we
-		 * cannot assert here that ip->i_delayed_blks == 0.
-		 */
-	}
 
-	lock = xfs_ilock_map_shared(ip);
+		lock = xfs_ilock_data_map_shared(ip);
+	} else {
+		lock = xfs_ilock_attr_map_shared(ip);
+	}
 
 	/*
 	 * Don't let nex be bigger than the number of extents
@@ -738,7 +743,7 @@ xfs_getbmap(
  out_free_map:
 	kmem_free(map);
  out_unlock_ilock:
-	xfs_iunlock_map_shared(ip, lock);
+	xfs_iunlock(ip, lock);
  out_unlock_iolock:
 	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
@@ -1169,9 +1174,15 @@ xfs_zero_remaining_bytes(
 	xfs_buf_unlock(bp);
 
 	for (offset = startoff; offset <= endoff; offset = lastoffset + 1) {
+		uint lock_mode;
+
 		offset_fsb = XFS_B_TO_FSBT(mp, offset);
 		nimap = 1;
+
+		lock_mode = xfs_ilock_data_map_shared(ip);
 		error = xfs_bmapi_read(ip, offset_fsb, 1, &imap, &nimap, 0);
+		xfs_iunlock(ip, lock_mode);
+
 		if (error || nimap < 1)
 			break;
 		ASSERT(imap.br_blockcount >= 1);

fs/xfs/xfs_buf.c

@@ -1593,12 +1593,11 @@ xfs_free_buftarg(
 	kmem_free(btp);
 }
 
-STATIC int
-xfs_setsize_buftarg_flags(
+int
+xfs_setsize_buftarg(
 	xfs_buftarg_t		*btp,
 	unsigned int		blocksize,
-	unsigned int		sectorsize,
-	int			verbose)
+	unsigned int		sectorsize)
 {
 	btp->bt_bsize = blocksize;
 	btp->bt_sshift = ffs(sectorsize) - 1;
@@ -1619,26 +1618,17 @@ xfs_setsize_buftarg_flags(
 }
 
 /*
- *	When allocating the initial buffer target we have not yet
- *	read in the superblock, so don't know what sized sectors
- *	are being used at this early stage.  Play safe.
+ * When allocating the initial buffer target we have not yet
+ * read in the superblock, so don't know what sized sectors
+ * are being used at this early stage.  Play safe.
  */
 STATIC int
 xfs_setsize_buftarg_early(
 	xfs_buftarg_t		*btp,
 	struct block_device	*bdev)
 {
-	return xfs_setsize_buftarg_flags(btp,
-			PAGE_SIZE, bdev_logical_block_size(bdev), 0);
-}
-
-int
-xfs_setsize_buftarg(
-	xfs_buftarg_t		*btp,
-	unsigned int		blocksize,
-	unsigned int		sectorsize)
-{
-	return xfs_setsize_buftarg_flags(btp, blocksize, sectorsize, 1);
+	return xfs_setsize_buftarg(btp, PAGE_SIZE,
+				   bdev_logical_block_size(bdev));
 }
 
 xfs_buftarg_t *