initial merge with 3.2.72

69 files changed, 32434 insertions, 0 deletions

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
new file mode 100644
index 0000000..22c64ff
--- /dev/null
+++ b/fs/btrfs/backref.c
@@ -0,0 +1,776 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "backref.h"
+
+struct __data_ref {
+	struct list_head list;
+	u64 inum;
+	u64 root;
+	u64 extent_data_item_offset;
+};
+
+struct __shared_ref {
+	struct list_head list;
+	u64 disk_byte;
+};
+
+static int __inode_info(u64 inum, u64 ioff, u8 key_type,
+			struct btrfs_root *fs_root, struct btrfs_path *path,
+			struct btrfs_key *found_key)
+{
+	int ret;
+	struct btrfs_key key;
+	struct extent_buffer *eb;
+
+	key.type = key_type;
+	key.objectid = inum;
+	key.offset = ioff;
+
+	ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
+	if (ret < 0)
+		return ret;
+
+	eb = path->nodes[0];
+	if (ret && path->slots[0] >= btrfs_header_nritems(eb)) {
+		ret = btrfs_next_leaf(fs_root, path);
+		if (ret)
+			return ret;
+		eb = path->nodes[0];
+	}
+
+	btrfs_item_key_to_cpu(eb, found_key, path->slots[0]);
+	if (found_key->type != key.type || found_key->objectid != key.objectid)
+		return 1;
+
+	return 0;
+}
+
+/*
+ * this makes the path point to (inum INODE_ITEM ioff)
+ */
+int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+			struct btrfs_path *path)
+{
+	struct btrfs_key key;
+	return __inode_info(inum, ioff, BTRFS_INODE_ITEM_KEY, fs_root, path,
+				&key);
+}
+
+static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+				struct btrfs_path *path,
+				struct btrfs_key *found_key)
+{
+	return __inode_info(inum, ioff, BTRFS_INODE_REF_KEY, fs_root, path,
+				found_key);
+}
+
+/*
+ * this iterates to turn a btrfs_inode_ref into a full filesystem path. elements
+ * of the path are separated by '/' and the path is guaranteed to be
+ * 0-terminated. the path is only given within the current file system.
+ * Therefore, it never starts with a '/'. the caller is responsible to provide
+ * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
+ * the start point of the resulting string is returned. this pointer is within
+ * dest, normally.
+ * in case the path buffer would overflow, the pointer is decremented further
+ * as if output was written to the buffer, though no more output is actually
+ * generated. that way, the caller can determine how much space would be
+ * required for the path to fit into the buffer. in that case, the returned
+ * value will be smaller than dest. callers must check this!
+ */
+static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
+				struct btrfs_inode_ref *iref,
+				struct extent_buffer *eb_in, u64 parent,
+				char *dest, u32 size)
+{
+	u32 len;
+	int slot;
+	u64 next_inum;
+	int ret;
+	s64 bytes_left = size - 1;
+	struct extent_buffer *eb = eb_in;
+	struct btrfs_key found_key;
+
+	if (bytes_left >= 0)
+		dest[bytes_left] = '\0';
+
+	while (1) {
+		len = btrfs_inode_ref_name_len(eb, iref);
+		bytes_left -= len;
+		if (bytes_left >= 0)
+			read_extent_buffer(eb, dest + bytes_left,
+						(unsigned long)(iref + 1), len);
+		if (eb != eb_in)
+			free_extent_buffer(eb);
+		ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
+		if (ret)
+			break;
+		next_inum = found_key.offset;
+
+		/* regular exit ahead */
+		if (parent == next_inum)
+			break;
+
+		slot = path->slots[0];
+		eb = path->nodes[0];
+		/* make sure we can use eb after releasing the path */
+		if (eb != eb_in)
+			atomic_inc(&eb->refs);
+		btrfs_release_path(path);
+
+		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
+		parent = next_inum;
+		--bytes_left;
+		if (bytes_left >= 0)
+			dest[bytes_left] = '/';
+	}
+
+	btrfs_release_path(path);
+
+	if (ret)
+		return ERR_PTR(ret);
+
+	return dest + bytes_left;
+}
+
+/*
+ * this makes the path point to (logical EXTENT_ITEM *)
+ * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
+ * tree blocks and <0 on error.
+ */
+int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
+			struct btrfs_path *path, struct btrfs_key *found_key)
+{
+	int ret;
+	u64 flags;
+	u32 item_size;
+	struct extent_buffer *eb;
+	struct btrfs_extent_item *ei;
+	struct btrfs_key key;
+
+	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.objectid = logical;
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
+	if (ret < 0)
+		return ret;
+	ret = btrfs_previous_item(fs_info->extent_root, path,
+					0, BTRFS_EXTENT_ITEM_KEY);
+	if (ret < 0)
+		return ret;
+
+	btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
+	if (found_key->type != BTRFS_EXTENT_ITEM_KEY ||
+	    found_key->objectid > logical ||
+	    found_key->objectid + found_key->offset <= logical)
+		return -ENOENT;
+
+	eb = path->nodes[0];
+	item_size = btrfs_item_size_nr(eb, path->slots[0]);
+	BUG_ON(item_size < sizeof(*ei));
+
+	ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+	flags = btrfs_extent_flags(eb, ei);
+
+	if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+		return BTRFS_EXTENT_FLAG_TREE_BLOCK;
+	if (flags & BTRFS_EXTENT_FLAG_DATA)
+		return BTRFS_EXTENT_FLAG_DATA;
+
+	return -EIO;
+}
+
+/*
+ * helper function to iterate extent inline refs. ptr must point to a 0 value
+ * for the first call and may be modified. it is used to track state.
+ * if more refs exist, 0 is returned and the next call to
+ * __get_extent_inline_ref must pass the modified ptr parameter to get the
+ * next ref. after the last ref was processed, 1 is returned.
+ * returns <0 on error
+ */
+static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
+				struct btrfs_extent_item *ei, u32 item_size,
+				struct btrfs_extent_inline_ref **out_eiref,
+				int *out_type)
+{
+	unsigned long end;
+	u64 flags;
+	struct btrfs_tree_block_info *info;
+
+	if (!*ptr) {
+		/* first call */
+		flags = btrfs_extent_flags(eb, ei);
+		if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+			info = (struct btrfs_tree_block_info *)(ei + 1);
+			*out_eiref =
+				(struct btrfs_extent_inline_ref *)(info + 1);
+		} else {
+			*out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
+		}
+		*ptr = (unsigned long)*out_eiref;
+		if ((void *)*ptr >= (void *)ei + item_size)
+			return -ENOENT;
+	}
+
+	end = (unsigned long)ei + item_size;
+	*out_eiref = (struct btrfs_extent_inline_ref *)*ptr;
+	*out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
+
+	*ptr += btrfs_extent_inline_ref_size(*out_type);
+	WARN_ON(*ptr > end);
+	if (*ptr == end)
+		return 1; /* last */
+
+	return 0;
+}
+
+/*
+ * reads the tree block backref for an extent. tree level and root are returned
+ * through out_level and out_root. ptr must point to a 0 value for the first
+ * call and may be modified (see __get_extent_inline_ref comment).
+ * returns 0 if data was provided, 1 if there was no more data to provide or
+ * <0 on error.
+ */
+int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
+				struct btrfs_extent_item *ei, u32 item_size,
+				u64 *out_root, u8 *out_level)
+{
+	int ret;
+	int type;
+	struct btrfs_tree_block_info *info;
+	struct btrfs_extent_inline_ref *eiref;
+
+	if (*ptr == (unsigned long)-1)
+		return 1;
+
+	while (1) {
+		ret = __get_extent_inline_ref(ptr, eb, ei, item_size,
+						&eiref, &type);
+		if (ret < 0)
+			return ret;
+
+		if (type == BTRFS_TREE_BLOCK_REF_KEY ||
+		    type == BTRFS_SHARED_BLOCK_REF_KEY)
+			break;
+
+		if (ret == 1)
+			return 1;
+	}
+
+	/* we can treat both ref types equally here */
+	info = (struct btrfs_tree_block_info *)(ei + 1);
+	*out_root = btrfs_extent_inline_ref_offset(eb, eiref);
+	*out_level = btrfs_tree_block_level(eb, info);
+
+	if (ret == 1)
+		*ptr = (unsigned long)-1;
+
+	return 0;
+}
+
+static int __data_list_add(struct list_head *head, u64 inum,
+				u64 extent_data_item_offset, u64 root)
+{
+	struct __data_ref *ref;
+
+	ref = kmalloc(sizeof(*ref), GFP_NOFS);
+	if (!ref)
+		return -ENOMEM;
+
+	ref->inum = inum;
+	ref->extent_data_item_offset = extent_data_item_offset;
+	ref->root = root;
+	list_add_tail(&ref->list, head);
+
+	return 0;
+}
+
+static int __data_list_add_eb(struct list_head *head, struct extent_buffer *eb,
+				struct btrfs_extent_data_ref *dref)
+{
+	return __data_list_add(head, btrfs_extent_data_ref_objectid(eb, dref),
+				btrfs_extent_data_ref_offset(eb, dref),
+				btrfs_extent_data_ref_root(eb, dref));
+}
+
+static int __shared_list_add(struct list_head *head, u64 disk_byte)
+{
+	struct __shared_ref *ref;
+
+	ref = kmalloc(sizeof(*ref), GFP_NOFS);
+	if (!ref)
+		return -ENOMEM;
+
+	ref->disk_byte = disk_byte;
+	list_add_tail(&ref->list, head);
+
+	return 0;
+}
+
+static int __iter_shared_inline_ref_inodes(struct btrfs_fs_info *fs_info,
+					   u64 logical, u64 inum,
+					   u64 extent_data_item_offset,
+					   u64 extent_offset,
+					   struct btrfs_path *path,
+					   struct list_head *data_refs,
+					   iterate_extent_inodes_t *iterate,
+					   void *ctx)
+{
+	u64 ref_root;
+	u32 item_size;
+	struct btrfs_key key;
+	struct extent_buffer *eb;
+	struct btrfs_extent_item *ei;
+	struct btrfs_extent_inline_ref *eiref;
+	struct __data_ref *ref;
+	int ret;
+	int type;
+	int last;
+	unsigned long ptr = 0;
+
+	WARN_ON(!list_empty(data_refs));
+	ret = extent_from_logical(fs_info, logical, path, &key);
+	if (ret & BTRFS_EXTENT_FLAG_DATA)
+		ret = -EIO;
+	if (ret < 0)
+		goto out;
+
+	eb = path->nodes[0];
+	ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+	item_size = btrfs_item_size_nr(eb, path->slots[0]);
+
+	ret = 0;
+	ref_root = 0;
+	/*
+	 * as done in iterate_extent_inodes, we first build a list of refs to
+	 * iterate, then free the path and then iterate them to avoid deadlocks.
+	 */
+	do {
+		last = __get_extent_inline_ref(&ptr, eb, ei, item_size,
+						&eiref, &type);
+		if (last < 0) {
+			ret = last;
+			goto out;
+		}
+		if (type == BTRFS_TREE_BLOCK_REF_KEY ||
+		    type == BTRFS_SHARED_BLOCK_REF_KEY) {
+			ref_root = btrfs_extent_inline_ref_offset(eb, eiref);
+			ret = __data_list_add(data_refs, inum,
+						extent_data_item_offset,
+						ref_root);
+		}
+	} while (!ret && !last);
+
+	btrfs_release_path(path);
+
+	if (ref_root == 0) {
+		printk(KERN_ERR "btrfs: failed to find tree block ref "
+			"for shared data backref %llu\n", logical);
+		WARN_ON(1);
+		ret = -EIO;
+	}
+
+out:
+	while (!list_empty(data_refs)) {
+		ref = list_first_entry(data_refs, struct __data_ref, list);
+		list_del(&ref->list);
+		if (!ret)
+			ret = iterate(ref->inum, extent_offset +
+					ref->extent_data_item_offset,
+					ref->root, ctx);
+		kfree(ref);
+	}
+
+	return ret;
+}
+
+static int __iter_shared_inline_ref(struct btrfs_fs_info *fs_info,
+				    u64 logical, u64 orig_extent_item_objectid,
+				    u64 extent_offset, struct btrfs_path *path,
+				    struct list_head *data_refs,
+				    iterate_extent_inodes_t *iterate,
+				    void *ctx)
+{
+	u64 disk_byte;
+	struct btrfs_key key;
+	struct btrfs_file_extent_item *fi;
+	struct extent_buffer *eb;
+	int slot;
+	int nritems;
+	int ret;
+	int found = 0;
+
+	eb = read_tree_block(fs_info->tree_root, logical,
+				fs_info->tree_root->leafsize, 0);
+	if (!eb)
+		return -EIO;
+
+	/*
+	 * from the shared data ref, we only have the leaf but we need
+	 * the key. thus, we must look into all items and see that we
+	 * find one (some) with a reference to our extent item.
+	 */
+	nritems = btrfs_header_nritems(eb);
+	for (slot = 0; slot < nritems; ++slot) {
+		btrfs_item_key_to_cpu(eb, &key, slot);
+		if (key.type != BTRFS_EXTENT_DATA_KEY)
+			continue;
+		fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
+		if (!fi) {
+			free_extent_buffer(eb);
+			return -EIO;
+		}
+		disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
+		if (disk_byte != orig_extent_item_objectid) {
+			if (found)
+				break;
+			else
+				continue;
+		}
+		++found;
+		ret = __iter_shared_inline_ref_inodes(fs_info, logical,
+							key.objectid,
+							key.offset,
+							extent_offset, path,
+							data_refs,
+							iterate, ctx);
+		if (ret)
+			break;
+	}
+
+	if (!found) {
+		printk(KERN_ERR "btrfs: failed to follow shared data backref "
+			"to parent %llu\n", logical);
+		WARN_ON(1);
+		ret = -EIO;
+	}
+
+	free_extent_buffer(eb);
+	return ret;
+}
+
+/*
+ * calls iterate() for every inode that references the extent identified by
+ * the given parameters. will use the path given as a parameter and return it
+ * released.
+ * when the iterator function returns a non-zero value, iteration stops.
+ */
+int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
+				struct btrfs_path *path,
+				u64 extent_item_objectid,
+				u64 extent_offset,
+				iterate_extent_inodes_t *iterate, void *ctx)
+{
+	unsigned long ptr = 0;
+	int last;
+	int ret;
+	int type;
+	u64 logical;
+	u32 item_size;
+	struct btrfs_extent_inline_ref *eiref;
+	struct btrfs_extent_data_ref *dref;
+	struct extent_buffer *eb;
+	struct btrfs_extent_item *ei;
+	struct btrfs_key key;
+	struct list_head data_refs = LIST_HEAD_INIT(data_refs);
+	struct list_head shared_refs = LIST_HEAD_INIT(shared_refs);
+	struct __data_ref *ref_d;
+	struct __shared_ref *ref_s;
+
+	eb = path->nodes[0];
+	ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
+	item_size = btrfs_item_size_nr(eb, path->slots[0]);
+
+	/* first we iterate the inline refs, ... */
+	do {
+		last = __get_extent_inline_ref(&ptr, eb, ei, item_size,
+						&eiref, &type);
+		if (last == -ENOENT) {
+			ret = 0;
+			break;
+		}
+		if (last < 0) {
+			ret = last;
+			break;
+		}
+
+		if (type == BTRFS_EXTENT_DATA_REF_KEY) {
+			dref = (struct btrfs_extent_data_ref *)(&eiref->offset);
+			ret = __data_list_add_eb(&data_refs, eb, dref);
+		} else if (type == BTRFS_SHARED_DATA_REF_KEY) {
+			logical = btrfs_extent_inline_ref_offset(eb, eiref);
+			ret = __shared_list_add(&shared_refs, logical);
+		}
+	} while (!ret && !last);
+
+	/* ... then we proceed to in-tree references and ... */
+	while (!ret) {
+		++path->slots[0];
+		if (path->slots[0] > btrfs_header_nritems(eb)) {
+			ret = btrfs_next_leaf(fs_info->extent_root, path);
+			if (ret) {
+				if (ret == 1)
+					ret = 0; /* we're done */
+				break;
+			}
+			eb = path->nodes[0];
+		}
+		btrfs_item_key_to_cpu(eb, &key, path->slots[0]);
+		if (key.objectid != extent_item_objectid)
+			break;
+		if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
+			dref = btrfs_item_ptr(eb, path->slots[0],
+						struct btrfs_extent_data_ref);
+			ret = __data_list_add_eb(&data_refs, eb, dref);
+		} else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
+			ret = __shared_list_add(&shared_refs, key.offset);
+		}
+	}
+
+	btrfs_release_path(path);
+
+	/*
+	 * ... only at the very end we can process the refs we found. this is
+	 * because the iterator function we call is allowed to make tree lookups
+	 * and we have to avoid deadlocks. additionally, we need more tree
+	 * lookups ourselves for shared data refs.
+	 */
+	while (!list_empty(&data_refs)) {
+		ref_d = list_first_entry(&data_refs, struct __data_ref, list);
+		list_del(&ref_d->list);
+		if (!ret)
+			ret = iterate(ref_d->inum, extent_offset +
+					ref_d->extent_data_item_offset,
+					ref_d->root, ctx);
+		kfree(ref_d);
+	}
+
+	while (!list_empty(&shared_refs)) {
+		ref_s = list_first_entry(&shared_refs, struct __shared_ref,
+					list);
+		list_del(&ref_s->list);
+		if (!ret)
+			ret = __iter_shared_inline_ref(fs_info,
+							ref_s->disk_byte,
+							extent_item_objectid,
+							extent_offset, path,
+							&data_refs,
+							iterate, ctx);
+		kfree(ref_s);
+	}
+
+	return ret;
+}
+
+int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
+				struct btrfs_path *path,
+				iterate_extent_inodes_t *iterate, void *ctx)
+{
+	int ret;
+	u64 offset;
+	struct btrfs_key found_key;
+
+	ret = extent_from_logical(fs_info, logical, path,
+					&found_key);
+	if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+		ret = -EINVAL;
+	if (ret < 0)
+		return ret;
+
+	offset = logical - found_key.objectid;
+	ret = iterate_extent_inodes(fs_info, path, found_key.objectid,
+					offset, iterate, ctx);
+
+	return ret;
+}
+
+static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
+				struct btrfs_path *path,
+				iterate_irefs_t *iterate, void *ctx)
+{
+	int ret;
+	int slot;
+	u32 cur;
+	u32 len;
+	u32 name_len;
+	u64 parent = 0;
+	int found = 0;
+	struct extent_buffer *eb;
+	struct btrfs_item *item;
+	struct btrfs_inode_ref *iref;
+	struct btrfs_key found_key;
+
+	while (1) {
+		ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
+					&found_key);
+		if (ret < 0)
+			break;
+		if (ret) {
+			ret = found ? 0 : -ENOENT;
+			break;
+		}
+		++found;
+
+		parent = found_key.offset;
+		slot = path->slots[0];
+		eb = path->nodes[0];
+		/* make sure we can use eb after releasing the path */
+		atomic_inc(&eb->refs);
+		btrfs_release_path(path);
+
+		item = btrfs_item_nr(eb, slot);
+		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
+
+		for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
+			name_len = btrfs_inode_ref_name_len(eb, iref);
+			/* path must be released before calling iterate()! */
+			ret = iterate(parent, iref, eb, ctx);
+			if (ret) {
+				free_extent_buffer(eb);
+				break;
+			}
+			len = sizeof(*iref) + name_len;
+			iref = (struct btrfs_inode_ref *)((char *)iref + len);
+		}
+		free_extent_buffer(eb);
+	}
+
+	btrfs_release_path(path);
+
+	return ret;
+}
+
+/*
+ * returns 0 if the path could be dumped (probably truncated)
+ * returns <0 in case of an error
+ */
+static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref,
+				struct extent_buffer *eb, void *ctx)
+{
+	struct inode_fs_paths *ipath = ctx;
+	char *fspath;
+	char *fspath_min;
+	int i = ipath->fspath->elem_cnt;
+	const int s_ptr = sizeof(char *);
+	u32 bytes_left;
+
+	bytes_left = ipath->fspath->bytes_left > s_ptr ?
+					ipath->fspath->bytes_left - s_ptr : 0;
+
+	fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
+	fspath = iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb,
+				inum, fspath_min, bytes_left);
+	if (IS_ERR(fspath))
+		return PTR_ERR(fspath);
+
+	if (fspath > fspath_min) {
+		ipath->fspath->val[i] = (u64)(unsigned long)fspath;
+		++ipath->fspath->elem_cnt;
+		ipath->fspath->bytes_left = fspath - fspath_min;
+	} else {
+		++ipath->fspath->elem_missed;
+		ipath->fspath->bytes_missing += fspath_min - fspath;
+		ipath->fspath->bytes_left = 0;
+	}
+
+	return 0;
+}
+
+/*
+ * this dumps all file system paths to the inode into the ipath struct, provided
+ * is has been created large enough. each path is zero-terminated and accessed
+ * from ipath->fspath->val[i].
+ * when it returns, there are ipath->fspath->elem_cnt number of paths available
+ * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
+ * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
+ * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
+ * have been needed to return all paths.
+ */
+int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
+{
+	return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
+				inode_to_path, ipath);
+}
+
+/*
+ * allocates space to return multiple file system paths for an inode.
+ * total_bytes to allocate are passed, note that space usable for actual path
+ * information will be total_bytes - sizeof(struct inode_fs_paths).
+ * the returned pointer must be freed with free_ipath() in the end.
+ */
+struct btrfs_data_container *init_data_container(u32 total_bytes)
+{
+	struct btrfs_data_container *data;
+	size_t alloc_bytes;
+
+	alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
+	data = kmalloc(alloc_bytes, GFP_NOFS);
+	if (!data)
+		return ERR_PTR(-ENOMEM);
+
+	if (total_bytes >= sizeof(*data)) {
+		data->bytes_left = total_bytes - sizeof(*data);
+		data->bytes_missing = 0;
+	} else {
+		data->bytes_missing = sizeof(*data) - total_bytes;
+		data->bytes_left = 0;
+	}
+
+	data->elem_cnt = 0;
+	data->elem_missed = 0;
+
+	return data;
+}
+
+/*
+ * allocates space to return multiple file system paths for an inode.
+ * total_bytes to allocate are passed, note that space usable for actual path
+ * information will be total_bytes - sizeof(struct inode_fs_paths).
+ * the returned pointer must be freed with free_ipath() in the end.
+ */
+struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
+					struct btrfs_path *path)
+{
+	struct inode_fs_paths *ifp;
+	struct btrfs_data_container *fspath;
+
+	fspath = init_data_container(total_bytes);
+	if (IS_ERR(fspath))
+		return (void *)fspath;
+
+	ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
+	if (!ifp) {
+		kfree(fspath);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	ifp->btrfs_path = path;
+	ifp->fspath = fspath;
+	ifp->fs_root = fs_root;
+
+	return ifp;
+}
+
+void free_ipath(struct inode_fs_paths *ipath)
+{
+	kfree(ipath);
+}
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
new file mode 100644
index 0000000..9261883
--- /dev/null
+++ b/fs/btrfs/backref.h
@@ -0,0 +1,62 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_BACKREF__
+#define __BTRFS_BACKREF__
+
+#include "ioctl.h"
+
+struct inode_fs_paths {
+	struct btrfs_path		*btrfs_path;
+	struct btrfs_root		*fs_root;
+	struct btrfs_data_container	*fspath;
+};
+
+typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root,
+		void *ctx);
+typedef int (iterate_irefs_t)(u64 parent, struct btrfs_inode_ref *iref,
+				struct extent_buffer *eb, void *ctx);
+
+int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
+			struct btrfs_path *path);
+
+int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
+			struct btrfs_path *path, struct btrfs_key *found_key);
+
+int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
+				struct btrfs_extent_item *ei, u32 item_size,
+				u64 *out_root, u8 *out_level);
+
+int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
+				struct btrfs_path *path,
+				u64 extent_item_objectid,
+				u64 extent_offset,
+				iterate_extent_inodes_t *iterate, void *ctx);
+
+int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
+				struct btrfs_path *path,
+				iterate_extent_inodes_t *iterate, void *ctx);
+
+int paths_from_inode(u64 inum, struct inode_fs_paths *ipath);
+
+struct btrfs_data_container *init_data_container(u32 total_bytes);
+struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
+					struct btrfs_path *path);
+void free_ipath(struct inode_fs_paths *ipath);
+
+#endif
diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c
new file mode 100644
index 0000000..2373b39
--- /dev/null
+++ b/fs/btrfs/reada.c
@@ -0,0 +1,951 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/writeback.h>
+#include <linux/blkdev.h>
+#include <linux/rbtree.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include "ctree.h"
+#include "volumes.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+#undef DEBUG
+
+/*
+ * This is the implementation for the generic read ahead framework.
+ *
+ * To trigger a readahead, btrfs_reada_add must be called. It will start
+ * a read ahead for the given range [start, end) on tree root. The returned
+ * handle can either be used to wait on the readahead to finish
+ * (btrfs_reada_wait), or to send it to the background (btrfs_reada_detach).
+ *
+ * The read ahead works as follows:
+ * On btrfs_reada_add, the root of the tree is inserted into a radix_tree.
+ * reada_start_machine will then search for extents to prefetch and trigger
+ * some reads. When a read finishes for a node, all contained node/leaf
+ * pointers that lie in the given range will also be enqueued. The reads will
+ * be triggered in sequential order, thus giving a big win over a naive
+ * enumeration. It will also make use of multi-device layouts. Each disk
+ * will have its on read pointer and all disks will by utilized in parallel.
+ * Also will no two disks read both sides of a mirror simultaneously, as this
+ * would waste seeking capacity. Instead both disks will read different parts
+ * of the filesystem.
+ * Any number of readaheads can be started in parallel. The read order will be
+ * determined globally, i.e. 2 parallel readaheads will normally finish faster
+ * than the 2 started one after another.
+ */
+
+#define MAX_MIRRORS 2
+#define MAX_IN_FLIGHT 6
+
+struct reada_extctl {
+	struct list_head	list;
+	struct reada_control	*rc;
+	u64			generation;
+};
+
+struct reada_extent {
+	u64			logical;
+	struct btrfs_key	top;
+	u32			blocksize;
+	int			err;
+	struct list_head	extctl;
+	struct kref		refcnt;
+	spinlock_t		lock;
+	struct reada_zone	*zones[MAX_MIRRORS];
+	int			nzones;
+	struct btrfs_device	*scheduled_for;
+};
+
+struct reada_zone {
+	u64			start;
+	u64			end;
+	u64			elems;
+	struct list_head	list;
+	spinlock_t		lock;
+	int			locked;
+	struct btrfs_device	*device;
+	struct btrfs_device	*devs[MAX_MIRRORS]; /* full list, incl self */
+	int			ndevs;
+	struct kref		refcnt;
+};
+
+struct reada_machine_work {
+	struct btrfs_work	work;
+	struct btrfs_fs_info	*fs_info;
+};
+
+static void reada_extent_put(struct btrfs_fs_info *, struct reada_extent *);
+static void reada_control_release(struct kref *kref);
+static void reada_zone_release(struct kref *kref);
+static void reada_start_machine(struct btrfs_fs_info *fs_info);
+static void __reada_start_machine(struct btrfs_fs_info *fs_info);
+
+static int reada_add_block(struct reada_control *rc, u64 logical,
+			   struct btrfs_key *top, int level, u64 generation);
+
+/* recurses */
+/* in case of err, eb might be NULL */
+static int __readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
+			    u64 start, int err)
+{
+	int level = 0;
+	int nritems;
+	int i;
+	u64 bytenr;
+	u64 generation;
+	struct reada_extent *re;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct list_head list;
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	struct btrfs_device *for_dev;
+
+	if (eb)
+		level = btrfs_header_level(eb);
+
+	/* find extent */
+	spin_lock(&fs_info->reada_lock);
+	re = radix_tree_lookup(&fs_info->reada_tree, index);
+	if (re)
+		kref_get(&re->refcnt);
+	spin_unlock(&fs_info->reada_lock);
+
+	if (!re)
+		return -1;
+
+	spin_lock(&re->lock);
+	/*
+	 * just take the full list from the extent. afterwards we
+	 * don't need the lock anymore
+	 */
+	list_replace_init(&re->extctl, &list);
+	for_dev = re->scheduled_for;
+	re->scheduled_for = NULL;
+	spin_unlock(&re->lock);
+
+	if (err == 0) {
+		nritems = level ? btrfs_header_nritems(eb) : 0;
+		generation = btrfs_header_generation(eb);
+		/*
+		 * FIXME: currently we just set nritems to 0 if this is a leaf,
+		 * effectively ignoring the content. In a next step we could
+		 * trigger more readahead depending from the content, e.g.
+		 * fetch the checksums for the extents in the leaf.
+		 */
+	} else {
+		/*
+		 * this is the error case, the extent buffer has not been
+		 * read correctly. We won't access anything from it and
+		 * just cleanup our data structures. Effectively this will
+		 * cut the branch below this node from read ahead.
+		 */
+		nritems = 0;
+		generation = 0;
+	}
+
+	for (i = 0; i < nritems; i++) {
+		struct reada_extctl *rec;
+		u64 n_gen;
+		struct btrfs_key key;
+		struct btrfs_key next_key;
+
+		btrfs_node_key_to_cpu(eb, &key, i);
+		if (i + 1 < nritems)
+			btrfs_node_key_to_cpu(eb, &next_key, i + 1);
+		else
+			next_key = re->top;
+		bytenr = btrfs_node_blockptr(eb, i);
+		n_gen = btrfs_node_ptr_generation(eb, i);
+
+		list_for_each_entry(rec, &list, list) {
+			struct reada_control *rc = rec->rc;
+
+			/*
+			 * if the generation doesn't match, just ignore this
+			 * extctl. This will probably cut off a branch from
+			 * prefetch. Alternatively one could start a new (sub-)
+			 * prefetch for this branch, starting again from root.
+			 * FIXME: move the generation check out of this loop
+			 */
+#ifdef DEBUG
+			if (rec->generation != generation) {
+				printk(KERN_DEBUG "generation mismatch for "
+						"(%llu,%d,%llu) %llu != %llu\n",
+				       key.objectid, key.type, key.offset,
+				       rec->generation, generation);
+			}
+#endif
+			if (rec->generation == generation &&
+			    btrfs_comp_cpu_keys(&key, &rc->key_end) < 0 &&
+			    btrfs_comp_cpu_keys(&next_key, &rc->key_start) > 0)
+				reada_add_block(rc, bytenr, &next_key,
+						level - 1, n_gen);
+		}
+	}
+	/*
+	 * free extctl records
+	 */
+	while (!list_empty(&list)) {
+		struct reada_control *rc;
+		struct reada_extctl *rec;
+
+		rec = list_first_entry(&list, struct reada_extctl, list);
+		list_del(&rec->list);
+		rc = rec->rc;
+		kfree(rec);
+
+		kref_get(&rc->refcnt);
+		if (atomic_dec_and_test(&rc->elems)) {
+			kref_put(&rc->refcnt, reada_control_release);
+			wake_up(&rc->wait);
+		}
+		kref_put(&rc->refcnt, reada_control_release);
+
+		reada_extent_put(fs_info, re);	/* one ref for each entry */
+	}
+	reada_extent_put(fs_info, re);	/* our ref */
+	if (for_dev)
+		atomic_dec(&for_dev->reada_in_flight);
+
+	return 0;
+}
+
+/*
+ * start is passed separately in case eb in NULL, which may be the case with
+ * failed I/O
+ */
+int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
+			 u64 start, int err)
+{
+	int ret;
+
+	ret = __readahead_hook(root, eb, start, err);
+
+	reada_start_machine(root->fs_info);
+
+	return ret;
+}
+
+static struct reada_zone *reada_find_zone(struct btrfs_fs_info *fs_info,
+					  struct btrfs_device *dev, u64 logical,
+					  struct btrfs_bio *bbio)
+{
+	int ret;
+	int looped = 0;
+	struct reada_zone *zone;
+	struct btrfs_block_group_cache *cache = NULL;
+	u64 start;
+	u64 end;
+	int i;
+
+again:
+	zone = NULL;
+	spin_lock(&fs_info->reada_lock);
+	ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone,
+				     logical >> PAGE_CACHE_SHIFT, 1);
+	if (ret == 1)
+		kref_get(&zone->refcnt);
+	spin_unlock(&fs_info->reada_lock);
+
+	if (ret == 1) {
+		if (logical >= zone->start && logical < zone->end)
+			return zone;
+		spin_lock(&fs_info->reada_lock);
+		kref_put(&zone->refcnt, reada_zone_release);
+		spin_unlock(&fs_info->reada_lock);
+	}
+
+	if (looped)
+		return NULL;
+
+	cache = btrfs_lookup_block_group(fs_info, logical);
+	if (!cache)
+		return NULL;
+
+	start = cache->key.objectid;
+	end = start + cache->key.offset - 1;
+	btrfs_put_block_group(cache);
+
+	zone = kzalloc(sizeof(*zone), GFP_NOFS);
+	if (!zone)
+		return NULL;
+
+	zone->start = start;
+	zone->end = end;
+	INIT_LIST_HEAD(&zone->list);
+	spin_lock_init(&zone->lock);
+	zone->locked = 0;
+	kref_init(&zone->refcnt);
+	zone->elems = 0;
+	zone->device = dev; /* our device always sits at index 0 */
+	for (i = 0; i < bbio->num_stripes; ++i) {
+		/* bounds have already been checked */
+		zone->devs[i] = bbio->stripes[i].dev;
+	}
+	zone->ndevs = bbio->num_stripes;
+
+	spin_lock(&fs_info->reada_lock);
+	ret = radix_tree_insert(&dev->reada_zones,
+				(unsigned long)zone->end >> PAGE_CACHE_SHIFT,
+				zone);
+	spin_unlock(&fs_info->reada_lock);
+
+	if (ret) {
+		kfree(zone);
+		looped = 1;
+		goto again;
+	}
+
+	return zone;
+}
+
+static struct reada_extent *reada_find_extent(struct btrfs_root *root,
+					      u64 logical,
+					      struct btrfs_key *top, int level)
+{
+	int ret;
+	int looped = 0;
+	struct reada_extent *re = NULL;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
+	struct btrfs_bio *bbio = NULL;
+	struct btrfs_device *dev;
+	u32 blocksize;
+	u64 length;
+	int nzones = 0;
+	int i;
+	unsigned long index = logical >> PAGE_CACHE_SHIFT;
+
+again:
+	spin_lock(&fs_info->reada_lock);
+	re = radix_tree_lookup(&fs_info->reada_tree, index);
+	if (re)
+		kref_get(&re->refcnt);
+	spin_unlock(&fs_info->reada_lock);
+
+	if (re || looped)
+		return re;
+
+	re = kzalloc(sizeof(*re), GFP_NOFS);
+	if (!re)
+		return NULL;
+
+	blocksize = btrfs_level_size(root, level);
+	re->logical = logical;
+	re->blocksize = blocksize;
+	re->top = *top;
+	INIT_LIST_HEAD(&re->extctl);
+	spin_lock_init(&re->lock);
+	kref_init(&re->refcnt);
+
+	/*
+	 * map block
+	 */
+	length = blocksize;
+	ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, &bbio, 0);
+	if (ret || !bbio || length < blocksize)
+		goto error;
+
+	if (bbio->num_stripes > MAX_MIRRORS) {
+		printk(KERN_ERR "btrfs readahead: more than %d copies not "
+				"supported", MAX_MIRRORS);
+		goto error;
+	}
+
+	for (nzones = 0; nzones < bbio->num_stripes; ++nzones) {
+		struct reada_zone *zone;
+
+		dev = bbio->stripes[nzones].dev;
+		zone = reada_find_zone(fs_info, dev, logical, bbio);
+		if (!zone)
+			break;
+
+		re->zones[nzones] = zone;
+		spin_lock(&zone->lock);
+		if (!zone->elems)
+			kref_get(&zone->refcnt);
+		++zone->elems;
+		spin_unlock(&zone->lock);
+		spin_lock(&fs_info->reada_lock);
+		kref_put(&zone->refcnt, reada_zone_release);
+		spin_unlock(&fs_info->reada_lock);
+	}
+	re->nzones = nzones;
+	if (nzones == 0) {
+		/* not a single zone found, error and out */
+		goto error;
+	}
+
+	/* insert extent in reada_tree + all per-device trees, all or nothing */
+	spin_lock(&fs_info->reada_lock);
+	ret = radix_tree_insert(&fs_info->reada_tree, index, re);
+	if (ret) {
+		spin_unlock(&fs_info->reada_lock);
+		if (ret != -ENOMEM) {
+			/* someone inserted the extent in the meantime */
+			looped = 1;
+		}
+		goto error;
+	}
+	for (i = 0; i < nzones; ++i) {
+		dev = bbio->stripes[i].dev;
+		ret = radix_tree_insert(&dev->reada_extents, index, re);
+		if (ret) {
+			while (--i >= 0) {
+				dev = bbio->stripes[i].dev;
+				BUG_ON(dev == NULL);
+				radix_tree_delete(&dev->reada_extents, index);
+			}
+			BUG_ON(fs_info == NULL);
+			radix_tree_delete(&fs_info->reada_tree, index);
+			spin_unlock(&fs_info->reada_lock);
+			goto error;
+		}
+	}
+	spin_unlock(&fs_info->reada_lock);
+
+	kfree(bbio);
+	return re;
+
+error:
+	while (nzones) {
+		struct reada_zone *zone;
+
+		--nzones;
+		zone = re->zones[nzones];
+		kref_get(&zone->refcnt);
+		spin_lock(&zone->lock);
+		--zone->elems;
+		if (zone->elems == 0) {
+			/*
+			 * no fs_info->reada_lock needed, as this can't be
+			 * the last ref
+			 */
+			kref_put(&zone->refcnt, reada_zone_release);
+		}
+		spin_unlock(&zone->lock);
+
+		spin_lock(&fs_info->reada_lock);
+		kref_put(&zone->refcnt, reada_zone_release);
+		spin_unlock(&fs_info->reada_lock);
+	}
+	kfree(bbio);
+	kfree(re);
+	if (looped)
+		goto again;
+	return NULL;
+}
+
+static void reada_kref_dummy(struct kref *kr)
+{
+}
+
+static void reada_extent_put(struct btrfs_fs_info *fs_info,
+			     struct reada_extent *re)
+{
+	int i;
+	unsigned long index = re->logical >> PAGE_CACHE_SHIFT;
+
+	spin_lock(&fs_info->reada_lock);
+	if (!kref_put(&re->refcnt, reada_kref_dummy)) {
+		spin_unlock(&fs_info->reada_lock);
+		return;
+	}
+
+	radix_tree_delete(&fs_info->reada_tree, index);
+	for (i = 0; i < re->nzones; ++i) {
+		struct reada_zone *zone = re->zones[i];
+
+		radix_tree_delete(&zone->device->reada_extents, index);
+	}
+
+	spin_unlock(&fs_info->reada_lock);
+
+	for (i = 0; i < re->nzones; ++i) {
+		struct reada_zone *zone = re->zones[i];
+
+		kref_get(&zone->refcnt);
+		spin_lock(&zone->lock);
+		--zone->elems;
+		if (zone->elems == 0) {
+			/* no fs_info->reada_lock needed, as this can't be
+			 * the last ref */
+			kref_put(&zone->refcnt, reada_zone_release);
+		}
+		spin_unlock(&zone->lock);
+
+		spin_lock(&fs_info->reada_lock);
+		kref_put(&zone->refcnt, reada_zone_release);
+		spin_unlock(&fs_info->reada_lock);
+	}
+	if (re->scheduled_for)
+		atomic_dec(&re->scheduled_for->reada_in_flight);
+
+	kfree(re);
+}
+
+static void reada_zone_release(struct kref *kref)
+{
+	struct reada_zone *zone = container_of(kref, struct reada_zone, refcnt);
+
+	radix_tree_delete(&zone->device->reada_zones,
+			  zone->end >> PAGE_CACHE_SHIFT);
+
+	kfree(zone);
+}
+
+static void reada_control_release(struct kref *kref)
+{
+	struct reada_control *rc = container_of(kref, struct reada_control,
+						refcnt);
+
+	kfree(rc);
+}
+
+static int reada_add_block(struct reada_control *rc, u64 logical,
+			   struct btrfs_key *top, int level, u64 generation)
+{
+	struct btrfs_root *root = rc->root;
+	struct reada_extent *re;
+	struct reada_extctl *rec;
+
+	re = reada_find_extent(root, logical, top, level); /* takes one ref */
+	if (!re)
+		return -1;
+
+	rec = kzalloc(sizeof(*rec), GFP_NOFS);
+	if (!rec) {
+		reada_extent_put(root->fs_info, re);
+		return -1;
+	}
+
+	rec->rc = rc;
+	rec->generation = generation;
+	atomic_inc(&rc->elems);
+
+	spin_lock(&re->lock);
+	list_add_tail(&rec->list, &re->extctl);
+	spin_unlock(&re->lock);
+
+	/* leave the ref on the extent */
+
+	return 0;
+}
+
+/*
+ * called with fs_info->reada_lock held
+ */
+static void reada_peer_zones_set_lock(struct reada_zone *zone, int lock)
+{
+	int i;
+	unsigned long index = zone->end >> PAGE_CACHE_SHIFT;
+
+	for (i = 0; i < zone->ndevs; ++i) {
+		struct reada_zone *peer;
+		peer = radix_tree_lookup(&zone->devs[i]->reada_zones, index);
+		if (peer && peer->device != zone->device)
+			peer->locked = lock;
+	}
+}
+
+/*
+ * called with fs_info->reada_lock held
+ */
+static int reada_pick_zone(struct btrfs_device *dev)
+{
+	struct reada_zone *top_zone = NULL;
+	struct reada_zone *top_locked_zone = NULL;
+	u64 top_elems = 0;
+	u64 top_locked_elems = 0;
+	unsigned long index = 0;
+	int ret;
+
+	if (dev->reada_curr_zone) {
+		reada_peer_zones_set_lock(dev->reada_curr_zone, 0);
+		kref_put(&dev->reada_curr_zone->refcnt, reada_zone_release);
+		dev->reada_curr_zone = NULL;
+	}
+	/* pick the zone with the most elements */
+	while (1) {
+		struct reada_zone *zone;
+
+		ret = radix_tree_gang_lookup(&dev->reada_zones,
+					     (void **)&zone, index, 1);
+		if (ret == 0)
+			break;
+		index = (zone->end >> PAGE_CACHE_SHIFT) + 1;
+		if (zone->locked) {
+			if (zone->elems > top_locked_elems) {
+				top_locked_elems = zone->elems;
+				top_locked_zone = zone;
+			}
+		} else {
+			if (zone->elems > top_elems) {
+				top_elems = zone->elems;
+				top_zone = zone;
+			}
+		}
+	}
+	if (top_zone)
+		dev->reada_curr_zone = top_zone;
+	else if (top_locked_zone)
+		dev->reada_curr_zone = top_locked_zone;
+	else
+		return 0;
+
+	dev->reada_next = dev->reada_curr_zone->start;
+	kref_get(&dev->reada_curr_zone->refcnt);
+	reada_peer_zones_set_lock(dev->reada_curr_zone, 1);
+
+	return 1;
+}
+
+static int reada_start_machine_dev(struct btrfs_fs_info *fs_info,
+				   struct btrfs_device *dev)
+{
+	struct reada_extent *re = NULL;
+	int mirror_num = 0;
+	struct extent_buffer *eb = NULL;
+	u64 logical;
+	u32 blocksize;
+	int ret;
+	int i;
+	int need_kick = 0;
+
+	spin_lock(&fs_info->reada_lock);
+	if (dev->reada_curr_zone == NULL) {
+		ret = reada_pick_zone(dev);
+		if (!ret) {
+			spin_unlock(&fs_info->reada_lock);
+			return 0;
+		}
+	}
+	/*
+	 * FIXME currently we issue the reads one extent at a time. If we have
+	 * a contiguous block of extents, we could also coagulate them or use
+	 * plugging to speed things up
+	 */
+	ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re,
+				     dev->reada_next >> PAGE_CACHE_SHIFT, 1);
+	if (ret == 0 || re->logical >= dev->reada_curr_zone->end) {
+		ret = reada_pick_zone(dev);
+		if (!ret) {
+			spin_unlock(&fs_info->reada_lock);
+			return 0;
+		}
+		re = NULL;
+		ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re,
+					dev->reada_next >> PAGE_CACHE_SHIFT, 1);
+	}
+	if (ret == 0) {
+		spin_unlock(&fs_info->reada_lock);
+		return 0;
+	}
+	dev->reada_next = re->logical + re->blocksize;
+	kref_get(&re->refcnt);
+
+	spin_unlock(&fs_info->reada_lock);
+
+	/*
+	 * find mirror num
+	 */
+	for (i = 0; i < re->nzones; ++i) {
+		if (re->zones[i]->device == dev) {
+			mirror_num = i + 1;
+			break;
+		}
+	}
+	logical = re->logical;
+	blocksize = re->blocksize;
+
+	spin_lock(&re->lock);
+	if (re->scheduled_for == NULL) {
+		re->scheduled_for = dev;
+		need_kick = 1;
+	}
+	spin_unlock(&re->lock);
+
+	reada_extent_put(fs_info, re);
+
+	if (!need_kick)
+		return 0;
+
+	atomic_inc(&dev->reada_in_flight);
+	ret = reada_tree_block_flagged(fs_info->extent_root, logical, blocksize,
+			 mirror_num, &eb);
+	if (ret)
+		__readahead_hook(fs_info->extent_root, NULL, logical, ret);
+	else if (eb)
+		__readahead_hook(fs_info->extent_root, eb, eb->start, ret);
+
+	if (eb)
+		free_extent_buffer(eb);
+
+	return 1;
+
+}
+
+static void reada_start_machine_worker(struct btrfs_work *work)
+{
+	struct reada_machine_work *rmw;
+	struct btrfs_fs_info *fs_info;
+
+	rmw = container_of(work, struct reada_machine_work, work);
+	fs_info = rmw->fs_info;
+
+	kfree(rmw);
+
+	__reada_start_machine(fs_info);
+}
+
+static void __reada_start_machine(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_device *device;
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	u64 enqueued;
+	u64 total = 0;
+	int i;
+
+	do {
+		enqueued = 0;
+		list_for_each_entry(device, &fs_devices->devices, dev_list) {
+			if (atomic_read(&device->reada_in_flight) <
+			    MAX_IN_FLIGHT)
+				enqueued += reada_start_machine_dev(fs_info,
+								    device);
+		}
+		total += enqueued;
+	} while (enqueued && total < 10000);
+
+	if (enqueued == 0)
+		return;
+
+	/*
+	 * If everything is already in the cache, this is effectively single
+	 * threaded. To a) not hold the caller for too long and b) to utilize
+	 * more cores, we broke the loop above after 10000 iterations and now
+	 * enqueue to workers to finish it. This will distribute the load to
+	 * the cores.
+	 */
+	for (i = 0; i < 2; ++i)
+		reada_start_machine(fs_info);
+}
+
+static void reada_start_machine(struct btrfs_fs_info *fs_info)
+{
+	struct reada_machine_work *rmw;
+
+	rmw = kzalloc(sizeof(*rmw), GFP_NOFS);
+	if (!rmw) {
+		/* FIXME we cannot handle this properly right now */
+		BUG();
+	}
+	rmw->work.func = reada_start_machine_worker;
+	rmw->fs_info = fs_info;
+
+	btrfs_queue_worker(&fs_info->readahead_workers, &rmw->work);
+}
+
+#ifdef DEBUG
+static void dump_devs(struct btrfs_fs_info *fs_info, int all)
+{
+	struct btrfs_device *device;
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	unsigned long index;
+	int ret;
+	int i;
+	int j;
+	int cnt;
+
+	spin_lock(&fs_info->reada_lock);
+	list_for_each_entry(device, &fs_devices->devices, dev_list) {
+		printk(KERN_DEBUG "dev %lld has %d in flight\n", device->devid,
+			atomic_read(&device->reada_in_flight));
+		index = 0;
+		while (1) {
+			struct reada_zone *zone;
+			ret = radix_tree_gang_lookup(&device->reada_zones,
+						     (void **)&zone, index, 1);
+			if (ret == 0)
+				break;
+			printk(KERN_DEBUG "  zone %llu-%llu elems %llu locked "
+				"%d devs", zone->start, zone->end, zone->elems,
+				zone->locked);
+			for (j = 0; j < zone->ndevs; ++j) {
+				printk(KERN_CONT " %lld",
+					zone->devs[j]->devid);
+			}
+			if (device->reada_curr_zone == zone)
+				printk(KERN_CONT " curr off %llu",
+					device->reada_next - zone->start);
+			printk(KERN_CONT "\n");
+			index = (zone->end >> PAGE_CACHE_SHIFT) + 1;
+		}
+		cnt = 0;
+		index = 0;
+		while (all) {
+			struct reada_extent *re = NULL;
+
+			ret = radix_tree_gang_lookup(&device->reada_extents,
+						     (void **)&re, index, 1);
+			if (ret == 0)
+				break;
+			printk(KERN_DEBUG
+				"  re: logical %llu size %u empty %d for %lld",
+				re->logical, re->blocksize,
+				list_empty(&re->extctl), re->scheduled_for ?
+				re->scheduled_for->devid : -1);
+
+			for (i = 0; i < re->nzones; ++i) {
+				printk(KERN_CONT " zone %llu-%llu devs",
+					re->zones[i]->start,
+					re->zones[i]->end);
+				for (j = 0; j < re->zones[i]->ndevs; ++j) {
+					printk(KERN_CONT " %lld",
+						re->zones[i]->devs[j]->devid);
+				}
+			}
+			printk(KERN_CONT "\n");
+			index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+			if (++cnt > 15)
+				break;
+		}
+	}
+
+	index = 0;
+	cnt = 0;
+	while (all) {
+		struct reada_extent *re = NULL;
+
+		ret = radix_tree_gang_lookup(&fs_info->reada_tree, (void **)&re,
+					     index, 1);
+		if (ret == 0)
+			break;
+		if (!re->scheduled_for) {
+			index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+			continue;
+		}
+		printk(KERN_DEBUG
+			"re: logical %llu size %u list empty %d for %lld",
+			re->logical, re->blocksize, list_empty(&re->extctl),
+			re->scheduled_for ? re->scheduled_for->devid : -1);
+		for (i = 0; i < re->nzones; ++i) {
+			printk(KERN_CONT " zone %llu-%llu devs",
+				re->zones[i]->start,
+				re->zones[i]->end);
+			for (i = 0; i < re->nzones; ++i) {
+				printk(KERN_CONT " zone %llu-%llu devs",
+					re->zones[i]->start,
+					re->zones[i]->end);
+				for (j = 0; j < re->zones[i]->ndevs; ++j) {
+					printk(KERN_CONT " %lld",
+						re->zones[i]->devs[j]->devid);
+				}
+			}
+		}
+		printk(KERN_CONT "\n");
+		index = (re->logical >> PAGE_CACHE_SHIFT) + 1;
+	}
+	spin_unlock(&fs_info->reada_lock);
+}
+#endif
+
+/*
+ * interface
+ */
+struct reada_control *btrfs_reada_add(struct btrfs_root *root,
+			struct btrfs_key *key_start, struct btrfs_key *key_end)
+{
+	struct reada_control *rc;
+	u64 start;
+	u64 generation;
+	int level;
+	struct extent_buffer *node;
+	static struct btrfs_key max_key = {
+		.objectid = (u64)-1,
+		.type = (u8)-1,
+		.offset = (u64)-1
+	};
+
+	rc = kzalloc(sizeof(*rc), GFP_NOFS);
+	if (!rc)
+		return ERR_PTR(-ENOMEM);
+
+	rc->root = root;
+	rc->key_start = *key_start;
+	rc->key_end = *key_end;
+	atomic_set(&rc->elems, 0);
+	init_waitqueue_head(&rc->wait);
+	kref_init(&rc->refcnt);
+	kref_get(&rc->refcnt); /* one ref for having elements */
+
+	node = btrfs_root_node(root);
+	start = node->start;
+	level = btrfs_header_level(node);
+	generation = btrfs_header_generation(node);
+	free_extent_buffer(node);
+
+	reada_add_block(rc, start, &max_key, level, generation);
+
+	reada_start_machine(root->fs_info);
+
+	return rc;
+}
+
+#ifdef DEBUG
+int btrfs_reada_wait(void *handle)
+{
+	struct reada_control *rc = handle;
+
+	while (atomic_read(&rc->elems)) {
+		wait_event_timeout(rc->wait, atomic_read(&rc->elems) == 0,
+				   5 * HZ);
+		dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0);
+	}
+
+	dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0);
+
+	kref_put(&rc->refcnt, reada_control_release);
+
+	return 0;
+}
+#else
+int btrfs_reada_wait(void *handle)
+{
+	struct reada_control *rc = handle;
+
+	while (atomic_read(&rc->elems)) {
+		wait_event(rc->wait, atomic_read(&rc->elems) == 0);
+	}
+
+	kref_put(&rc->refcnt, reada_control_release);
+
+	return 0;
+}
+#endif
+
+void btrfs_reada_detach(void *handle)
+{
+	struct reada_control *rc = handle;
+
+	kref_put(&rc->refcnt, reada_control_release);
+}
diff --git a/fs/exofs/Kconfig.ore b/fs/exofs/Kconfig.ore
new file mode 100644
index 0000000..1ca7fb7
--- /dev/null
+++ b/fs/exofs/Kconfig.ore
@@ -0,0 +1,12 @@
+# ORE - Objects Raid Engine (libore.ko)
+#
+# Note ORE needs to "select ASYNC_XOR". So Not to force multiple selects
+# for every ORE user we do it like this. Any user should add itself here
+# at the "depends on EXOFS_FS || ..." with an ||. The dependencies are
+# selected here, and we default to "ON". So in effect it is like been
+# selected by any of the users.
+config ORE
+	tristate
+	depends on EXOFS_FS || PNFS_OBJLAYOUT
+	select ASYNC_XOR
+	default SCSI_OSD_ULD
diff --git a/fs/exofs/ore.c b/fs/exofs/ore.c
new file mode 100644
index 0000000..df4a10f
--- /dev/null
+++ b/fs/exofs/ore.c
@@ -0,0 +1,1133 @@
+/*
+ * Copyright (C) 2005, 2006
+ * Avishay Traeger (avishay@gmail.com)
+ * Copyright (C) 2008, 2009
+ * Boaz Harrosh <bharrosh@panasas.com>
+ *
+ * This file is part of exofs.
+ *
+ * exofs is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation.  Since it is based on ext2, and the only
+ * valid version of GPL for the Linux kernel is version 2, the only valid
+ * version of GPL for exofs is version 2.
+ *
+ * exofs is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with exofs; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <asm/div64.h>
+#include <linux/lcm.h>
+
+#include "ore_raid.h"
+
+MODULE_AUTHOR("Boaz Harrosh <bharrosh@panasas.com>");
+MODULE_DESCRIPTION("Objects Raid Engine ore.ko");
+MODULE_LICENSE("GPL");
+
+/* ore_verify_layout does a couple of things:
+ * 1. Given a minimum number of needed parameters fixes up the rest of the
+ *    members to be operatonals for the ore. The needed parameters are those
+ *    that are defined by the pnfs-objects layout STD.
+ * 2. Check to see if the current ore code actually supports these parameters
+ *    for example stripe_unit must be a multple of the system PAGE_SIZE,
+ *    and etc...
+ * 3. Cache some havily used calculations that will be needed by users.
+ */
+
+enum { BIO_MAX_PAGES_KMALLOC =
+		(PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),};
+
+int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
+{
+	u64 stripe_length;
+
+	switch (layout->raid_algorithm) {
+	case PNFS_OSD_RAID_0:
+		layout->parity = 0;
+		break;
+	case PNFS_OSD_RAID_5:
+		layout->parity = 1;
+		break;
+	case PNFS_OSD_RAID_PQ:
+	case PNFS_OSD_RAID_4:
+	default:
+		ORE_ERR("Only RAID_0/5 for now\n");
+		return -EINVAL;
+	}
+	if (0 != (layout->stripe_unit & ~PAGE_MASK)) {
+		ORE_ERR("Stripe Unit(0x%llx)"
+			  " must be Multples of PAGE_SIZE(0x%lx)\n",
+			  _LLU(layout->stripe_unit), PAGE_SIZE);
+		return -EINVAL;
+	}
+	if (layout->group_width) {
+		if (!layout->group_depth) {
+			ORE_ERR("group_depth == 0 && group_width != 0\n");
+			return -EINVAL;
+		}
+		if (total_comps < (layout->group_width * layout->mirrors_p1)) {
+			ORE_ERR("Data Map wrong, "
+				"numdevs=%d < group_width=%d * mirrors=%d\n",
+				total_comps, layout->group_width,
+				layout->mirrors_p1);
+			return -EINVAL;
+		}
+		layout->group_count = total_comps / layout->mirrors_p1 /
+						layout->group_width;
+	} else {
+		if (layout->group_depth) {
+			printk(KERN_NOTICE "Warning: group_depth ignored "
+				"group_width == 0 && group_depth == %lld\n",
+				_LLU(layout->group_depth));
+		}
+		layout->group_width = total_comps / layout->mirrors_p1;
+		layout->group_depth = -1;
+		layout->group_count = 1;
+	}
+
+	stripe_length = (u64)layout->group_width * layout->stripe_unit;
+	if (stripe_length >= (1ULL << 32)) {
+		ORE_ERR("Stripe_length(0x%llx) >= 32bit is not supported\n",
+			_LLU(stripe_length));
+		return -EINVAL;
+	}
+
+	layout->max_io_length =
+		(BIO_MAX_PAGES_KMALLOC * PAGE_SIZE - layout->stripe_unit) *
+					(layout->group_width - layout->parity);
+	if (layout->parity) {
+		unsigned stripe_length =
+				(layout->group_width - layout->parity) *
+				layout->stripe_unit;
+
+		layout->max_io_length /= stripe_length;
+		layout->max_io_length *= stripe_length;
+	}
+	return 0;
+}
+EXPORT_SYMBOL(ore_verify_layout);
+
+static u8 *_ios_cred(struct ore_io_state *ios, unsigned index)
+{
+	return ios->oc->comps[index & ios->oc->single_comp].cred;
+}
+
+static struct osd_obj_id *_ios_obj(struct ore_io_state *ios, unsigned index)
+{
+	return &ios->oc->comps[index & ios->oc->single_comp].obj;
+}
+
+static struct osd_dev *_ios_od(struct ore_io_state *ios, unsigned index)
+{
+	ORE_DBGMSG2("oc->first_dev=%d oc->numdevs=%d i=%d oc->ods=%p\n",
+		    ios->oc->first_dev, ios->oc->numdevs, index,
+		    ios->oc->ods);
+
+	return ore_comp_dev(ios->oc, index);
+}
+
+int  _ore_get_io_state(struct ore_layout *layout,
+			struct ore_components *oc, unsigned numdevs,
+			unsigned sgs_per_dev, unsigned num_par_pages,
+			struct ore_io_state **pios)
+{
+	struct ore_io_state *ios;
+	struct page **pages;
+	struct osd_sg_entry *sgilist;
+	struct __alloc_all_io_state {
+		struct ore_io_state ios;
+		struct ore_per_dev_state per_dev[numdevs];
+		union {
+			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
+			struct page *pages[num_par_pages];
+		};
+	} *_aios;
+
+	if (likely(sizeof(*_aios) <= PAGE_SIZE)) {
+		_aios = kzalloc(sizeof(*_aios), GFP_KERNEL);
+		if (unlikely(!_aios)) {
+			ORE_DBGMSG("Failed kzalloc bytes=%zd\n",
+				   sizeof(*_aios));
+			*pios = NULL;
+			return -ENOMEM;
+		}
+		pages = num_par_pages ? _aios->pages : NULL;
+		sgilist = sgs_per_dev ? _aios->sglist : NULL;
+		ios = &_aios->ios;
+	} else {
+		struct __alloc_small_io_state {
+			struct ore_io_state ios;
+			struct ore_per_dev_state per_dev[numdevs];
+		} *_aio_small;
+		union __extra_part {
+			struct osd_sg_entry sglist[sgs_per_dev * numdevs];
+			struct page *pages[num_par_pages];
+		} *extra_part;
+
+		_aio_small = kzalloc(sizeof(*_aio_small), GFP_KERNEL);
+		if (unlikely(!_aio_small)) {
+			ORE_DBGMSG("Failed alloc first part bytes=%zd\n",
+				   sizeof(*_aio_small));
+			*pios = NULL;
+			return -ENOMEM;
+		}
+		extra_part = kzalloc(sizeof(*extra_part), GFP_KERNEL);
+		if (unlikely(!extra_part)) {
+			ORE_DBGMSG("Failed alloc second part bytes=%zd\n",
+				   sizeof(*extra_part));
+			kfree(_aio_small);
+			*pios = NULL;
+			return -ENOMEM;
+		}
+
+		pages = num_par_pages ? extra_part->pages : NULL;
+		sgilist = sgs_per_dev ? extra_part->sglist : NULL;
+		/* In this case the per_dev[0].sgilist holds the pointer to
+		 * be freed
+		 */
+		ios = &_aio_small->ios;
+		ios->extra_part_alloc = true;
+	}
+
+	if (pages) {
+		ios->parity_pages = pages;
+		ios->max_par_pages = num_par_pages;
+	}
+	if (sgilist) {
+		unsigned d;
+
+		for (d = 0; d < numdevs; ++d) {
+			ios->per_dev[d].sglist = sgilist;
+			sgilist += sgs_per_dev;
+		}
+		ios->sgs_per_dev = sgs_per_dev;
+	}
+
+	ios->layout = layout;
+	ios->oc = oc;
+	*pios = ios;
+	return 0;
+}
+
+/* Allocate an io_state for only a single group of devices
+ *
+ * If a user needs to call ore_read/write() this version must be used becase it
+ * allocates extra stuff for striping and raid.
+ * The ore might decide to only IO less then @length bytes do to alignmets
+ * and constrains as follows:
+ * - The IO cannot cross group boundary.
+ * - In raid5/6 The end of the IO must align at end of a stripe eg.
+ *   (@offset + @length) % strip_size == 0. Or the complete range is within a
+ *   single stripe.
+ * - Memory condition only permitted a shorter IO. (A user can use @length=~0
+ *   And check the returned ios->length for max_io_size.)
+ *
+ * The caller must check returned ios->length (and/or ios->nr_pages) and
+ * re-issue these pages that fall outside of ios->length
+ */
+int  ore_get_rw_state(struct ore_layout *layout, struct ore_components *oc,
+		      bool is_reading, u64 offset, u64 length,
+		      struct ore_io_state **pios)
+{
+	struct ore_io_state *ios;
+	unsigned numdevs = layout->group_width * layout->mirrors_p1;
+	unsigned sgs_per_dev = 0, max_par_pages = 0;
+	int ret;
+
+	if (layout->parity && length) {
+		unsigned data_devs = layout->group_width - layout->parity;
+		unsigned stripe_size = layout->stripe_unit * data_devs;
+		unsigned pages_in_unit = layout->stripe_unit / PAGE_SIZE;
+		u32 remainder;
+		u64 num_stripes;
+		u64 num_raid_units;
+
+		num_stripes = div_u64_rem(length, stripe_size, &remainder);
+		if (remainder)
+			++num_stripes;
+
+		num_raid_units =  num_stripes * layout->parity;
+
+		if (is_reading) {
+			/* For reads add per_dev sglist array */
+			/* TODO: Raid 6 we need twice more. Actually:
+			*         num_stripes / LCMdP(W,P);
+			*         if (W%P != 0) num_stripes *= parity;
+			*/
+
+			/* first/last seg is split */
+			num_raid_units += layout->group_width;
+			sgs_per_dev = div_u64(num_raid_units, data_devs) + 2;
+		} else {
+			/* For Writes add parity pages array. */
+			max_par_pages = num_raid_units * pages_in_unit *
+						sizeof(struct page *);
+		}
+	}
+
+	ret = _ore_get_io_state(layout, oc, numdevs, sgs_per_dev, max_par_pages,
+				pios);
+	if (unlikely(ret))
+		return ret;
+
+	ios = *pios;
+	ios->reading = is_reading;
+	ios->offset = offset;
+
+	if (length) {
+		ore_calc_stripe_info(layout, offset, length, &ios->si);
+		ios->length = ios->si.length;
+		ios->nr_pages = ((ios->offset & (PAGE_SIZE - 1)) +
+				 ios->length + PAGE_SIZE - 1) / PAGE_SIZE;
+		if (layout->parity)
+			_ore_post_alloc_raid_stuff(ios);
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(ore_get_rw_state);
+
+/* Allocate an io_state for all the devices in the comps array
+ *
+ * This version of io_state allocation is used mostly by create/remove
+ * and trunc where we currently need all the devices. The only wastful
+ * bit is the read/write_attributes with no IO. Those sites should
+ * be converted to use ore_get_rw_state() with length=0
+ */
+int  ore_get_io_state(struct ore_layout *layout, struct ore_components *oc,
+		      struct ore_io_state **pios)
+{
+	return _ore_get_io_state(layout, oc, oc->numdevs, 0, 0, pios);
+}
+EXPORT_SYMBOL(ore_get_io_state);
+
+void ore_put_io_state(struct ore_io_state *ios)
+{
+	if (ios) {
+		unsigned i;
+
+		for (i = 0; i < ios->numdevs; i++) {
+			struct ore_per_dev_state *per_dev = &ios->per_dev[i];
+
+			if (per_dev->or)
+				osd_end_request(per_dev->or);
+			if (per_dev->bio)
+				bio_put(per_dev->bio);
+		}
+
+		_ore_free_raid_stuff(ios);
+		kfree(ios);
+	}
+}
+EXPORT_SYMBOL(ore_put_io_state);
+
+static void _sync_done(struct ore_io_state *ios, void *p)
+{
+	struct completion *waiting = p;
+
+	complete(waiting);
+}
+
+static void _last_io(struct kref *kref)
+{
+	struct ore_io_state *ios = container_of(
+					kref, struct ore_io_state, kref);
+
+	ios->done(ios, ios->private);
+}
+
+static void _done_io(struct osd_request *or, void *p)
+{
+	struct ore_io_state *ios = p;
+
+	kref_put(&ios->kref, _last_io);
+}
+
+int ore_io_execute(struct ore_io_state *ios)
+{
+	DECLARE_COMPLETION_ONSTACK(wait);
+	bool sync = (ios->done == NULL);
+	int i, ret;
+
+	if (sync) {
+		ios->done = _sync_done;
+		ios->private = &wait;
+	}
+
+	for (i = 0; i < ios->numdevs; i++) {
+		struct osd_request *or = ios->per_dev[i].or;
+		if (unlikely(!or))
+			continue;
+
+		ret = osd_finalize_request(or, 0, _ios_cred(ios, i), NULL);
+		if (unlikely(ret)) {
+			ORE_DBGMSG("Failed to osd_finalize_request() => %d\n",
+				     ret);
+			return ret;
+		}
+	}
+
+	kref_init(&ios->kref);
+
+	for (i = 0; i < ios->numdevs; i++) {
+		struct osd_request *or = ios->per_dev[i].or;
+		if (unlikely(!or))
+			continue;
+
+		kref_get(&ios->kref);
+		osd_execute_request_async(or, _done_io, ios);
+	}
+
+	kref_put(&ios->kref, _last_io);
+	ret = 0;
+
+	if (sync) {
+		wait_for_completion(&wait);
+		ret = ore_check_io(ios, NULL);
+	}
+	return ret;
+}
+
+static void _clear_bio(struct bio *bio)
+{
+	struct bio_vec *bv;
+	unsigned i;
+
+	bio_for_each_segment_all(bv, bio, i) {
+		unsigned this_count = bv->bv_len;
+
+		if (likely(PAGE_SIZE == this_count))
+			clear_highpage(bv->bv_page);
+		else
+			zero_user(bv->bv_page, bv->bv_offset, this_count);
+	}
+}
+
+int ore_check_io(struct ore_io_state *ios, ore_on_dev_error on_dev_error)
+{
+	enum osd_err_priority acumulated_osd_err = 0;
+	int acumulated_lin_err = 0;
+	int i;
+
+	for (i = 0; i < ios->numdevs; i++) {
+		struct osd_sense_info osi;
+		struct ore_per_dev_state *per_dev = &ios->per_dev[i];
+		struct osd_request *or = per_dev->or;
+		int ret;
+
+		if (unlikely(!or))
+			continue;
+
+		ret = osd_req_decode_sense(or, &osi);
+		if (likely(!ret))
+			continue;
+
+		if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) {
+			/* start read offset passed endof file */
+			_clear_bio(per_dev->bio);
+			ORE_DBGMSG("start read offset passed end of file "
+				"offset=0x%llx, length=0x%llx\n",
+				_LLU(per_dev->offset),
+				_LLU(per_dev->length));
+
+			continue; /* we recovered */
+		}
+
+		if (on_dev_error) {
+			u64 residual = ios->reading ?
+					or->in.residual : or->out.residual;
+			u64 offset = (ios->offset + ios->length) - residual;
+			unsigned dev = per_dev->dev - ios->oc->first_dev;
+			struct ore_dev *od = ios->oc->ods[dev];
+
+			on_dev_error(ios, od, dev, osi.osd_err_pri,
+				     offset, residual);
+		}
+		if (osi.osd_err_pri >= acumulated_osd_err) {
+			acumulated_osd_err = osi.osd_err_pri;
+			acumulated_lin_err = ret;
+		}
+	}
+
+	return acumulated_lin_err;
+}
+EXPORT_SYMBOL(ore_check_io);
+
+/*
+ * L - logical offset into the file
+ *
+ * D - number of Data devices
+ *	D = group_width - parity
+ *
+ * U - The number of bytes in a stripe within a group
+ *	U =  stripe_unit * D
+ *
+ * T - The number of bytes striped within a group of component objects
+ *     (before advancing to the next group)
+ *	T = U * group_depth
+ *
+ * S - The number of bytes striped across all component objects
+ *     before the pattern repeats
+ *	S = T * group_count
+ *
+ * M - The "major" (i.e., across all components) cycle number
+ *	M = L / S
+ *
+ * G - Counts the groups from the beginning of the major cycle
+ *	G = (L - (M * S)) / T	[or (L % S) / T]
+ *
+ * H - The byte offset within the group
+ *	H = (L - (M * S)) % T	[or (L % S) % T]
+ *
+ * N - The "minor" (i.e., across the group) stripe number
+ *	N = H / U
+ *
+ * C - The component index coresponding to L
+ *
+ *	C = (H - (N * U)) / stripe_unit + G * D
+ *	[or (L % U) / stripe_unit + G * D]
+ *
+ * O - The component offset coresponding to L
+ *	O = L % stripe_unit + N * stripe_unit + M * group_depth * stripe_unit
+ *
+ * LCMdP – Parity cycle: Lowest Common Multiple of group_width, parity
+ *          divide by parity
+ *	LCMdP = lcm(group_width, parity) / parity
+ *
+ * R - The parity Rotation stripe
+ *     (Note parity cycle always starts at a group's boundary)
+ *	R = N % LCMdP
+ *
+ * I = the first parity device index
+ *	I = (group_width + group_width - R*parity - parity) % group_width
+ *
+ * Craid - The component index Rotated
+ *	Craid = (group_width + C - R*parity) % group_width
+ *      (We add the group_width to avoid negative numbers modulo math)
+ */
+void ore_calc_stripe_info(struct ore_layout *layout, u64 file_offset,
+			  u64 length, struct ore_striping_info *si)
+{
+	u32	stripe_unit = layout->stripe_unit;
+	u32	group_width = layout->group_width;
+	u64	group_depth = layout->group_depth;
+	u32	parity      = layout->parity;
+
+	u32	D = group_width - parity;
+	u32	U = D * stripe_unit;
+	u64	T = U * group_depth;
+	u64	S = T * layout->group_count;
+	u64	M = div64_u64(file_offset, S);
+
+	/*
+	G = (L - (M * S)) / T
+	H = (L - (M * S)) % T
+	*/
+	u64	LmodS = file_offset - M * S;
+	u32	G = div64_u64(LmodS, T);
+	u64	H = LmodS - G * T;
+
+	u32	N = div_u64(H, U);
+	u32	Nlast;
+
+	/* "H - (N * U)" is just "H % U" so it's bound to u32 */
+	u32	C = (u32)(H - (N * U)) / stripe_unit + G * group_width;
+
+	div_u64_rem(file_offset, stripe_unit, &si->unit_off);
+
+	si->obj_offset = si->unit_off + (N * stripe_unit) +
+				  (M * group_depth * stripe_unit);
+
+	if (parity) {
+		u32 LCMdP = lcm(group_width, parity) / parity;
+		/* R     = N % LCMdP; */
+		u32 RxP   = (N % LCMdP) * parity;
+		u32 first_dev = C - C % group_width;
+
+		si->par_dev = (group_width + group_width - parity - RxP) %
+			      group_width + first_dev;
+		si->dev = (group_width + C - RxP) % group_width + first_dev;
+		si->bytes_in_stripe = U;
+		si->first_stripe_start = M * S + G * T + N * U;
+	} else {
+		/* Make the math correct see _prepare_one_group */
+		si->par_dev = group_width;
+		si->dev = C;
+	}
+
+	si->dev *= layout->mirrors_p1;
+	si->par_dev *= layout->mirrors_p1;
+	si->offset = file_offset;
+	si->length = T - H;
+	if (si->length > length)
+		si->length = length;
+
+	Nlast = div_u64(H + si->length + U - 1, U);
+	si->maxdevUnits = Nlast - N;
+
+	si->M = M;
+}
+EXPORT_SYMBOL(ore_calc_stripe_info);
+
+int _ore_add_stripe_unit(struct ore_io_state *ios,  unsigned *cur_pg,
+			 unsigned pgbase, struct page **pages,
+			 struct ore_per_dev_state *per_dev, int cur_len)
+{
+	unsigned pg = *cur_pg;
+	struct request_queue *q =
+			osd_request_queue(_ios_od(ios, per_dev->dev));
+	unsigned len = cur_len;
+	int ret;
+
+	if (per_dev->bio == NULL) {
+		unsigned bio_size;
+
+		if (!ios->reading) {
+			bio_size = ios->si.maxdevUnits;
+		} else {
+			bio_size = (ios->si.maxdevUnits + 1) *
+			     (ios->layout->group_width - ios->layout->parity) /
+			     ios->layout->group_width;
+		}
+		bio_size *= (ios->layout->stripe_unit / PAGE_SIZE);
+
+		per_dev->bio = bio_kmalloc(GFP_KERNEL, bio_size);
+		if (unlikely(!per_dev->bio)) {
+			ORE_DBGMSG("Failed to allocate BIO size=%u\n",
+				     bio_size);
+			ret = -ENOMEM;
+			goto out;
+		}
+	}
+
+	while (cur_len > 0) {
+		unsigned pglen = min_t(unsigned, PAGE_SIZE - pgbase, cur_len);
+		unsigned added_len;
+
+		cur_len -= pglen;
+
+		added_len = bio_add_pc_page(q, per_dev->bio, pages[pg],
+					    pglen, pgbase);
+		if (unlikely(pglen != added_len)) {
+			/* If bi_vcnt == bi_max then this is a SW BUG */
+			ORE_DBGMSG("Failed bio_add_pc_page bi_vcnt=0x%x "
+				   "bi_max=0x%x BIO_MAX=0x%x cur_len=0x%x\n",
+				   per_dev->bio->bi_vcnt,
+				   per_dev->bio->bi_max_vecs,
+				   BIO_MAX_PAGES_KMALLOC, cur_len);
+			ret = -ENOMEM;
+			goto out;
+		}
+		_add_stripe_page(ios->sp2d, &ios->si, pages[pg]);
+
+		pgbase = 0;
+		++pg;
+	}
+	BUG_ON(cur_len);
+
+	per_dev->length += len;
+	*cur_pg = pg;
+	ret = 0;
+out:	/* we fail the complete unit on an error eg don't advance
+	 * per_dev->length and cur_pg. This means that we might have a bigger
+	 * bio than the CDB requested length (per_dev->length). That's fine
+	 * only the oposite is fatal.
+	 */
+	return ret;
+}
+
+static int _prepare_for_striping(struct ore_io_state *ios)
+{
+	struct ore_striping_info *si = &ios->si;
+	unsigned stripe_unit = ios->layout->stripe_unit;
+	unsigned mirrors_p1 = ios->layout->mirrors_p1;
+	unsigned group_width = ios->layout->group_width;
+	unsigned devs_in_group = group_width * mirrors_p1;
+	unsigned dev = si->dev;
+	unsigned first_dev = dev - (dev % devs_in_group);
+	unsigned dev_order;
+	unsigned cur_pg = ios->pages_consumed;
+	u64 length = ios->length;
+	int ret = 0;
+
+	if (!ios->pages) {
+		ios->numdevs = ios->layout->mirrors_p1;
+		return 0;
+	}
+
+	BUG_ON(length > si->length);
+
+	dev_order = _dev_order(devs_in_group, mirrors_p1, si->par_dev, dev);
+	si->cur_comp = dev_order;
+	si->cur_pg = si->unit_off / PAGE_SIZE;
+
+	while (length) {
+		unsigned comp = dev - first_dev;
+		struct ore_per_dev_state *per_dev = &ios->per_dev[comp];
+		unsigned cur_len, page_off = 0;
+
+		if (!per_dev->length) {
+			per_dev->dev = dev;
+			if (dev == si->dev) {
+				WARN_ON(dev == si->par_dev);
+				per_dev->offset = si->obj_offset;
+				cur_len = stripe_unit - si->unit_off;
+				page_off = si->unit_off & ~PAGE_MASK;
+				BUG_ON(page_off && (page_off != ios->pgbase));
+			} else {
+				if (si->cur_comp > dev_order)
+					per_dev->offset =
+						si->obj_offset - si->unit_off;
+				else /* si->cur_comp < dev_order */
+					per_dev->offset =
+						si->obj_offset + stripe_unit -
+								   si->unit_off;
+				cur_len = stripe_unit;
+			}
+		} else {
+			cur_len = stripe_unit;
+		}
+		if (cur_len >= length)
+			cur_len = length;
+
+		ret = _ore_add_stripe_unit(ios, &cur_pg, page_off, ios->pages,
+					   per_dev, cur_len);
+		if (unlikely(ret))
+			goto out;
+
+		dev += mirrors_p1;
+		dev = (dev % devs_in_group) + first_dev;
+
+		length -= cur_len;
+
+		si->cur_comp = (si->cur_comp + 1) % group_width;
+		if (unlikely((dev == si->par_dev) || (!length && ios->sp2d))) {
+			if (!length && ios->sp2d) {
+				/* If we are writing and this is the very last
+				 * stripe. then operate on parity dev.
+				 */
+				dev = si->par_dev;
+			}
+			if (ios->sp2d)
+				/* In writes cur_len just means if it's the
+				 * last one. See _ore_add_parity_unit.
+				 */
+				cur_len = length;
+			per_dev = &ios->per_dev[dev - first_dev];
+			if (!per_dev->length) {
+				/* Only/always the parity unit of the first
+				 * stripe will be empty. So this is a chance to
+				 * initialize the per_dev info.
+				 */
+				per_dev->dev = dev;
+				per_dev->offset = si->obj_offset - si->unit_off;
+			}
+
+			ret = _ore_add_parity_unit(ios, si, per_dev, cur_len);
+			if (unlikely(ret))
+					goto out;
+
+			/* Rotate next par_dev backwards with wraping */
+			si->par_dev = (devs_in_group + si->par_dev -
+				       ios->layout->parity * mirrors_p1) %
+				      devs_in_group + first_dev;
+			/* Next stripe, start fresh */
+			si->cur_comp = 0;
+			si->cur_pg = 0;
+		}
+	}
+out:
+	ios->numdevs = devs_in_group;
+	ios->pages_consumed = cur_pg;
+	return ret;
+}
+
+int ore_create(struct ore_io_state *ios)
+{
+	int i, ret;
+
+	for (i = 0; i < ios->oc->numdevs; i++) {
+		struct osd_request *or;
+
+		or = osd_start_request(_ios_od(ios, i), GFP_KERNEL);
+		if (unlikely(!or)) {
+			ORE_ERR("%s: osd_start_request failed\n", __func__);
+			ret = -ENOMEM;
+			goto out;
+		}
+		ios->per_dev[i].or = or;
+		ios->numdevs++;
+
+		osd_req_create_object(or, _ios_obj(ios, i));
+	}
+	ret = ore_io_execute(ios);
+
+out:
+	return ret;
+}
+EXPORT_SYMBOL(ore_create);
+
+int ore_remove(struct ore_io_state *ios)
+{
+	int i, ret;
+
+	for (i = 0; i < ios->oc->numdevs; i++) {
+		struct osd_request *or;
+
+		or = osd_start_request(_ios_od(ios, i), GFP_KERNEL);
+		if (unlikely(!or)) {
+			ORE_ERR("%s: osd_start_request failed\n", __func__);
+			ret = -ENOMEM;
+			goto out;
+		}
+		ios->per_dev[i].or = or;
+		ios->numdevs++;
+
+		osd_req_remove_object(or, _ios_obj(ios, i));
+	}
+	ret = ore_io_execute(ios);
+
+out:
+	return ret;
+}
+EXPORT_SYMBOL(ore_remove);
+
+static int _write_mirror(struct ore_io_state *ios, int cur_comp)
+{
+	struct ore_per_dev_state *master_dev = &ios->per_dev[cur_comp];
+	unsigned dev = ios->per_dev[cur_comp].dev;
+	unsigned last_comp = cur_comp + ios->layout->mirrors_p1;
+	int ret = 0;
+
+	if (ios->pages && !master_dev->length)
+		return 0; /* Just an empty slot */
+
+	for (; cur_comp < last_comp; ++cur_comp, ++dev) {
+		struct ore_per_dev_state *per_dev = &ios->per_dev[cur_comp];
+		struct osd_request *or;
+
+		or = osd_start_request(_ios_od(ios, dev), GFP_KERNEL);
+		if (unlikely(!or)) {
+			ORE_ERR("%s: osd_start_request failed\n", __func__);
+			ret = -ENOMEM;
+			goto out;
+		}
+		per_dev->or = or;
+
+		if (ios->pages) {
+			struct bio *bio;
+
+			if (per_dev != master_dev) {
+				bio = bio_kmalloc(GFP_KERNEL,
+						  master_dev->bio->bi_max_vecs);
+				if (unlikely(!bio)) {
+					ORE_DBGMSG(
+					      "Failed to allocate BIO size=%u\n",
+					      master_dev->bio->bi_max_vecs);
+					ret = -ENOMEM;
+					goto out;
+				}
+
+				__bio_clone(bio, master_dev->bio);
+				bio->bi_bdev = NULL;
+				bio->bi_next = NULL;
+				per_dev->offset = master_dev->offset;
+				per_dev->length = master_dev->length;
+				per_dev->bio =  bio;
+				per_dev->dev = dev;
+			} else {
+				bio = master_dev->bio;
+				/* FIXME: bio_set_dir() */
+				bio->bi_rw |= REQ_WRITE;
+			}
+
+			osd_req_write(or, _ios_obj(ios, cur_comp),
+				      per_dev->offset, bio, per_dev->length);
+			ORE_DBGMSG("write(0x%llx) offset=0x%llx "
+				      "length=0x%llx dev=%d\n",
+				     _LLU(_ios_obj(ios, cur_comp)->id),
+				     _LLU(per_dev->offset),
+				     _LLU(per_dev->length), dev);
+		} else if (ios->kern_buff) {
+			per_dev->offset = ios->si.obj_offset;
+			per_dev->dev = ios->si.dev + dev;
+
+			/* no cross device without page array */
+			BUG_ON((ios->layout->group_width > 1) &&
+			       (ios->si.unit_off + ios->length >
+				ios->layout->stripe_unit));
+
+			ret = osd_req_write_kern(or, _ios_obj(ios, cur_comp),
+						 per_dev->offset,
+						 ios->kern_buff, ios->length);
+			if (unlikely(ret))
+				goto out;
+			ORE_DBGMSG2("write_kern(0x%llx) offset=0x%llx "
+				      "length=0x%llx dev=%d\n",
+				     _LLU(_ios_obj(ios, cur_comp)->id),
+				     _LLU(per_dev->offset),
+				     _LLU(ios->length), per_dev->dev);
+		} else {
+			osd_req_set_attributes(or, _ios_obj(ios, cur_comp));
+			ORE_DBGMSG2("obj(0x%llx) set_attributes=%d dev=%d\n",
+				     _LLU(_ios_obj(ios, cur_comp)->id),
+				     ios->out_attr_len, dev);
+		}
+
+		if (ios->out_attr)
+			osd_req_add_set_attr_list(or, ios->out_attr,
+						  ios->out_attr_len);
+
+		if (ios->in_attr)
+			osd_req_add_get_attr_list(or, ios->in_attr,
+						  ios->in_attr_len);
+	}
+
+out:
+	return ret;
+}
+
+int ore_write(struct ore_io_state *ios)
+{
+	int i;
+	int ret;
+
+	if (unlikely(ios->sp2d && !ios->r4w)) {
+		/* A library is attempting a RAID-write without providing
+		 * a pages lock interface.
+		 */
+		WARN_ON_ONCE(1);
+		return -ENOTSUPP;
+	}
+
+	ret = _prepare_for_striping(ios);
+	if (unlikely(ret))
+		return ret;
+
+	for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) {
+		ret = _write_mirror(ios, i);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	ret = ore_io_execute(ios);
+	return ret;
+}
+EXPORT_SYMBOL(ore_write);
+
+int _ore_read_mirror(struct ore_io_state *ios, unsigned cur_comp)
+{
+	struct osd_request *or;
+	struct ore_per_dev_state *per_dev = &ios->per_dev[cur_comp];
+	struct osd_obj_id *obj = _ios_obj(ios, cur_comp);
+	unsigned first_dev = (unsigned)obj->id;
+
+	if (ios->pages && !per_dev->length)
+		return 0; /* Just an empty slot */
+
+	first_dev = per_dev->dev + first_dev % ios->layout->mirrors_p1;
+	or = osd_start_request(_ios_od(ios, first_dev), GFP_KERNEL);
+	if (unlikely(!or)) {
+		ORE_ERR("%s: osd_start_request failed\n", __func__);
+		return -ENOMEM;
+	}
+	per_dev->or = or;
+
+	if (ios->pages) {
+		if (per_dev->cur_sg) {
+			/* finalize the last sg_entry */
+			_ore_add_sg_seg(per_dev, 0, false);
+			if (unlikely(!per_dev->cur_sg))
+				return 0; /* Skip parity only device */
+
+			osd_req_read_sg(or, obj, per_dev->bio,
+					per_dev->sglist, per_dev->cur_sg);
+		} else {
+			/* The no raid case */
+			osd_req_read(or, obj, per_dev->offset,
+				     per_dev->bio, per_dev->length);
+		}
+
+		ORE_DBGMSG("read(0x%llx) offset=0x%llx length=0x%llx"
+			     " dev=%d sg_len=%d\n", _LLU(obj->id),
+			     _LLU(per_dev->offset), _LLU(per_dev->length),
+			     first_dev, per_dev->cur_sg);
+	} else {
+		BUG_ON(ios->kern_buff);
+
+		osd_req_get_attributes(or, obj);
+		ORE_DBGMSG2("obj(0x%llx) get_attributes=%d dev=%d\n",
+			      _LLU(obj->id),
+			      ios->in_attr_len, first_dev);
+	}
+	if (ios->out_attr)
+		osd_req_add_set_attr_list(or, ios->out_attr, ios->out_attr_len);
+
+	if (ios->in_attr)
+		osd_req_add_get_attr_list(or, ios->in_attr, ios->in_attr_len);
+
+	return 0;
+}
+
+int ore_read(struct ore_io_state *ios)
+{
+	int i;
+	int ret;
+
+	ret = _prepare_for_striping(ios);
+	if (unlikely(ret))
+		return ret;
+
+	for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) {
+		ret = _ore_read_mirror(ios, i);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	ret = ore_io_execute(ios);
+	return ret;
+}
+EXPORT_SYMBOL(ore_read);
+
+int extract_attr_from_ios(struct ore_io_state *ios, struct osd_attr *attr)
+{
+	struct osd_attr cur_attr = {.attr_page = 0}; /* start with zeros */
+	void *iter = NULL;
+	int nelem;
+
+	do {
+		nelem = 1;
+		osd_req_decode_get_attr_list(ios->per_dev[0].or,
+					     &cur_attr, &nelem, &iter);
+		if ((cur_attr.attr_page == attr->attr_page) &&
+		    (cur_attr.attr_id == attr->attr_id)) {
+			attr->len = cur_attr.len;
+			attr->val_ptr = cur_attr.val_ptr;
+			return 0;
+		}
+	} while (iter);
+
+	return -EIO;
+}
+EXPORT_SYMBOL(extract_attr_from_ios);
+
+static int _truncate_mirrors(struct ore_io_state *ios, unsigned cur_comp,
+			     struct osd_attr *attr)
+{
+	int last_comp = cur_comp + ios->layout->mirrors_p1;
+
+	for (; cur_comp < last_comp; ++cur_comp) {
+		struct ore_per_dev_state *per_dev = &ios->per_dev[cur_comp];
+		struct osd_request *or;
+
+		or = osd_start_request(_ios_od(ios, cur_comp), GFP_KERNEL);
+		if (unlikely(!or)) {
+			ORE_ERR("%s: osd_start_request failed\n", __func__);
+			return -ENOMEM;
+		}
+		per_dev->or = or;
+
+		osd_req_set_attributes(or, _ios_obj(ios, cur_comp));
+		osd_req_add_set_attr_list(or, attr, 1);
+	}
+
+	return 0;
+}
+
+struct _trunc_info {
+	struct ore_striping_info si;
+	u64 prev_group_obj_off;
+	u64 next_group_obj_off;
+
+	unsigned first_group_dev;
+	unsigned nex_group_dev;
+};
+
+static void _calc_trunk_info(struct ore_layout *layout, u64 file_offset,
+			     struct _trunc_info *ti)
+{
+	unsigned stripe_unit = layout->stripe_unit;
+
+	ore_calc_stripe_info(layout, file_offset, 0, &ti->si);
+
+	ti->prev_group_obj_off = ti->si.M * stripe_unit;
+	ti->next_group_obj_off = ti->si.M ? (ti->si.M - 1) * stripe_unit : 0;
+
+	ti->first_group_dev = ti->si.dev - (ti->si.dev % layout->group_width);
+	ti->nex_group_dev = ti->first_group_dev + layout->group_width;
+}
+
+int ore_truncate(struct ore_layout *layout, struct ore_components *oc,
+		   u64 size)
+{
+	struct ore_io_state *ios;
+	struct exofs_trunc_attr {
+		struct osd_attr attr;
+		__be64 newsize;
+	} *size_attrs;
+	struct _trunc_info ti;
+	int i, ret;
+
+	ret = ore_get_io_state(layout, oc, &ios);
+	if (unlikely(ret))
+		return ret;
+
+	_calc_trunk_info(ios->layout, size, &ti);
+
+	size_attrs = kcalloc(ios->oc->numdevs, sizeof(*size_attrs),
+			     GFP_KERNEL);
+	if (unlikely(!size_attrs)) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ios->numdevs = ios->oc->numdevs;
+
+	for (i = 0; i < ios->numdevs; ++i) {
+		struct exofs_trunc_attr *size_attr = &size_attrs[i];
+		u64 obj_size;
+
+		if (i < ti.first_group_dev)
+			obj_size = ti.prev_group_obj_off;
+		else if (i >= ti.nex_group_dev)
+			obj_size = ti.next_group_obj_off;
+		else if (i < ti.si.dev) /* dev within this group */
+			obj_size = ti.si.obj_offset +
+				      ios->layout->stripe_unit - ti.si.unit_off;
+		else if (i == ti.si.dev)
+			obj_size = ti.si.obj_offset;
+		else /* i > ti.dev */
+			obj_size = ti.si.obj_offset - ti.si.unit_off;
+
+		size_attr->newsize = cpu_to_be64(obj_size);
+		size_attr->attr = g_attr_logical_length;
+		size_attr->attr.val_ptr = &size_attr->newsize;
+
+		ORE_DBGMSG2("trunc(0x%llx) obj_offset=0x%llx dev=%d\n",
+			     _LLU(oc->comps->obj.id), _LLU(obj_size), i);
+		ret = _truncate_mirrors(ios, i * ios->layout->mirrors_p1,
+					&size_attr->attr);
+		if (unlikely(ret))
+			goto out;
+	}
+	ret = ore_io_execute(ios);
+
+out:
+	kfree(size_attrs);
+	ore_put_io_state(ios);
+	return ret;
+}
+EXPORT_SYMBOL(ore_truncate);
+
+const struct osd_attr g_attr_logical_length = ATTR_DEF(
+	OSD_APAGE_OBJECT_INFORMATION, OSD_ATTR_OI_LOGICAL_LENGTH, 8);
+EXPORT_SYMBOL(g_attr_logical_length);
diff --git a/fs/exofs/ore_raid.c b/fs/exofs/ore_raid.c
new file mode 100644
index 0000000..2c64826
--- /dev/null
+++ b/fs/exofs/ore_raid.c
@@ -0,0 +1,717 @@
+/*
+ * Copyright (C) 2011
+ * Boaz Harrosh <bharrosh@panasas.com>
+ *
+ * This file is part of the objects raid engine (ore).
+ *
+ * It is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with "ore". If not, write to the Free Software Foundation, Inc:
+ *	"Free Software Foundation <info@fsf.org>"
+ */
+
+#include <linux/gfp.h>
+#include <linux/async_tx.h>
+
+#include "ore_raid.h"
+
+#undef ORE_DBGMSG2
+#define ORE_DBGMSG2 ORE_DBGMSG
+
+struct page *_raid_page_alloc(void)
+{
+	return alloc_page(GFP_KERNEL);
+}
+
+void _raid_page_free(struct page *p)
+{
+	__free_page(p);
+}
+
+/* This struct is forward declare in ore_io_state, but is private to here.
+ * It is put on ios->sp2d for RAID5/6 writes only. See _gen_xor_unit.
+ *
+ * __stripe_pages_2d is a 2d array of pages, and it is also a corner turn.
+ * Ascending page index access is sp2d(p-minor, c-major). But storage is
+ * sp2d[p-minor][c-major], so it can be properlly presented to the async-xor
+ * API.
+ */
+struct __stripe_pages_2d {
+	/* Cache some hot path repeated calculations */
+	unsigned parity;
+	unsigned data_devs;
+	unsigned pages_in_unit;
+
+	bool needed ;
+
+	/* Array size is pages_in_unit (layout->stripe_unit / PAGE_SIZE) */
+	struct __1_page_stripe {
+		bool alloc;
+		unsigned write_count;
+		struct async_submit_ctl submit;
+		struct dma_async_tx_descriptor *tx;
+
+		/* The size of this array is data_devs + parity */
+		struct page **pages;
+		struct page **scribble;
+		/* bool array, size of this array is data_devs */
+		char *page_is_read;
+	} _1p_stripes[];
+};
+
+/* This can get bigger then a page. So support multiple page allocations
+ * _sp2d_free should be called even if _sp2d_alloc fails (by returning
+ * none-zero).
+ */
+static int _sp2d_alloc(unsigned pages_in_unit, unsigned group_width,
+		       unsigned parity, struct __stripe_pages_2d **psp2d)
+{
+	struct __stripe_pages_2d *sp2d;
+	unsigned data_devs = group_width - parity;
+	struct _alloc_all_bytes {
+		struct __alloc_stripe_pages_2d {
+			struct __stripe_pages_2d sp2d;
+			struct __1_page_stripe _1p_stripes[pages_in_unit];
+		} __asp2d;
+		struct __alloc_1p_arrays {
+			struct page *pages[group_width];
+			struct page *scribble[group_width];
+			char page_is_read[data_devs];
+		} __a1pa[pages_in_unit];
+	} *_aab;
+	struct __alloc_1p_arrays *__a1pa;
+	struct __alloc_1p_arrays *__a1pa_end;
+	const unsigned sizeof__a1pa = sizeof(_aab->__a1pa[0]);
+	unsigned num_a1pa, alloc_size, i;
+
+	/* FIXME: check these numbers in ore_verify_layout */
+	BUG_ON(sizeof(_aab->__asp2d) > PAGE_SIZE);
+	BUG_ON(sizeof__a1pa > PAGE_SIZE);
+
+	if (sizeof(*_aab) > PAGE_SIZE) {
+		num_a1pa = (PAGE_SIZE - sizeof(_aab->__asp2d)) / sizeof__a1pa;
+		alloc_size = sizeof(_aab->__asp2d) + sizeof__a1pa * num_a1pa;
+	} else {
+		num_a1pa = pages_in_unit;
+		alloc_size = sizeof(*_aab);
+	}
+
+	_aab = kzalloc(alloc_size, GFP_KERNEL);
+	if (unlikely(!_aab)) {
+		ORE_DBGMSG("!! Failed to alloc sp2d size=%d\n", alloc_size);
+		return -ENOMEM;
+	}
+
+	sp2d = &_aab->__asp2d.sp2d;
+	*psp2d = sp2d; /* From here Just call _sp2d_free */
+
+	__a1pa = _aab->__a1pa;
+	__a1pa_end = __a1pa + num_a1pa;
+
+	for (i = 0; i < pages_in_unit; ++i) {
+		if (unlikely(__a1pa >= __a1pa_end)) {
+			num_a1pa = min_t(unsigned, PAGE_SIZE / sizeof__a1pa,
+							pages_in_unit - i);
+
+			__a1pa = kzalloc(num_a1pa * sizeof__a1pa, GFP_KERNEL);
+			if (unlikely(!__a1pa)) {
+				ORE_DBGMSG("!! Failed to _alloc_1p_arrays=%d\n",
+					   num_a1pa);
+				return -ENOMEM;
+			}
+			__a1pa_end = __a1pa + num_a1pa;
+			/* First *pages is marked for kfree of the buffer */
+			sp2d->_1p_stripes[i].alloc = true;
+		}
+
+		sp2d->_1p_stripes[i].pages = __a1pa->pages;
+		sp2d->_1p_stripes[i].scribble = __a1pa->scribble ;
+		sp2d->_1p_stripes[i].page_is_read = __a1pa->page_is_read;
+		++__a1pa;
+	}
+
+	sp2d->parity = parity;
+	sp2d->data_devs = data_devs;
+	sp2d->pages_in_unit = pages_in_unit;
+	return 0;
+}
+
+static void _sp2d_reset(struct __stripe_pages_2d *sp2d,
+			const struct _ore_r4w_op *r4w, void *priv)
+{
+	unsigned data_devs = sp2d->data_devs;
+	unsigned group_width = data_devs + sp2d->parity;
+	unsigned p;
+
+	if (!sp2d->needed)
+		return;
+
+	for (p = 0; p < sp2d->pages_in_unit; p++) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+
+		if (_1ps->write_count < group_width) {
+			unsigned c;
+
+			for (c = 0; c < data_devs; c++)
+				if (_1ps->page_is_read[c]) {
+					struct page *page = _1ps->pages[c];
+
+					r4w->put_page(priv, page);
+					_1ps->page_is_read[c] = false;
+				}
+		}
+
+		memset(_1ps->pages, 0, group_width * sizeof(*_1ps->pages));
+		_1ps->write_count = 0;
+		_1ps->tx = NULL;
+	}
+
+	sp2d->needed = false;
+}
+
+static void _sp2d_free(struct __stripe_pages_2d *sp2d)
+{
+	unsigned i;
+
+	if (!sp2d)
+		return;
+
+	for (i = 0; i < sp2d->pages_in_unit; ++i) {
+		if (sp2d->_1p_stripes[i].alloc)
+			kfree(sp2d->_1p_stripes[i].pages);
+	}
+
+	kfree(sp2d);
+}
+
+static unsigned _sp2d_min_pg(struct __stripe_pages_2d *sp2d)
+{
+	unsigned p;
+
+	for (p = 0; p < sp2d->pages_in_unit; p++) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+
+		if (_1ps->write_count)
+			return p;
+	}
+
+	return ~0;
+}
+
+static unsigned _sp2d_max_pg(struct __stripe_pages_2d *sp2d)
+{
+	unsigned p;
+
+	for (p = sp2d->pages_in_unit - 1; p >= 0; --p) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+
+		if (_1ps->write_count)
+			return p;
+	}
+
+	return ~0;
+}
+
+static void _gen_xor_unit(struct __stripe_pages_2d *sp2d)
+{
+	unsigned p;
+	for (p = 0; p < sp2d->pages_in_unit; p++) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+
+		if (!_1ps->write_count)
+			continue;
+
+		init_async_submit(&_1ps->submit,
+			ASYNC_TX_XOR_ZERO_DST | ASYNC_TX_ACK,
+			NULL,
+			NULL, NULL,
+			(addr_conv_t *)_1ps->scribble);
+
+		/* TODO: raid6 */
+		_1ps->tx = async_xor(_1ps->pages[sp2d->data_devs], _1ps->pages,
+				     0, sp2d->data_devs, PAGE_SIZE,
+				     &_1ps->submit);
+	}
+
+	for (p = 0; p < sp2d->pages_in_unit; p++) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+		/* NOTE: We wait for HW synchronously (I don't have such HW
+		 * to test with.) Is parallelism needed with today's multi
+		 * cores?
+		 */
+		async_tx_issue_pending(_1ps->tx);
+	}
+}
+
+void _ore_add_stripe_page(struct __stripe_pages_2d *sp2d,
+		       struct ore_striping_info *si, struct page *page)
+{
+	struct __1_page_stripe *_1ps;
+
+	sp2d->needed = true;
+
+	_1ps = &sp2d->_1p_stripes[si->cur_pg];
+	_1ps->pages[si->cur_comp] = page;
+	++_1ps->write_count;
+
+	si->cur_pg = (si->cur_pg + 1) % sp2d->pages_in_unit;
+	/* si->cur_comp is advanced outside at main loop */
+}
+
+void _ore_add_sg_seg(struct ore_per_dev_state *per_dev, unsigned cur_len,
+		     bool not_last)
+{
+	struct osd_sg_entry *sge;
+
+	ORE_DBGMSG("dev=%d cur_len=0x%x not_last=%d cur_sg=%d "
+		     "offset=0x%llx length=0x%x last_sgs_total=0x%x\n",
+		     per_dev->dev, cur_len, not_last, per_dev->cur_sg,
+		     _LLU(per_dev->offset), per_dev->length,
+		     per_dev->last_sgs_total);
+
+	if (!per_dev->cur_sg) {
+		sge = per_dev->sglist;
+
+		/* First time we prepare two entries */
+		if (per_dev->length) {
+			++per_dev->cur_sg;
+			sge->offset = per_dev->offset;
+			sge->len = per_dev->length;
+		} else {
+			/* Here the parity is the first unit of this object.
+			 * This happens every time we reach a parity device on
+			 * the same stripe as the per_dev->offset. We need to
+			 * just skip this unit.
+			 */
+			per_dev->offset += cur_len;
+			return;
+		}
+	} else {
+		/* finalize the last one */
+		sge = &per_dev->sglist[per_dev->cur_sg - 1];
+		sge->len = per_dev->length - per_dev->last_sgs_total;
+	}
+
+	if (not_last) {
+		/* Partly prepare the next one */
+		struct osd_sg_entry *next_sge = sge + 1;
+
+		++per_dev->cur_sg;
+		next_sge->offset = sge->offset + sge->len + cur_len;
+		/* Save cur len so we know how mutch was added next time */
+		per_dev->last_sgs_total = per_dev->length;
+		next_sge->len = 0;
+	} else if (!sge->len) {
+		/* Optimize for when the last unit is a parity */
+		--per_dev->cur_sg;
+	}
+}
+
+static int _alloc_read_4_write(struct ore_io_state *ios)
+{
+	struct ore_layout *layout = ios->layout;
+	int ret;
+	/* We want to only read those pages not in cache so worst case
+	 * is a stripe populated with every other page
+	 */
+	unsigned sgs_per_dev = ios->sp2d->pages_in_unit + 2;
+
+	ret = _ore_get_io_state(layout, ios->oc,
+				layout->group_width * layout->mirrors_p1,
+				sgs_per_dev, 0, &ios->ios_read_4_write);
+	return ret;
+}
+
+/* @si contains info of the to-be-inserted page. Update of @si should be
+ * maintained by caller. Specificaly si->dev, si->obj_offset, ...
+ */
+static int _add_to_r4w(struct ore_io_state *ios, struct ore_striping_info *si,
+		       struct page *page, unsigned pg_len)
+{
+	struct request_queue *q;
+	struct ore_per_dev_state *per_dev;
+	struct ore_io_state *read_ios;
+	unsigned first_dev = si->dev - (si->dev %
+			  (ios->layout->group_width * ios->layout->mirrors_p1));
+	unsigned comp = si->dev - first_dev;
+	unsigned added_len;
+
+	if (!ios->ios_read_4_write) {
+		int ret = _alloc_read_4_write(ios);
+
+		if (unlikely(ret))
+			return ret;
+	}
+
+	read_ios = ios->ios_read_4_write;
+	read_ios->numdevs = ios->layout->group_width * ios->layout->mirrors_p1;
+
+	per_dev = &read_ios->per_dev[comp];
+	if (!per_dev->length) {
+		per_dev->bio = bio_kmalloc(GFP_KERNEL,
+					   ios->sp2d->pages_in_unit);
+		if (unlikely(!per_dev->bio)) {
+			ORE_DBGMSG("Failed to allocate BIO size=%u\n",
+				     ios->sp2d->pages_in_unit);
+			return -ENOMEM;
+		}
+		per_dev->offset = si->obj_offset;
+		per_dev->dev = si->dev;
+	} else if (si->obj_offset != (per_dev->offset + per_dev->length)) {
+		u64 gap = si->obj_offset - (per_dev->offset + per_dev->length);
+
+		_ore_add_sg_seg(per_dev, gap, true);
+	}
+	q = osd_request_queue(ore_comp_dev(read_ios->oc, per_dev->dev));
+	added_len = bio_add_pc_page(q, per_dev->bio, page, pg_len,
+				    si->obj_offset % PAGE_SIZE);
+	if (unlikely(added_len != pg_len)) {
+		ORE_DBGMSG("Failed to bio_add_pc_page bi_vcnt=%d\n",
+			      per_dev->bio->bi_vcnt);
+		return -ENOMEM;
+	}
+
+	per_dev->length += pg_len;
+	return 0;
+}
+
+/* read the beginning of an unaligned first page */
+static int _add_to_r4w_first_page(struct ore_io_state *ios, struct page *page)
+{
+	struct ore_striping_info si;
+	unsigned pg_len;
+
+	ore_calc_stripe_info(ios->layout, ios->offset, 0, &si);
+
+	pg_len = si.obj_offset % PAGE_SIZE;
+	si.obj_offset -= pg_len;
+
+	ORE_DBGMSG("offset=0x%llx len=0x%x index=0x%lx dev=%x\n",
+		   _LLU(si.obj_offset), pg_len, page->index, si.dev);
+
+	return _add_to_r4w(ios, &si, page, pg_len);
+}
+
+/* read the end of an incomplete last page */
+static int _add_to_r4w_last_page(struct ore_io_state *ios, u64 *offset)
+{
+	struct ore_striping_info si;
+	struct page *page;
+	unsigned pg_len, p, c;
+
+	ore_calc_stripe_info(ios->layout, *offset, 0, &si);
+
+	p = si.unit_off / PAGE_SIZE;
+	c = _dev_order(ios->layout->group_width * ios->layout->mirrors_p1,
+		       ios->layout->mirrors_p1, si.par_dev, si.dev);
+	page = ios->sp2d->_1p_stripes[p].pages[c];
+
+	pg_len = PAGE_SIZE - (si.unit_off % PAGE_SIZE);
+	*offset += pg_len;
+
+	ORE_DBGMSG("p=%d, c=%d next-offset=0x%llx len=0x%x dev=%x par_dev=%d\n",
+		   p, c, _LLU(*offset), pg_len, si.dev, si.par_dev);
+
+	BUG_ON(!page);
+
+	return _add_to_r4w(ios, &si, page, pg_len);
+}
+
+static void _mark_read4write_pages_uptodate(struct ore_io_state *ios, int ret)
+{
+	struct bio_vec *bv;
+	unsigned i, d;
+
+	/* loop on all devices all pages */
+	for (d = 0; d < ios->numdevs; d++) {
+		struct bio *bio = ios->per_dev[d].bio;
+
+		if (!bio)
+			continue;
+
+		bio_for_each_segment_all(bv, bio, i) {
+			struct page *page = bv->bv_page;
+
+			SetPageUptodate(page);
+			if (PageError(page))
+				ClearPageError(page);
+		}
+	}
+}
+
+/* read_4_write is hacked to read the start of the first stripe and/or
+ * the end of the last stripe. If needed, with an sg-gap at each device/page.
+ * It is assumed to be called after the to_be_written pages of the first stripe
+ * are populating ios->sp2d[][]
+ *
+ * NOTE: We call ios->r4w->lock_fn for all pages needed for parity calculations
+ * These pages are held at sp2d[p].pages[c] but with
+ * sp2d[p].page_is_read[c] = true. At _sp2d_reset these pages are
+ * ios->r4w->lock_fn(). The ios->r4w->lock_fn might signal that the page is
+ * @uptodate=true, so we don't need to read it, only unlock, after IO.
+ *
+ * TODO: The read_4_write should calc a need_to_read_pages_count, if bigger then
+ * to-be-written count, we should consider the xor-in-place mode.
+ * need_to_read_pages_count is the actual number of pages not present in cache.
+ * maybe "devs_in_group - ios->sp2d[p].write_count" is a good enough
+ * approximation? In this mode the read pages are put in the empty places of
+ * ios->sp2d[p][*], xor is calculated the same way. These pages are
+ * allocated/freed and don't go through cache
+ */
+static int _read_4_write_first_stripe(struct ore_io_state *ios)
+{
+	struct ore_striping_info read_si;
+	struct __stripe_pages_2d *sp2d = ios->sp2d;
+	u64 offset = ios->si.first_stripe_start;
+	unsigned c, p, min_p = sp2d->pages_in_unit, max_p = -1;
+
+	if (offset == ios->offset) /* Go to start collect $200 */
+		goto read_last_stripe;
+
+	min_p = _sp2d_min_pg(sp2d);
+	max_p = _sp2d_max_pg(sp2d);
+
+	ORE_DBGMSG("stripe_start=0x%llx ios->offset=0x%llx min_p=%d max_p=%d\n",
+		   offset, ios->offset, min_p, max_p);
+
+	for (c = 0; ; c++) {
+		ore_calc_stripe_info(ios->layout, offset, 0, &read_si);
+		read_si.obj_offset += min_p * PAGE_SIZE;
+		offset += min_p * PAGE_SIZE;
+		for (p = min_p; p <= max_p; p++) {
+			struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+			struct page **pp = &_1ps->pages[c];
+			bool uptodate;
+
+			if (*pp) {
+				if (ios->offset % PAGE_SIZE)
+					/* Read the remainder of the page */
+					_add_to_r4w_first_page(ios, *pp);
+				/* to-be-written pages start here */
+				goto read_last_stripe;
+			}
+
+			*pp = ios->r4w->get_page(ios->private, offset,
+						 &uptodate);
+			if (unlikely(!*pp))
+				return -ENOMEM;
+
+			if (!uptodate)
+				_add_to_r4w(ios, &read_si, *pp, PAGE_SIZE);
+
+			/* Mark read-pages to be cache_released */
+			_1ps->page_is_read[c] = true;
+			read_si.obj_offset += PAGE_SIZE;
+			offset += PAGE_SIZE;
+		}
+		offset += (sp2d->pages_in_unit - p) * PAGE_SIZE;
+	}
+
+read_last_stripe:
+	return 0;
+}
+
+static int _read_4_write_last_stripe(struct ore_io_state *ios)
+{
+	struct ore_striping_info read_si;
+	struct __stripe_pages_2d *sp2d = ios->sp2d;
+	u64 offset;
+	u64 last_stripe_end;
+	unsigned bytes_in_stripe = ios->si.bytes_in_stripe;
+	unsigned c, p, min_p = sp2d->pages_in_unit, max_p = -1;
+
+	offset = ios->offset + ios->length;
+	if (offset % PAGE_SIZE)
+		_add_to_r4w_last_page(ios, &offset);
+		/* offset will be aligned to next page */
+
+	last_stripe_end = div_u64(offset + bytes_in_stripe - 1, bytes_in_stripe)
+				 * bytes_in_stripe;
+	if (offset == last_stripe_end) /* Optimize for the aligned case */
+		goto read_it;
+
+	ore_calc_stripe_info(ios->layout, offset, 0, &read_si);
+	p = read_si.unit_off / PAGE_SIZE;
+	c = _dev_order(ios->layout->group_width * ios->layout->mirrors_p1,
+		       ios->layout->mirrors_p1, read_si.par_dev, read_si.dev);
+
+	if (min_p == sp2d->pages_in_unit) {
+		/* Didn't do it yet */
+		min_p = _sp2d_min_pg(sp2d);
+		max_p = _sp2d_max_pg(sp2d);
+	}
+
+	ORE_DBGMSG("offset=0x%llx stripe_end=0x%llx min_p=%d max_p=%d\n",
+		   offset, last_stripe_end, min_p, max_p);
+
+	while (offset < last_stripe_end) {
+		struct __1_page_stripe *_1ps = &sp2d->_1p_stripes[p];
+
+		if ((min_p <= p) && (p <= max_p)) {
+			struct page *page;
+			bool uptodate;
+
+			BUG_ON(_1ps->pages[c]);
+			page = ios->r4w->get_page(ios->private, offset,
+						  &uptodate);
+			if (unlikely(!page))
+				return -ENOMEM;
+
+			_1ps->pages[c] = page;
+			/* Mark read-pages to be cache_released */
+			_1ps->page_is_read[c] = true;
+			if (!uptodate)
+				_add_to_r4w(ios, &read_si, page, PAGE_SIZE);
+		}
+
+		offset += PAGE_SIZE;
+		if (p == (sp2d->pages_in_unit - 1)) {
+			++c;
+			p = 0;
+			ore_calc_stripe_info(ios->layout, offset, 0, &read_si);
+		} else {
+			read_si.obj_offset += PAGE_SIZE;
+			++p;
+		}
+	}
+
+read_it:
+	return 0;
+}
+
+static int _read_4_write_execute(struct ore_io_state *ios)
+{
+	struct ore_io_state *ios_read;
+	unsigned i;
+	int ret;
+
+	ios_read = ios->ios_read_4_write;
+	if (!ios_read)
+		return 0;
+
+	/* FIXME: Ugly to signal _sbi_read_mirror that we have bio(s). Change
+	 * to check for per_dev->bio
+	 */
+	ios_read->pages = ios->pages;
+
+	/* Now read these devices */
+	for (i = 0; i < ios_read->numdevs; i += ios_read->layout->mirrors_p1) {
+		ret = _ore_read_mirror(ios_read, i);
+		if (unlikely(ret))
+			return ret;
+	}
+
+	ret = ore_io_execute(ios_read); /* Synchronus execution */
+	if (unlikely(ret)) {
+		ORE_DBGMSG("!! ore_io_execute => %d\n", ret);
+		return ret;
+	}
+
+	_mark_read4write_pages_uptodate(ios_read, ret);
+	ore_put_io_state(ios_read);
+	ios->ios_read_4_write = NULL; /* Might need a reuse at last stripe */
+	return 0;
+}
+
+/* In writes @cur_len means length left. .i.e cur_len==0 is the last parity U */
+int _ore_add_parity_unit(struct ore_io_state *ios,
+			    struct ore_striping_info *si,
+			    struct ore_per_dev_state *per_dev,
+			    unsigned cur_len)
+{
+	if (ios->reading) {
+		if (per_dev->cur_sg >= ios->sgs_per_dev) {
+			ORE_DBGMSG("cur_sg(%d) >= sgs_per_dev(%d)\n" ,
+				per_dev->cur_sg, ios->sgs_per_dev);
+			return -ENOMEM;
+		}
+		_ore_add_sg_seg(per_dev, cur_len, true);
+	} else {
+		struct __stripe_pages_2d *sp2d = ios->sp2d;
+		struct page **pages = ios->parity_pages + ios->cur_par_page;
+		unsigned num_pages;
+		unsigned array_start = 0;
+		unsigned i;
+		int ret;
+
+		si->cur_pg = _sp2d_min_pg(sp2d);
+		num_pages  = _sp2d_max_pg(sp2d) + 1 - si->cur_pg;
+
+		if (!cur_len) /* If last stripe operate on parity comp */
+			si->cur_comp = sp2d->data_devs;
+
+		if (!per_dev->length) {
+			per_dev->offset += si->cur_pg * PAGE_SIZE;
+			/* If first stripe, Read in all read4write pages
+			 * (if needed) before we calculate the first parity.
+			 */
+			_read_4_write_first_stripe(ios);
+		}
+		if (!cur_len) /* If last stripe r4w pages of last stripe */
+			_read_4_write_last_stripe(ios);
+		_read_4_write_execute(ios);
+
+		for (i = 0; i < num_pages; i++) {
+			pages[i] = _raid_page_alloc();
+			if (unlikely(!pages[i]))
+				return -ENOMEM;
+
+			++(ios->cur_par_page);
+		}
+
+		BUG_ON(si->cur_comp != sp2d->data_devs);
+		BUG_ON(si->cur_pg + num_pages > sp2d->pages_in_unit);
+
+		ret = _ore_add_stripe_unit(ios,  &array_start, 0, pages,
+					   per_dev, num_pages * PAGE_SIZE);
+		if (unlikely(ret))
+			return ret;
+
+		/* TODO: raid6 if (last_parity_dev) */
+		_gen_xor_unit(sp2d);
+		_sp2d_reset(sp2d, ios->r4w, ios->private);
+	}
+	return 0;
+}
+
+int _ore_post_alloc_raid_stuff(struct ore_io_state *ios)
+{
+	if (ios->parity_pages) {
+		struct ore_layout *layout = ios->layout;
+		unsigned pages_in_unit = layout->stripe_unit / PAGE_SIZE;
+
+		if (_sp2d_alloc(pages_in_unit, layout->group_width,
+				layout->parity, &ios->sp2d)) {
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+void _ore_free_raid_stuff(struct ore_io_state *ios)
+{
+	if (ios->sp2d) { /* writing and raid */
+		unsigned i;
+
+		for (i = 0; i < ios->cur_par_page; i++) {
+			struct page *page = ios->parity_pages[i];
+
+			if (page)
+				_raid_page_free(page);
+		}
+		if (ios->extra_part_alloc)
+			kfree(ios->parity_pages);
+		/* If IO returned an error pages might need unlocking */
+		_sp2d_reset(ios->sp2d, ios->r4w, ios->private);
+		_sp2d_free(ios->sp2d);
+	} else {
+		/* Will only be set if raid reading && sglist is big */
+		if (ios->extra_part_alloc)
+			kfree(ios->per_dev[0].sglist);
+	}
+	if (ios->ios_read_4_write)
+		ore_put_io_state(ios->ios_read_4_write);
+}
diff --git a/fs/exofs/ore_raid.h b/fs/exofs/ore_raid.h
new file mode 100644
index 0000000..2ffd2c3
--- /dev/null
+++ b/fs/exofs/ore_raid.h
@@ -0,0 +1,79 @@
+/*
+ * Copyright (C) from 2011
+ * Boaz Harrosh <bharrosh@panasas.com>
+ *
+ * This file is part of the objects raid engine (ore).
+ *
+ * It is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with "ore". If not, write to the Free Software Foundation, Inc:
+ *	"Free Software Foundation <info@fsf.org>"
+ */
+
+#include <scsi/osd_ore.h>
+
+#define ORE_ERR(fmt, a...) printk(KERN_ERR "ore: " fmt, ##a)
+
+#ifdef CONFIG_EXOFS_DEBUG
+#define ORE_DBGMSG(fmt, a...) \
+	printk(KERN_NOTICE "ore @%s:%d: " fmt, __func__, __LINE__, ##a)
+#else
+#define ORE_DBGMSG(fmt, a...) \
+	do { if (0) printk(fmt, ##a); } while (0)
+#endif
+
+/* u64 has problems with printk this will cast it to unsigned long long */
+#define _LLU(x) (unsigned long long)(x)
+
+#define ORE_DBGMSG2(M...) do {} while (0)
+/* #define ORE_DBGMSG2 ORE_DBGMSG */
+
+/* Calculate the component order in a stripe. eg the logical data unit
+ * address within the stripe of @dev given the @par_dev of this stripe.
+ */
+static inline unsigned _dev_order(unsigned devs_in_group, unsigned mirrors_p1,
+				  unsigned par_dev, unsigned dev)
+{
+	unsigned first_dev = dev - dev % devs_in_group;
+
+	dev -= first_dev;
+	par_dev -= first_dev;
+
+	if (devs_in_group == par_dev) /* The raid 0 case */
+		return dev / mirrors_p1;
+	/* raid4/5/6 case */
+	return ((devs_in_group + dev - par_dev - mirrors_p1) % devs_in_group) /
+	       mirrors_p1;
+}
+
+/* ios_raid.c stuff needed by ios.c */
+int _ore_post_alloc_raid_stuff(struct ore_io_state *ios);
+void _ore_free_raid_stuff(struct ore_io_state *ios);
+
+void _ore_add_sg_seg(struct ore_per_dev_state *per_dev, unsigned cur_len,
+		 bool not_last);
+int _ore_add_parity_unit(struct ore_io_state *ios, struct ore_striping_info *si,
+		     struct ore_per_dev_state *per_dev, unsigned cur_len);
+void _ore_add_stripe_page(struct __stripe_pages_2d *sp2d,
+		       struct ore_striping_info *si, struct page *page);
+static inline void _add_stripe_page(struct __stripe_pages_2d *sp2d,
+				struct ore_striping_info *si, struct page *page)
+{
+	if (!sp2d) /* Inline the fast path */
+		return; /* Hay no raid stuff */
+	_ore_add_stripe_page(sp2d, si, page);
+}
+
+/* ios.c stuff needed by ios_raid.c */
+int  _ore_get_io_state(struct ore_layout *layout,
+			struct ore_components *oc, unsigned numdevs,
+			unsigned sgs_per_dev, unsigned num_par_pages,
+			struct ore_io_state **pios);
+int _ore_add_stripe_unit(struct ore_io_state *ios,  unsigned *cur_pg,
+		unsigned pgbase, struct page **pages,
+		struct ore_per_dev_state *per_dev, int cur_len);
+int _ore_read_mirror(struct ore_io_state *ios, unsigned cur_comp);
+int ore_io_execute(struct ore_io_state *ios);
diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c
new file mode 100644
index 0000000..ae1425a
--- /dev/null
+++ b/fs/ext4/indirect.c
@@ -0,0 +1,1506 @@
+/*
+ *  linux/fs/ext4/indirect.c
+ *
+ *  from
+ *
+ *  linux/fs/ext4/inode.c
+ *
+ * Copyright (C) 1992, 1993, 1994, 1995
+ * Remy Card (card@masi.ibp.fr)
+ * Laboratoire MASI - Institut Blaise Pascal
+ * Universite Pierre et Marie Curie (Paris VI)
+ *
+ *  from
+ *
+ *  linux/fs/minix/inode.c
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ *
+ *  Goal-directed block allocation by Stephen Tweedie
+ *	(sct@redhat.com), 1993, 1998
+ */
+
+#include <linux/module.h>
+#include "ext4_jbd2.h"
+#include "truncate.h"
+
+#include <trace/events/ext4.h>
+
+typedef struct {
+	__le32	*p;
+	__le32	key;
+	struct buffer_head *bh;
+} Indirect;
+
+static inline void add_chain(Indirect *p, struct buffer_head *bh, __le32 *v)
+{
+	p->key = *(p->p = v);
+	p->bh = bh;
+}
+
+/**
+ *	ext4_block_to_path - parse the block number into array of offsets
+ *	@inode: inode in question (we are only interested in its superblock)
+ *	@i_block: block number to be parsed
+ *	@offsets: array to store the offsets in
+ *	@boundary: set this non-zero if the referred-to block is likely to be
+ *	       followed (on disk) by an indirect block.
+ *
+ *	To store the locations of file's data ext4 uses a data structure common
+ *	for UNIX filesystems - tree of pointers anchored in the inode, with
+ *	data blocks at leaves and indirect blocks in intermediate nodes.
+ *	This function translates the block number into path in that tree -
+ *	return value is the path length and @offsets[n] is the offset of
+ *	pointer to (n+1)th node in the nth one. If @block is out of range
+ *	(negative or too large) warning is printed and zero returned.
+ *
+ *	Note: function doesn't find node addresses, so no IO is needed. All
+ *	we need to know is the capacity of indirect blocks (taken from the
+ *	inode->i_sb).
+ */
+
+/*
+ * Portability note: the last comparison (check that we fit into triple
+ * indirect block) is spelled differently, because otherwise on an
+ * architecture with 32-bit longs and 8Kb pages we might get into trouble
+ * if our filesystem had 8Kb blocks. We might use long long, but that would
+ * kill us on x86. Oh, well, at least the sign propagation does not matter -
+ * i_block would have to be negative in the very beginning, so we would not
+ * get there at all.
+ */
+
+static int ext4_block_to_path(struct inode *inode,
+			      ext4_lblk_t i_block,
+			      ext4_lblk_t offsets[4], int *boundary)
+{
+	int ptrs = EXT4_ADDR_PER_BLOCK(inode->i_sb);
+	int ptrs_bits = EXT4_ADDR_PER_BLOCK_BITS(inode->i_sb);
+	const long direct_blocks = EXT4_NDIR_BLOCKS,
+		indirect_blocks = ptrs,
+		double_blocks = (1 << (ptrs_bits * 2));
+	int n = 0;
+	int final = 0;
+
+	if (i_block < direct_blocks) {
+		offsets[n++] = i_block;
+		final = direct_blocks;
+	} else if ((i_block -= direct_blocks) < indirect_blocks) {
+		offsets[n++] = EXT4_IND_BLOCK;
+		offsets[n++] = i_block;
+		final = ptrs;
+	} else if ((i_block -= indirect_blocks) < double_blocks) {
+		offsets[n++] = EXT4_DIND_BLOCK;
+		offsets[n++] = i_block >> ptrs_bits;
+		offsets[n++] = i_block & (ptrs - 1);
+		final = ptrs;
+	} else if (((i_block -= double_blocks) >> (ptrs_bits * 2)) < ptrs) {
+		offsets[n++] = EXT4_TIND_BLOCK;
+		offsets[n++] = i_block >> (ptrs_bits * 2);
+		offsets[n++] = (i_block >> ptrs_bits) & (ptrs - 1);
+		offsets[n++] = i_block & (ptrs - 1);
+		final = ptrs;
+	} else {
+		ext4_warning(inode->i_sb, "block %lu > max in inode %lu",
+			     i_block + direct_blocks +
+			     indirect_blocks + double_blocks, inode->i_ino);
+	}
+	if (boundary)
+		*boundary = final - 1 - (i_block & (ptrs - 1));
+	return n;
+}
+
+/**
+ *	ext4_get_branch - read the chain of indirect blocks leading to data
+ *	@inode: inode in question
+ *	@depth: depth of the chain (1 - direct pointer, etc.)
+ *	@offsets: offsets of pointers in inode/indirect blocks
+ *	@chain: place to store the result
+ *	@err: here we store the error value
+ *
+ *	Function fills the array of triples <key, p, bh> and returns %NULL
+ *	if everything went OK or the pointer to the last filled triple
+ *	(incomplete one) otherwise. Upon the return chain[i].key contains
+ *	the number of (i+1)-th block in the chain (as it is stored in memory,
+ *	i.e. little-endian 32-bit), chain[i].p contains the address of that
+ *	number (it points into struct inode for i==0 and into the bh->b_data
+ *	for i>0) and chain[i].bh points to the buffer_head of i-th indirect
+ *	block for i>0 and NULL for i==0. In other words, it holds the block
+ *	numbers of the chain, addresses they were taken from (and where we can
+ *	verify that chain did not change) and buffer_heads hosting these
+ *	numbers.
+ *
+ *	Function stops when it stumbles upon zero pointer (absent block)
+ *		(pointer to last triple returned, *@err == 0)
+ *	or when it gets an IO error reading an indirect block
+ *		(ditto, *@err == -EIO)
+ *	or when it reads all @depth-1 indirect blocks successfully and finds
+ *	the whole chain, all way to the data (returns %NULL, *err == 0).
+ *
+ *      Need to be called with
+ *      down_read(&EXT4_I(inode)->i_data_sem)
+ */
+static Indirect *ext4_get_branch(struct inode *inode, int depth,
+				 ext4_lblk_t  *offsets,
+				 Indirect chain[4], int *err)
+{
+	struct super_block *sb = inode->i_sb;
+	Indirect *p = chain;
+	struct buffer_head *bh;
+	int ret = -EIO;
+
+	*err = 0;
+	/* i_data is not going away, no lock needed */
+	add_chain(chain, NULL, EXT4_I(inode)->i_data + *offsets);
+	if (!p->key)
+		goto no_block;
+	while (--depth) {
+		bh = sb_getblk(sb, le32_to_cpu(p->key));
+		if (unlikely(!bh)) {
+			ret = -ENOMEM;
+			goto failure;
+		}
+
+		if (!bh_uptodate_or_lock(bh)) {
+			if (bh_submit_read(bh) < 0) {
+				put_bh(bh);
+				goto failure;
+			}
+			/* validate block references */
+			if (ext4_check_indirect_blockref(inode, bh)) {
+				put_bh(bh);
+				goto failure;
+			}
+		}
+
+		add_chain(++p, bh, (__le32 *)bh->b_data + *++offsets);
+		/* Reader: end */
+		if (!p->key)
+			goto no_block;
+	}
+	return NULL;
+
+failure:
+	*err = ret;
+no_block:
+	return p;
+}
+
+/**
+ *	ext4_find_near - find a place for allocation with sufficient locality
+ *	@inode: owner
+ *	@ind: descriptor of indirect block.
+ *
+ *	This function returns the preferred place for block allocation.
+ *	It is used when heuristic for sequential allocation fails.
+ *	Rules are:
+ *	  + if there is a block to the left of our position - allocate near it.
+ *	  + if pointer will live in indirect block - allocate near that block.
+ *	  + if pointer will live in inode - allocate in the same
+ *	    cylinder group.
+ *
+ * In the latter case we colour the starting block by the callers PID to
+ * prevent it from clashing with concurrent allocations for a different inode
+ * in the same block group.   The PID is used here so that functionally related
+ * files will be close-by on-disk.
+ *
+ *	Caller must make sure that @ind is valid and will stay that way.
+ */
+static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	__le32 *start = ind->bh ? (__le32 *) ind->bh->b_data : ei->i_data;
+	__le32 *p;
+
+	/* Try to find previous block */
+	for (p = ind->p - 1; p >= start; p--) {
+		if (*p)
+			return le32_to_cpu(*p);
+	}
+
+	/* No such thing, so let's try location of indirect block */
+	if (ind->bh)
+		return ind->bh->b_blocknr;
+
+	/*
+	 * It is going to be referred to from the inode itself? OK, just put it
+	 * into the same cylinder group then.
+	 */
+	return ext4_inode_to_goal_block(inode);
+}
+
+/**
+ *	ext4_find_goal - find a preferred place for allocation.
+ *	@inode: owner
+ *	@block:  block we want
+ *	@partial: pointer to the last triple within a chain
+ *
+ *	Normally this function find the preferred place for block allocation,
+ *	returns it.
+ *	Because this is only used for non-extent files, we limit the block nr
+ *	to 32 bits.
+ */
+static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
+				   Indirect *partial)
+{
+	ext4_fsblk_t goal;
+
+	/*
+	 * XXX need to get goal block from mballoc's data structures
+	 */
+
+	goal = ext4_find_near(inode, partial);
+	goal = goal & EXT4_MAX_BLOCK_FILE_PHYS;
+	return goal;
+}
+
+/**
+ *	ext4_blks_to_allocate - Look up the block map and count the number
+ *	of direct blocks need to be allocated for the given branch.
+ *
+ *	@branch: chain of indirect blocks
+ *	@k: number of blocks need for indirect blocks
+ *	@blks: number of data blocks to be mapped.
+ *	@blocks_to_boundary:  the offset in the indirect block
+ *
+ *	return the total number of blocks to be allocate, including the
+ *	direct and indirect blocks.
+ */
+static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
+				 int blocks_to_boundary)
+{
+	unsigned int count = 0;
+
+	/*
+	 * Simple case, [t,d]Indirect block(s) has not allocated yet
+	 * then it's clear blocks on that path have not allocated
+	 */
+	if (k > 0) {
+		/* right now we don't handle cross boundary allocation */
+		if (blks < blocks_to_boundary + 1)
+			count += blks;
+		else
+			count += blocks_to_boundary + 1;
+		return count;
+	}
+
+	count++;
+	while (count < blks && count <= blocks_to_boundary &&
+		le32_to_cpu(*(branch[0].p + count)) == 0) {
+		count++;
+	}
+	return count;
+}
+
+/**
+ *	ext4_alloc_blocks: multiple allocate blocks needed for a branch
+ *	@handle: handle for this transaction
+ *	@inode: inode which needs allocated blocks
+ *	@iblock: the logical block to start allocated at
+ *	@goal: preferred physical block of allocation
+ *	@indirect_blks: the number of blocks need to allocate for indirect
+ *			blocks
+ *	@blks: number of desired blocks
+ *	@new_blocks: on return it will store the new block numbers for
+ *	the indirect blocks(if needed) and the first direct block,
+ *	@err: on return it will store the error code
+ *
+ *	This function will return the number of blocks allocated as
+ *	requested by the passed-in parameters.
+ */
+static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
+			     ext4_lblk_t iblock, ext4_fsblk_t goal,
+			     int indirect_blks, int blks,
+			     ext4_fsblk_t new_blocks[4], int *err)
+{
+	struct ext4_allocation_request ar;
+	int target, i;
+	unsigned long count = 0, blk_allocated = 0;
+	int index = 0;
+	ext4_fsblk_t current_block = 0;
+	int ret = 0;
+
+	/*
+	 * Here we try to allocate the requested multiple blocks at once,
+	 * on a best-effort basis.
+	 * To build a branch, we should allocate blocks for
+	 * the indirect blocks(if not allocated yet), and at least
+	 * the first direct block of this branch.  That's the
+	 * minimum number of blocks need to allocate(required)
+	 */
+	/* first we try to allocate the indirect blocks */
+	target = indirect_blks;
+	while (target > 0) {
+		count = target;
+		/* allocating blocks for indirect blocks and direct blocks */
+		current_block = ext4_new_meta_blocks(handle, inode, goal,
+						     0, &count, err);
+		if (*err)
+			goto failed_out;
+
+		if (unlikely(current_block + count > EXT4_MAX_BLOCK_FILE_PHYS)) {
+			EXT4_ERROR_INODE(inode,
+					 "current_block %llu + count %lu > %d!",
+					 current_block, count,
+					 EXT4_MAX_BLOCK_FILE_PHYS);
+			*err = -EIO;
+			goto failed_out;
+		}
+
+		target -= count;
+		/* allocate blocks for indirect blocks */
+		while (index < indirect_blks && count) {
+			new_blocks[index++] = current_block++;
+			count--;
+		}
+		if (count > 0) {
+			/*
+			 * save the new block number
+			 * for the first direct block
+			 */
+			new_blocks[index] = current_block;
+			printk(KERN_INFO "%s returned more blocks than "
+						"requested\n", __func__);
+			WARN_ON(1);
+			break;
+		}
+	}
+
+	target = blks - count ;
+	blk_allocated = count;
+	if (!target)
+		goto allocated;
+	/* Now allocate data blocks */
+	memset(&ar, 0, sizeof(ar));
+	ar.inode = inode;
+	ar.goal = goal;
+	ar.len = target;
+	ar.logical = iblock;
+	if (S_ISREG(inode->i_mode))
+		/* enable in-core preallocation only for regular files */
+		ar.flags = EXT4_MB_HINT_DATA;
+
+	current_block = ext4_mb_new_blocks(handle, &ar, err);
+	if (unlikely(current_block + ar.len > EXT4_MAX_BLOCK_FILE_PHYS)) {
+		EXT4_ERROR_INODE(inode,
+				 "current_block %llu + ar.len %d > %d!",
+				 current_block, ar.len,
+				 EXT4_MAX_BLOCK_FILE_PHYS);
+		*err = -EIO;
+		goto failed_out;
+	}
+
+	if (*err && (target == blks)) {
+		/*
+		 * if the allocation failed and we didn't allocate
+		 * any blocks before
+		 */
+		goto failed_out;
+	}
+	if (!*err) {
+		if (target == blks) {
+			/*
+			 * save the new block number
+			 * for the first direct block
+			 */
+			new_blocks[index] = current_block;
+		}
+		blk_allocated += ar.len;
+	}
+allocated:
+	/* total number of blocks allocated for direct blocks */
+	ret = blk_allocated;
+	*err = 0;
+	return ret;
+failed_out:
+	for (i = 0; i < index; i++)
+		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1, 0);
+	return ret;
+}
+
+/**
+ *	ext4_alloc_branch - allocate and set up a chain of blocks.
+ *	@handle: handle for this transaction
+ *	@inode: owner
+ *	@indirect_blks: number of allocated indirect blocks
+ *	@blks: number of allocated direct blocks
+ *	@goal: preferred place for allocation
+ *	@offsets: offsets (in the blocks) to store the pointers to next.
+ *	@branch: place to store the chain in.
+ *
+ *	This function allocates blocks, zeroes out all but the last one,
+ *	links them into chain and (if we are synchronous) writes them to disk.
+ *	In other words, it prepares a branch that can be spliced onto the
+ *	inode. It stores the information about that chain in the branch[], in
+ *	the same format as ext4_get_branch() would do. We are calling it after
+ *	we had read the existing part of chain and partial points to the last
+ *	triple of that (one with zero ->key). Upon the exit we have the same
+ *	picture as after the successful ext4_get_block(), except that in one
+ *	place chain is disconnected - *branch->p is still zero (we did not
+ *	set the last link), but branch->key contains the number that should
+ *	be placed into *branch->p to fill that gap.
+ *
+ *	If allocation fails we free all blocks we've allocated (and forget
+ *	their buffer_heads) and return the error value the from failed
+ *	ext4_alloc_block() (normally -ENOSPC). Otherwise we set the chain
+ *	as described above and return 0.
+ */
+static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
+			     ext4_lblk_t iblock, int indirect_blks,
+			     int *blks, ext4_fsblk_t goal,
+			     ext4_lblk_t *offsets, Indirect *branch)
+{
+	int blocksize = inode->i_sb->s_blocksize;
+	int i, n = 0;
+	int err = 0;
+	struct buffer_head *bh;
+	int num;
+	ext4_fsblk_t new_blocks[4];
+	ext4_fsblk_t current_block;
+
+	num = ext4_alloc_blocks(handle, inode, iblock, goal, indirect_blks,
+				*blks, new_blocks, &err);
+	if (err)
+		return err;
+
+	branch[0].key = cpu_to_le32(new_blocks[0]);
+	/*
+	 * metadata blocks and data blocks are allocated.
+	 */
+	for (n = 1; n <= indirect_blks;  n++) {
+		/*
+		 * Get buffer_head for parent block, zero it out
+		 * and set the pointer to new one, then send
+		 * parent to disk.
+		 */
+		bh = sb_getblk(inode->i_sb, new_blocks[n-1]);
+		if (unlikely(!bh)) {
+			err = -ENOMEM;
+			goto failed;
+		}
+
+		branch[n].bh = bh;
+		lock_buffer(bh);
+		BUFFER_TRACE(bh, "call get_create_access");
+		err = ext4_journal_get_create_access(handle, bh);
+		if (err) {
+			/* Don't brelse(bh) here; it's done in
+			 * ext4_journal_forget() below */
+			unlock_buffer(bh);
+			goto failed;
+		}
+
+		memset(bh->b_data, 0, blocksize);
+		branch[n].p = (__le32 *) bh->b_data + offsets[n];
+		branch[n].key = cpu_to_le32(new_blocks[n]);
+		*branch[n].p = branch[n].key;
+		if (n == indirect_blks) {
+			current_block = new_blocks[n];
+			/*
+			 * End of chain, update the last new metablock of
+			 * the chain to point to the new allocated
+			 * data blocks numbers
+			 */
+			for (i = 1; i < num; i++)
+				*(branch[n].p + i) = cpu_to_le32(++current_block);
+		}
+		BUFFER_TRACE(bh, "marking uptodate");
+		set_buffer_uptodate(bh);
+		unlock_buffer(bh);
+
+		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
+		if (err)
+			goto failed;
+	}
+	*blks = num;
+	return err;
+failed:
+	/* Allocation failed, free what we already allocated */
+	ext4_free_blocks(handle, inode, NULL, new_blocks[0], 1, 0);
+	for (i = 1; i <= n ; i++) {
+		/*
+		 * branch[i].bh is newly allocated, so there is no
+		 * need to revoke the block, which is why we don't
+		 * need to set EXT4_FREE_BLOCKS_METADATA.
+		 */
+		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1,
+				 EXT4_FREE_BLOCKS_FORGET);
+	}
+	for (i = n+1; i < indirect_blks; i++)
+		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1, 0);
+
+	ext4_free_blocks(handle, inode, NULL, new_blocks[i], num, 0);
+
+	return err;
+}
+
+/**
+ * ext4_splice_branch - splice the allocated branch onto inode.
+ * @handle: handle for this transaction
+ * @inode: owner
+ * @block: (logical) number of block we are adding
+ * @chain: chain of indirect blocks (with a missing link - see
+ *	ext4_alloc_branch)
+ * @where: location of missing link
+ * @num:   number of indirect blocks we are adding
+ * @blks:  number of direct blocks we are adding
+ *
+ * This function fills the missing link and does all housekeeping needed in
+ * inode (->i_blocks, etc.). In case of success we end up with the full
+ * chain to new block and return 0.
+ */
+static int ext4_splice_branch(handle_t *handle, struct inode *inode,
+			      ext4_lblk_t block, Indirect *where, int num,
+			      int blks)
+{
+	int i;
+	int err = 0;
+	ext4_fsblk_t current_block;
+
+	/*
+	 * If we're splicing into a [td]indirect block (as opposed to the
+	 * inode) then we need to get write access to the [td]indirect block
+	 * before the splice.
+	 */
+	if (where->bh) {
+		BUFFER_TRACE(where->bh, "get_write_access");
+		err = ext4_journal_get_write_access(handle, where->bh);
+		if (err)
+			goto err_out;
+	}
+	/* That's it */
+
+	*where->p = where->key;
+
+	/*
+	 * Update the host buffer_head or inode to point to more just allocated
+	 * direct blocks blocks
+	 */
+	if (num == 0 && blks > 1) {
+		current_block = le32_to_cpu(where->key) + 1;
+		for (i = 1; i < blks; i++)
+			*(where->p + i) = cpu_to_le32(current_block++);
+	}
+
+	/* We are done with atomic stuff, now do the rest of housekeeping */
+	/* had we spliced it onto indirect block? */
+	if (where->bh) {
+		/*
+		 * If we spliced it onto an indirect block, we haven't
+		 * altered the inode.  Note however that if it is being spliced
+		 * onto an indirect block at the very end of the file (the
+		 * file is growing) then we *will* alter the inode to reflect
+		 * the new i_size.  But that is not done here - it is done in
+		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
+		 */
+		jbd_debug(5, "splicing indirect only\n");
+		BUFFER_TRACE(where->bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, where->bh);
+		if (err)
+			goto err_out;
+	} else {
+		/*
+		 * OK, we spliced it into the inode itself on a direct block.
+		 */
+		ext4_mark_inode_dirty(handle, inode);
+		jbd_debug(5, "splicing direct\n");
+	}
+	return err;
+
+err_out:
+	for (i = 1; i <= num; i++) {
+		/*
+		 * branch[i].bh is newly allocated, so there is no
+		 * need to revoke the block, which is why we don't
+		 * need to set EXT4_FREE_BLOCKS_METADATA.
+		 */
+		ext4_free_blocks(handle, inode, where[i].bh, 0, 1,
+				 EXT4_FREE_BLOCKS_FORGET);
+	}
+	ext4_free_blocks(handle, inode, NULL, le32_to_cpu(where[num].key),
+			 blks, 0);
+
+	return err;
+}
+
+/*
+ * The ext4_ind_map_blocks() function handles non-extents inodes
+ * (i.e., using the traditional indirect/double-indirect i_blocks
+ * scheme) for ext4_map_blocks().
+ *
+ * Allocation strategy is simple: if we have to allocate something, we will
+ * have to go the whole way to leaf. So let's do it before attaching anything
+ * to tree, set linkage between the newborn blocks, write them if sync is
+ * required, recheck the path, free and repeat if check fails, otherwise
+ * set the last missing link (that will protect us from any truncate-generated
+ * removals - all blocks on the path are immune now) and possibly force the
+ * write on the parent block.
+ * That has a nice additional property: no special recovery from the failed
+ * allocations is needed - we simply release blocks and do not touch anything
+ * reachable from inode.
+ *
+ * `handle' can be NULL if create == 0.
+ *
+ * return > 0, # of blocks mapped or allocated.
+ * return = 0, if plain lookup failed.
+ * return < 0, error case.
+ *
+ * The ext4_ind_get_blocks() function should be called with
+ * down_write(&EXT4_I(inode)->i_data_sem) if allocating filesystem
+ * blocks (i.e., flags has EXT4_GET_BLOCKS_CREATE set) or
+ * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system
+ * blocks.
+ */
+int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
+			struct ext4_map_blocks *map,
+			int flags)
+{
+	int err = -EIO;
+	ext4_lblk_t offsets[4];
+	Indirect chain[4];
+	Indirect *partial;
+	ext4_fsblk_t goal;
+	int indirect_blks;
+	int blocks_to_boundary = 0;
+	int depth;
+	int count = 0;
+	ext4_fsblk_t first_block = 0;
+
+	trace_ext4_ind_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
+	J_ASSERT(!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)));
+	J_ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
+	depth = ext4_block_to_path(inode, map->m_lblk, offsets,
+				   &blocks_to_boundary);
+
+	if (depth == 0)
+		goto out;
+
+	partial = ext4_get_branch(inode, depth, offsets, chain, &err);
+
+	/* Simplest case - block found, no allocation needed */
+	if (!partial) {
+		first_block = le32_to_cpu(chain[depth - 1].key);
+		count++;
+		/*map more blocks*/
+		while (count < map->m_len && count <= blocks_to_boundary) {
+			ext4_fsblk_t blk;
+
+			blk = le32_to_cpu(*(chain[depth-1].p + count));
+
+			if (blk == first_block + count)
+				count++;
+			else
+				break;
+		}
+		goto got_it;
+	}
+
+	/* Next simple case - plain lookup or failed read of indirect block */
+	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0 || err == -EIO)
+		goto cleanup;
+
+	/*
+	 * Okay, we need to do block allocation.
+	*/
+	if (EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
+				       EXT4_FEATURE_RO_COMPAT_BIGALLOC)) {
+		EXT4_ERROR_INODE(inode, "Can't allocate blocks for "
+				 "non-extent mapped inodes with bigalloc");
+		return -EUCLEAN;
+	}
+
+	goal = ext4_find_goal(inode, map->m_lblk, partial);
+
+	/* the number of blocks need to allocate for [d,t]indirect blocks */
+	indirect_blks = (chain + depth) - partial - 1;
+
+	/*
+	 * Next look up the indirect map to count the totoal number of
+	 * direct blocks to allocate for this branch.
+	 */
+	count = ext4_blks_to_allocate(partial, indirect_blks,
+				      map->m_len, blocks_to_boundary);
+	/*
+	 * Block out ext4_truncate while we alter the tree
+	 */
+	err = ext4_alloc_branch(handle, inode, map->m_lblk, indirect_blks,
+				&count, goal,
+				offsets + (partial - chain), partial);
+
+	/*
+	 * The ext4_splice_branch call will free and forget any buffers
+	 * on the new chain if there is a failure, but that risks using
+	 * up transaction credits, especially for bitmaps where the
+	 * credits cannot be returned.  Can we handle this somehow?  We
+	 * may need to return -EAGAIN upwards in the worst case.  --sct
+	 */
+	if (!err)
+		err = ext4_splice_branch(handle, inode, map->m_lblk,
+					 partial, indirect_blks, count);
+	if (err)
+		goto cleanup;
+
+	map->m_flags |= EXT4_MAP_NEW;
+
+	ext4_update_inode_fsync_trans(handle, inode, 1);
+got_it:
+	map->m_flags |= EXT4_MAP_MAPPED;
+	map->m_pblk = le32_to_cpu(chain[depth-1].key);
+	map->m_len = count;
+	if (count > blocks_to_boundary)
+		map->m_flags |= EXT4_MAP_BOUNDARY;
+	err = count;
+	/* Clean up and exit */
+	partial = chain + depth - 1;	/* the whole chain */
+cleanup:
+	while (partial > chain) {
+		BUFFER_TRACE(partial->bh, "call brelse");
+		brelse(partial->bh);
+		partial--;
+	}
+out:
+	trace_ext4_ind_map_blocks_exit(inode, map->m_lblk,
+				map->m_pblk, map->m_len, err);
+	return err;
+}
+
+/*
+ * O_DIRECT for ext3 (or indirect map) based files
+ *
+ * If the O_DIRECT write will extend the file then add this inode to the
+ * orphan list.  So recovery will truncate it back to the original size
+ * if the machine crashes during the write.
+ *
+ * If the O_DIRECT write is intantiating holes inside i_size and the machine
+ * crashes then stale disk data _may_ be exposed inside the file. But current
+ * VFS code falls back into buffered path in that case so we are safe.
+ */
+ssize_t ext4_ind_direct_IO(int rw, struct kiocb *iocb,
+			   const struct iovec *iov, loff_t offset,
+			   unsigned long nr_segs)
+{
+	struct file *file = iocb->ki_filp;
+	struct inode *inode = file->f_mapping->host;
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	handle_t *handle;
+	ssize_t ret;
+	int orphan = 0;
+	size_t count = iov_length(iov, nr_segs);
+	int retries = 0;
+
+	if (rw == WRITE) {
+		loff_t final_size = offset + count;
+
+		if (final_size > inode->i_size) {
+			/* Credits for sb + inode write */
+			handle = ext4_journal_start(inode, 2);
+			if (IS_ERR(handle)) {
+				ret = PTR_ERR(handle);
+				goto out;
+			}
+			ret = ext4_orphan_add(handle, inode);
+			if (ret) {
+				ext4_journal_stop(handle);
+				goto out;
+			}
+			orphan = 1;
+			ei->i_disksize = inode->i_size;
+			ext4_journal_stop(handle);
+		}
+	}
+
+retry:
+	if (rw == READ && ext4_should_dioread_nolock(inode)) {
+		if (unlikely(!list_empty(&ei->i_completed_io_list))) {
+			mutex_lock(&inode->i_mutex);
+			ext4_flush_completed_IO(inode);
+			mutex_unlock(&inode->i_mutex);
+		}
+		ret = __blockdev_direct_IO(rw, iocb, inode,
+				 inode->i_sb->s_bdev, iov,
+				 offset, nr_segs,
+				 ext4_get_block, NULL, NULL, 0);
+	} else {
+		ret = blockdev_direct_IO(rw, iocb, inode, iov,
+				 offset, nr_segs, ext4_get_block);
+
+		if (unlikely((rw & WRITE) && ret < 0)) {
+			loff_t isize = i_size_read(inode);
+			loff_t end = offset + iov_length(iov, nr_segs);
+
+			if (end > isize)
+				ext4_truncate_failed_write(inode);
+		}
+	}
+	if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
+		goto retry;
+
+	if (orphan) {
+		int err;
+
+		/* Credits for sb + inode write */
+		handle = ext4_journal_start(inode, 2);
+		if (IS_ERR(handle)) {
+			/* This is really bad luck. We've written the data
+			 * but cannot extend i_size. Bail out and pretend
+			 * the write failed... */
+			ret = PTR_ERR(handle);
+			if (inode->i_nlink)
+				ext4_orphan_del(NULL, inode);
+
+			goto out;
+		}
+		if (inode->i_nlink)
+			ext4_orphan_del(handle, inode);
+		if (ret > 0) {
+			loff_t end = offset + ret;
+			if (end > inode->i_size) {
+				ei->i_disksize = end;
+				i_size_write(inode, end);
+				/*
+				 * We're going to return a positive `ret'
+				 * here due to non-zero-length I/O, so there's
+				 * no way of reporting error returns from
+				 * ext4_mark_inode_dirty() to userspace.  So
+				 * ignore it.
+				 */
+				ext4_mark_inode_dirty(handle, inode);
+			}
+		}
+		err = ext4_journal_stop(handle);
+		if (ret == 0)
+			ret = err;
+	}
+out:
+	return ret;
+}
+
+/*
+ * Calculate the number of metadata blocks need to reserve
+ * to allocate a new block at @lblocks for non extent file based file
+ */
+int ext4_ind_calc_metadata_amount(struct inode *inode, sector_t lblock)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	sector_t dind_mask = ~((sector_t)EXT4_ADDR_PER_BLOCK(inode->i_sb) - 1);
+	int blk_bits;
+
+	if (lblock < EXT4_NDIR_BLOCKS)
+		return 0;
+
+	lblock -= EXT4_NDIR_BLOCKS;
+
+	if (ei->i_da_metadata_calc_len &&
+	    (lblock & dind_mask) == ei->i_da_metadata_calc_last_lblock) {
+		ei->i_da_metadata_calc_len++;
+		return 0;
+	}
+	ei->i_da_metadata_calc_last_lblock = lblock & dind_mask;
+	ei->i_da_metadata_calc_len = 1;
+	blk_bits = order_base_2(lblock);
+	return (blk_bits / EXT4_ADDR_PER_BLOCK_BITS(inode->i_sb)) + 1;
+}
+
+int ext4_ind_trans_blocks(struct inode *inode, int nrblocks, int chunk)
+{
+	int indirects;
+
+	/* if nrblocks are contiguous */
+	if (chunk) {
+		/*
+		 * With N contiguous data blocks, we need at most
+		 * N/EXT4_ADDR_PER_BLOCK(inode->i_sb) + 1 indirect blocks,
+		 * 2 dindirect blocks, and 1 tindirect block
+		 */
+		return DIV_ROUND_UP(nrblocks,
+				    EXT4_ADDR_PER_BLOCK(inode->i_sb)) + 4;
+	}
+	/*
+	 * if nrblocks are not contiguous, worse case, each block touch
+	 * a indirect block, and each indirect block touch a double indirect
+	 * block, plus a triple indirect block
+	 */
+	indirects = nrblocks * 2 + 1;
+	return indirects;
+}
+
+/*
+ * Truncate transactions can be complex and absolutely huge.  So we need to
+ * be able to restart the transaction at a conventient checkpoint to make
+ * sure we don't overflow the journal.
+ *
+ * start_transaction gets us a new handle for a truncate transaction,
+ * and extend_transaction tries to extend the existing one a bit.  If
+ * extend fails, we need to propagate the failure up and restart the
+ * transaction in the top-level truncate loop. --sct
+ */
+static handle_t *start_transaction(struct inode *inode)
+{
+	handle_t *result;
+
+	result = ext4_journal_start(inode, ext4_blocks_for_truncate(inode));
+	if (!IS_ERR(result))
+		return result;
+
+	ext4_std_error(inode->i_sb, PTR_ERR(result));
+	return result;
+}
+
+/*
+ * Try to extend this transaction for the purposes of truncation.
+ *
+ * Returns 0 if we managed to create more room.  If we can't create more
+ * room, and the transaction must be restarted we return 1.
+ */
+static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
+{
+	if (!ext4_handle_valid(handle))
+		return 0;
+	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
+		return 0;
+	if (!ext4_journal_extend(handle, ext4_blocks_for_truncate(inode)))
+		return 0;
+	return 1;
+}
+
+/*
+ * Probably it should be a library function... search for first non-zero word
+ * or memcmp with zero_page, whatever is better for particular architecture.
+ * Linus?
+ */
+static inline int all_zeroes(__le32 *p, __le32 *q)
+{
+	while (p < q)
+		if (*p++)
+			return 0;
+	return 1;
+}
+
+/**
+ *	ext4_find_shared - find the indirect blocks for partial truncation.
+ *	@inode:	  inode in question
+ *	@depth:	  depth of the affected branch
+ *	@offsets: offsets of pointers in that branch (see ext4_block_to_path)
+ *	@chain:	  place to store the pointers to partial indirect blocks
+ *	@top:	  place to the (detached) top of branch
+ *
+ *	This is a helper function used by ext4_truncate().
+ *
+ *	When we do truncate() we may have to clean the ends of several
+ *	indirect blocks but leave the blocks themselves alive. Block is
+ *	partially truncated if some data below the new i_size is referred
+ *	from it (and it is on the path to the first completely truncated
+ *	data block, indeed).  We have to free the top of that path along
+ *	with everything to the right of the path. Since no allocation
+ *	past the truncation point is possible until ext4_truncate()
+ *	finishes, we may safely do the latter, but top of branch may
+ *	require special attention - pageout below the truncation point
+ *	might try to populate it.
+ *
+ *	We atomically detach the top of branch from the tree, store the
+ *	block number of its root in *@top, pointers to buffer_heads of
+ *	partially truncated blocks - in @chain[].bh and pointers to
+ *	their last elements that should not be removed - in
+ *	@chain[].p. Return value is the pointer to last filled element
+ *	of @chain.
+ *
+ *	The work left to caller to do the actual freeing of subtrees:
+ *		a) free the subtree starting from *@top
+ *		b) free the subtrees whose roots are stored in
+ *			(@chain[i].p+1 .. end of @chain[i].bh->b_data)
+ *		c) free the subtrees growing from the inode past the @chain[0].
+ *			(no partially truncated stuff there).  */
+
+static Indirect *ext4_find_shared(struct inode *inode, int depth,
+				  ext4_lblk_t offsets[4], Indirect chain[4],
+				  __le32 *top)
+{
+	Indirect *partial, *p;
+	int k, err;
+
+	*top = 0;
+	/* Make k index the deepest non-null offset + 1 */
+	for (k = depth; k > 1 && !offsets[k-1]; k--)
+		;
+	partial = ext4_get_branch(inode, k, offsets, chain, &err);
+	/* Writer: pointers */
+	if (!partial)
+		partial = chain + k-1;
+	/*
+	 * If the branch acquired continuation since we've looked at it -
+	 * fine, it should all survive and (new) top doesn't belong to us.
+	 */
+	if (!partial->key && *partial->p)
+		/* Writer: end */
+		goto no_top;
+	for (p = partial; (p > chain) && all_zeroes((__le32 *) p->bh->b_data, p->p); p--)
+		;
+	/*
+	 * OK, we've found the last block that must survive. The rest of our
+	 * branch should be detached before unlocking. However, if that rest
+	 * of branch is all ours and does not grow immediately from the inode
+	 * it's easier to cheat and just decrement partial->p.
+	 */
+	if (p == chain + k - 1 && p > chain) {
+		p->p--;
+	} else {
+		*top = *p->p;
+		/* Nope, don't do this in ext4.  Must leave the tree intact */
+#if 0
+		*p->p = 0;
+#endif
+	}
+	/* Writer: end */
+
+	while (partial > p) {
+		brelse(partial->bh);
+		partial--;
+	}
+no_top:
+	return partial;
+}
+
+/*
+ * Zero a number of block pointers in either an inode or an indirect block.
+ * If we restart the transaction we must again get write access to the
+ * indirect block for further modification.
+ *
+ * We release `count' blocks on disk, but (last - first) may be greater
+ * than `count' because there can be holes in there.
+ *
+ * Return 0 on success, 1 on invalid block range
+ * and < 0 on fatal error.
+ */
+static int ext4_clear_blocks(handle_t *handle, struct inode *inode,
+			     struct buffer_head *bh,
+			     ext4_fsblk_t block_to_free,
+			     unsigned long count, __le32 *first,
+			     __le32 *last)
+{
+	__le32 *p;
+	int	flags = EXT4_FREE_BLOCKS_FORGET | EXT4_FREE_BLOCKS_VALIDATED;
+	int	err;
+
+	if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
+		flags |= EXT4_FREE_BLOCKS_METADATA;
+
+	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), block_to_free,
+				   count)) {
+		EXT4_ERROR_INODE(inode, "attempt to clear invalid "
+				 "blocks %llu len %lu",
+				 (unsigned long long) block_to_free, count);
+		return 1;
+	}
+
+	if (try_to_extend_transaction(handle, inode)) {
+		if (bh) {
+			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+			err = ext4_handle_dirty_metadata(handle, inode, bh);
+			if (unlikely(err))
+				goto out_err;
+		}
+		err = ext4_mark_inode_dirty(handle, inode);
+		if (unlikely(err))
+			goto out_err;
+		err = ext4_truncate_restart_trans(handle, inode,
+					ext4_blocks_for_truncate(inode));
+		if (unlikely(err))
+			goto out_err;
+		if (bh) {
+			BUFFER_TRACE(bh, "retaking write access");
+			err = ext4_journal_get_write_access(handle, bh);
+			if (unlikely(err))
+				goto out_err;
+		}
+	}
+
+	for (p = first; p < last; p++)
+		*p = 0;
+
+	ext4_free_blocks(handle, inode, NULL, block_to_free, count, flags);
+	return 0;
+out_err:
+	ext4_std_error(inode->i_sb, err);
+	return err;
+}
+
+/**
+ * ext4_free_data - free a list of data blocks
+ * @handle:	handle for this transaction
+ * @inode:	inode we are dealing with
+ * @this_bh:	indirect buffer_head which contains *@first and *@last
+ * @first:	array of block numbers
+ * @last:	points immediately past the end of array
+ *
+ * We are freeing all blocks referred from that array (numbers are stored as
+ * little-endian 32-bit) and updating @inode->i_blocks appropriately.
+ *
+ * We accumulate contiguous runs of blocks to free.  Conveniently, if these
+ * blocks are contiguous then releasing them at one time will only affect one
+ * or two bitmap blocks (+ group descriptor(s) and superblock) and we won't
+ * actually use a lot of journal space.
+ *
+ * @this_bh will be %NULL if @first and @last point into the inode's direct
+ * block pointers.
+ */
+static void ext4_free_data(handle_t *handle, struct inode *inode,
+			   struct buffer_head *this_bh,
+			   __le32 *first, __le32 *last)
+{
+	ext4_fsblk_t block_to_free = 0;    /* Starting block # of a run */
+	unsigned long count = 0;	    /* Number of blocks in the run */
+	__le32 *block_to_free_p = NULL;	    /* Pointer into inode/ind
+					       corresponding to
+					       block_to_free */
+	ext4_fsblk_t nr;		    /* Current block # */
+	__le32 *p;			    /* Pointer into inode/ind
+					       for current block */
+	int err = 0;
+
+	if (this_bh) {				/* For indirect block */
+		BUFFER_TRACE(this_bh, "get_write_access");
+		err = ext4_journal_get_write_access(handle, this_bh);
+		/* Important: if we can't update the indirect pointers
+		 * to the blocks, we can't free them. */
+		if (err)
+			return;
+	}
+
+	for (p = first; p < last; p++) {
+		nr = le32_to_cpu(*p);
+		if (nr) {
+			/* accumulate blocks to free if they're contiguous */
+			if (count == 0) {
+				block_to_free = nr;
+				block_to_free_p = p;
+				count = 1;
+			} else if (nr == block_to_free + count) {
+				count++;
+			} else {
+				err = ext4_clear_blocks(handle, inode, this_bh,
+						        block_to_free, count,
+						        block_to_free_p, p);
+				if (err)
+					break;
+				block_to_free = nr;
+				block_to_free_p = p;
+				count = 1;
+			}
+		}
+	}
+
+	if (!err && count > 0)
+		err = ext4_clear_blocks(handle, inode, this_bh, block_to_free,
+					count, block_to_free_p, p);
+	if (err < 0)
+		/* fatal error */
+		return;
+
+	if (this_bh) {
+		BUFFER_TRACE(this_bh, "call ext4_handle_dirty_metadata");
+
+		/*
+		 * The buffer head should have an attached journal head at this
+		 * point. However, if the data is corrupted and an indirect
+		 * block pointed to itself, it would have been detached when
+		 * the block was cleared. Check for this instead of OOPSing.
+		 */
+		if ((EXT4_JOURNAL(inode) == NULL) || bh2jh(this_bh))
+			ext4_handle_dirty_metadata(handle, inode, this_bh);
+		else
+			EXT4_ERROR_INODE(inode,
+					 "circular indirect block detected at "
+					 "block %llu",
+				(unsigned long long) this_bh->b_blocknr);
+	}
+}
+
+/**
+ *	ext4_free_branches - free an array of branches
+ *	@handle: JBD handle for this transaction
+ *	@inode:	inode we are dealing with
+ *	@parent_bh: the buffer_head which contains *@first and *@last
+ *	@first:	array of block numbers
+ *	@last:	pointer immediately past the end of array
+ *	@depth:	depth of the branches to free
+ *
+ *	We are freeing all blocks referred from these branches (numbers are
+ *	stored as little-endian 32-bit) and updating @inode->i_blocks
+ *	appropriately.
+ */
+static void ext4_free_branches(handle_t *handle, struct inode *inode,
+			       struct buffer_head *parent_bh,
+			       __le32 *first, __le32 *last, int depth)
+{
+	ext4_fsblk_t nr;
+	__le32 *p;
+
+	if (ext4_handle_is_aborted(handle))
+		return;
+
+	if (depth--) {
+		struct buffer_head *bh;
+		int addr_per_block = EXT4_ADDR_PER_BLOCK(inode->i_sb);
+		p = last;
+		while (--p >= first) {
+			nr = le32_to_cpu(*p);
+			if (!nr)
+				continue;		/* A hole */
+
+			if (!ext4_data_block_valid(EXT4_SB(inode->i_sb),
+						   nr, 1)) {
+				EXT4_ERROR_INODE(inode,
+						 "invalid indirect mapped "
+						 "block %lu (level %d)",
+						 (unsigned long) nr, depth);
+				break;
+			}
+
+			/* Go read the buffer for the next level down */
+			bh = sb_bread(inode->i_sb, nr);
+
+			/*
+			 * A read failure? Report error and clear slot
+			 * (should be rare).
+			 */
+			if (!bh) {
+				EXT4_ERROR_INODE_BLOCK(inode, nr,
+						       "Read failure");
+				continue;
+			}
+
+			/* This zaps the entire block.  Bottom up. */
+			BUFFER_TRACE(bh, "free child branches");
+			ext4_free_branches(handle, inode, bh,
+					(__le32 *) bh->b_data,
+					(__le32 *) bh->b_data + addr_per_block,
+					depth);
+			brelse(bh);
+
+			/*
+			 * Everything below this this pointer has been
+			 * released.  Now let this top-of-subtree go.
+			 *
+			 * We want the freeing of this indirect block to be
+			 * atomic in the journal with the updating of the
+			 * bitmap block which owns it.  So make some room in
+			 * the journal.
+			 *
+			 * We zero the parent pointer *after* freeing its
+			 * pointee in the bitmaps, so if extend_transaction()
+			 * for some reason fails to put the bitmap changes and
+			 * the release into the same transaction, recovery
+			 * will merely complain about releasing a free block,
+			 * rather than leaking blocks.
+			 */
+			if (ext4_handle_is_aborted(handle))
+				return;
+			if (try_to_extend_transaction(handle, inode)) {
+				ext4_mark_inode_dirty(handle, inode);
+				ext4_truncate_restart_trans(handle, inode,
+					    ext4_blocks_for_truncate(inode));
+			}
+
+			/*
+			 * The forget flag here is critical because if
+			 * we are journaling (and not doing data
+			 * journaling), we have to make sure a revoke
+			 * record is written to prevent the journal
+			 * replay from overwriting the (former)
+			 * indirect block if it gets reallocated as a
+			 * data block.  This must happen in the same
+			 * transaction where the data blocks are
+			 * actually freed.
+			 */
+			ext4_free_blocks(handle, inode, NULL, nr, 1,
+					 EXT4_FREE_BLOCKS_METADATA|
+					 EXT4_FREE_BLOCKS_FORGET);
+
+			if (parent_bh) {
+				/*
+				 * The block which we have just freed is
+				 * pointed to by an indirect block: journal it
+				 */
+				BUFFER_TRACE(parent_bh, "get_write_access");
+				if (!ext4_journal_get_write_access(handle,
+								   parent_bh)){
+					*p = 0;
+					BUFFER_TRACE(parent_bh,
+					"call ext4_handle_dirty_metadata");
+					ext4_handle_dirty_metadata(handle,
+								   inode,
+								   parent_bh);
+				}
+			}
+		}
+	} else {
+		/* We have reached the bottom of the tree. */
+		BUFFER_TRACE(parent_bh, "free data blocks");
+		ext4_free_data(handle, inode, parent_bh, first, last);
+	}
+}
+
+void ext4_ind_truncate(struct inode *inode)
+{
+	handle_t *handle;
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	__le32 *i_data = ei->i_data;
+	int addr_per_block = EXT4_ADDR_PER_BLOCK(inode->i_sb);
+	struct address_space *mapping = inode->i_mapping;
+	ext4_lblk_t offsets[4];
+	Indirect chain[4];
+	Indirect *partial;
+	__le32 nr = 0;
+	int n = 0;
+	ext4_lblk_t last_block, max_block;
+	loff_t page_len;
+	unsigned blocksize = inode->i_sb->s_blocksize;
+	int err;
+
+	handle = start_transaction(inode);
+	if (IS_ERR(handle))
+		return;		/* AKPM: return what? */
+
+	last_block = (inode->i_size + blocksize-1)
+					>> EXT4_BLOCK_SIZE_BITS(inode->i_sb);
+	max_block = (EXT4_SB(inode->i_sb)->s_bitmap_maxbytes + blocksize-1)
+					>> EXT4_BLOCK_SIZE_BITS(inode->i_sb);
+
+	if (inode->i_size % PAGE_CACHE_SIZE != 0) {
+		page_len = PAGE_CACHE_SIZE -
+			(inode->i_size & (PAGE_CACHE_SIZE - 1));
+
+		err = ext4_discard_partial_page_buffers(handle,
+			mapping, inode->i_size, page_len, 0);
+
+		if (err)
+			goto out_stop;
+	}
+
+	if (last_block != max_block) {
+		n = ext4_block_to_path(inode, last_block, offsets, NULL);
+		if (n == 0)
+			goto out_stop;	/* error */
+	}
+
+	/*
+	 * OK.  This truncate is going to happen.  We add the inode to the
+	 * orphan list, so that if this truncate spans multiple transactions,
+	 * and we crash, we will resume the truncate when the filesystem
+	 * recovers.  It also marks the inode dirty, to catch the new size.
+	 *
+	 * Implication: the file must always be in a sane, consistent
+	 * truncatable state while each transaction commits.
+	 */
+	if (ext4_orphan_add(handle, inode))
+		goto out_stop;
+
+	/*
+	 * From here we block out all ext4_get_block() callers who want to
+	 * modify the block allocation tree.
+	 */
+	down_write(&ei->i_data_sem);
+
+	ext4_discard_preallocations(inode);
+
+	/*
+	 * The orphan list entry will now protect us from any crash which
+	 * occurs before the truncate completes, so it is now safe to propagate
+	 * the new, shorter inode size (held for now in i_size) into the
+	 * on-disk inode. We do this via i_disksize, which is the value which
+	 * ext4 *really* writes onto the disk inode.
+	 */
+	ei->i_disksize = inode->i_size;
+
+	if (last_block == max_block) {
+		/*
+		 * It is unnecessary to free any data blocks if last_block is
+		 * equal to the indirect block limit.
+		 */
+		goto out_unlock;
+	} else if (n == 1) {		/* direct blocks */
+		ext4_free_data(handle, inode, NULL, i_data+offsets[0],
+			       i_data + EXT4_NDIR_BLOCKS);
+		goto do_indirects;
+	}
+
+	partial = ext4_find_shared(inode, n, offsets, chain, &nr);
+	/* Kill the top of shared branch (not detached) */
+	if (nr) {
+		if (partial == chain) {
+			/* Shared branch grows from the inode */
+			ext4_free_branches(handle, inode, NULL,
+					   &nr, &nr+1, (chain+n-1) - partial);
+			*partial->p = 0;
+			/*
+			 * We mark the inode dirty prior to restart,
+			 * and prior to stop.  No need for it here.
+			 */
+		} else {
+			/* Shared branch grows from an indirect block */
+			BUFFER_TRACE(partial->bh, "get_write_access");
+			ext4_free_branches(handle, inode, partial->bh,
+					partial->p,
+					partial->p+1, (chain+n-1) - partial);
+		}
+	}
+	/* Clear the ends of indirect blocks on the shared branch */
+	while (partial > chain) {
+		ext4_free_branches(handle, inode, partial->bh, partial->p + 1,
+				   (__le32*)partial->bh->b_data+addr_per_block,
+				   (chain+n-1) - partial);
+		BUFFER_TRACE(partial->bh, "call brelse");
+		brelse(partial->bh);
+		partial--;
+	}
+do_indirects:
+	/* Kill the remaining (whole) subtrees */
+	switch (offsets[0]) {
+	default:
+		nr = i_data[EXT4_IND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 1);
+			i_data[EXT4_IND_BLOCK] = 0;
+		}
+	case EXT4_IND_BLOCK:
+		nr = i_data[EXT4_DIND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 2);
+			i_data[EXT4_DIND_BLOCK] = 0;
+		}
+	case EXT4_DIND_BLOCK:
+		nr = i_data[EXT4_TIND_BLOCK];
+		if (nr) {
+			ext4_free_branches(handle, inode, NULL, &nr, &nr+1, 3);
+			i_data[EXT4_TIND_BLOCK] = 0;
+		}
+	case EXT4_TIND_BLOCK:
+		;
+	}
+
+out_unlock:
+	up_write(&ei->i_data_sem);
+	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
+	ext4_mark_inode_dirty(handle, inode);
+
+	/*
+	 * In a multi-transaction truncate, we only make the final transaction
+	 * synchronous
+	 */
+	if (IS_SYNC(inode))
+		ext4_handle_sync(handle);
+out_stop:
+	/*
+	 * If this was a simple ftruncate(), and the file will remain alive
+	 * then we need to clear up the orphan record which we created above.
+	 * However, if this was a real unlink then we were called by
+	 * ext4_delete_inode(), and we allow that function to clean up the
+	 * orphan info for us.
+	 */
+	if (inode->i_nlink)
+		ext4_orphan_del(handle, inode);
+
+	ext4_journal_stop(handle);
+	trace_ext4_truncate_exit(inode);
+}
+
diff --git a/fs/ext4/truncate.h b/fs/ext4/truncate.h
new file mode 100644
index 0000000..011ba66
--- /dev/null
+++ b/fs/ext4/truncate.h
@@ -0,0 +1,43 @@
+/*
+ * linux/fs/ext4/truncate.h
+ *
+ * Common inline functions needed for truncate support
+ */
+
+/*
+ * Truncate blocks that were not used by write. We have to truncate the
+ * pagecache as well so that corresponding buffers get properly unmapped.
+ */
+static inline void ext4_truncate_failed_write(struct inode *inode)
+{
+	truncate_inode_pages(inode->i_mapping, inode->i_size);
+	ext4_truncate(inode);
+}
+
+/*
+ * Work out how many blocks we need to proceed with the next chunk of a
+ * truncate transaction.
+ */
+static inline unsigned long ext4_blocks_for_truncate(struct inode *inode)
+{
+	ext4_lblk_t needed;
+
+	needed = inode->i_blocks >> (inode->i_sb->s_blocksize_bits - 9);
+
+	/* Give ourselves just enough room to cope with inodes in which
+	 * i_blocks is corrupt: we've seen disk corruptions in the past
+	 * which resulted in random data in an inode which looked enough
+	 * like a regular file for ext4 to try to delete it.  Things
+	 * will go a bit crazy if that happens, but at least we should
+	 * try not to panic the whole kernel. */
+	if (needed < 2)
+		needed = 2;
+
+	/* But we need to bound the transaction so we don't overflow the
+	 * journal. */
+	if (needed > EXT4_MAX_TRANS_DATA)
+		needed = EXT4_MAX_TRANS_DATA;
+
+	return EXT4_DATA_TRANS_BLOCKS(inode->i_sb) + needed;
+}
+
diff --git a/fs/nfs/blocklayout/Makefile b/fs/nfs/blocklayout/Makefile
new file mode 100644
index 0000000..d581550
--- /dev/null
+++ b/fs/nfs/blocklayout/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for the pNFS block layout driver kernel module
+#
+obj-$(CONFIG_PNFS_BLOCK) += blocklayoutdriver.o
+blocklayoutdriver-objs := blocklayout.o extents.o blocklayoutdev.o blocklayoutdm.o
diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
new file mode 100644
index 0000000..b17a81c
--- /dev/null
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -0,0 +1,1181 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayout.c
+ *
+ *  Module for the NFSv4.1 pNFS block layout driver.
+ *
+ *  Copyright (c) 2006 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/bio.h>		/* struct bio */
+#include <linux/buffer_head.h>	/* various write calls */
+#include <linux/prefetch.h>
+
+#include "../pnfs.h"
+#include "../internal.h"
+#include "blocklayout.h"
+
+#define NFSDBG_FACILITY	NFSDBG_PNFS_LD
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Andy Adamson <andros@citi.umich.edu>");
+MODULE_DESCRIPTION("The NFSv4.1 pNFS Block layout driver");
+
+struct dentry *bl_device_pipe;
+wait_queue_head_t bl_wq;
+
+static void print_page(struct page *page)
+{
+	dprintk("PRINTPAGE page %p\n", page);
+	dprintk("	PagePrivate %d\n", PagePrivate(page));
+	dprintk("	PageUptodate %d\n", PageUptodate(page));
+	dprintk("	PageError %d\n", PageError(page));
+	dprintk("	PageDirty %d\n", PageDirty(page));
+	dprintk("	PageReferenced %d\n", PageReferenced(page));
+	dprintk("	PageLocked %d\n", PageLocked(page));
+	dprintk("	PageWriteback %d\n", PageWriteback(page));
+	dprintk("	PageMappedToDisk %d\n", PageMappedToDisk(page));
+	dprintk("\n");
+}
+
+/* Given the be associated with isect, determine if page data needs to be
+ * initialized.
+ */
+static int is_hole(struct pnfs_block_extent *be, sector_t isect)
+{
+	if (be->be_state == PNFS_BLOCK_NONE_DATA)
+		return 1;
+	else if (be->be_state != PNFS_BLOCK_INVALID_DATA)
+		return 0;
+	else
+		return !bl_is_sector_init(be->be_inval, isect);
+}
+
+/* Given the be associated with isect, determine if page data can be
+ * written to disk.
+ */
+static int is_writable(struct pnfs_block_extent *be, sector_t isect)
+{
+	return (be->be_state == PNFS_BLOCK_READWRITE_DATA ||
+		be->be_state == PNFS_BLOCK_INVALID_DATA);
+}
+
+/* The data we are handed might be spread across several bios.  We need
+ * to track when the last one is finished.
+ */
+struct parallel_io {
+	struct kref refcnt;
+	struct rpc_call_ops call_ops;
+	void (*pnfs_callback) (void *data);
+	void *data;
+};
+
+static inline struct parallel_io *alloc_parallel(void *data)
+{
+	struct parallel_io *rv;
+
+	rv  = kmalloc(sizeof(*rv), GFP_NOFS);
+	if (rv) {
+		rv->data = data;
+		kref_init(&rv->refcnt);
+	}
+	return rv;
+}
+
+static inline void get_parallel(struct parallel_io *p)
+{
+	kref_get(&p->refcnt);
+}
+
+static void destroy_parallel(struct kref *kref)
+{
+	struct parallel_io *p = container_of(kref, struct parallel_io, refcnt);
+
+	dprintk("%s enter\n", __func__);
+	p->pnfs_callback(p->data);
+	kfree(p);
+}
+
+static inline void put_parallel(struct parallel_io *p)
+{
+	kref_put(&p->refcnt, destroy_parallel);
+}
+
+static struct bio *
+bl_submit_bio(int rw, struct bio *bio)
+{
+	if (bio) {
+		get_parallel(bio->bi_private);
+		dprintk("%s submitting %s bio %u@%llu\n", __func__,
+			rw == READ ? "read" : "write",
+			bio->bi_size, (unsigned long long)bio->bi_sector);
+		submit_bio(rw, bio);
+	}
+	return NULL;
+}
+
+static struct bio *bl_alloc_init_bio(int npg, sector_t isect,
+				     struct pnfs_block_extent *be,
+				     void (*end_io)(struct bio *, int err),
+				     struct parallel_io *par)
+{
+	struct bio *bio;
+
+	npg = min(npg, BIO_MAX_PAGES);
+	bio = bio_alloc(GFP_NOIO, npg);
+	if (!bio && (current->flags & PF_MEMALLOC)) {
+		while (!bio && (npg /= 2))
+			bio = bio_alloc(GFP_NOIO, npg);
+	}
+
+	if (bio) {
+		bio->bi_sector = isect - be->be_f_offset + be->be_v_offset;
+		bio->bi_bdev = be->be_mdev;
+		bio->bi_end_io = end_io;
+		bio->bi_private = par;
+	}
+	return bio;
+}
+
+static struct bio *do_add_page_to_bio(struct bio *bio, int npg, int rw,
+				      sector_t isect, struct page *page,
+				      struct pnfs_block_extent *be,
+				      void (*end_io)(struct bio *, int err),
+				      struct parallel_io *par,
+				      unsigned int offset, int len)
+{
+	isect = isect + (offset >> SECTOR_SHIFT);
+	dprintk("%s: npg %d rw %d isect %llu offset %u len %d\n", __func__,
+		npg, rw, (unsigned long long)isect, offset, len);
+retry:
+	if (!bio) {
+		bio = bl_alloc_init_bio(npg, isect, be, end_io, par);
+		if (!bio)
+			return ERR_PTR(-ENOMEM);
+	}
+	if (bio_add_page(bio, page, len, offset) < len) {
+		bio = bl_submit_bio(rw, bio);
+		goto retry;
+	}
+	return bio;
+}
+
+static struct bio *bl_add_page_to_bio(struct bio *bio, int npg, int rw,
+				      sector_t isect, struct page *page,
+				      struct pnfs_block_extent *be,
+				      void (*end_io)(struct bio *, int err),
+				      struct parallel_io *par)
+{
+	return do_add_page_to_bio(bio, npg, rw, isect, page, be,
+				  end_io, par, 0, PAGE_CACHE_SIZE);
+}
+
+/* This is basically copied from mpage_end_io_read */
+static void bl_end_io_read(struct bio *bio, int err)
+{
+	struct parallel_io *par = bio->bi_private;
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct nfs_read_data *rdata = (struct nfs_read_data *)par->data;
+
+	do {
+		struct page *page = bvec->bv_page;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+		if (uptodate)
+			SetPageUptodate(page);
+	} while (bvec >= bio->bi_io_vec);
+	if (!uptodate) {
+		if (!rdata->pnfs_error)
+			rdata->pnfs_error = -EIO;
+		pnfs_set_lo_fail(rdata->lseg);
+	}
+	bio_put(bio);
+	put_parallel(par);
+}
+
+static void bl_read_cleanup(struct work_struct *work)
+{
+	struct rpc_task *task;
+	struct nfs_read_data *rdata;
+	dprintk("%s enter\n", __func__);
+	task = container_of(work, struct rpc_task, u.tk_work);
+	rdata = container_of(task, struct nfs_read_data, task);
+	pnfs_ld_read_done(rdata);
+}
+
+static void
+bl_end_par_io_read(void *data)
+{
+	struct nfs_read_data *rdata = data;
+
+	INIT_WORK(&rdata->task.u.tk_work, bl_read_cleanup);
+	schedule_work(&rdata->task.u.tk_work);
+}
+
+/* We don't want normal .rpc_call_done callback used, so we replace it
+ * with this stub.
+ */
+static void bl_rpc_do_nothing(struct rpc_task *task, void *calldata)
+{
+	return;
+}
+
+static enum pnfs_try_status
+bl_read_pagelist(struct nfs_read_data *rdata)
+{
+	int i, hole;
+	struct bio *bio = NULL;
+	struct pnfs_block_extent *be = NULL, *cow_read = NULL;
+	sector_t isect, extent_length = 0;
+	struct parallel_io *par;
+	loff_t f_offset = rdata->args.offset;
+	size_t count = rdata->args.count;
+	struct page **pages = rdata->args.pages;
+	int pg_index = rdata->args.pgbase >> PAGE_CACHE_SHIFT;
+
+	dprintk("%s enter nr_pages %u offset %lld count %Zd\n", __func__,
+	       rdata->npages, f_offset, count);
+
+	par = alloc_parallel(rdata);
+	if (!par)
+		goto use_mds;
+	par->call_ops = *rdata->mds_ops;
+	par->call_ops.rpc_call_done = bl_rpc_do_nothing;
+	par->pnfs_callback = bl_end_par_io_read;
+	/* At this point, we can no longer jump to use_mds */
+
+	isect = (sector_t) (f_offset >> SECTOR_SHIFT);
+	/* Code assumes extents are page-aligned */
+	for (i = pg_index; i < rdata->npages; i++) {
+		if (!extent_length) {
+			/* We've used up the previous extent */
+			bl_put_extent(be);
+			bl_put_extent(cow_read);
+			bio = bl_submit_bio(READ, bio);
+			/* Get the next one */
+			be = bl_find_get_extent(BLK_LSEG2EXT(rdata->lseg),
+					     isect, &cow_read);
+			if (!be) {
+				rdata->pnfs_error = -EIO;
+				goto out;
+			}
+			extent_length = be->be_length -
+				(isect - be->be_f_offset);
+			if (cow_read) {
+				sector_t cow_length = cow_read->be_length -
+					(isect - cow_read->be_f_offset);
+				extent_length = min(extent_length, cow_length);
+			}
+		}
+		hole = is_hole(be, isect);
+		if (hole && !cow_read) {
+			bio = bl_submit_bio(READ, bio);
+			/* Fill hole w/ zeroes w/o accessing device */
+			dprintk("%s Zeroing page for hole\n", __func__);
+			zero_user_segment(pages[i], 0, PAGE_CACHE_SIZE);
+			print_page(pages[i]);
+			SetPageUptodate(pages[i]);
+		} else {
+			struct pnfs_block_extent *be_read;
+
+			be_read = (hole && cow_read) ? cow_read : be;
+			bio = bl_add_page_to_bio(bio, rdata->npages - i, READ,
+						 isect, pages[i], be_read,
+						 bl_end_io_read, par);
+			if (IS_ERR(bio)) {
+				rdata->pnfs_error = PTR_ERR(bio);
+				bio = NULL;
+				goto out;
+			}
+		}
+		isect += PAGE_CACHE_SECTORS;
+		extent_length -= PAGE_CACHE_SECTORS;
+	}
+	if ((isect << SECTOR_SHIFT) >= rdata->inode->i_size) {
+		rdata->res.eof = 1;
+		rdata->res.count = rdata->inode->i_size - f_offset;
+	} else {
+		rdata->res.count = (isect << SECTOR_SHIFT) - f_offset;
+	}
+out:
+	bl_put_extent(be);
+	bl_put_extent(cow_read);
+	bl_submit_bio(READ, bio);
+	put_parallel(par);
+	return PNFS_ATTEMPTED;
+
+ use_mds:
+	dprintk("Giving up and using normal NFS\n");
+	return PNFS_NOT_ATTEMPTED;
+}
+
+static void mark_extents_written(struct pnfs_block_layout *bl,
+				 __u64 offset, __u32 count)
+{
+	sector_t isect, end;
+	struct pnfs_block_extent *be;
+
+	dprintk("%s(%llu, %u)\n", __func__, offset, count);
+	if (count == 0)
+		return;
+	isect = (offset & (long)(PAGE_CACHE_MASK)) >> SECTOR_SHIFT;
+	end = (offset + count + PAGE_CACHE_SIZE - 1) & (long)(PAGE_CACHE_MASK);
+	end >>= SECTOR_SHIFT;
+	while (isect < end) {
+		sector_t len;
+		be = bl_find_get_extent(bl, isect, NULL);
+		BUG_ON(!be); /* FIXME */
+		len = min(end, be->be_f_offset + be->be_length) - isect;
+		if (be->be_state == PNFS_BLOCK_INVALID_DATA)
+			bl_mark_for_commit(be, isect, len); /* What if fails? */
+		isect += len;
+		bl_put_extent(be);
+	}
+}
+
+static void bl_end_io_write_zero(struct bio *bio, int err)
+{
+	struct parallel_io *par = bio->bi_private;
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct nfs_write_data *wdata = (struct nfs_write_data *)par->data;
+
+	do {
+		struct page *page = bvec->bv_page;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+		/* This is the zeroing page we added */
+		end_page_writeback(page);
+		page_cache_release(page);
+	} while (bvec >= bio->bi_io_vec);
+	if (!uptodate) {
+		if (!wdata->pnfs_error)
+			wdata->pnfs_error = -EIO;
+		pnfs_set_lo_fail(wdata->lseg);
+	}
+	bio_put(bio);
+	put_parallel(par);
+}
+
+/* This is basically copied from mpage_end_io_read */
+static void bl_end_io_write(struct bio *bio, int err)
+{
+	struct parallel_io *par = bio->bi_private;
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct nfs_write_data *wdata = (struct nfs_write_data *)par->data;
+
+	if (!uptodate) {
+		if (!wdata->pnfs_error)
+			wdata->pnfs_error = -EIO;
+		pnfs_set_lo_fail(wdata->lseg);
+	}
+	bio_put(bio);
+	put_parallel(par);
+}
+
+/* Function scheduled for call during bl_end_par_io_write,
+ * it marks sectors as written and extends the commitlist.
+ */
+static void bl_write_cleanup(struct work_struct *work)
+{
+	struct rpc_task *task;
+	struct nfs_write_data *wdata;
+	dprintk("%s enter\n", __func__);
+	task = container_of(work, struct rpc_task, u.tk_work);
+	wdata = container_of(task, struct nfs_write_data, task);
+	if (!wdata->pnfs_error) {
+		/* Marks for LAYOUTCOMMIT */
+		mark_extents_written(BLK_LSEG2EXT(wdata->lseg),
+				     wdata->args.offset, wdata->args.count);
+	}
+	pnfs_ld_write_done(wdata);
+}
+
+/* Called when last of bios associated with a bl_write_pagelist call finishes */
+static void bl_end_par_io_write(void *data)
+{
+	struct nfs_write_data *wdata = data;
+
+	wdata->task.tk_status = 0;
+	wdata->verf.committed = NFS_FILE_SYNC;
+	INIT_WORK(&wdata->task.u.tk_work, bl_write_cleanup);
+	schedule_work(&wdata->task.u.tk_work);
+}
+
+/* FIXME STUB - mark intersection of layout and page as bad, so is not
+ * used again.
+ */
+static void mark_bad_read(void)
+{
+	return;
+}
+
+/*
+ * map_block:  map a requested I/0 block (isect) into an offset in the LVM
+ * block_device
+ */
+static void
+map_block(struct buffer_head *bh, sector_t isect, struct pnfs_block_extent *be)
+{
+	dprintk("%s enter be=%p\n", __func__, be);
+
+	set_buffer_mapped(bh);
+	bh->b_bdev = be->be_mdev;
+	bh->b_blocknr = (isect - be->be_f_offset + be->be_v_offset) >>
+	    (be->be_mdev->bd_inode->i_blkbits - SECTOR_SHIFT);
+
+	dprintk("%s isect %llu, bh->b_blocknr %ld, using bsize %Zd\n",
+		__func__, (unsigned long long)isect, (long)bh->b_blocknr,
+		bh->b_size);
+	return;
+}
+
+static void
+bl_read_single_end_io(struct bio *bio, int error)
+{
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct page *page = bvec->bv_page;
+
+	/* Only one page in bvec */
+	unlock_page(page);
+}
+
+static int
+bl_do_readpage_sync(struct page *page, struct pnfs_block_extent *be,
+		    unsigned int offset, unsigned int len)
+{
+	struct bio *bio;
+	struct page *shadow_page;
+	sector_t isect;
+	char *kaddr, *kshadow_addr;
+	int ret = 0;
+
+	dprintk("%s: offset %u len %u\n", __func__, offset, len);
+
+	shadow_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
+	if (shadow_page == NULL)
+		return -ENOMEM;
+
+	bio = bio_alloc(GFP_NOIO, 1);
+	if (bio == NULL)
+		return -ENOMEM;
+
+	isect = (page->index << PAGE_CACHE_SECTOR_SHIFT) +
+		(offset / SECTOR_SIZE);
+
+	bio->bi_sector = isect - be->be_f_offset + be->be_v_offset;
+	bio->bi_bdev = be->be_mdev;
+	bio->bi_end_io = bl_read_single_end_io;
+
+	lock_page(shadow_page);
+	if (bio_add_page(bio, shadow_page,
+			 SECTOR_SIZE, round_down(offset, SECTOR_SIZE)) == 0) {
+		unlock_page(shadow_page);
+		bio_put(bio);
+		return -EIO;
+	}
+
+	submit_bio(READ, bio);
+	wait_on_page_locked(shadow_page);
+	if (unlikely(!test_bit(BIO_UPTODATE, &bio->bi_flags))) {
+		ret = -EIO;
+	} else {
+		kaddr = kmap_atomic(page);
+		kshadow_addr = kmap_atomic(shadow_page);
+		memcpy(kaddr + offset, kshadow_addr + offset, len);
+		kunmap_atomic(kshadow_addr);
+		kunmap_atomic(kaddr);
+	}
+	__free_page(shadow_page);
+	bio_put(bio);
+
+	return ret;
+}
+
+static int
+bl_read_partial_page_sync(struct page *page, struct pnfs_block_extent *be,
+			  unsigned int dirty_offset, unsigned int dirty_len,
+			  bool full_page)
+{
+	int ret = 0;
+	unsigned int start, end;
+
+	if (full_page) {
+		start = 0;
+		end = PAGE_CACHE_SIZE;
+	} else {
+		start = round_down(dirty_offset, SECTOR_SIZE);
+		end = round_up(dirty_offset + dirty_len, SECTOR_SIZE);
+	}
+
+	dprintk("%s: offset %u len %d\n", __func__, dirty_offset, dirty_len);
+	if (!be) {
+		zero_user_segments(page, start, dirty_offset,
+				   dirty_offset + dirty_len, end);
+		if (start == 0 && end == PAGE_CACHE_SIZE &&
+		    trylock_page(page)) {
+			SetPageUptodate(page);
+			unlock_page(page);
+		}
+		return ret;
+	}
+
+	if (start != dirty_offset)
+		ret = bl_do_readpage_sync(page, be, start, dirty_offset - start);
+
+	if (!ret && (dirty_offset + dirty_len < end))
+		ret = bl_do_readpage_sync(page, be, dirty_offset + dirty_len,
+					  end - dirty_offset - dirty_len);
+
+	return ret;
+}
+
+/* Given an unmapped page, zero it or read in page for COW, page is locked
+ * by caller.
+ */
+static int
+init_page_for_write(struct page *page, struct pnfs_block_extent *cow_read)
+{
+	struct buffer_head *bh = NULL;
+	int ret = 0;
+	sector_t isect;
+
+	dprintk("%s enter, %p\n", __func__, page);
+	BUG_ON(PageUptodate(page));
+	if (!cow_read) {
+		zero_user_segment(page, 0, PAGE_SIZE);
+		SetPageUptodate(page);
+		goto cleanup;
+	}
+
+	bh = alloc_page_buffers(page, PAGE_CACHE_SIZE, 0);
+	if (!bh) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	isect = (sector_t) page->index << PAGE_CACHE_SECTOR_SHIFT;
+	map_block(bh, isect, cow_read);
+	if (!bh_uptodate_or_lock(bh))
+		ret = bh_submit_read(bh);
+	if (ret)
+		goto cleanup;
+	SetPageUptodate(page);
+
+cleanup:
+	if (bh)
+		free_buffer_head(bh);
+	if (ret) {
+		/* Need to mark layout with bad read...should now
+		 * just use nfs4 for reads and writes.
+		 */
+		mark_bad_read();
+	}
+	return ret;
+}
+
+static enum pnfs_try_status
+bl_write_pagelist(struct nfs_write_data *wdata, int sync)
+{
+	int i, ret, npg_zero, pg_index, last = 0;
+	struct bio *bio = NULL;
+	struct pnfs_block_extent *be = NULL, *cow_read = NULL;
+	sector_t isect, last_isect = 0, extent_length = 0;
+	struct parallel_io *par;
+	loff_t offset = wdata->args.offset;
+	size_t count = wdata->args.count;
+	unsigned int pg_offset, pg_len, saved_len;
+	struct page **pages = wdata->args.pages;
+	struct page *page;
+	pgoff_t index;
+	u64 temp;
+	int npg_per_block =
+	    NFS_SERVER(wdata->inode)->pnfs_blksize >> PAGE_CACHE_SHIFT;
+
+	dprintk("%s enter, %Zu@%lld\n", __func__, count, offset);
+	/* At this point, wdata->pages is a (sequential) list of nfs_pages.
+	 * We want to write each, and if there is an error set pnfs_error
+	 * to have it redone using nfs.
+	 */
+	par = alloc_parallel(wdata);
+	if (!par)
+		return PNFS_NOT_ATTEMPTED;
+	par->call_ops = *wdata->mds_ops;
+	par->call_ops.rpc_call_done = bl_rpc_do_nothing;
+	par->pnfs_callback = bl_end_par_io_write;
+	/* At this point, have to be more careful with error handling */
+
+	isect = (sector_t) ((offset & (long)PAGE_CACHE_MASK) >> SECTOR_SHIFT);
+	be = bl_find_get_extent(BLK_LSEG2EXT(wdata->lseg), isect, &cow_read);
+	if (!be || !is_writable(be, isect)) {
+		dprintk("%s no matching extents!\n", __func__);
+		wdata->pnfs_error = -EINVAL;
+		goto out;
+	}
+
+	/* First page inside INVALID extent */
+	if (be->be_state == PNFS_BLOCK_INVALID_DATA) {
+		temp = offset >> PAGE_CACHE_SHIFT;
+		npg_zero = do_div(temp, npg_per_block);
+		isect = (sector_t) (((offset - npg_zero * PAGE_CACHE_SIZE) &
+				     (long)PAGE_CACHE_MASK) >> SECTOR_SHIFT);
+		extent_length = be->be_length - (isect - be->be_f_offset);
+
+fill_invalid_ext:
+		dprintk("%s need to zero %d pages\n", __func__, npg_zero);
+		for (;npg_zero > 0; npg_zero--) {
+			if (bl_is_sector_init(be->be_inval, isect)) {
+				dprintk("isect %llu already init\n",
+					(unsigned long long)isect);
+				goto next_page;
+			}
+			/* page ref released in bl_end_io_write_zero */
+			index = isect >> PAGE_CACHE_SECTOR_SHIFT;
+			dprintk("%s zero %dth page: index %lu isect %llu\n",
+				__func__, npg_zero, index,
+				(unsigned long long)isect);
+			page =
+			    find_or_create_page(wdata->inode->i_mapping, index,
+						GFP_NOFS);
+			if (!page) {
+				dprintk("%s oom\n", __func__);
+				wdata->pnfs_error = -ENOMEM;
+				goto out;
+			}
+
+			/* PageDirty: Other will write this out
+			 * PageWriteback: Other is writing this out
+			 * PageUptodate: It was read before
+			 * sector_initialized: already written out
+			 */
+			if (PageDirty(page) || PageWriteback(page)) {
+				print_page(page);
+				unlock_page(page);
+				page_cache_release(page);
+				goto next_page;
+			}
+			if (!PageUptodate(page)) {
+				/* New page, readin or zero it */
+				init_page_for_write(page, cow_read);
+			}
+			set_page_writeback(page);
+			unlock_page(page);
+
+			ret = bl_mark_sectors_init(be->be_inval, isect,
+						       PAGE_CACHE_SECTORS,
+						       NULL);
+			if (unlikely(ret)) {
+				dprintk("%s bl_mark_sectors_init fail %d\n",
+					__func__, ret);
+				end_page_writeback(page);
+				page_cache_release(page);
+				wdata->pnfs_error = ret;
+				goto out;
+			}
+			bio = bl_add_page_to_bio(bio, npg_zero, WRITE,
+						 isect, page, be,
+						 bl_end_io_write_zero, par);
+			if (IS_ERR(bio)) {
+				wdata->pnfs_error = PTR_ERR(bio);
+				bio = NULL;
+				goto out;
+			}
+			/* FIXME: This should be done in bi_end_io */
+			mark_extents_written(BLK_LSEG2EXT(wdata->lseg),
+					     page->index << PAGE_CACHE_SHIFT,
+					     PAGE_CACHE_SIZE);
+next_page:
+			isect += PAGE_CACHE_SECTORS;
+			extent_length -= PAGE_CACHE_SECTORS;
+		}
+		if (last)
+			goto write_done;
+	}
+	bio = bl_submit_bio(WRITE, bio);
+
+	/* Middle pages */
+	pg_index = wdata->args.pgbase >> PAGE_CACHE_SHIFT;
+	for (i = pg_index; i < wdata->npages; i++) {
+		if (!extent_length) {
+			/* We've used up the previous extent */
+			bl_put_extent(be);
+			bl_put_extent(cow_read);
+			bio = bl_submit_bio(WRITE, bio);
+			/* Get the next one */
+			be = bl_find_get_extent(BLK_LSEG2EXT(wdata->lseg),
+					     isect, &cow_read);
+			if (!be || !is_writable(be, isect)) {
+				wdata->pnfs_error = -EINVAL;
+				goto out;
+			}
+			extent_length = be->be_length -
+			    (isect - be->be_f_offset);
+		}
+
+		dprintk("%s offset %lld count %Zu\n", __func__, offset, count);
+		pg_offset = offset & ~PAGE_CACHE_MASK;
+		if (pg_offset + count > PAGE_CACHE_SIZE)
+			pg_len = PAGE_CACHE_SIZE - pg_offset;
+		else
+			pg_len = count;
+
+		saved_len = pg_len;
+		if (be->be_state == PNFS_BLOCK_INVALID_DATA &&
+		    !bl_is_sector_init(be->be_inval, isect)) {
+			ret = bl_read_partial_page_sync(pages[i], cow_read,
+							pg_offset, pg_len, true);
+			if (ret) {
+				dprintk("%s bl_read_partial_page_sync fail %d\n",
+					__func__, ret);
+				wdata->pnfs_error = ret;
+				goto out;
+			}
+
+			ret = bl_mark_sectors_init(be->be_inval, isect,
+						       PAGE_CACHE_SECTORS,
+						       NULL);
+			if (unlikely(ret)) {
+				dprintk("%s bl_mark_sectors_init fail %d\n",
+					__func__, ret);
+				wdata->pnfs_error = ret;
+				goto out;
+			}
+
+			/* Expand to full page write */
+			pg_offset = 0;
+			pg_len = PAGE_CACHE_SIZE;
+		} else if  ((pg_offset & (SECTOR_SIZE - 1)) ||
+			    (pg_len & (SECTOR_SIZE - 1))){
+			/* ahh, nasty case. We have to do sync full sector
+			 * read-modify-write cycles.
+			 */
+			unsigned int saved_offset = pg_offset;
+			ret = bl_read_partial_page_sync(pages[i], be, pg_offset,
+							pg_len, false);
+			pg_offset = round_down(pg_offset, SECTOR_SIZE);
+			pg_len = round_up(saved_offset + pg_len, SECTOR_SIZE)
+				 - pg_offset;
+		}
+
+
+		bio = do_add_page_to_bio(bio, wdata->npages - i, WRITE,
+					 isect, pages[i], be,
+					 bl_end_io_write, par,
+					 pg_offset, pg_len);
+		if (IS_ERR(bio)) {
+			wdata->pnfs_error = PTR_ERR(bio);
+			bio = NULL;
+			goto out;
+		}
+		offset += saved_len;
+		count -= saved_len;
+		isect += PAGE_CACHE_SECTORS;
+		last_isect = isect;
+		extent_length -= PAGE_CACHE_SECTORS;
+	}
+
+	/* Last page inside INVALID extent */
+	if (be->be_state == PNFS_BLOCK_INVALID_DATA) {
+		bio = bl_submit_bio(WRITE, bio);
+		temp = last_isect >> PAGE_CACHE_SECTOR_SHIFT;
+		npg_zero = npg_per_block - do_div(temp, npg_per_block);
+		if (npg_zero < npg_per_block) {
+			last = 1;
+			goto fill_invalid_ext;
+		}
+	}
+
+write_done:
+	wdata->res.count = wdata->args.count;
+out:
+	bl_put_extent(be);
+	bl_put_extent(cow_read);
+	bl_submit_bio(WRITE, bio);
+	put_parallel(par);
+	return PNFS_ATTEMPTED;
+}
+
+/* FIXME - range ignored */
+static void
+release_extents(struct pnfs_block_layout *bl, struct pnfs_layout_range *range)
+{
+	int i;
+	struct pnfs_block_extent *be;
+
+	spin_lock(&bl->bl_ext_lock);
+	for (i = 0; i < EXTENT_LISTS; i++) {
+		while (!list_empty(&bl->bl_extents[i])) {
+			be = list_first_entry(&bl->bl_extents[i],
+					      struct pnfs_block_extent,
+					      be_node);
+			list_del(&be->be_node);
+			bl_put_extent(be);
+		}
+	}
+	spin_unlock(&bl->bl_ext_lock);
+}
+
+static void
+release_inval_marks(struct pnfs_inval_markings *marks)
+{
+	struct pnfs_inval_tracking *pos, *temp;
+
+	list_for_each_entry_safe(pos, temp, &marks->im_tree.mtt_stub, it_link) {
+		list_del(&pos->it_link);
+		kfree(pos);
+	}
+	return;
+}
+
+static void bl_free_layout_hdr(struct pnfs_layout_hdr *lo)
+{
+	struct pnfs_block_layout *bl = BLK_LO2EXT(lo);
+
+	dprintk("%s enter\n", __func__);
+	release_extents(bl, NULL);
+	release_inval_marks(&bl->bl_inval);
+	kfree(bl);
+}
+
+static struct pnfs_layout_hdr *bl_alloc_layout_hdr(struct inode *inode,
+						   gfp_t gfp_flags)
+{
+	struct pnfs_block_layout *bl;
+
+	dprintk("%s enter\n", __func__);
+	bl = kzalloc(sizeof(*bl), gfp_flags);
+	if (!bl)
+		return NULL;
+	spin_lock_init(&bl->bl_ext_lock);
+	INIT_LIST_HEAD(&bl->bl_extents[0]);
+	INIT_LIST_HEAD(&bl->bl_extents[1]);
+	INIT_LIST_HEAD(&bl->bl_commit);
+	INIT_LIST_HEAD(&bl->bl_committing);
+	bl->bl_count = 0;
+	bl->bl_blocksize = NFS_SERVER(inode)->pnfs_blksize >> SECTOR_SHIFT;
+	BL_INIT_INVAL_MARKS(&bl->bl_inval, bl->bl_blocksize);
+	return &bl->bl_layout;
+}
+
+static void bl_free_lseg(struct pnfs_layout_segment *lseg)
+{
+	dprintk("%s enter\n", __func__);
+	kfree(lseg);
+}
+
+/* We pretty much ignore lseg, and store all data layout wide, so we
+ * can correctly merge.
+ */
+static struct pnfs_layout_segment *bl_alloc_lseg(struct pnfs_layout_hdr *lo,
+						 struct nfs4_layoutget_res *lgr,
+						 gfp_t gfp_flags)
+{
+	struct pnfs_layout_segment *lseg;
+	int status;
+
+	dprintk("%s enter\n", __func__);
+	lseg = kzalloc(sizeof(*lseg), gfp_flags);
+	if (!lseg)
+		return ERR_PTR(-ENOMEM);
+	status = nfs4_blk_process_layoutget(lo, lgr, gfp_flags);
+	if (status) {
+		/* We don't want to call the full-blown bl_free_lseg,
+		 * since on error extents were not touched.
+		 */
+		kfree(lseg);
+		return ERR_PTR(status);
+	}
+	return lseg;
+}
+
+static void
+bl_encode_layoutcommit(struct pnfs_layout_hdr *lo, struct xdr_stream *xdr,
+		       const struct nfs4_layoutcommit_args *arg)
+{
+	dprintk("%s enter\n", __func__);
+	encode_pnfs_block_layoutupdate(BLK_LO2EXT(lo), xdr, arg);
+}
+
+static void
+bl_cleanup_layoutcommit(struct nfs4_layoutcommit_data *lcdata)
+{
+	struct pnfs_layout_hdr *lo = NFS_I(lcdata->args.inode)->layout;
+
+	dprintk("%s enter\n", __func__);
+	clean_pnfs_block_layoutupdate(BLK_LO2EXT(lo), &lcdata->args, lcdata->res.status);
+}
+
+static void free_blk_mountid(struct block_mount_id *mid)
+{
+	if (mid) {
+		struct pnfs_block_dev *dev, *tmp;
+
+		/* No need to take bm_lock as we are last user freeing bm_devlist */
+		list_for_each_entry_safe(dev, tmp, &mid->bm_devlist, bm_node) {
+			list_del(&dev->bm_node);
+			bl_free_block_dev(dev);
+		}
+		kfree(mid);
+	}
+}
+
+/* This is mostly copied from the filelayout's get_device_info function.
+ * It seems much of this should be at the generic pnfs level.
+ */
+static struct pnfs_block_dev *
+nfs4_blk_get_deviceinfo(struct nfs_server *server, const struct nfs_fh *fh,
+			struct nfs4_deviceid *d_id)
+{
+	struct pnfs_device *dev;
+	struct pnfs_block_dev *rv;
+	u32 max_resp_sz;
+	int max_pages;
+	struct page **pages = NULL;
+	int i, rc;
+
+	/*
+	 * Use the session max response size as the basis for setting
+	 * GETDEVICEINFO's maxcount
+	 */
+	max_resp_sz = server->nfs_client->cl_session->fc_attrs.max_resp_sz;
+	max_pages = nfs_page_array_len(0, max_resp_sz);
+	dprintk("%s max_resp_sz %u max_pages %d\n",
+		__func__, max_resp_sz, max_pages);
+
+	dev = kmalloc(sizeof(*dev), GFP_NOFS);
+	if (!dev) {
+		dprintk("%s kmalloc failed\n", __func__);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	pages = kzalloc(max_pages * sizeof(struct page *), GFP_NOFS);
+	if (pages == NULL) {
+		kfree(dev);
+		return ERR_PTR(-ENOMEM);
+	}
+	for (i = 0; i < max_pages; i++) {
+		pages[i] = alloc_page(GFP_NOFS);
+		if (!pages[i]) {
+			rv = ERR_PTR(-ENOMEM);
+			goto out_free;
+		}
+	}
+
+	memcpy(&dev->dev_id, d_id, sizeof(*d_id));
+	dev->layout_type = LAYOUT_BLOCK_VOLUME;
+	dev->pages = pages;
+	dev->pgbase = 0;
+	dev->pglen = PAGE_SIZE * max_pages;
+	dev->mincount = 0;
+
+	dprintk("%s: dev_id: %s\n", __func__, dev->dev_id.data);
+	rc = nfs4_proc_getdeviceinfo(server, dev);
+	dprintk("%s getdevice info returns %d\n", __func__, rc);
+	if (rc) {
+		rv = ERR_PTR(rc);
+		goto out_free;
+	}
+
+	rv = nfs4_blk_decode_device(server, dev);
+ out_free:
+	for (i = 0; i < max_pages; i++)
+		__free_page(pages[i]);
+	kfree(pages);
+	kfree(dev);
+	return rv;
+}
+
+static int
+bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh)
+{
+	struct block_mount_id *b_mt_id = NULL;
+	struct pnfs_devicelist *dlist = NULL;
+	struct pnfs_block_dev *bdev;
+	LIST_HEAD(block_disklist);
+	int status, i;
+
+	dprintk("%s enter\n", __func__);
+
+	if (server->pnfs_blksize == 0) {
+		dprintk("%s Server did not return blksize\n", __func__);
+		return -EINVAL;
+	}
+	b_mt_id = kzalloc(sizeof(struct block_mount_id), GFP_NOFS);
+	if (!b_mt_id) {
+		status = -ENOMEM;
+		goto out_error;
+	}
+	/* Initialize nfs4 block layout mount id */
+	spin_lock_init(&b_mt_id->bm_lock);
+	INIT_LIST_HEAD(&b_mt_id->bm_devlist);
+
+	dlist = kmalloc(sizeof(struct pnfs_devicelist), GFP_NOFS);
+	if (!dlist) {
+		status = -ENOMEM;
+		goto out_error;
+	}
+	dlist->eof = 0;
+	while (!dlist->eof) {
+		status = nfs4_proc_getdevicelist(server, fh, dlist);
+		if (status)
+			goto out_error;
+		dprintk("%s GETDEVICELIST numdevs=%i, eof=%i\n",
+			__func__, dlist->num_devs, dlist->eof);
+		for (i = 0; i < dlist->num_devs; i++) {
+			bdev = nfs4_blk_get_deviceinfo(server, fh,
+						       &dlist->dev_id[i]);
+			if (IS_ERR(bdev)) {
+				status = PTR_ERR(bdev);
+				goto out_error;
+			}
+			spin_lock(&b_mt_id->bm_lock);
+			list_add(&bdev->bm_node, &b_mt_id->bm_devlist);
+			spin_unlock(&b_mt_id->bm_lock);
+		}
+	}
+	dprintk("%s SUCCESS\n", __func__);
+	server->pnfs_ld_data = b_mt_id;
+
+ out_return:
+	kfree(dlist);
+	return status;
+
+ out_error:
+	free_blk_mountid(b_mt_id);
+	goto out_return;
+}
+
+static int
+bl_clear_layoutdriver(struct nfs_server *server)
+{
+	struct block_mount_id *b_mt_id = server->pnfs_ld_data;
+
+	dprintk("%s enter\n", __func__);
+	free_blk_mountid(b_mt_id);
+	dprintk("%s RETURNS\n", __func__);
+	return 0;
+}
+
+static const struct nfs_pageio_ops bl_pg_read_ops = {
+	.pg_init = pnfs_generic_pg_init_read,
+	.pg_test = pnfs_generic_pg_test,
+	.pg_doio = pnfs_generic_pg_readpages,
+};
+
+static const struct nfs_pageio_ops bl_pg_write_ops = {
+	.pg_init = pnfs_generic_pg_init_write,
+	.pg_test = pnfs_generic_pg_test,
+	.pg_doio = pnfs_generic_pg_writepages,
+};
+
+static struct pnfs_layoutdriver_type blocklayout_type = {
+	.id				= LAYOUT_BLOCK_VOLUME,
+	.name				= "LAYOUT_BLOCK_VOLUME",
+	.owner				= THIS_MODULE,
+	.read_pagelist			= bl_read_pagelist,
+	.write_pagelist			= bl_write_pagelist,
+	.alloc_layout_hdr		= bl_alloc_layout_hdr,
+	.free_layout_hdr		= bl_free_layout_hdr,
+	.alloc_lseg			= bl_alloc_lseg,
+	.free_lseg			= bl_free_lseg,
+	.encode_layoutcommit		= bl_encode_layoutcommit,
+	.cleanup_layoutcommit		= bl_cleanup_layoutcommit,
+	.set_layoutdriver		= bl_set_layoutdriver,
+	.clear_layoutdriver		= bl_clear_layoutdriver,
+	.pg_read_ops			= &bl_pg_read_ops,
+	.pg_write_ops			= &bl_pg_write_ops,
+};
+
+static const struct rpc_pipe_ops bl_upcall_ops = {
+	.upcall		= rpc_pipe_generic_upcall,
+	.downcall	= bl_pipe_downcall,
+	.destroy_msg	= bl_pipe_destroy_msg,
+};
+
+static int __init nfs4blocklayout_init(void)
+{
+	struct vfsmount *mnt;
+	struct path path;
+	int ret;
+
+	dprintk("%s: NFSv4 Block Layout Driver Registering...\n", __func__);
+
+	ret = pnfs_register_layoutdriver(&blocklayout_type);
+	if (ret)
+		goto out;
+
+	init_waitqueue_head(&bl_wq);
+
+	mnt = rpc_get_mount();
+	if (IS_ERR(mnt)) {
+		ret = PTR_ERR(mnt);
+		goto out_remove;
+	}
+
+	ret = vfs_path_lookup(mnt->mnt_root,
+			      mnt,
+			      NFS_PIPE_DIRNAME, 0, &path);
+	if (ret)
+		goto out_putrpc;
+
+	bl_device_pipe = rpc_mkpipe(path.dentry, "blocklayout", NULL,
+				    &bl_upcall_ops, 0);
+	path_put(&path);
+	if (IS_ERR(bl_device_pipe)) {
+		ret = PTR_ERR(bl_device_pipe);
+		goto out_putrpc;
+	}
+out:
+	return ret;
+
+out_putrpc:
+	rpc_put_mount();
+out_remove:
+	pnfs_unregister_layoutdriver(&blocklayout_type);
+	return ret;
+}
+
+static void __exit nfs4blocklayout_exit(void)
+{
+	dprintk("%s: NFSv4 Block Layout Driver Unregistering...\n",
+	       __func__);
+
+	pnfs_unregister_layoutdriver(&blocklayout_type);
+	rpc_unlink(bl_device_pipe);
+	rpc_put_mount();
+}
+
+MODULE_ALIAS("nfs-layouttype4-3");
+
+module_init(nfs4blocklayout_init);
+module_exit(nfs4blocklayout_exit);
diff --git a/fs/nfs/blocklayout/blocklayout.h b/fs/nfs/blocklayout/blocklayout.h
new file mode 100644
index 0000000..519a9de
--- /dev/null
+++ b/fs/nfs/blocklayout/blocklayout.h
@@ -0,0 +1,206 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayout.h
+ *
+ *  Module for the NFSv4.1 pNFS block layout driver.
+ *
+ *  Copyright (c) 2006 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+#ifndef FS_NFS_NFS4BLOCKLAYOUT_H
+#define FS_NFS_NFS4BLOCKLAYOUT_H
+
+#include <linux/device-mapper.h>
+#include <linux/nfs_fs.h>
+#include <linux/sunrpc/rpc_pipe_fs.h>
+
+#include "../pnfs.h"
+
+#define PAGE_CACHE_SECTORS (PAGE_CACHE_SIZE >> SECTOR_SHIFT)
+#define PAGE_CACHE_SECTOR_SHIFT (PAGE_CACHE_SHIFT - SECTOR_SHIFT)
+#define SECTOR_SIZE (1 << SECTOR_SHIFT)
+
+struct block_mount_id {
+	spinlock_t			bm_lock;    /* protects list */
+	struct list_head		bm_devlist; /* holds pnfs_block_dev */
+};
+
+struct pnfs_block_dev {
+	struct list_head		bm_node;
+	struct nfs4_deviceid		bm_mdevid;    /* associated devid */
+	struct block_device		*bm_mdev;     /* meta device itself */
+};
+
+enum exstate4 {
+	PNFS_BLOCK_READWRITE_DATA	= 0,
+	PNFS_BLOCK_READ_DATA		= 1,
+	PNFS_BLOCK_INVALID_DATA		= 2, /* mapped, but data is invalid */
+	PNFS_BLOCK_NONE_DATA		= 3  /* unmapped, it's a hole */
+};
+
+#define MY_MAX_TAGS (15) /* tag bitnums used must be less than this */
+
+struct my_tree {
+	sector_t		mtt_step_size;	/* Internal sector alignment */
+	struct list_head	mtt_stub; /* Should be a radix tree */
+};
+
+struct pnfs_inval_markings {
+	spinlock_t	im_lock;
+	struct my_tree	im_tree;	/* Sectors that need LAYOUTCOMMIT */
+	sector_t	im_block_size;	/* Server blocksize in sectors */
+};
+
+struct pnfs_inval_tracking {
+	struct list_head it_link;
+	int		 it_sector;
+	int		 it_tags;
+};
+
+/* sector_t fields are all in 512-byte sectors */
+struct pnfs_block_extent {
+	struct kref	be_refcnt;
+	struct list_head be_node;	/* link into lseg list */
+	struct nfs4_deviceid be_devid;  /* FIXME: could use device cache instead */
+	struct block_device *be_mdev;
+	sector_t	be_f_offset;	/* the starting offset in the file */
+	sector_t	be_length;	/* the size of the extent */
+	sector_t	be_v_offset;	/* the starting offset in the volume */
+	enum exstate4	be_state;	/* the state of this extent */
+	struct pnfs_inval_markings *be_inval; /* tracks INVAL->RW transition */
+};
+
+/* Shortened extent used by LAYOUTCOMMIT */
+struct pnfs_block_short_extent {
+	struct list_head bse_node;
+	struct nfs4_deviceid bse_devid;
+	struct block_device *bse_mdev;
+	sector_t	bse_f_offset;	/* the starting offset in the file */
+	sector_t	bse_length;	/* the size of the extent */
+};
+
+static inline void
+BL_INIT_INVAL_MARKS(struct pnfs_inval_markings *marks, sector_t blocksize)
+{
+	spin_lock_init(&marks->im_lock);
+	INIT_LIST_HEAD(&marks->im_tree.mtt_stub);
+	marks->im_block_size = blocksize;
+	marks->im_tree.mtt_step_size = min((sector_t)PAGE_CACHE_SECTORS,
+					   blocksize);
+}
+
+enum extentclass4 {
+	RW_EXTENT       = 0, /* READWRTE and INVAL */
+	RO_EXTENT       = 1, /* READ and NONE */
+	EXTENT_LISTS    = 2,
+};
+
+static inline int bl_choose_list(enum exstate4 state)
+{
+	if (state == PNFS_BLOCK_READ_DATA || state == PNFS_BLOCK_NONE_DATA)
+		return RO_EXTENT;
+	else
+		return RW_EXTENT;
+}
+
+struct pnfs_block_layout {
+	struct pnfs_layout_hdr bl_layout;
+	struct pnfs_inval_markings bl_inval; /* tracks INVAL->RW transition */
+	spinlock_t		bl_ext_lock;   /* Protects list manipulation */
+	struct list_head	bl_extents[EXTENT_LISTS]; /* R and RW extents */
+	struct list_head	bl_commit;	/* Needs layout commit */
+	struct list_head	bl_committing;	/* Layout committing */
+	unsigned int		bl_count;	/* entries in bl_commit */
+	sector_t		bl_blocksize;  /* Server blocksize in sectors */
+};
+
+#define BLK_ID(lo) ((struct block_mount_id *)(NFS_SERVER(lo->plh_inode)->pnfs_ld_data))
+
+static inline struct pnfs_block_layout *
+BLK_LO2EXT(struct pnfs_layout_hdr *lo)
+{
+	return container_of(lo, struct pnfs_block_layout, bl_layout);
+}
+
+static inline struct pnfs_block_layout *
+BLK_LSEG2EXT(struct pnfs_layout_segment *lseg)
+{
+	return BLK_LO2EXT(lseg->pls_layout);
+}
+
+struct bl_dev_msg {
+	int32_t status;
+	uint32_t major, minor;
+};
+
+struct bl_msg_hdr {
+	u8  type;
+	u16 totallen; /* length of entire message, including hdr itself */
+};
+
+extern struct dentry *bl_device_pipe;
+extern wait_queue_head_t bl_wq;
+
+#define BL_DEVICE_UMOUNT               0x0 /* Umount--delete devices */
+#define BL_DEVICE_MOUNT                0x1 /* Mount--create devices*/
+#define BL_DEVICE_REQUEST_INIT         0x0 /* Start request */
+#define BL_DEVICE_REQUEST_PROC         0x1 /* User level process succeeds */
+#define BL_DEVICE_REQUEST_ERR          0x2 /* User level process fails */
+
+/* blocklayoutdev.c */
+ssize_t bl_pipe_downcall(struct file *, const char __user *, size_t);
+void bl_pipe_destroy_msg(struct rpc_pipe_msg *);
+struct block_device *nfs4_blkdev_get(dev_t dev);
+int nfs4_blkdev_put(struct block_device *bdev);
+struct pnfs_block_dev *nfs4_blk_decode_device(struct nfs_server *server,
+						struct pnfs_device *dev);
+int nfs4_blk_process_layoutget(struct pnfs_layout_hdr *lo,
+				struct nfs4_layoutget_res *lgr, gfp_t gfp_flags);
+
+/* blocklayoutdm.c */
+void bl_free_block_dev(struct pnfs_block_dev *bdev);
+
+/* extents.c */
+struct pnfs_block_extent *
+bl_find_get_extent(struct pnfs_block_layout *bl, sector_t isect,
+		struct pnfs_block_extent **cow_read);
+int bl_mark_sectors_init(struct pnfs_inval_markings *marks,
+			     sector_t offset, sector_t length,
+			     sector_t **pages);
+void bl_put_extent(struct pnfs_block_extent *be);
+struct pnfs_block_extent *bl_alloc_extent(void);
+int bl_is_sector_init(struct pnfs_inval_markings *marks, sector_t isect);
+int encode_pnfs_block_layoutupdate(struct pnfs_block_layout *bl,
+				   struct xdr_stream *xdr,
+				   const struct nfs4_layoutcommit_args *arg);
+void clean_pnfs_block_layoutupdate(struct pnfs_block_layout *bl,
+				   const struct nfs4_layoutcommit_args *arg,
+				   int status);
+int bl_add_merge_extent(struct pnfs_block_layout *bl,
+			 struct pnfs_block_extent *new);
+int bl_mark_for_commit(struct pnfs_block_extent *be,
+			sector_t offset, sector_t length);
+
+#endif /* FS_NFS_NFS4BLOCKLAYOUT_H */
diff --git a/fs/nfs/blocklayout/blocklayoutdev.c b/fs/nfs/blocklayout/blocklayoutdev.c
new file mode 100644
index 0000000..d08ba91
--- /dev/null
+++ b/fs/nfs/blocklayout/blocklayoutdev.c
@@ -0,0 +1,391 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayoutdev.c
+ *
+ *  Device operations for the pnfs nfs4 file layout driver.
+ *
+ *  Copyright (c) 2006 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+#include <linux/module.h>
+#include <linux/buffer_head.h> /* __bread */
+
+#include <linux/genhd.h>
+#include <linux/blkdev.h>
+#include <linux/hash.h>
+
+#include "blocklayout.h"
+
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+
+static int decode_sector_number(__be32 **rp, sector_t *sp)
+{
+	uint64_t s;
+
+	*rp = xdr_decode_hyper(*rp, &s);
+	if (s & 0x1ff) {
+		printk(KERN_WARNING "%s: sector not aligned\n", __func__);
+		return -1;
+	}
+	*sp = s >> SECTOR_SHIFT;
+	return 0;
+}
+
+/* Open a block_device by device number. */
+struct block_device *nfs4_blkdev_get(dev_t dev)
+{
+	struct block_device *bd;
+
+	dprintk("%s enter\n", __func__);
+	bd = blkdev_get_by_dev(dev, FMODE_READ, NULL);
+	if (IS_ERR(bd))
+		goto fail;
+	return bd;
+fail:
+	dprintk("%s failed to open device : %ld\n",
+			__func__, PTR_ERR(bd));
+	return NULL;
+}
+
+/*
+ * Release the block device
+ */
+int nfs4_blkdev_put(struct block_device *bdev)
+{
+	dprintk("%s for device %d:%d\n", __func__, MAJOR(bdev->bd_dev),
+			MINOR(bdev->bd_dev));
+	return blkdev_put(bdev, FMODE_READ);
+}
+
+static struct bl_dev_msg bl_mount_reply;
+
+ssize_t bl_pipe_downcall(struct file *filp, const char __user *src,
+			 size_t mlen)
+{
+	if (mlen != sizeof (struct bl_dev_msg))
+		return -EINVAL;
+
+	if (copy_from_user(&bl_mount_reply, src, mlen) != 0)
+		return -EFAULT;
+
+	wake_up(&bl_wq);
+
+	return mlen;
+}
+
+void bl_pipe_destroy_msg(struct rpc_pipe_msg *msg)
+{
+	if (msg->errno >= 0)
+		return;
+	wake_up(&bl_wq);
+}
+
+/*
+ * Decodes pnfs_block_deviceaddr4 which is XDR encoded in dev->dev_addr_buf.
+ */
+struct pnfs_block_dev *
+nfs4_blk_decode_device(struct nfs_server *server,
+		       struct pnfs_device *dev)
+{
+	struct pnfs_block_dev *rv;
+	struct block_device *bd = NULL;
+	struct rpc_pipe_msg msg;
+	struct bl_msg_hdr bl_msg = {
+		.type = BL_DEVICE_MOUNT,
+		.totallen = dev->mincount,
+	};
+	uint8_t *dataptr;
+	DECLARE_WAITQUEUE(wq, current);
+	struct bl_dev_msg *reply = &bl_mount_reply;
+	int offset, len, i, rc;
+
+	dprintk("%s CREATING PIPEFS MESSAGE\n", __func__);
+	dprintk("%s: deviceid: %s, mincount: %d\n", __func__, dev->dev_id.data,
+		dev->mincount);
+
+	memset(&msg, 0, sizeof(msg));
+	msg.data = kzalloc(sizeof(bl_msg) + dev->mincount, GFP_NOFS);
+	if (!msg.data) {
+		rv = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	memcpy(msg.data, &bl_msg, sizeof(bl_msg));
+	dataptr = (uint8_t *) msg.data;
+	len = dev->mincount;
+	offset = sizeof(bl_msg);
+	for (i = 0; len > 0; i++) {
+		memcpy(&dataptr[offset], page_address(dev->pages[i]),
+				len < PAGE_CACHE_SIZE ? len : PAGE_CACHE_SIZE);
+		len -= PAGE_CACHE_SIZE;
+		offset += PAGE_CACHE_SIZE;
+	}
+	msg.len = sizeof(bl_msg) + dev->mincount;
+
+	dprintk("%s CALLING USERSPACE DAEMON\n", __func__);
+	add_wait_queue(&bl_wq, &wq);
+	rc = rpc_queue_upcall(bl_device_pipe->d_inode, &msg);
+	if (rc < 0) {
+		remove_wait_queue(&bl_wq, &wq);
+		rv = ERR_PTR(rc);
+		goto out;
+	}
+
+	set_current_state(TASK_UNINTERRUPTIBLE);
+	schedule();
+	__set_current_state(TASK_RUNNING);
+	remove_wait_queue(&bl_wq, &wq);
+
+	if (reply->status != BL_DEVICE_REQUEST_PROC) {
+		dprintk("%s failed to open device: %d\n",
+			__func__, reply->status);
+		rv = ERR_PTR(-EINVAL);
+		goto out;
+	}
+
+	bd = nfs4_blkdev_get(MKDEV(reply->major, reply->minor));
+	if (IS_ERR(bd)) {
+		rc = PTR_ERR(bd);
+		dprintk("%s failed to open device : %d\n", __func__, rc);
+		rv = ERR_PTR(rc);
+		goto out;
+	}
+
+	rv = kzalloc(sizeof(*rv), GFP_NOFS);
+	if (!rv) {
+		rv = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	rv->bm_mdev = bd;
+	memcpy(&rv->bm_mdevid, &dev->dev_id, sizeof(struct nfs4_deviceid));
+	dprintk("%s Created device %s with bd_block_size %u\n",
+		__func__,
+		bd->bd_disk->disk_name,
+		bd->bd_block_size);
+
+out:
+	kfree(msg.data);
+	return rv;
+}
+
+/* Map deviceid returned by the server to constructed block_device */
+static struct block_device *translate_devid(struct pnfs_layout_hdr *lo,
+					    struct nfs4_deviceid *id)
+{
+	struct block_device *rv = NULL;
+	struct block_mount_id *mid;
+	struct pnfs_block_dev *dev;
+
+	dprintk("%s enter, lo=%p, id=%p\n", __func__, lo, id);
+	mid = BLK_ID(lo);
+	spin_lock(&mid->bm_lock);
+	list_for_each_entry(dev, &mid->bm_devlist, bm_node) {
+		if (memcmp(id->data, dev->bm_mdevid.data,
+			   NFS4_DEVICEID4_SIZE) == 0) {
+			rv = dev->bm_mdev;
+			goto out;
+		}
+	}
+ out:
+	spin_unlock(&mid->bm_lock);
+	dprintk("%s returning %p\n", __func__, rv);
+	return rv;
+}
+
+/* Tracks info needed to ensure extents in layout obey constraints of spec */
+struct layout_verification {
+	u32 mode;	/* R or RW */
+	u64 start;	/* Expected start of next non-COW extent */
+	u64 inval;	/* Start of INVAL coverage */
+	u64 cowread;	/* End of COW read coverage */
+};
+
+/* Verify the extent meets the layout requirements of the pnfs-block draft,
+ * section 2.3.1.
+ */
+static int verify_extent(struct pnfs_block_extent *be,
+			 struct layout_verification *lv)
+{
+	if (lv->mode == IOMODE_READ) {
+		if (be->be_state == PNFS_BLOCK_READWRITE_DATA ||
+		    be->be_state == PNFS_BLOCK_INVALID_DATA)
+			return -EIO;
+		if (be->be_f_offset != lv->start)
+			return -EIO;
+		lv->start += be->be_length;
+		return 0;
+	}
+	/* lv->mode == IOMODE_RW */
+	if (be->be_state == PNFS_BLOCK_READWRITE_DATA) {
+		if (be->be_f_offset != lv->start)
+			return -EIO;
+		if (lv->cowread > lv->start)
+			return -EIO;
+		lv->start += be->be_length;
+		lv->inval = lv->start;
+		return 0;
+	} else if (be->be_state == PNFS_BLOCK_INVALID_DATA) {
+		if (be->be_f_offset != lv->start)
+			return -EIO;
+		lv->start += be->be_length;
+		return 0;
+	} else if (be->be_state == PNFS_BLOCK_READ_DATA) {
+		if (be->be_f_offset > lv->start)
+			return -EIO;
+		if (be->be_f_offset < lv->inval)
+			return -EIO;
+		if (be->be_f_offset < lv->cowread)
+			return -EIO;
+		/* It looks like you might want to min this with lv->start,
+		 * but you really don't.
+		 */
+		lv->inval = lv->inval + be->be_length;
+		lv->cowread = be->be_f_offset + be->be_length;
+		return 0;
+	} else
+		return -EIO;
+}
+
+/* XDR decode pnfs_block_layout4 structure */
+int
+nfs4_blk_process_layoutget(struct pnfs_layout_hdr *lo,
+			   struct nfs4_layoutget_res *lgr, gfp_t gfp_flags)
+{
+	struct pnfs_block_layout *bl = BLK_LO2EXT(lo);
+	int i, status = -EIO;
+	uint32_t count;
+	struct pnfs_block_extent *be = NULL, *save;
+	struct xdr_stream stream;
+	struct xdr_buf buf;
+	struct page *scratch;
+	__be32 *p;
+	struct layout_verification lv = {
+		.mode = lgr->range.iomode,
+		.start = lgr->range.offset >> SECTOR_SHIFT,
+		.inval = lgr->range.offset >> SECTOR_SHIFT,
+		.cowread = lgr->range.offset >> SECTOR_SHIFT,
+	};
+	LIST_HEAD(extents);
+
+	dprintk("---> %s\n", __func__);
+
+	scratch = alloc_page(gfp_flags);
+	if (!scratch)
+		return -ENOMEM;
+
+	xdr_init_decode_pages(&stream, &buf, lgr->layoutp->pages, lgr->layoutp->len);
+	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
+
+	p = xdr_inline_decode(&stream, 4);
+	if (unlikely(!p))
+		goto out_err;
+
+	count = be32_to_cpup(p++);
+
+	dprintk("%s enter, number of extents %i\n", __func__, count);
+	p = xdr_inline_decode(&stream, (28 + NFS4_DEVICEID4_SIZE) * count);
+	if (unlikely(!p))
+		goto out_err;
+
+	/* Decode individual extents, putting them in temporary
+	 * staging area until whole layout is decoded to make error
+	 * recovery easier.
+	 */
+	for (i = 0; i < count; i++) {
+		be = bl_alloc_extent();
+		if (!be) {
+			status = -ENOMEM;
+			goto out_err;
+		}
+		memcpy(&be->be_devid, p, NFS4_DEVICEID4_SIZE);
+		p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE);
+		be->be_mdev = translate_devid(lo, &be->be_devid);
+		if (!be->be_mdev)
+			goto out_err;
+
+		/* The next three values are read in as bytes,
+		 * but stored as 512-byte sector lengths
+		 */
+		if (decode_sector_number(&p, &be->be_f_offset) < 0)
+			goto out_err;
+		if (decode_sector_number(&p, &be->be_length) < 0)
+			goto out_err;
+		if (decode_sector_number(&p, &be->be_v_offset) < 0)
+			goto out_err;
+		be->be_state = be32_to_cpup(p++);
+		if (be->be_state == PNFS_BLOCK_INVALID_DATA)
+			be->be_inval = &bl->bl_inval;
+		if (verify_extent(be, &lv)) {
+			dprintk("%s verify failed\n", __func__);
+			goto out_err;
+		}
+		list_add_tail(&be->be_node, &extents);
+	}
+	if (lgr->range.offset + lgr->range.length !=
+			lv.start << SECTOR_SHIFT) {
+		dprintk("%s Final length mismatch\n", __func__);
+		be = NULL;
+		goto out_err;
+	}
+	if (lv.start < lv.cowread) {
+		dprintk("%s Final uncovered COW extent\n", __func__);
+		be = NULL;
+		goto out_err;
+	}
+	/* Extents decoded properly, now try to merge them in to
+	 * existing layout extents.
+	 */
+	spin_lock(&bl->bl_ext_lock);
+	list_for_each_entry_safe(be, save, &extents, be_node) {
+		list_del(&be->be_node);
+		status = bl_add_merge_extent(bl, be);
+		if (status) {
+			spin_unlock(&bl->bl_ext_lock);
+			/* This is a fairly catastrophic error, as the
+			 * entire layout extent lists are now corrupted.
+			 * We should have some way to distinguish this.
+			 */
+			be = NULL;
+			goto out_err;
+		}
+	}
+	spin_unlock(&bl->bl_ext_lock);
+	status = 0;
+ out:
+	__free_page(scratch);
+	dprintk("%s returns %i\n", __func__, status);
+	return status;
+
+ out_err:
+	bl_put_extent(be);
+	while (!list_empty(&extents)) {
+		be = list_first_entry(&extents, struct pnfs_block_extent,
+				      be_node);
+		list_del(&be->be_node);
+		bl_put_extent(be);
+	}
+	goto out;
+}
diff --git a/fs/nfs/blocklayout/blocklayoutdm.c b/fs/nfs/blocklayout/blocklayoutdm.c
new file mode 100644
index 0000000..7326e6e
--- /dev/null
+++ b/fs/nfs/blocklayout/blocklayoutdm.c
@@ -0,0 +1,111 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayoutdm.c
+ *
+ *  Module for the NFSv4.1 pNFS block layout driver.
+ *
+ *  Copyright (c) 2007 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Fred Isaman <iisaman@umich.edu>
+ *  Andy Adamson <andros@citi.umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+
+#include <linux/genhd.h> /* gendisk - used in a dprintk*/
+#include <linux/sched.h>
+#include <linux/hash.h>
+
+#include "blocklayout.h"
+
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+
+static void dev_remove(dev_t dev)
+{
+	struct rpc_pipe_msg msg;
+	struct bl_dev_msg bl_umount_request;
+	struct bl_msg_hdr bl_msg = {
+		.type = BL_DEVICE_UMOUNT,
+		.totallen = sizeof(bl_umount_request),
+	};
+	uint8_t *dataptr;
+	DECLARE_WAITQUEUE(wq, current);
+
+	dprintk("Entering %s\n", __func__);
+
+	memset(&msg, 0, sizeof(msg));
+	msg.len = sizeof(bl_msg) + bl_msg.totallen;
+	msg.data = kzalloc(msg.len, GFP_NOFS);
+	if (!msg.data)
+		goto out;
+
+	memset(&bl_umount_request, 0, sizeof(bl_umount_request));
+	bl_umount_request.major = MAJOR(dev);
+	bl_umount_request.minor = MINOR(dev);
+
+	memcpy(msg.data, &bl_msg, sizeof(bl_msg));
+	dataptr = (uint8_t *) msg.data;
+	memcpy(&dataptr[sizeof(bl_msg)], &bl_umount_request, sizeof(bl_umount_request));
+
+	add_wait_queue(&bl_wq, &wq);
+	if (rpc_queue_upcall(bl_device_pipe->d_inode, &msg) < 0) {
+		remove_wait_queue(&bl_wq, &wq);
+		goto out;
+	}
+
+	set_current_state(TASK_UNINTERRUPTIBLE);
+	schedule();
+	__set_current_state(TASK_RUNNING);
+	remove_wait_queue(&bl_wq, &wq);
+
+out:
+	kfree(msg.data);
+}
+
+/*
+ * Release meta device
+ */
+static void nfs4_blk_metadev_release(struct pnfs_block_dev *bdev)
+{
+	int rv;
+
+	dprintk("%s Releasing\n", __func__);
+	rv = nfs4_blkdev_put(bdev->bm_mdev);
+	if (rv)
+		printk(KERN_ERR "%s nfs4_blkdev_put returns %d\n",
+				__func__, rv);
+
+	dev_remove(bdev->bm_mdev->bd_dev);
+}
+
+void bl_free_block_dev(struct pnfs_block_dev *bdev)
+{
+	if (bdev) {
+		if (bdev->bm_mdev) {
+			dprintk("%s Removing DM device: %d:%d\n",
+				__func__,
+				MAJOR(bdev->bm_mdev->bd_dev),
+				MINOR(bdev->bm_mdev->bd_dev));
+			nfs4_blk_metadev_release(bdev);
+		}
+		kfree(bdev);
+	}
+}
diff --git a/fs/nfs/blocklayout/extents.c b/fs/nfs/blocklayout/extents.c
new file mode 100644
index 0000000..bb15ccc
--- /dev/null
+++ b/fs/nfs/blocklayout/extents.c
@@ -0,0 +1,936 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayout.h
+ *
+ *  Module for the NFSv4.1 pNFS block layout driver.
+ *
+ *  Copyright (c) 2006 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+
+#include "blocklayout.h"
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+
+/* Bit numbers */
+#define EXTENT_INITIALIZED 0
+#define EXTENT_WRITTEN     1
+#define EXTENT_IN_COMMIT   2
+#define INTERNAL_EXISTS    MY_MAX_TAGS
+#define INTERNAL_MASK      ((1 << INTERNAL_EXISTS) - 1)
+
+/* Returns largest t<=s s.t. t%base==0 */
+static inline sector_t normalize(sector_t s, int base)
+{
+	sector_t tmp = s; /* Since do_div modifies its argument */
+	return s - sector_div(tmp, base);
+}
+
+static inline sector_t normalize_up(sector_t s, int base)
+{
+	return normalize(s + base - 1, base);
+}
+
+/* Complete stub using list while determine API wanted */
+
+/* Returns tags, or negative */
+static int32_t _find_entry(struct my_tree *tree, u64 s)
+{
+	struct pnfs_inval_tracking *pos;
+
+	dprintk("%s(%llu) enter\n", __func__, s);
+	list_for_each_entry_reverse(pos, &tree->mtt_stub, it_link) {
+		if (pos->it_sector > s)
+			continue;
+		else if (pos->it_sector == s)
+			return pos->it_tags & INTERNAL_MASK;
+		else
+			break;
+	}
+	return -ENOENT;
+}
+
+static inline
+int _has_tag(struct my_tree *tree, u64 s, int32_t tag)
+{
+	int32_t tags;
+
+	dprintk("%s(%llu, %i) enter\n", __func__, s, tag);
+	s = normalize(s, tree->mtt_step_size);
+	tags = _find_entry(tree, s);
+	if ((tags < 0) || !(tags & (1 << tag)))
+		return 0;
+	else
+		return 1;
+}
+
+/* Creates entry with tag, or if entry already exists, unions tag to it.
+ * If storage is not NULL, newly created entry will use it.
+ * Returns number of entries added, or negative on error.
+ */
+static int _add_entry(struct my_tree *tree, u64 s, int32_t tag,
+		      struct pnfs_inval_tracking *storage)
+{
+	int found = 0;
+	struct pnfs_inval_tracking *pos;
+
+	dprintk("%s(%llu, %i, %p) enter\n", __func__, s, tag, storage);
+	list_for_each_entry_reverse(pos, &tree->mtt_stub, it_link) {
+		if (pos->it_sector > s)
+			continue;
+		else if (pos->it_sector == s) {
+			found = 1;
+			break;
+		} else
+			break;
+	}
+	if (found) {
+		pos->it_tags |= (1 << tag);
+		return 0;
+	} else {
+		struct pnfs_inval_tracking *new;
+		if (storage)
+			new = storage;
+		else {
+			new = kmalloc(sizeof(*new), GFP_NOFS);
+			if (!new)
+				return -ENOMEM;
+		}
+		new->it_sector = s;
+		new->it_tags = (1 << tag);
+		list_add(&new->it_link, &pos->it_link);
+		return 1;
+	}
+}
+
+/* XXXX Really want option to not create */
+/* Over range, unions tag with existing entries, else creates entry with tag */
+static int _set_range(struct my_tree *tree, int32_t tag, u64 s, u64 length)
+{
+	u64 i;
+
+	dprintk("%s(%i, %llu, %llu) enter\n", __func__, tag, s, length);
+	for (i = normalize(s, tree->mtt_step_size); i < s + length;
+	     i += tree->mtt_step_size)
+		if (_add_entry(tree, i, tag, NULL))
+			return -ENOMEM;
+	return 0;
+}
+
+/* Ensure that future operations on given range of tree will not malloc */
+static int _preload_range(struct pnfs_inval_markings *marks,
+		u64 offset, u64 length)
+{
+	u64 start, end, s;
+	int count, i, used = 0, status = -ENOMEM;
+	struct pnfs_inval_tracking **storage;
+	struct my_tree  *tree = &marks->im_tree;
+
+	dprintk("%s(%llu, %llu) enter\n", __func__, offset, length);
+	start = normalize(offset, tree->mtt_step_size);
+	end = normalize_up(offset + length, tree->mtt_step_size);
+	count = (int)(end - start) / (int)tree->mtt_step_size;
+
+	/* Pre-malloc what memory we might need */
+	storage = kcalloc(count, sizeof(*storage), GFP_NOFS);
+	if (!storage)
+		return -ENOMEM;
+	for (i = 0; i < count; i++) {
+		storage[i] = kmalloc(sizeof(struct pnfs_inval_tracking),
+				     GFP_NOFS);
+		if (!storage[i])
+			goto out_cleanup;
+	}
+
+	spin_lock(&marks->im_lock);
+	for (s = start; s < end; s += tree->mtt_step_size)
+		used += _add_entry(tree, s, INTERNAL_EXISTS, storage[used]);
+	spin_unlock(&marks->im_lock);
+
+	status = 0;
+
+ out_cleanup:
+	for (i = used; i < count; i++) {
+		if (!storage[i])
+			break;
+		kfree(storage[i]);
+	}
+	kfree(storage);
+	return status;
+}
+
+static void set_needs_init(sector_t *array, sector_t offset)
+{
+	sector_t *p = array;
+
+	dprintk("%s enter\n", __func__);
+	if (!p)
+		return;
+	while (*p < offset)
+		p++;
+	if (*p == offset)
+		return;
+	else if (*p == ~0) {
+		*p++ = offset;
+		*p = ~0;
+		return;
+	} else {
+		sector_t *save = p;
+		dprintk("%s Adding %llu\n", __func__, (u64)offset);
+		while (*p != ~0)
+			p++;
+		p++;
+		memmove(save + 1, save, (char *)p - (char *)save);
+		*save = offset;
+		return;
+	}
+}
+
+/* We are relying on page lock to serialize this */
+int bl_is_sector_init(struct pnfs_inval_markings *marks, sector_t isect)
+{
+	int rv;
+
+	spin_lock(&marks->im_lock);
+	rv = _has_tag(&marks->im_tree, isect, EXTENT_INITIALIZED);
+	spin_unlock(&marks->im_lock);
+	return rv;
+}
+
+/* Assume start, end already sector aligned */
+static int
+_range_has_tag(struct my_tree *tree, u64 start, u64 end, int32_t tag)
+{
+	struct pnfs_inval_tracking *pos;
+	u64 expect = 0;
+
+	dprintk("%s(%llu, %llu, %i) enter\n", __func__, start, end, tag);
+	list_for_each_entry_reverse(pos, &tree->mtt_stub, it_link) {
+		if (pos->it_sector >= end)
+			continue;
+		if (!expect) {
+			if ((pos->it_sector == end - tree->mtt_step_size) &&
+			    (pos->it_tags & (1 << tag))) {
+				expect = pos->it_sector - tree->mtt_step_size;
+				if (pos->it_sector < tree->mtt_step_size || expect < start)
+					return 1;
+				continue;
+			} else {
+				return 0;
+			}
+		}
+		if (pos->it_sector != expect || !(pos->it_tags & (1 << tag)))
+			return 0;
+		expect -= tree->mtt_step_size;
+		if (expect < start)
+			return 1;
+	}
+	return 0;
+}
+
+static int is_range_written(struct pnfs_inval_markings *marks,
+			    sector_t start, sector_t end)
+{
+	int rv;
+
+	spin_lock(&marks->im_lock);
+	rv = _range_has_tag(&marks->im_tree, start, end, EXTENT_WRITTEN);
+	spin_unlock(&marks->im_lock);
+	return rv;
+}
+
+/* Marks sectors in [offest, offset_length) as having been initialized.
+ * All lengths are step-aligned, where step is min(pagesize, blocksize).
+ * Notes where partial block is initialized, and helps prepare it for
+ * complete initialization later.
+ */
+/* Currently assumes offset is page-aligned */
+int bl_mark_sectors_init(struct pnfs_inval_markings *marks,
+			     sector_t offset, sector_t length,
+			     sector_t **pages)
+{
+	sector_t s, start, end;
+	sector_t *array = NULL; /* Pages to mark */
+
+	dprintk("%s(offset=%llu,len=%llu) enter\n",
+		__func__, (u64)offset, (u64)length);
+	s = max((sector_t) 3,
+		2 * (marks->im_block_size / (PAGE_CACHE_SECTORS)));
+	dprintk("%s set max=%llu\n", __func__, (u64)s);
+	if (pages) {
+		array = kmalloc(s * sizeof(sector_t), GFP_NOFS);
+		if (!array)
+			goto outerr;
+		array[0] = ~0;
+	}
+
+	start = normalize(offset, marks->im_block_size);
+	end = normalize_up(offset + length, marks->im_block_size);
+	if (_preload_range(marks, start, end - start))
+		goto outerr;
+
+	spin_lock(&marks->im_lock);
+
+	for (s = normalize_up(start, PAGE_CACHE_SECTORS);
+	     s < offset; s += PAGE_CACHE_SECTORS) {
+		dprintk("%s pre-area pages\n", __func__);
+		/* Portion of used block is not initialized */
+		if (!_has_tag(&marks->im_tree, s, EXTENT_INITIALIZED))
+			set_needs_init(array, s);
+	}
+	if (_set_range(&marks->im_tree, EXTENT_INITIALIZED, offset, length))
+		goto out_unlock;
+	for (s = normalize_up(offset + length, PAGE_CACHE_SECTORS);
+	     s < end; s += PAGE_CACHE_SECTORS) {
+		dprintk("%s post-area pages\n", __func__);
+		if (!_has_tag(&marks->im_tree, s, EXTENT_INITIALIZED))
+			set_needs_init(array, s);
+	}
+
+	spin_unlock(&marks->im_lock);
+
+	if (pages) {
+		if (array[0] == ~0) {
+			kfree(array);
+			*pages = NULL;
+		} else
+			*pages = array;
+	}
+	return 0;
+
+ out_unlock:
+	spin_unlock(&marks->im_lock);
+ outerr:
+	if (pages) {
+		kfree(array);
+		*pages = NULL;
+	}
+	return -ENOMEM;
+}
+
+/* Marks sectors in [offest, offset+length) as having been written to disk.
+ * All lengths should be block aligned.
+ */
+static int mark_written_sectors(struct pnfs_inval_markings *marks,
+				sector_t offset, sector_t length)
+{
+	int status;
+
+	dprintk("%s(offset=%llu,len=%llu) enter\n", __func__,
+		(u64)offset, (u64)length);
+	spin_lock(&marks->im_lock);
+	status = _set_range(&marks->im_tree, EXTENT_WRITTEN, offset, length);
+	spin_unlock(&marks->im_lock);
+	return status;
+}
+
+static void print_short_extent(struct pnfs_block_short_extent *be)
+{
+	dprintk("PRINT SHORT EXTENT extent %p\n", be);
+	if (be) {
+		dprintk("        be_f_offset %llu\n", (u64)be->bse_f_offset);
+		dprintk("        be_length   %llu\n", (u64)be->bse_length);
+	}
+}
+
+static void print_clist(struct list_head *list, unsigned int count)
+{
+	struct pnfs_block_short_extent *be;
+	unsigned int i = 0;
+
+	ifdebug(FACILITY) {
+		printk(KERN_DEBUG "****************\n");
+		printk(KERN_DEBUG "Extent list looks like:\n");
+		list_for_each_entry(be, list, bse_node) {
+			i++;
+			print_short_extent(be);
+		}
+		if (i != count)
+			printk(KERN_DEBUG "\n\nExpected %u entries\n\n\n", count);
+		printk(KERN_DEBUG "****************\n");
+	}
+}
+
+/* Note: In theory, we should do more checking that devid's match between
+ * old and new, but if they don't, the lists are too corrupt to salvage anyway.
+ */
+/* Note this is very similar to bl_add_merge_extent */
+static void add_to_commitlist(struct pnfs_block_layout *bl,
+			      struct pnfs_block_short_extent *new)
+{
+	struct list_head *clist = &bl->bl_commit;
+	struct pnfs_block_short_extent *old, *save;
+	sector_t end = new->bse_f_offset + new->bse_length;
+
+	dprintk("%s enter\n", __func__);
+	print_short_extent(new);
+	print_clist(clist, bl->bl_count);
+	bl->bl_count++;
+	/* Scan for proper place to insert, extending new to the left
+	 * as much as possible.
+	 */
+	list_for_each_entry_safe(old, save, clist, bse_node) {
+		if (new->bse_f_offset < old->bse_f_offset)
+			break;
+		if (end <= old->bse_f_offset + old->bse_length) {
+			/* Range is already in list */
+			bl->bl_count--;
+			kfree(new);
+			return;
+		} else if (new->bse_f_offset <=
+				old->bse_f_offset + old->bse_length) {
+			/* new overlaps or abuts existing be */
+			if (new->bse_mdev == old->bse_mdev) {
+				/* extend new to fully replace old */
+				new->bse_length += new->bse_f_offset -
+						old->bse_f_offset;
+				new->bse_f_offset = old->bse_f_offset;
+				list_del(&old->bse_node);
+				bl->bl_count--;
+				kfree(old);
+			}
+		}
+	}
+	/* Note that if we never hit the above break, old will not point to a
+	 * valid extent.  However, in that case &old->bse_node==list.
+	 */
+	list_add_tail(&new->bse_node, &old->bse_node);
+	/* Scan forward for overlaps.  If we find any, extend new and
+	 * remove the overlapped extent.
+	 */
+	old = list_prepare_entry(new, clist, bse_node);
+	list_for_each_entry_safe_continue(old, save, clist, bse_node) {
+		if (end < old->bse_f_offset)
+			break;
+		/* new overlaps or abuts old */
+		if (new->bse_mdev == old->bse_mdev) {
+			if (end < old->bse_f_offset + old->bse_length) {
+				/* extend new to fully cover old */
+				end = old->bse_f_offset + old->bse_length;
+				new->bse_length = end - new->bse_f_offset;
+			}
+			list_del(&old->bse_node);
+			bl->bl_count--;
+			kfree(old);
+		}
+	}
+	dprintk("%s: after merging\n", __func__);
+	print_clist(clist, bl->bl_count);
+}
+
+/* Note the range described by offset, length is guaranteed to be contained
+ * within be.
+ */
+int bl_mark_for_commit(struct pnfs_block_extent *be,
+		    sector_t offset, sector_t length)
+{
+	sector_t new_end, end = offset + length;
+	struct pnfs_block_short_extent *new;
+	struct pnfs_block_layout *bl = container_of(be->be_inval,
+						    struct pnfs_block_layout,
+						    bl_inval);
+
+	new = kmalloc(sizeof(*new), GFP_NOFS);
+	if (!new)
+		return -ENOMEM;
+
+	mark_written_sectors(be->be_inval, offset, length);
+	/* We want to add the range to commit list, but it must be
+	 * block-normalized, and verified that the normalized range has
+	 * been entirely written to disk.
+	 */
+	new->bse_f_offset = offset;
+	offset = normalize(offset, bl->bl_blocksize);
+	if (offset < new->bse_f_offset) {
+		if (is_range_written(be->be_inval, offset, new->bse_f_offset))
+			new->bse_f_offset = offset;
+		else
+			new->bse_f_offset = offset + bl->bl_blocksize;
+	}
+	new_end = normalize_up(end, bl->bl_blocksize);
+	if (end < new_end) {
+		if (is_range_written(be->be_inval, end, new_end))
+			end = new_end;
+		else
+			end = new_end - bl->bl_blocksize;
+	}
+	if (end <= new->bse_f_offset) {
+		kfree(new);
+		return 0;
+	}
+	new->bse_length = end - new->bse_f_offset;
+	new->bse_devid = be->be_devid;
+	new->bse_mdev = be->be_mdev;
+
+	spin_lock(&bl->bl_ext_lock);
+	/* new will be freed, either by add_to_commitlist if it decides not
+	 * to use it, or after LAYOUTCOMMIT uses it in the commitlist.
+	 */
+	add_to_commitlist(bl, new);
+	spin_unlock(&bl->bl_ext_lock);
+	return 0;
+}
+
+static void print_bl_extent(struct pnfs_block_extent *be)
+{
+	dprintk("PRINT EXTENT extent %p\n", be);
+	if (be) {
+		dprintk("        be_f_offset %llu\n", (u64)be->be_f_offset);
+		dprintk("        be_length   %llu\n", (u64)be->be_length);
+		dprintk("        be_v_offset %llu\n", (u64)be->be_v_offset);
+		dprintk("        be_state    %d\n", be->be_state);
+	}
+}
+
+static void
+destroy_extent(struct kref *kref)
+{
+	struct pnfs_block_extent *be;
+
+	be = container_of(kref, struct pnfs_block_extent, be_refcnt);
+	dprintk("%s be=%p\n", __func__, be);
+	kfree(be);
+}
+
+void
+bl_put_extent(struct pnfs_block_extent *be)
+{
+	if (be) {
+		dprintk("%s enter %p (%i)\n", __func__, be,
+			atomic_read(&be->be_refcnt.refcount));
+		kref_put(&be->be_refcnt, destroy_extent);
+	}
+}
+
+struct pnfs_block_extent *bl_alloc_extent(void)
+{
+	struct pnfs_block_extent *be;
+
+	be = kmalloc(sizeof(struct pnfs_block_extent), GFP_NOFS);
+	if (!be)
+		return NULL;
+	INIT_LIST_HEAD(&be->be_node);
+	kref_init(&be->be_refcnt);
+	be->be_inval = NULL;
+	return be;
+}
+
+static void print_elist(struct list_head *list)
+{
+	struct pnfs_block_extent *be;
+	dprintk("****************\n");
+	dprintk("Extent list looks like:\n");
+	list_for_each_entry(be, list, be_node) {
+		print_bl_extent(be);
+	}
+	dprintk("****************\n");
+}
+
+static inline int
+extents_consistent(struct pnfs_block_extent *old, struct pnfs_block_extent *new)
+{
+	/* Note this assumes new->be_f_offset >= old->be_f_offset */
+	return (new->be_state == old->be_state) &&
+		((new->be_state == PNFS_BLOCK_NONE_DATA) ||
+		 ((new->be_v_offset - old->be_v_offset ==
+		   new->be_f_offset - old->be_f_offset) &&
+		  new->be_mdev == old->be_mdev));
+}
+
+/* Adds new to appropriate list in bl, modifying new and removing existing
+ * extents as appropriate to deal with overlaps.
+ *
+ * See bl_find_get_extent for list constraints.
+ *
+ * Refcount on new is already set.  If end up not using it, or error out,
+ * need to put the reference.
+ *
+ * bl->bl_ext_lock is held by caller.
+ */
+int
+bl_add_merge_extent(struct pnfs_block_layout *bl,
+		     struct pnfs_block_extent *new)
+{
+	struct pnfs_block_extent *be, *tmp;
+	sector_t end = new->be_f_offset + new->be_length;
+	struct list_head *list;
+
+	dprintk("%s enter with be=%p\n", __func__, new);
+	print_bl_extent(new);
+	list = &bl->bl_extents[bl_choose_list(new->be_state)];
+	print_elist(list);
+
+	/* Scan for proper place to insert, extending new to the left
+	 * as much as possible.
+	 */
+	list_for_each_entry_safe_reverse(be, tmp, list, be_node) {
+		if (new->be_f_offset >= be->be_f_offset + be->be_length)
+			break;
+		if (new->be_f_offset >= be->be_f_offset) {
+			if (end <= be->be_f_offset + be->be_length) {
+				/* new is a subset of existing be*/
+				if (extents_consistent(be, new)) {
+					dprintk("%s: new is subset, ignoring\n",
+						__func__);
+					bl_put_extent(new);
+					return 0;
+				} else {
+					goto out_err;
+				}
+			} else {
+				/* |<--   be   -->|
+				 *          |<--   new   -->| */
+				if (extents_consistent(be, new)) {
+					/* extend new to fully replace be */
+					new->be_length += new->be_f_offset -
+						be->be_f_offset;
+					new->be_f_offset = be->be_f_offset;
+					new->be_v_offset = be->be_v_offset;
+					dprintk("%s: removing %p\n", __func__, be);
+					list_del(&be->be_node);
+					bl_put_extent(be);
+				} else {
+					goto out_err;
+				}
+			}
+		} else if (end >= be->be_f_offset + be->be_length) {
+			/* new extent overlap existing be */
+			if (extents_consistent(be, new)) {
+				/* extend new to fully replace be */
+				dprintk("%s: removing %p\n", __func__, be);
+				list_del(&be->be_node);
+				bl_put_extent(be);
+			} else {
+				goto out_err;
+			}
+		} else if (end > be->be_f_offset) {
+			/*           |<--   be   -->|
+			 *|<--   new   -->| */
+			if (extents_consistent(new, be)) {
+				/* extend new to fully replace be */
+				new->be_length += be->be_f_offset + be->be_length -
+					new->be_f_offset - new->be_length;
+				dprintk("%s: removing %p\n", __func__, be);
+				list_del(&be->be_node);
+				bl_put_extent(be);
+			} else {
+				goto out_err;
+			}
+		}
+	}
+	/* Note that if we never hit the above break, be will not point to a
+	 * valid extent.  However, in that case &be->be_node==list.
+	 */
+	list_add(&new->be_node, &be->be_node);
+	dprintk("%s: inserting new\n", __func__);
+	print_elist(list);
+	/* FIXME - The per-list consistency checks have all been done,
+	 * should now check cross-list consistency.
+	 */
+	return 0;
+
+ out_err:
+	bl_put_extent(new);
+	return -EIO;
+}
+
+/* Returns extent, or NULL.  If a second READ extent exists, it is returned
+ * in cow_read, if given.
+ *
+ * The extents are kept in two seperate ordered lists, one for READ and NONE,
+ * one for READWRITE and INVALID.  Within each list, we assume:
+ * 1. Extents are ordered by file offset.
+ * 2. For any given isect, there is at most one extents that matches.
+ */
+struct pnfs_block_extent *
+bl_find_get_extent(struct pnfs_block_layout *bl, sector_t isect,
+	    struct pnfs_block_extent **cow_read)
+{
+	struct pnfs_block_extent *be, *cow, *ret;
+	int i;
+
+	dprintk("%s enter with isect %llu\n", __func__, (u64)isect);
+	cow = ret = NULL;
+	spin_lock(&bl->bl_ext_lock);
+	for (i = 0; i < EXTENT_LISTS; i++) {
+		list_for_each_entry_reverse(be, &bl->bl_extents[i], be_node) {
+			if (isect >= be->be_f_offset + be->be_length)
+				break;
+			if (isect >= be->be_f_offset) {
+				/* We have found an extent */
+				dprintk("%s Get %p (%i)\n", __func__, be,
+					atomic_read(&be->be_refcnt.refcount));
+				kref_get(&be->be_refcnt);
+				if (!ret)
+					ret = be;
+				else if (be->be_state != PNFS_BLOCK_READ_DATA)
+					bl_put_extent(be);
+				else
+					cow = be;
+				break;
+			}
+		}
+		if (ret &&
+		    (!cow_read || ret->be_state != PNFS_BLOCK_INVALID_DATA))
+			break;
+	}
+	spin_unlock(&bl->bl_ext_lock);
+	if (cow_read)
+		*cow_read = cow;
+	print_bl_extent(ret);
+	return ret;
+}
+
+/* Similar to bl_find_get_extent, but called with lock held, and ignores cow */
+static struct pnfs_block_extent *
+bl_find_get_extent_locked(struct pnfs_block_layout *bl, sector_t isect)
+{
+	struct pnfs_block_extent *be, *ret = NULL;
+	int i;
+
+	dprintk("%s enter with isect %llu\n", __func__, (u64)isect);
+	for (i = 0; i < EXTENT_LISTS; i++) {
+		if (ret)
+			break;
+		list_for_each_entry_reverse(be, &bl->bl_extents[i], be_node) {
+			if (isect >= be->be_f_offset + be->be_length)
+				break;
+			if (isect >= be->be_f_offset) {
+				/* We have found an extent */
+				dprintk("%s Get %p (%i)\n", __func__, be,
+					atomic_read(&be->be_refcnt.refcount));
+				kref_get(&be->be_refcnt);
+				ret = be;
+				break;
+			}
+		}
+	}
+	print_bl_extent(ret);
+	return ret;
+}
+
+int
+encode_pnfs_block_layoutupdate(struct pnfs_block_layout *bl,
+			       struct xdr_stream *xdr,
+			       const struct nfs4_layoutcommit_args *arg)
+{
+	struct pnfs_block_short_extent *lce, *save;
+	unsigned int count = 0;
+	__be32 *p, *xdr_start;
+
+	dprintk("%s enter\n", __func__);
+	/* BUG - creation of bl_commit is buggy - need to wait for
+	 * entire block to be marked WRITTEN before it can be added.
+	 */
+	spin_lock(&bl->bl_ext_lock);
+	/* Want to adjust for possible truncate */
+	/* We now want to adjust argument range */
+
+	/* XDR encode the ranges found */
+	xdr_start = xdr_reserve_space(xdr, 8);
+	if (!xdr_start)
+		goto out;
+	list_for_each_entry_safe(lce, save, &bl->bl_commit, bse_node) {
+		p = xdr_reserve_space(xdr, 7 * 4 + sizeof(lce->bse_devid.data));
+		if (!p)
+			break;
+		p = xdr_encode_opaque_fixed(p, lce->bse_devid.data, NFS4_DEVICEID4_SIZE);
+		p = xdr_encode_hyper(p, lce->bse_f_offset << SECTOR_SHIFT);
+		p = xdr_encode_hyper(p, lce->bse_length << SECTOR_SHIFT);
+		p = xdr_encode_hyper(p, 0LL);
+		*p++ = cpu_to_be32(PNFS_BLOCK_READWRITE_DATA);
+		list_del(&lce->bse_node);
+		list_add_tail(&lce->bse_node, &bl->bl_committing);
+		bl->bl_count--;
+		count++;
+	}
+	xdr_start[0] = cpu_to_be32((xdr->p - xdr_start - 1) * 4);
+	xdr_start[1] = cpu_to_be32(count);
+out:
+	spin_unlock(&bl->bl_ext_lock);
+	dprintk("%s found %i ranges\n", __func__, count);
+	return 0;
+}
+
+/* Helper function to set_to_rw that initialize a new extent */
+static void
+_prep_new_extent(struct pnfs_block_extent *new,
+		 struct pnfs_block_extent *orig,
+		 sector_t offset, sector_t length, int state)
+{
+	kref_init(&new->be_refcnt);
+	/* don't need to INIT_LIST_HEAD(&new->be_node) */
+	memcpy(&new->be_devid, &orig->be_devid, sizeof(struct nfs4_deviceid));
+	new->be_mdev = orig->be_mdev;
+	new->be_f_offset = offset;
+	new->be_length = length;
+	new->be_v_offset = orig->be_v_offset - orig->be_f_offset + offset;
+	new->be_state = state;
+	new->be_inval = orig->be_inval;
+}
+
+/* Tries to merge be with extent in front of it in list.
+ * Frees storage if not used.
+ */
+static struct pnfs_block_extent *
+_front_merge(struct pnfs_block_extent *be, struct list_head *head,
+	     struct pnfs_block_extent *storage)
+{
+	struct pnfs_block_extent *prev;
+
+	if (!storage)
+		goto no_merge;
+	if (&be->be_node == head || be->be_node.prev == head)
+		goto no_merge;
+	prev = list_entry(be->be_node.prev, struct pnfs_block_extent, be_node);
+	if ((prev->be_f_offset + prev->be_length != be->be_f_offset) ||
+	    !extents_consistent(prev, be))
+		goto no_merge;
+	_prep_new_extent(storage, prev, prev->be_f_offset,
+			 prev->be_length + be->be_length, prev->be_state);
+	list_replace(&prev->be_node, &storage->be_node);
+	bl_put_extent(prev);
+	list_del(&be->be_node);
+	bl_put_extent(be);
+	return storage;
+
+ no_merge:
+	kfree(storage);
+	return be;
+}
+
+static u64
+set_to_rw(struct pnfs_block_layout *bl, u64 offset, u64 length)
+{
+	u64 rv = offset + length;
+	struct pnfs_block_extent *be, *e1, *e2, *e3, *new, *old;
+	struct pnfs_block_extent *children[3];
+	struct pnfs_block_extent *merge1 = NULL, *merge2 = NULL;
+	int i = 0, j;
+
+	dprintk("%s(%llu, %llu)\n", __func__, offset, length);
+	/* Create storage for up to three new extents e1, e2, e3 */
+	e1 = kmalloc(sizeof(*e1), GFP_ATOMIC);
+	e2 = kmalloc(sizeof(*e2), GFP_ATOMIC);
+	e3 = kmalloc(sizeof(*e3), GFP_ATOMIC);
+	/* BUG - we are ignoring any failure */
+	if (!e1 || !e2 || !e3)
+		goto out_nosplit;
+
+	spin_lock(&bl->bl_ext_lock);
+	be = bl_find_get_extent_locked(bl, offset);
+	rv = be->be_f_offset + be->be_length;
+	if (be->be_state != PNFS_BLOCK_INVALID_DATA) {
+		spin_unlock(&bl->bl_ext_lock);
+		goto out_nosplit;
+	}
+	/* Add e* to children, bumping e*'s krefs */
+	if (be->be_f_offset != offset) {
+		_prep_new_extent(e1, be, be->be_f_offset,
+				 offset - be->be_f_offset,
+				 PNFS_BLOCK_INVALID_DATA);
+		children[i++] = e1;
+		print_bl_extent(e1);
+	} else
+		merge1 = e1;
+	_prep_new_extent(e2, be, offset,
+			 min(length, be->be_f_offset + be->be_length - offset),
+			 PNFS_BLOCK_READWRITE_DATA);
+	children[i++] = e2;
+	print_bl_extent(e2);
+	if (offset + length < be->be_f_offset + be->be_length) {
+		_prep_new_extent(e3, be, e2->be_f_offset + e2->be_length,
+				 be->be_f_offset + be->be_length -
+				 offset - length,
+				 PNFS_BLOCK_INVALID_DATA);
+		children[i++] = e3;
+		print_bl_extent(e3);
+	} else
+		merge2 = e3;
+
+	/* Remove be from list, and insert the e* */
+	/* We don't get refs on e*, since this list is the base reference
+	 * set when init'ed.
+	 */
+	if (i < 3)
+		children[i] = NULL;
+	new = children[0];
+	list_replace(&be->be_node, &new->be_node);
+	bl_put_extent(be);
+	new = _front_merge(new, &bl->bl_extents[RW_EXTENT], merge1);
+	for (j = 1; j < i; j++) {
+		old = new;
+		new = children[j];
+		list_add(&new->be_node, &old->be_node);
+	}
+	if (merge2) {
+		/* This is a HACK, should just create a _back_merge function */
+		new = list_entry(new->be_node.next,
+				 struct pnfs_block_extent, be_node);
+		new = _front_merge(new, &bl->bl_extents[RW_EXTENT], merge2);
+	}
+	spin_unlock(&bl->bl_ext_lock);
+
+	/* Since we removed the base reference above, be is now scheduled for
+	 * destruction.
+	 */
+	bl_put_extent(be);
+	dprintk("%s returns %llu after split\n", __func__, rv);
+	return rv;
+
+ out_nosplit:
+	kfree(e1);
+	kfree(e2);
+	kfree(e3);
+	dprintk("%s returns %llu without splitting\n", __func__, rv);
+	return rv;
+}
+
+void
+clean_pnfs_block_layoutupdate(struct pnfs_block_layout *bl,
+			      const struct nfs4_layoutcommit_args *arg,
+			      int status)
+{
+	struct pnfs_block_short_extent *lce, *save;
+
+	dprintk("%s status %d\n", __func__, status);
+	list_for_each_entry_safe(lce, save, &bl->bl_committing, bse_node) {
+		if (likely(!status)) {
+			u64 offset = lce->bse_f_offset;
+			u64 end = offset + lce->bse_length;
+
+			do {
+				offset = set_to_rw(bl, offset, end - offset);
+			} while (offset < end);
+			list_del(&lce->bse_node);
+
+			kfree(lce);
+		} else {
+			list_del(&lce->bse_node);
+			spin_lock(&bl->bl_ext_lock);
+			add_to_commitlist(bl, lce);
+			spin_unlock(&bl->bl_ext_lock);
+		}
+	}
+}
diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c
new file mode 100644
index 0000000..a907de5
--- /dev/null
+++ b/fs/xfs/kmem.c
@@ -0,0 +1,132 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include <linux/mm.h>
+#include <linux/highmem.h>
+#include <linux/slab.h>
+#include <linux/swap.h>
+#include <linux/blkdev.h>
+#include <linux/backing-dev.h>
+#include "time.h"
+#include "kmem.h"
+#include "xfs_message.h"
+
+/*
+ * Greedy allocation.  May fail and may return vmalloced memory.
+ *
+ * Must be freed using kmem_free_large.
+ */
+void *
+kmem_zalloc_greedy(size_t *size, size_t minsize, size_t maxsize)
+{
+	void		*ptr;
+	size_t		kmsize = maxsize;
+
+	while (!(ptr = kmem_zalloc_large(kmsize))) {
+		if ((kmsize >>= 1) <= minsize)
+			kmsize = minsize;
+	}
+	if (ptr)
+		*size = kmsize;
+	return ptr;
+}
+
+void *
+kmem_alloc(size_t size, unsigned int __nocast flags)
+{
+	int	retries = 0;
+	gfp_t	lflags = kmem_flags_convert(flags);
+	void	*ptr;
+
+	do {
+		ptr = kmalloc(size, lflags);
+		if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP)))
+			return ptr;
+		if (!(++retries % 100))
+			xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
+					__func__, lflags);
+		congestion_wait(BLK_RW_ASYNC, HZ/50);
+	} while (1);
+}
+
+void *
+kmem_zalloc(size_t size, unsigned int __nocast flags)
+{
+	void	*ptr;
+
+	ptr = kmem_alloc(size, flags);
+	if (ptr)
+		memset((char *)ptr, 0, (int)size);
+	return ptr;
+}
+
+void
+kmem_free(const void *ptr)
+{
+	if (!is_vmalloc_addr(ptr)) {
+		kfree(ptr);
+	} else {
+		vfree(ptr);
+	}
+}
+
+void *
+kmem_realloc(const void *ptr, size_t newsize, size_t oldsize,
+	     unsigned int __nocast flags)
+{
+	void	*new;
+
+	new = kmem_alloc(newsize, flags);
+	if (ptr) {
+		if (new)
+			memcpy(new, ptr,
+				((oldsize < newsize) ? oldsize : newsize));
+		kmem_free(ptr);
+	}
+	return new;
+}
+
+void *
+kmem_zone_alloc(kmem_zone_t *zone, unsigned int __nocast flags)
+{
+	int	retries = 0;
+	gfp_t	lflags = kmem_flags_convert(flags);
+	void	*ptr;
+
+	do {
+		ptr = kmem_cache_alloc(zone, lflags);
+		if (ptr || (flags & (KM_MAYFAIL|KM_NOSLEEP)))
+			return ptr;
+		if (!(++retries % 100))
+			xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
+					__func__, lflags);
+		congestion_wait(BLK_RW_ASYNC, HZ/50);
+	} while (1);
+}
+
+void *
+kmem_zone_zalloc(kmem_zone_t *zone, unsigned int __nocast flags)
+{
+	void	*ptr;
+
+	ptr = kmem_zone_alloc(zone, flags);
+	if (ptr)
+		memset((char *)ptr, 0, kmem_cache_size(zone));
+	return ptr;
+}
diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h
new file mode 100644
index 0000000..292eff1
--- /dev/null
+++ b/fs/xfs/kmem.h
@@ -0,0 +1,119 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SUPPORT_KMEM_H__
+#define __XFS_SUPPORT_KMEM_H__
+
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+
+/*
+ * General memory allocation interfaces
+ */
+
+#define KM_SLEEP	0x0001u
+#define KM_NOSLEEP	0x0002u
+#define KM_NOFS		0x0004u
+#define KM_MAYFAIL	0x0008u
+
+/*
+ * We use a special process flag to avoid recursive callbacks into
+ * the filesystem during transactions.  We will also issue our own
+ * warnings, so we explicitly skip any generic ones (silly of us).
+ */
+static inline gfp_t
+kmem_flags_convert(unsigned int __nocast flags)
+{
+	gfp_t	lflags;
+
+	BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL));
+
+	if (flags & KM_NOSLEEP) {
+		lflags = GFP_ATOMIC | __GFP_NOWARN;
+	} else {
+		lflags = GFP_KERNEL | __GFP_NOWARN;
+		if ((current->flags & PF_FSTRANS) || (flags & KM_NOFS))
+			lflags &= ~__GFP_FS;
+	}
+	return lflags;
+}
+
+extern void *kmem_alloc(size_t, unsigned int __nocast);
+extern void *kmem_zalloc(size_t, unsigned int __nocast);
+extern void *kmem_realloc(const void *, size_t, size_t, unsigned int __nocast);
+extern void  kmem_free(const void *);
+
+static inline void *kmem_zalloc_large(size_t size)
+{
+	return vzalloc(size);
+}
+static inline void kmem_free_large(void *ptr)
+{
+	vfree(ptr);
+}
+
+extern void *kmem_zalloc_greedy(size_t *, size_t, size_t);
+
+/*
+ * Zone interfaces
+ */
+
+#define KM_ZONE_HWALIGN	SLAB_HWCACHE_ALIGN
+#define KM_ZONE_RECLAIM	SLAB_RECLAIM_ACCOUNT
+#define KM_ZONE_SPREAD	SLAB_MEM_SPREAD
+
+#define kmem_zone	kmem_cache
+#define kmem_zone_t	struct kmem_cache
+
+static inline kmem_zone_t *
+kmem_zone_init(int size, char *zone_name)
+{
+	return kmem_cache_create(zone_name, size, 0, 0, NULL);
+}
+
+static inline kmem_zone_t *
+kmem_zone_init_flags(int size, char *zone_name, unsigned long flags,
+		     void (*construct)(void *))
+{
+	return kmem_cache_create(zone_name, size, 0, flags, construct);
+}
+
+static inline void
+kmem_zone_free(kmem_zone_t *zone, void *ptr)
+{
+	kmem_cache_free(zone, ptr);
+}
+
+static inline void
+kmem_zone_destroy(kmem_zone_t *zone)
+{
+	if (zone)
+		kmem_cache_destroy(zone);
+}
+
+extern void *kmem_zone_alloc(kmem_zone_t *, unsigned int __nocast);
+extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
+
+static inline int
+kmem_shake_allow(gfp_t gfp_mask)
+{
+	return ((gfp_mask & __GFP_WAIT) && (gfp_mask & __GFP_FS));
+}
+
+#endif /* __XFS_SUPPORT_KMEM_H__ */
diff --git a/fs/xfs/mrlock.h b/fs/xfs/mrlock.h
new file mode 100644
index 0000000..ff6a198
--- /dev/null
+++ b/fs/xfs/mrlock.h
@@ -0,0 +1,90 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SUPPORT_MRLOCK_H__
+#define __XFS_SUPPORT_MRLOCK_H__
+
+#include <linux/rwsem.h>
+
+typedef struct {
+	struct rw_semaphore	mr_lock;
+#ifdef DEBUG
+	int			mr_writer;
+#endif
+} mrlock_t;
+
+#ifdef DEBUG
+#define mrinit(mrp, name)	\
+	do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0)
+#else
+#define mrinit(mrp, name)	\
+	do { init_rwsem(&(mrp)->mr_lock); } while (0)
+#endif
+
+#define mrlock_init(mrp, t,n,s)	mrinit(mrp, n)
+#define mrfree(mrp)		do { } while (0)
+
+static inline void mraccess_nested(mrlock_t *mrp, int subclass)
+{
+	down_read_nested(&mrp->mr_lock, subclass);
+}
+
+static inline void mrupdate_nested(mrlock_t *mrp, int subclass)
+{
+	down_write_nested(&mrp->mr_lock, subclass);
+#ifdef DEBUG
+	mrp->mr_writer = 1;
+#endif
+}
+
+static inline int mrtryaccess(mrlock_t *mrp)
+{
+	return down_read_trylock(&mrp->mr_lock);
+}
+
+static inline int mrtryupdate(mrlock_t *mrp)
+{
+	if (!down_write_trylock(&mrp->mr_lock))
+		return 0;
+#ifdef DEBUG
+	mrp->mr_writer = 1;
+#endif
+	return 1;
+}
+
+static inline void mrunlock_excl(mrlock_t *mrp)
+{
+#ifdef DEBUG
+	mrp->mr_writer = 0;
+#endif
+	up_write(&mrp->mr_lock);
+}
+
+static inline void mrunlock_shared(mrlock_t *mrp)
+{
+	up_read(&mrp->mr_lock);
+}
+
+static inline void mrdemote(mrlock_t *mrp)
+{
+#ifdef DEBUG
+	mrp->mr_writer = 0;
+#endif
+	downgrade_write(&mrp->mr_lock);
+}
+
+#endif /* __XFS_SUPPORT_MRLOCK_H__ */
diff --git a/fs/xfs/time.h b/fs/xfs/time.h
new file mode 100644
index 0000000..387e695
--- /dev/null
+++ b/fs/xfs/time.h
@@ -0,0 +1,36 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SUPPORT_TIME_H__
+#define __XFS_SUPPORT_TIME_H__
+
+#include <linux/sched.h>
+#include <linux/time.h>
+
+typedef struct timespec timespec_t;
+
+static inline void delay(long ticks)
+{
+	schedule_timeout_uninterruptible(ticks);
+}
+
+static inline void nanotime(struct timespec *tvp)
+{
+	*tvp = CURRENT_TIME;
+}
+
+#endif /* __XFS_SUPPORT_TIME_H__ */
diff --git a/fs/xfs/uuid.c b/fs/xfs/uuid.c
new file mode 100644
index 0000000..b83f76b
--- /dev/null
+++ b/fs/xfs/uuid.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include <xfs.h>
+
+/* IRIX interpretation of an uuid_t */
+typedef struct {
+	__be32	uu_timelow;
+	__be16	uu_timemid;
+	__be16	uu_timehi;
+	__be16	uu_clockseq;
+	__be16	uu_node[3];
+} xfs_uu_t;
+
+/*
+ * uuid_getnodeuniq - obtain the node unique fields of a UUID.
+ *
+ * This is not in any way a standard or condoned UUID function;
+ * it just something that's needed for user-level file handles.
+ */
+void
+uuid_getnodeuniq(uuid_t *uuid, int fsid [2])
+{
+	xfs_uu_t *uup = (xfs_uu_t *)uuid;
+
+	fsid[0] = (be16_to_cpu(uup->uu_clockseq) << 16) |
+		   be16_to_cpu(uup->uu_timemid);
+	fsid[1] = be32_to_cpu(uup->uu_timelow);
+}
+
+int
+uuid_is_nil(uuid_t *uuid)
+{
+	int	i;
+	char	*cp = (char *)uuid;
+
+	if (uuid == NULL)
+		return 0;
+	/* implied check of version number here... */
+	for (i = 0; i < sizeof *uuid; i++)
+		if (*cp++) return 0;	/* not nil */
+	return 1;	/* is nil */
+}
+
+int
+uuid_equal(uuid_t *uuid1, uuid_t *uuid2)
+{
+	return memcmp(uuid1, uuid2, sizeof(uuid_t)) ? 0 : 1;
+}
diff --git a/fs/xfs/uuid.h b/fs/xfs/uuid.h
new file mode 100644
index 0000000..4732d71
--- /dev/null
+++ b/fs/xfs/uuid.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SUPPORT_UUID_H__
+#define __XFS_SUPPORT_UUID_H__
+
+typedef struct {
+	unsigned char	__u_bits[16];
+} uuid_t;
+
+extern int uuid_is_nil(uuid_t *uuid);
+extern int uuid_equal(uuid_t *uuid1, uuid_t *uuid2);
+extern void uuid_getnodeuniq(uuid_t *uuid, int fsid [2]);
+
+#endif	/* __XFS_SUPPORT_UUID_H__ */
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
new file mode 100644
index 0000000..ac702a6
--- /dev/null
+++ b/fs/xfs/xfs_acl.c
@@ -0,0 +1,422 @@
+/*
+ * Copyright (c) 2008, Christoph Hellwig
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_acl.h"
+#include "xfs_attr.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_vnodeops.h"
+#include "xfs_trace.h"
+#include <linux/slab.h>
+#include <linux/xattr.h>
+#include <linux/posix_acl_xattr.h>
+
+
+/*
+ * Locking scheme:
+ *  - all ACL updates are protected by inode->i_mutex, which is taken before
+ *    calling into this file.
+ */
+
+STATIC struct posix_acl *
+xfs_acl_from_disk(struct xfs_acl *aclp)
+{
+	struct posix_acl_entry *acl_e;
+	struct posix_acl *acl;
+	struct xfs_acl_entry *ace;
+	unsigned int count, i;
+
+	count = be32_to_cpu(aclp->acl_cnt);
+	if (count > XFS_ACL_MAX_ENTRIES)
+		return ERR_PTR(-EFSCORRUPTED);
+
+	acl = posix_acl_alloc(count, GFP_KERNEL);
+	if (!acl)
+		return ERR_PTR(-ENOMEM);
+
+	for (i = 0; i < count; i++) {
+		acl_e = &acl->a_entries[i];
+		ace = &aclp->acl_entry[i];
+
+		/*
+		 * The tag is 32 bits on disk and 16 bits in core.
+		 *
+		 * Because every access to it goes through the core
+		 * format first this is not a problem.
+		 */
+		acl_e->e_tag = be32_to_cpu(ace->ae_tag);
+		acl_e->e_perm = be16_to_cpu(ace->ae_perm);
+
+		switch (acl_e->e_tag) {
+		case ACL_USER:
+		case ACL_GROUP:
+			acl_e->e_id = be32_to_cpu(ace->ae_id);
+			break;
+		case ACL_USER_OBJ:
+		case ACL_GROUP_OBJ:
+		case ACL_MASK:
+		case ACL_OTHER:
+			acl_e->e_id = ACL_UNDEFINED_ID;
+			break;
+		default:
+			goto fail;
+		}
+	}
+	return acl;
+
+fail:
+	posix_acl_release(acl);
+	return ERR_PTR(-EINVAL);
+}
+
+STATIC void
+xfs_acl_to_disk(struct xfs_acl *aclp, const struct posix_acl *acl)
+{
+	const struct posix_acl_entry *acl_e;
+	struct xfs_acl_entry *ace;
+	int i;
+
+	aclp->acl_cnt = cpu_to_be32(acl->a_count);
+	for (i = 0; i < acl->a_count; i++) {
+		ace = &aclp->acl_entry[i];
+		acl_e = &acl->a_entries[i];
+
+		ace->ae_tag = cpu_to_be32(acl_e->e_tag);
+		ace->ae_id = cpu_to_be32(acl_e->e_id);
+		ace->ae_perm = cpu_to_be16(acl_e->e_perm);
+	}
+}
+
+struct posix_acl *
+xfs_get_acl(struct inode *inode, int type)
+{
+	struct xfs_inode *ip = XFS_I(inode);
+	struct posix_acl *acl;
+	struct xfs_acl *xfs_acl;
+	int len = sizeof(struct xfs_acl);
+	unsigned char *ea_name;
+	int error;
+
+	acl = get_cached_acl(inode, type);
+	if (acl != ACL_NOT_CACHED)
+		return acl;
+
+	trace_xfs_get_acl(ip);
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		ea_name = SGI_ACL_FILE;
+		break;
+	case ACL_TYPE_DEFAULT:
+		ea_name = SGI_ACL_DEFAULT;
+		break;
+	default:
+		BUG();
+	}
+
+	/*
+	 * If we have a cached ACLs value just return it, not need to
+	 * go out to the disk.
+	 */
+
+	xfs_acl = kzalloc(sizeof(struct xfs_acl), GFP_KERNEL);
+	if (!xfs_acl)
+		return ERR_PTR(-ENOMEM);
+
+	error = -xfs_attr_get(ip, ea_name, (unsigned char *)xfs_acl,
+							&len, ATTR_ROOT);
+	if (error) {
+		/*
+		 * If the attribute doesn't exist make sure we have a negative
+		 * cache entry, for any other error assume it is transient and
+		 * leave the cache entry as ACL_NOT_CACHED.
+		 */
+		if (error == -ENOATTR) {
+			acl = NULL;
+			goto out_update_cache;
+		}
+		goto out;
+	}
+
+	acl = xfs_acl_from_disk(xfs_acl);
+	if (IS_ERR(acl))
+		goto out;
+
+ out_update_cache:
+	set_cached_acl(inode, type, acl);
+ out:
+	kfree(xfs_acl);
+	return acl;
+}
+
+STATIC int
+xfs_set_acl(struct inode *inode, int type, struct posix_acl *acl)
+{
+	struct xfs_inode *ip = XFS_I(inode);
+	unsigned char *ea_name;
+	int error;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		ea_name = SGI_ACL_FILE;
+		break;
+	case ACL_TYPE_DEFAULT:
+		if (!S_ISDIR(inode->i_mode))
+			return acl ? -EACCES : 0;
+		ea_name = SGI_ACL_DEFAULT;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (acl) {
+		struct xfs_acl *xfs_acl;
+		int len;
+
+		xfs_acl = kzalloc(sizeof(struct xfs_acl), GFP_KERNEL);
+		if (!xfs_acl)
+			return -ENOMEM;
+
+		xfs_acl_to_disk(xfs_acl, acl);
+		len = sizeof(struct xfs_acl) -
+			(sizeof(struct xfs_acl_entry) *
+			 (XFS_ACL_MAX_ENTRIES - acl->a_count));
+
+		error = -xfs_attr_set(ip, ea_name, (unsigned char *)xfs_acl,
+				len, ATTR_ROOT);
+
+		kfree(xfs_acl);
+	} else {
+		/*
+		 * A NULL ACL argument means we want to remove the ACL.
+		 */
+		error = -xfs_attr_remove(ip, ea_name, ATTR_ROOT);
+
+		/*
+		 * If the attribute didn't exist to start with that's fine.
+		 */
+		if (error == -ENOATTR)
+			error = 0;
+	}
+
+	if (!error)
+		set_cached_acl(inode, type, acl);
+	return error;
+}
+
+static int
+xfs_set_mode(struct inode *inode, umode_t mode)
+{
+	int error = 0;
+
+	if (mode != inode->i_mode) {
+		struct iattr iattr;
+
+		iattr.ia_valid = ATTR_MODE | ATTR_CTIME;
+		iattr.ia_mode = mode;
+		iattr.ia_ctime = current_fs_time(inode->i_sb);
+
+		error = -xfs_setattr_nonsize(XFS_I(inode), &iattr, XFS_ATTR_NOACL);
+	}
+
+	return error;
+}
+
+static int
+xfs_acl_exists(struct inode *inode, unsigned char *name)
+{
+	int len = sizeof(struct xfs_acl);
+
+	return (xfs_attr_get(XFS_I(inode), name, NULL, &len,
+			    ATTR_ROOT|ATTR_KERNOVAL) == 0);
+}
+
+int
+posix_acl_access_exists(struct inode *inode)
+{
+	return xfs_acl_exists(inode, SGI_ACL_FILE);
+}
+
+int
+posix_acl_default_exists(struct inode *inode)
+{
+	if (!S_ISDIR(inode->i_mode))
+		return 0;
+	return xfs_acl_exists(inode, SGI_ACL_DEFAULT);
+}
+
+/*
+ * No need for i_mutex because the inode is not yet exposed to the VFS.
+ */
+int
+xfs_inherit_acl(struct inode *inode, struct posix_acl *acl)
+{
+	umode_t mode = inode->i_mode;
+	int error = 0, inherit = 0;
+
+	if (S_ISDIR(inode->i_mode)) {
+		error = xfs_set_acl(inode, ACL_TYPE_DEFAULT, acl);
+		if (error)
+			goto out;
+	}
+
+	error = posix_acl_create(&acl, GFP_KERNEL, &mode);
+	if (error < 0)
+		return error;
+
+	/*
+	 * If posix_acl_create returns a positive value we need to
+	 * inherit a permission that can't be represented using the Unix
+	 * mode bits and we actually need to set an ACL.
+	 */
+	if (error > 0)
+		inherit = 1;
+
+	error = xfs_set_mode(inode, mode);
+	if (error)
+		goto out;
+
+	if (inherit)
+		error = xfs_set_acl(inode, ACL_TYPE_ACCESS, acl);
+
+out:
+	posix_acl_release(acl);
+	return error;
+}
+
+int
+xfs_acl_chmod(struct inode *inode)
+{
+	struct posix_acl *acl;
+	int error;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	acl = xfs_get_acl(inode, ACL_TYPE_ACCESS);
+	if (IS_ERR(acl) || !acl)
+		return PTR_ERR(acl);
+
+	error = posix_acl_chmod(&acl, GFP_KERNEL, inode->i_mode);
+	if (error)
+		return error;
+
+	error = xfs_set_acl(inode, ACL_TYPE_ACCESS, acl);
+	posix_acl_release(acl);
+	return error;
+}
+
+static int
+xfs_xattr_acl_get(struct dentry *dentry, const char *name,
+		void *value, size_t size, int type)
+{
+	struct posix_acl *acl;
+	int error;
+
+	acl = xfs_get_acl(dentry->d_inode, type);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl == NULL)
+		return -ENODATA;
+
+	error = posix_acl_to_xattr(acl, value, size);
+	posix_acl_release(acl);
+
+	return error;
+}
+
+static int
+xfs_xattr_acl_set(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags, int type)
+{
+	struct inode *inode = dentry->d_inode;
+	struct posix_acl *acl = NULL;
+	int error = 0;
+
+	if (flags & XATTR_CREATE)
+		return -EINVAL;
+	if (type == ACL_TYPE_DEFAULT && !S_ISDIR(inode->i_mode))
+		return value ? -EACCES : 0;
+	if ((current_fsuid() != inode->i_uid) && !capable(CAP_FOWNER))
+		return -EPERM;
+
+	if (!value)
+		goto set_acl;
+
+	acl = posix_acl_from_xattr(value, size);
+	if (!acl) {
+		/*
+		 * acl_set_file(3) may request that we set default ACLs with
+		 * zero length -- defend (gracefully) against that here.
+		 */
+		goto out;
+	}
+	if (IS_ERR(acl)) {
+		error = PTR_ERR(acl);
+		goto out;
+	}
+
+	error = posix_acl_valid(acl);
+	if (error)
+		goto out_release;
+
+	error = -EINVAL;
+	if (acl->a_count > XFS_ACL_MAX_ENTRIES)
+		goto out_release;
+
+	if (type == ACL_TYPE_ACCESS) {
+		umode_t mode = inode->i_mode;
+		error = posix_acl_equiv_mode(acl, &mode);
+
+		if (error <= 0) {
+			posix_acl_release(acl);
+			acl = NULL;
+
+			if (error < 0)
+				return error;
+		}
+
+		error = xfs_set_mode(inode, mode);
+		if (error)
+			goto out_release;
+	}
+
+ set_acl:
+	error = xfs_set_acl(inode, type, acl);
+ out_release:
+	posix_acl_release(acl);
+ out:
+	return error;
+}
+
+const struct xattr_handler xfs_xattr_acl_access_handler = {
+	.prefix	= POSIX_ACL_XATTR_ACCESS,
+	.flags	= ACL_TYPE_ACCESS,
+	.get	= xfs_xattr_acl_get,
+	.set	= xfs_xattr_acl_set,
+};
+
+const struct xattr_handler xfs_xattr_acl_default_handler = {
+	.prefix	= POSIX_ACL_XATTR_DEFAULT,
+	.flags	= ACL_TYPE_DEFAULT,
+	.get	= xfs_xattr_acl_get,
+	.set	= xfs_xattr_acl_set,
+};
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
new file mode 100644
index 0000000..c2b06d4
--- /dev/null
+++ b/fs/xfs/xfs_aops.c
@@ -0,0 +1,1466 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_trans.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_error.h"
+#include "xfs_rw.h"
+#include "xfs_iomap.h"
+#include "xfs_vnodeops.h"
+#include "xfs_trace.h"
+#include "xfs_bmap.h"
+#include <linux/gfp.h>
+#include <linux/mpage.h>
+#include <linux/pagevec.h>
+#include <linux/writeback.h>
+
+void
+xfs_count_page_state(
+	struct page		*page,
+	int			*delalloc,
+	int			*unwritten)
+{
+	struct buffer_head	*bh, *head;
+
+	*delalloc = *unwritten = 0;
+
+	bh = head = page_buffers(page);
+	do {
+		if (buffer_unwritten(bh))
+			(*unwritten) = 1;
+		else if (buffer_delay(bh))
+			(*delalloc) = 1;
+	} while ((bh = bh->b_this_page) != head);
+}
+
+STATIC struct block_device *
+xfs_find_bdev_for_inode(
+	struct inode		*inode)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+
+	if (XFS_IS_REALTIME_INODE(ip))
+		return mp->m_rtdev_targp->bt_bdev;
+	else
+		return mp->m_ddev_targp->bt_bdev;
+}
+
+/*
+ * We're now finished for good with this ioend structure.
+ * Update the page state via the associated buffer_heads,
+ * release holds on the inode and bio, and finally free
+ * up memory.  Do not use the ioend after this.
+ */
+STATIC void
+xfs_destroy_ioend(
+	xfs_ioend_t		*ioend)
+{
+	struct buffer_head	*bh, *next;
+
+	for (bh = ioend->io_buffer_head; bh; bh = next) {
+		next = bh->b_private;
+		bh->b_end_io(bh, !ioend->io_error);
+	}
+
+	if (ioend->io_iocb) {
+		inode_dio_done(ioend->io_inode);
+		if (ioend->io_isasync) {
+			aio_complete(ioend->io_iocb, ioend->io_error ?
+					ioend->io_error : ioend->io_result, 0);
+		}
+	}
+
+	mempool_free(ioend, xfs_ioend_pool);
+}
+
+/*
+ * If the end of the current ioend is beyond the current EOF,
+ * return the new EOF value, otherwise zero.
+ */
+STATIC xfs_fsize_t
+xfs_ioend_new_eof(
+	xfs_ioend_t		*ioend)
+{
+	xfs_inode_t		*ip = XFS_I(ioend->io_inode);
+	xfs_fsize_t		isize;
+	xfs_fsize_t		bsize;
+
+	bsize = ioend->io_offset + ioend->io_size;
+	isize = MAX(ip->i_size, ip->i_new_size);
+	isize = MIN(isize, bsize);
+	return isize > ip->i_d.di_size ? isize : 0;
+}
+
+/*
+ * Fast and loose check if this write could update the on-disk inode size.
+ */
+static inline bool xfs_ioend_is_append(struct xfs_ioend *ioend)
+{
+	return ioend->io_offset + ioend->io_size >
+		XFS_I(ioend->io_inode)->i_d.di_size;
+}
+
+/*
+ * Update on-disk file size now that data has been written to disk.  The
+ * current in-memory file size is i_size.  If a write is beyond eof i_new_size
+ * will be the intended file size until i_size is updated.  If this write does
+ * not extend all the way to the valid file size then restrict this update to
+ * the end of the write.
+ *
+ * This function does not block as blocking on the inode lock in IO completion
+ * can lead to IO completion order dependency deadlocks.. If it can't get the
+ * inode ilock it will return EAGAIN. Callers must handle this.
+ */
+STATIC int
+xfs_setfilesize(
+	xfs_ioend_t		*ioend)
+{
+	xfs_inode_t		*ip = XFS_I(ioend->io_inode);
+	xfs_fsize_t		isize;
+
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL))
+		return EAGAIN;
+
+	isize = xfs_ioend_new_eof(ioend);
+	if (isize) {
+		trace_xfs_setfilesize(ip, ioend->io_offset, ioend->io_size);
+		ip->i_d.di_size = isize;
+		xfs_mark_inode_dirty(ip);
+	}
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+/*
+ * Schedule IO completion handling on the final put of an ioend.
+ *
+ * If there is no work to do we might as well call it a day and free the
+ * ioend right now.
+ */
+STATIC void
+xfs_finish_ioend(
+	struct xfs_ioend	*ioend)
+{
+	if (atomic_dec_and_test(&ioend->io_remaining)) {
+		if (ioend->io_type == IO_UNWRITTEN)
+			queue_work(xfsconvertd_workqueue, &ioend->io_work);
+		else if (xfs_ioend_is_append(ioend))
+			queue_work(xfsdatad_workqueue, &ioend->io_work);
+		else
+			xfs_destroy_ioend(ioend);
+	}
+}
+
+/*
+ * IO write completion.
+ */
+STATIC void
+xfs_end_io(
+	struct work_struct *work)
+{
+	xfs_ioend_t	*ioend = container_of(work, xfs_ioend_t, io_work);
+	struct xfs_inode *ip = XFS_I(ioend->io_inode);
+	int		error = 0;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+		ioend->io_error = -EIO;
+		goto done;
+	}
+	if (ioend->io_error)
+		goto done;
+
+	/*
+	 * For unwritten extents we need to issue transactions to convert a
+	 * range to normal written extens after the data I/O has finished.
+	 */
+	if (ioend->io_type == IO_UNWRITTEN) {
+		error = xfs_iomap_write_unwritten(ip, ioend->io_offset,
+						 ioend->io_size);
+		if (error) {
+			ioend->io_error = -error;
+			goto done;
+		}
+	}
+
+	/*
+	 * We might have to update the on-disk file size after extending
+	 * writes.
+	 */
+	error = xfs_setfilesize(ioend);
+	ASSERT(!error || error == EAGAIN);
+
+done:
+	/*
+	 * If we didn't complete processing of the ioend, requeue it to the
+	 * tail of the workqueue for another attempt later. Otherwise destroy
+	 * it.
+	 */
+	if (error == EAGAIN) {
+		atomic_inc(&ioend->io_remaining);
+		xfs_finish_ioend(ioend);
+		/* ensure we don't spin on blocked ioends */
+		delay(1);
+	} else {
+		xfs_destroy_ioend(ioend);
+	}
+}
+
+/*
+ * Call IO completion handling in caller context on the final put of an ioend.
+ */
+STATIC void
+xfs_finish_ioend_sync(
+	struct xfs_ioend	*ioend)
+{
+	if (atomic_dec_and_test(&ioend->io_remaining))
+		xfs_end_io(&ioend->io_work);
+}
+
+/*
+ * Allocate and initialise an IO completion structure.
+ * We need to track unwritten extent write completion here initially.
+ * We'll need to extend this for updating the ondisk inode size later
+ * (vs. incore size).
+ */
+STATIC xfs_ioend_t *
+xfs_alloc_ioend(
+	struct inode		*inode,
+	unsigned int		type)
+{
+	xfs_ioend_t		*ioend;
+
+	ioend = mempool_alloc(xfs_ioend_pool, GFP_NOFS);
+
+	/*
+	 * Set the count to 1 initially, which will prevent an I/O
+	 * completion callback from happening before we have started
+	 * all the I/O from calling the completion routine too early.
+	 */
+	atomic_set(&ioend->io_remaining, 1);
+	ioend->io_isasync = 0;
+	ioend->io_error = 0;
+	ioend->io_list = NULL;
+	ioend->io_type = type;
+	ioend->io_inode = inode;
+	ioend->io_buffer_head = NULL;
+	ioend->io_buffer_tail = NULL;
+	ioend->io_offset = 0;
+	ioend->io_size = 0;
+	ioend->io_iocb = NULL;
+	ioend->io_result = 0;
+
+	INIT_WORK(&ioend->io_work, xfs_end_io);
+	return ioend;
+}
+
+STATIC int
+xfs_map_blocks(
+	struct inode		*inode,
+	loff_t			offset,
+	struct xfs_bmbt_irec	*imap,
+	int			type,
+	int			nonblocking)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	ssize_t			count = 1 << inode->i_blkbits;
+	xfs_fileoff_t		offset_fsb, end_fsb;
+	int			error = 0;
+	int			bmapi_flags = XFS_BMAPI_ENTIRE;
+	int			nimaps = 1;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	if (type == IO_UNWRITTEN)
+		bmapi_flags |= XFS_BMAPI_IGSTATE;
+
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
+		if (nonblocking)
+			return -XFS_ERROR(EAGAIN);
+		xfs_ilock(ip, XFS_ILOCK_SHARED);
+	}
+
+	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
+	       (ip->i_df.if_flags & XFS_IFEXTENTS));
+	ASSERT(offset <= mp->m_maxioffset);
+
+	if (offset + count > mp->m_maxioffset)
+		count = mp->m_maxioffset - offset;
+	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
+				imap, &nimaps, bmapi_flags);
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (error)
+		return -XFS_ERROR(error);
+
+	if (type == IO_DELALLOC &&
+	    (!nimaps || isnullstartblock(imap->br_startblock))) {
+		error = xfs_iomap_write_allocate(ip, offset, count, imap);
+		if (!error)
+			trace_xfs_map_blocks_alloc(ip, offset, count, type, imap);
+		return -XFS_ERROR(error);
+	}
+
+#ifdef DEBUG
+	if (type == IO_UNWRITTEN) {
+		ASSERT(nimaps);
+		ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+		ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+	}
+#endif
+	if (nimaps)
+		trace_xfs_map_blocks_found(ip, offset, count, type, imap);
+	return 0;
+}
+
+STATIC int
+xfs_imap_valid(
+	struct inode		*inode,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	offset >>= inode->i_blkbits;
+
+	return offset >= imap->br_startoff &&
+		offset < imap->br_startoff + imap->br_blockcount;
+}
+
+/*
+ * BIO completion handler for buffered IO.
+ */
+STATIC void
+xfs_end_bio(
+	struct bio		*bio,
+	int			error)
+{
+	xfs_ioend_t		*ioend = bio->bi_private;
+
+	ASSERT(atomic_read(&bio->bi_cnt) >= 1);
+	if (!ioend->io_error && !test_bit(BIO_UPTODATE, &bio->bi_flags))
+		ioend->io_error = error;
+
+	/* Toss bio and pass work off to an xfsdatad thread */
+	bio->bi_private = NULL;
+	bio->bi_end_io = NULL;
+	bio_put(bio);
+
+	xfs_finish_ioend(ioend);
+}
+
+STATIC void
+xfs_submit_ioend_bio(
+	struct writeback_control *wbc,
+	xfs_ioend_t		*ioend,
+	struct bio		*bio)
+{
+	atomic_inc(&ioend->io_remaining);
+	bio->bi_private = ioend;
+	bio->bi_end_io = xfs_end_bio;
+
+	/*
+	 * If the I/O is beyond EOF we mark the inode dirty immediately
+	 * but don't update the inode size until I/O completion.
+	 */
+	if (xfs_ioend_new_eof(ioend))
+		xfs_mark_inode_dirty(XFS_I(ioend->io_inode));
+
+	submit_bio(wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE, bio);
+}
+
+STATIC struct bio *
+xfs_alloc_ioend_bio(
+	struct buffer_head	*bh)
+{
+	int			nvecs = bio_get_nr_vecs(bh->b_bdev);
+	struct bio		*bio = bio_alloc(GFP_NOIO, nvecs);
+
+	ASSERT(bio->bi_private == NULL);
+	bio->bi_sector = bh->b_blocknr * (bh->b_size >> 9);
+	bio->bi_bdev = bh->b_bdev;
+	return bio;
+}
+
+STATIC void
+xfs_start_buffer_writeback(
+	struct buffer_head	*bh)
+{
+	ASSERT(buffer_mapped(bh));
+	ASSERT(buffer_locked(bh));
+	ASSERT(!buffer_delay(bh));
+	ASSERT(!buffer_unwritten(bh));
+
+	mark_buffer_async_write(bh);
+	set_buffer_uptodate(bh);
+	clear_buffer_dirty(bh);
+}
+
+STATIC void
+xfs_start_page_writeback(
+	struct page		*page,
+	int			clear_dirty,
+	int			buffers)
+{
+	ASSERT(PageLocked(page));
+	ASSERT(!PageWriteback(page));
+	if (clear_dirty)
+		clear_page_dirty_for_io(page);
+	set_page_writeback(page);
+	unlock_page(page);
+	/* If no buffers on the page are to be written, finish it here */
+	if (!buffers)
+		end_page_writeback(page);
+}
+
+static inline int bio_add_buffer(struct bio *bio, struct buffer_head *bh)
+{
+	return bio_add_page(bio, bh->b_page, bh->b_size, bh_offset(bh));
+}
+
+/*
+ * Submit all of the bios for all of the ioends we have saved up, covering the
+ * initial writepage page and also any probed pages.
+ *
+ * Because we may have multiple ioends spanning a page, we need to start
+ * writeback on all the buffers before we submit them for I/O. If we mark the
+ * buffers as we got, then we can end up with a page that only has buffers
+ * marked async write and I/O complete on can occur before we mark the other
+ * buffers async write.
+ *
+ * The end result of this is that we trip a bug in end_page_writeback() because
+ * we call it twice for the one page as the code in end_buffer_async_write()
+ * assumes that all buffers on the page are started at the same time.
+ *
+ * The fix is two passes across the ioend list - one to start writeback on the
+ * buffer_heads, and then submit them for I/O on the second pass.
+ */
+STATIC void
+xfs_submit_ioend(
+	struct writeback_control *wbc,
+	xfs_ioend_t		*ioend)
+{
+	xfs_ioend_t		*head = ioend;
+	xfs_ioend_t		*next;
+	struct buffer_head	*bh;
+	struct bio		*bio;
+	sector_t		lastblock = 0;
+
+	/* Pass 1 - start writeback */
+	do {
+		next = ioend->io_list;
+		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private)
+			xfs_start_buffer_writeback(bh);
+	} while ((ioend = next) != NULL);
+
+	/* Pass 2 - submit I/O */
+	ioend = head;
+	do {
+		next = ioend->io_list;
+		bio = NULL;
+
+		for (bh = ioend->io_buffer_head; bh; bh = bh->b_private) {
+
+			if (!bio) {
+ retry:
+				bio = xfs_alloc_ioend_bio(bh);
+			} else if (bh->b_blocknr != lastblock + 1) {
+				xfs_submit_ioend_bio(wbc, ioend, bio);
+				goto retry;
+			}
+
+			if (bio_add_buffer(bio, bh) != bh->b_size) {
+				xfs_submit_ioend_bio(wbc, ioend, bio);
+				goto retry;
+			}
+
+			lastblock = bh->b_blocknr;
+		}
+		if (bio)
+			xfs_submit_ioend_bio(wbc, ioend, bio);
+		xfs_finish_ioend(ioend);
+	} while ((ioend = next) != NULL);
+}
+
+/*
+ * Cancel submission of all buffer_heads so far in this endio.
+ * Toss the endio too.  Only ever called for the initial page
+ * in a writepage request, so only ever one page.
+ */
+STATIC void
+xfs_cancel_ioend(
+	xfs_ioend_t		*ioend)
+{
+	xfs_ioend_t		*next;
+	struct buffer_head	*bh, *next_bh;
+
+	do {
+		next = ioend->io_list;
+		bh = ioend->io_buffer_head;
+		do {
+			next_bh = bh->b_private;
+			clear_buffer_async_write(bh);
+			unlock_buffer(bh);
+		} while ((bh = next_bh) != NULL);
+
+		mempool_free(ioend, xfs_ioend_pool);
+	} while ((ioend = next) != NULL);
+}
+
+/*
+ * Test to see if we've been building up a completion structure for
+ * earlier buffers -- if so, we try to append to this ioend if we
+ * can, otherwise we finish off any current ioend and start another.
+ * Return true if we've finished the given ioend.
+ */
+STATIC void
+xfs_add_to_ioend(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	xfs_off_t		offset,
+	unsigned int		type,
+	xfs_ioend_t		**result,
+	int			need_ioend)
+{
+	xfs_ioend_t		*ioend = *result;
+
+	if (!ioend || need_ioend || type != ioend->io_type) {
+		xfs_ioend_t	*previous = *result;
+
+		ioend = xfs_alloc_ioend(inode, type);
+		ioend->io_offset = offset;
+		ioend->io_buffer_head = bh;
+		ioend->io_buffer_tail = bh;
+		if (previous)
+			previous->io_list = ioend;
+		*result = ioend;
+	} else {
+		ioend->io_buffer_tail->b_private = bh;
+		ioend->io_buffer_tail = bh;
+	}
+
+	bh->b_private = NULL;
+	ioend->io_size += bh->b_size;
+}
+
+STATIC void
+xfs_map_buffer(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	sector_t		bn;
+	struct xfs_mount	*m = XFS_I(inode)->i_mount;
+	xfs_off_t		iomap_offset = XFS_FSB_TO_B(m, imap->br_startoff);
+	xfs_daddr_t		iomap_bn = xfs_fsb_to_db(XFS_I(inode), imap->br_startblock);
+
+	ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+	ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+
+	bn = (iomap_bn >> (inode->i_blkbits - BBSHIFT)) +
+	      ((offset - iomap_offset) >> inode->i_blkbits);
+
+	ASSERT(bn || XFS_IS_REALTIME_INODE(XFS_I(inode)));
+
+	bh->b_blocknr = bn;
+	set_buffer_mapped(bh);
+}
+
+STATIC void
+xfs_map_at_offset(
+	struct inode		*inode,
+	struct buffer_head	*bh,
+	struct xfs_bmbt_irec	*imap,
+	xfs_off_t		offset)
+{
+	ASSERT(imap->br_startblock != HOLESTARTBLOCK);
+	ASSERT(imap->br_startblock != DELAYSTARTBLOCK);
+
+	xfs_map_buffer(inode, bh, imap, offset);
+	set_buffer_mapped(bh);
+	clear_buffer_delay(bh);
+	clear_buffer_unwritten(bh);
+}
+
+/*
+ * Test if a given page is suitable for writing as part of an unwritten
+ * or delayed allocate extent.
+ */
+STATIC int
+xfs_is_delayed_page(
+	struct page		*page,
+	unsigned int		type)
+{
+	if (PageWriteback(page))
+		return 0;
+
+	if (page->mapping && page_has_buffers(page)) {
+		struct buffer_head	*bh, *head;
+		int			acceptable = 0;
+
+		bh = head = page_buffers(page);
+		do {
+			if (buffer_unwritten(bh))
+				acceptable = (type == IO_UNWRITTEN);
+			else if (buffer_delay(bh))
+				acceptable = (type == IO_DELALLOC);
+			else if (buffer_dirty(bh) && buffer_mapped(bh))
+				acceptable = (type == IO_OVERWRITE);
+			else
+				break;
+		} while ((bh = bh->b_this_page) != head);
+
+		if (acceptable)
+			return 1;
+	}
+
+	return 0;
+}
+
+/*
+ * Allocate & map buffers for page given the extent map. Write it out.
+ * except for the original page of a writepage, this is called on
+ * delalloc/unwritten pages only, for the original page it is possible
+ * that the page has no mapping at all.
+ */
+STATIC int
+xfs_convert_page(
+	struct inode		*inode,
+	struct page		*page,
+	loff_t			tindex,
+	struct xfs_bmbt_irec	*imap,
+	xfs_ioend_t		**ioendp,
+	struct writeback_control *wbc)
+{
+	struct buffer_head	*bh, *head;
+	xfs_off_t		end_offset;
+	unsigned long		p_offset;
+	unsigned int		type;
+	int			len, page_dirty;
+	int			count = 0, done = 0, uptodate = 1;
+ 	xfs_off_t		offset = page_offset(page);
+
+	if (page->index != tindex)
+		goto fail;
+	if (!trylock_page(page))
+		goto fail;
+	if (PageWriteback(page))
+		goto fail_unlock_page;
+	if (page->mapping != inode->i_mapping)
+		goto fail_unlock_page;
+	if (!xfs_is_delayed_page(page, (*ioendp)->io_type))
+		goto fail_unlock_page;
+
+	/*
+	 * page_dirty is initially a count of buffers on the page before
+	 * EOF and is decremented as we move each into a cleanable state.
+	 *
+	 * Derivation:
+	 *
+	 * End offset is the highest offset that this page should represent.
+	 * If we are on the last page, (end_offset & (PAGE_CACHE_SIZE - 1))
+	 * will evaluate non-zero and be less than PAGE_CACHE_SIZE and
+	 * hence give us the correct page_dirty count. On any other page,
+	 * it will be zero and in that case we need page_dirty to be the
+	 * count of buffers on the page.
+	 */
+	end_offset = min_t(unsigned long long,
+			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
+			i_size_read(inode));
+
+	len = 1 << inode->i_blkbits;
+	p_offset = min_t(unsigned long, end_offset & (PAGE_CACHE_SIZE - 1),
+					PAGE_CACHE_SIZE);
+	p_offset = p_offset ? roundup(p_offset, len) : PAGE_CACHE_SIZE;
+	page_dirty = p_offset / len;
+
+	bh = head = page_buffers(page);
+	do {
+		if (offset >= end_offset)
+			break;
+		if (!buffer_uptodate(bh))
+			uptodate = 0;
+		if (!(PageUptodate(page) || buffer_uptodate(bh))) {
+			done = 1;
+			continue;
+		}
+
+		if (buffer_unwritten(bh) || buffer_delay(bh) ||
+		    buffer_mapped(bh)) {
+			if (buffer_unwritten(bh))
+				type = IO_UNWRITTEN;
+			else if (buffer_delay(bh))
+				type = IO_DELALLOC;
+			else
+				type = IO_OVERWRITE;
+
+			if (!xfs_imap_valid(inode, imap, offset)) {
+				done = 1;
+				continue;
+			}
+
+			lock_buffer(bh);
+			if (type != IO_OVERWRITE)
+				xfs_map_at_offset(inode, bh, imap, offset);
+			xfs_add_to_ioend(inode, bh, offset, type,
+					 ioendp, done);
+
+			page_dirty--;
+			count++;
+		} else {
+			done = 1;
+		}
+	} while (offset += len, (bh = bh->b_this_page) != head);
+
+	if (uptodate && bh == head)
+		SetPageUptodate(page);
+
+	if (count) {
+		if (--wbc->nr_to_write <= 0 &&
+		    wbc->sync_mode == WB_SYNC_NONE)
+			done = 1;
+	}
+	xfs_start_page_writeback(page, !page_dirty, count);
+
+	return done;
+ fail_unlock_page:
+	unlock_page(page);
+ fail:
+	return 1;
+}
+
+/*
+ * Convert & write out a cluster of pages in the same extent as defined
+ * by mp and following the start page.
+ */
+STATIC void
+xfs_cluster_write(
+	struct inode		*inode,
+	pgoff_t			tindex,
+	struct xfs_bmbt_irec	*imap,
+	xfs_ioend_t		**ioendp,
+	struct writeback_control *wbc,
+	pgoff_t			tlast)
+{
+	struct pagevec		pvec;
+	int			done = 0, i;
+
+	pagevec_init(&pvec, 0);
+	while (!done && tindex <= tlast) {
+		unsigned len = min_t(pgoff_t, PAGEVEC_SIZE, tlast - tindex + 1);
+
+		if (!pagevec_lookup(&pvec, inode->i_mapping, tindex, len))
+			break;
+
+		for (i = 0; i < pagevec_count(&pvec); i++) {
+			done = xfs_convert_page(inode, pvec.pages[i], tindex++,
+					imap, ioendp, wbc);
+			if (done)
+				break;
+		}
+
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+}
+
+STATIC void
+xfs_vm_invalidatepage(
+	struct page		*page,
+	unsigned long		offset)
+{
+	trace_xfs_invalidatepage(page->mapping->host, page, offset);
+	block_invalidatepage(page, offset);
+}
+
+/*
+ * If the page has delalloc buffers on it, we need to punch them out before we
+ * invalidate the page. If we don't, we leave a stale delalloc mapping on the
+ * inode that can trip a BUG() in xfs_get_blocks() later on if a direct IO read
+ * is done on that same region - the delalloc extent is returned when none is
+ * supposed to be there.
+ *
+ * We prevent this by truncating away the delalloc regions on the page before
+ * invalidating it. Because they are delalloc, we can do this without needing a
+ * transaction. Indeed - if we get ENOSPC errors, we have to be able to do this
+ * truncation without a transaction as there is no space left for block
+ * reservation (typically why we see a ENOSPC in writeback).
+ *
+ * This is not a performance critical path, so for now just do the punching a
+ * buffer head at a time.
+ */
+STATIC void
+xfs_aops_discard_page(
+	struct page		*page)
+{
+	struct inode		*inode = page->mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct buffer_head	*bh, *head;
+	loff_t			offset = page_offset(page);
+
+	if (!xfs_is_delayed_page(page, IO_DELALLOC))
+		goto out_invalidate;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		goto out_invalidate;
+
+	xfs_alert(ip->i_mount,
+		"page discard on page %p, inode 0x%llx, offset %llu.",
+			page, ip->i_ino, offset);
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	bh = head = page_buffers(page);
+	do {
+		int		error;
+		xfs_fileoff_t	start_fsb;
+
+		if (!buffer_delay(bh))
+			goto next_buffer;
+
+		start_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
+		error = xfs_bmap_punch_delalloc_range(ip, start_fsb, 1);
+		if (error) {
+			/* something screwed, just bail */
+			if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+				xfs_alert(ip->i_mount,
+			"page discard unable to remove delalloc mapping.");
+			}
+			break;
+		}
+next_buffer:
+		offset += 1 << inode->i_blkbits;
+
+	} while ((bh = bh->b_this_page) != head);
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+out_invalidate:
+	xfs_vm_invalidatepage(page, 0);
+	return;
+}
+
+/*
+ * Write out a dirty page.
+ *
+ * For delalloc space on the page we need to allocate space and flush it.
+ * For unwritten space on the page we need to start the conversion to
+ * regular allocated space.
+ * For any other dirty buffer heads on the page we should flush them.
+ */
+STATIC int
+xfs_vm_writepage(
+	struct page		*page,
+	struct writeback_control *wbc)
+{
+	struct inode		*inode = page->mapping->host;
+	struct buffer_head	*bh, *head;
+	struct xfs_bmbt_irec	imap;
+	xfs_ioend_t		*ioend = NULL, *iohead = NULL;
+	loff_t			offset;
+	unsigned int		type;
+	__uint64_t              end_offset;
+	pgoff_t                 end_index, last_index;
+	ssize_t			len;
+	int			err, imap_valid = 0, uptodate = 1;
+	int			count = 0;
+	int			nonblocking = 0;
+
+	trace_xfs_writepage(inode, page, 0);
+
+	ASSERT(page_has_buffers(page));
+
+	/*
+	 * Refuse to write the page out if we are called from reclaim context.
+	 *
+	 * This avoids stack overflows when called from deeply used stacks in
+	 * random callers for direct reclaim or memcg reclaim.  We explicitly
+	 * allow reclaim from kswapd as the stack usage there is relatively low.
+	 *
+	 * This should never happen except in the case of a VM regression so
+	 * warn about it.
+	 */
+	if (WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
+			PF_MEMALLOC))
+		goto redirty;
+
+	/*
+	 * Given that we do not allow direct reclaim to call us, we should
+	 * never be called while in a filesystem transaction.
+	 */
+	if (WARN_ON(current->flags & PF_FSTRANS))
+		goto redirty;
+
+	/* Is this page beyond the end of the file? */
+	offset = i_size_read(inode);
+	end_index = offset >> PAGE_CACHE_SHIFT;
+	last_index = (offset - 1) >> PAGE_CACHE_SHIFT;
+	if (page->index >= end_index) {
+		if ((page->index >= end_index + 1) ||
+		    !(i_size_read(inode) & (PAGE_CACHE_SIZE - 1))) {
+			unlock_page(page);
+			return 0;
+		}
+	}
+
+	end_offset = min_t(unsigned long long,
+			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
+			offset);
+	len = 1 << inode->i_blkbits;
+
+	bh = head = page_buffers(page);
+	offset = page_offset(page);
+	type = IO_OVERWRITE;
+
+	if (wbc->sync_mode == WB_SYNC_NONE)
+		nonblocking = 1;
+
+	do {
+		int new_ioend = 0;
+
+		if (offset >= end_offset)
+			break;
+		if (!buffer_uptodate(bh))
+			uptodate = 0;
+
+		/*
+		 * set_page_dirty dirties all buffers in a page, independent
+		 * of their state.  The dirty state however is entirely
+		 * meaningless for holes (!mapped && uptodate), so skip
+		 * buffers covering holes here.
+		 */
+		if (!buffer_mapped(bh) && buffer_uptodate(bh)) {
+			imap_valid = 0;
+			continue;
+		}
+
+		if (buffer_unwritten(bh)) {
+			if (type != IO_UNWRITTEN) {
+				type = IO_UNWRITTEN;
+				imap_valid = 0;
+			}
+		} else if (buffer_delay(bh)) {
+			if (type != IO_DELALLOC) {
+				type = IO_DELALLOC;
+				imap_valid = 0;
+			}
+		} else if (buffer_uptodate(bh)) {
+			if (type != IO_OVERWRITE) {
+				type = IO_OVERWRITE;
+				imap_valid = 0;
+			}
+		} else {
+			if (PageUptodate(page)) {
+				ASSERT(buffer_mapped(bh));
+				imap_valid = 0;
+			}
+			continue;
+		}
+
+		if (imap_valid)
+			imap_valid = xfs_imap_valid(inode, &imap, offset);
+		if (!imap_valid) {
+			/*
+			 * If we didn't have a valid mapping then we need to
+			 * put the new mapping into a separate ioend structure.
+			 * This ensures non-contiguous extents always have
+			 * separate ioends, which is particularly important
+			 * for unwritten extent conversion at I/O completion
+			 * time.
+			 */
+			new_ioend = 1;
+			err = xfs_map_blocks(inode, offset, &imap, type,
+					     nonblocking);
+			if (err)
+				goto error;
+			imap_valid = xfs_imap_valid(inode, &imap, offset);
+		}
+		if (imap_valid) {
+			lock_buffer(bh);
+			if (type != IO_OVERWRITE)
+				xfs_map_at_offset(inode, bh, &imap, offset);
+			xfs_add_to_ioend(inode, bh, offset, type, &ioend,
+					 new_ioend);
+			count++;
+		}
+
+		if (!iohead)
+			iohead = ioend;
+
+	} while (offset += len, ((bh = bh->b_this_page) != head));
+
+	if (uptodate && bh == head)
+		SetPageUptodate(page);
+
+	xfs_start_page_writeback(page, 1, count);
+
+	if (ioend && imap_valid) {
+		xfs_off_t		end_index;
+
+		end_index = imap.br_startoff + imap.br_blockcount;
+
+		/* to bytes */
+		end_index <<= inode->i_blkbits;
+
+		/* to pages */
+		end_index = (end_index - 1) >> PAGE_CACHE_SHIFT;
+
+		/* check against file size */
+		if (end_index > last_index)
+			end_index = last_index;
+
+		xfs_cluster_write(inode, page->index + 1, &imap, &ioend,
+				  wbc, end_index);
+	}
+
+	if (iohead)
+		xfs_submit_ioend(wbc, iohead);
+
+	return 0;
+
+error:
+	if (iohead)
+		xfs_cancel_ioend(iohead);
+
+	if (err == -EAGAIN)
+		goto redirty;
+
+	xfs_aops_discard_page(page);
+	ClearPageUptodate(page);
+	unlock_page(page);
+	return err;
+
+redirty:
+	redirty_page_for_writepage(wbc, page);
+	unlock_page(page);
+	return 0;
+}
+
+STATIC int
+xfs_vm_writepages(
+	struct address_space	*mapping,
+	struct writeback_control *wbc)
+{
+	xfs_iflags_clear(XFS_I(mapping->host), XFS_ITRUNCATED);
+	return generic_writepages(mapping, wbc);
+}
+
+/*
+ * Called to move a page into cleanable state - and from there
+ * to be released. The page should already be clean. We always
+ * have buffer heads in this call.
+ *
+ * Returns 1 if the page is ok to release, 0 otherwise.
+ */
+STATIC int
+xfs_vm_releasepage(
+	struct page		*page,
+	gfp_t			gfp_mask)
+{
+	int			delalloc, unwritten;
+
+	trace_xfs_releasepage(page->mapping->host, page, 0);
+
+	xfs_count_page_state(page, &delalloc, &unwritten);
+
+	if (WARN_ON(delalloc))
+		return 0;
+	if (WARN_ON(unwritten))
+		return 0;
+
+	return try_to_free_buffers(page);
+}
+
+STATIC int
+__xfs_get_blocks(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create,
+	int			direct)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_fileoff_t		offset_fsb, end_fsb;
+	int			error = 0;
+	int			lockmode = 0;
+	struct xfs_bmbt_irec	imap;
+	int			nimaps = 1;
+	xfs_off_t		offset;
+	ssize_t			size;
+	int			new = 0;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	offset = (xfs_off_t)iblock << inode->i_blkbits;
+	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
+	size = bh_result->b_size;
+
+	if (!create && direct && offset >= i_size_read(inode))
+		return 0;
+
+	if (create) {
+		lockmode = XFS_ILOCK_EXCL;
+		xfs_ilock(ip, lockmode);
+	} else {
+		lockmode = xfs_ilock_map_shared(ip);
+	}
+
+	ASSERT(offset <= mp->m_maxioffset);
+	if (offset + size > mp->m_maxioffset)
+		size = mp->m_maxioffset - offset;
+	end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + size);
+	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+
+	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb,
+				&imap, &nimaps, XFS_BMAPI_ENTIRE);
+	if (error)
+		goto out_unlock;
+
+	if (create &&
+	    (!nimaps ||
+	     (imap.br_startblock == HOLESTARTBLOCK ||
+	      imap.br_startblock == DELAYSTARTBLOCK))) {
+		if (direct) {
+			error = xfs_iomap_write_direct(ip, offset, size,
+						       &imap, nimaps);
+		} else {
+			error = xfs_iomap_write_delay(ip, offset, size, &imap);
+		}
+		if (error)
+			goto out_unlock;
+
+		trace_xfs_get_blocks_alloc(ip, offset, size, 0, &imap);
+	} else if (nimaps) {
+		trace_xfs_get_blocks_found(ip, offset, size, 0, &imap);
+	} else {
+		trace_xfs_get_blocks_notfound(ip, offset, size);
+		goto out_unlock;
+	}
+	xfs_iunlock(ip, lockmode);
+
+	if (imap.br_startblock != HOLESTARTBLOCK &&
+	    imap.br_startblock != DELAYSTARTBLOCK) {
+		/*
+		 * For unwritten extents do not report a disk address on
+		 * the read case (treat as if we're reading into a hole).
+		 */
+		if (create || !ISUNWRITTEN(&imap))
+			xfs_map_buffer(inode, bh_result, &imap, offset);
+		if (create && ISUNWRITTEN(&imap)) {
+			if (direct)
+				bh_result->b_private = inode;
+			set_buffer_unwritten(bh_result);
+		}
+	}
+
+	/*
+	 * If this is a realtime file, data may be on a different device.
+	 * to that pointed to from the buffer_head b_bdev currently.
+	 */
+	bh_result->b_bdev = xfs_find_bdev_for_inode(inode);
+
+	/*
+	 * If we previously allocated a block out beyond eof and we are now
+	 * coming back to use it then we will need to flag it as new even if it
+	 * has a disk address.
+	 *
+	 * With sub-block writes into unwritten extents we also need to mark
+	 * the buffer as new so that the unwritten parts of the buffer gets
+	 * correctly zeroed.
+	 */
+	if (create &&
+	    ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) ||
+	     (offset >= i_size_read(inode)) ||
+	     (new || ISUNWRITTEN(&imap))))
+		set_buffer_new(bh_result);
+
+	if (imap.br_startblock == DELAYSTARTBLOCK) {
+		BUG_ON(direct);
+		if (create) {
+			set_buffer_uptodate(bh_result);
+			set_buffer_mapped(bh_result);
+			set_buffer_delay(bh_result);
+		}
+	}
+
+	/*
+	 * If this is O_DIRECT or the mpage code calling tell them how large
+	 * the mapping is, so that we can avoid repeated get_blocks calls.
+	 */
+	if (direct || size > (1 << inode->i_blkbits)) {
+		xfs_off_t		mapping_size;
+
+		mapping_size = imap.br_startoff + imap.br_blockcount - iblock;
+		mapping_size <<= inode->i_blkbits;
+
+		ASSERT(mapping_size > 0);
+		if (mapping_size > size)
+			mapping_size = size;
+		if (mapping_size > LONG_MAX)
+			mapping_size = LONG_MAX;
+
+		bh_result->b_size = mapping_size;
+	}
+
+	return 0;
+
+out_unlock:
+	xfs_iunlock(ip, lockmode);
+	return -error;
+}
+
+int
+xfs_get_blocks(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create)
+{
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 0);
+}
+
+STATIC int
+xfs_get_blocks_direct(
+	struct inode		*inode,
+	sector_t		iblock,
+	struct buffer_head	*bh_result,
+	int			create)
+{
+	return __xfs_get_blocks(inode, iblock, bh_result, create, 1);
+}
+
+/*
+ * Complete a direct I/O write request.
+ *
+ * If the private argument is non-NULL __xfs_get_blocks signals us that we
+ * need to issue a transaction to convert the range from unwritten to written
+ * extents.  In case this is regular synchronous I/O we just call xfs_end_io
+ * to do this and we are done.  But in case this was a successful AIO
+ * request this handler is called from interrupt context, from which we
+ * can't start transactions.  In that case offload the I/O completion to
+ * the workqueues we also use for buffered I/O completion.
+ */
+STATIC void
+xfs_end_io_direct_write(
+	struct kiocb		*iocb,
+	loff_t			offset,
+	ssize_t			size,
+	void			*private,
+	int			ret,
+	bool			is_async)
+{
+	struct xfs_ioend	*ioend = iocb->private;
+
+	/*
+	 * blockdev_direct_IO can return an error even after the I/O
+	 * completion handler was called.  Thus we need to protect
+	 * against double-freeing.
+	 */
+	iocb->private = NULL;
+
+	ioend->io_offset = offset;
+	ioend->io_size = size;
+	ioend->io_iocb = iocb;
+	ioend->io_result = ret;
+	if (private && size > 0)
+		ioend->io_type = IO_UNWRITTEN;
+
+	if (is_async) {
+		ioend->io_isasync = 1;
+		xfs_finish_ioend(ioend);
+	} else {
+		xfs_finish_ioend_sync(ioend);
+	}
+}
+
+STATIC ssize_t
+xfs_vm_direct_IO(
+	int			rw,
+	struct kiocb		*iocb,
+	const struct iovec	*iov,
+	loff_t			offset,
+	unsigned long		nr_segs)
+{
+	struct inode		*inode = iocb->ki_filp->f_mapping->host;
+	struct block_device	*bdev = xfs_find_bdev_for_inode(inode);
+	ssize_t			ret;
+
+	if (rw & WRITE) {
+		iocb->private = xfs_alloc_ioend(inode, IO_DIRECT);
+
+		ret = __blockdev_direct_IO(rw, iocb, inode, bdev, iov,
+					    offset, nr_segs,
+					    xfs_get_blocks_direct,
+					    xfs_end_io_direct_write, NULL, 0);
+		if (ret != -EIOCBQUEUED && iocb->private)
+			xfs_destroy_ioend(iocb->private);
+	} else {
+		ret = __blockdev_direct_IO(rw, iocb, inode, bdev, iov,
+					    offset, nr_segs,
+					    xfs_get_blocks_direct,
+					    NULL, NULL, 0);
+	}
+
+	return ret;
+}
+
+STATIC void
+xfs_vm_write_failed(
+	struct address_space	*mapping,
+	loff_t			to)
+{
+	struct inode		*inode = mapping->host;
+
+	if (to > inode->i_size) {
+		/*
+		 * punch out the delalloc blocks we have already allocated. We
+		 * don't call xfs_setattr() to do this as we may be in the
+		 * middle of a multi-iovec write and so the vfs inode->i_size
+		 * will not match the xfs ip->i_size and so it will zero too
+		 * much. Hence we jus truncate the page cache to zero what is
+		 * necessary and punch the delalloc blocks directly.
+		 */
+		struct xfs_inode	*ip = XFS_I(inode);
+		xfs_fileoff_t		start_fsb;
+		xfs_fileoff_t		end_fsb;
+		int			error;
+
+		truncate_pagecache(inode, to, inode->i_size);
+
+		/*
+		 * Check if there are any blocks that are outside of i_size
+		 * that need to be trimmed back.
+		 */
+		start_fsb = XFS_B_TO_FSB(ip->i_mount, inode->i_size) + 1;
+		end_fsb = XFS_B_TO_FSB(ip->i_mount, to);
+		if (end_fsb <= start_fsb)
+			return;
+
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		error = xfs_bmap_punch_delalloc_range(ip, start_fsb,
+							end_fsb - start_fsb);
+		if (error) {
+			/* something screwed, just bail */
+			if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+				xfs_alert(ip->i_mount,
+			"xfs_vm_write_failed: unable to clean up ino %lld",
+						ip->i_ino);
+			}
+		}
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	}
+}
+
+STATIC int
+xfs_vm_write_begin(
+	struct file		*file,
+	struct address_space	*mapping,
+	loff_t			pos,
+	unsigned		len,
+	unsigned		flags,
+	struct page		**pagep,
+	void			**fsdata)
+{
+	int			ret;
+
+	ret = block_write_begin(mapping, pos, len, flags | AOP_FLAG_NOFS,
+				pagep, xfs_get_blocks);
+	if (unlikely(ret))
+		xfs_vm_write_failed(mapping, pos + len);
+	return ret;
+}
+
+STATIC int
+xfs_vm_write_end(
+	struct file		*file,
+	struct address_space	*mapping,
+	loff_t			pos,
+	unsigned		len,
+	unsigned		copied,
+	struct page		*page,
+	void			*fsdata)
+{
+	int			ret;
+
+	ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata);
+	if (unlikely(ret < len))
+		xfs_vm_write_failed(mapping, pos + len);
+	return ret;
+}
+
+STATIC sector_t
+xfs_vm_bmap(
+	struct address_space	*mapping,
+	sector_t		block)
+{
+	struct inode		*inode = (struct inode *)mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+
+	trace_xfs_vm_bmap(XFS_I(inode));
+	xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	return generic_block_bmap(mapping, block, xfs_get_blocks);
+}
+
+STATIC int
+xfs_vm_readpage(
+	struct file		*unused,
+	struct page		*page)
+{
+	return mpage_readpage(page, xfs_get_blocks);
+}
+
+STATIC int
+xfs_vm_readpages(
+	struct file		*unused,
+	struct address_space	*mapping,
+	struct list_head	*pages,
+	unsigned		nr_pages)
+{
+	return mpage_readpages(mapping, pages, nr_pages, xfs_get_blocks);
+}
+
+const struct address_space_operations xfs_address_space_operations = {
+	.readpage		= xfs_vm_readpage,
+	.readpages		= xfs_vm_readpages,
+	.writepage		= xfs_vm_writepage,
+	.writepages		= xfs_vm_writepages,
+	.releasepage		= xfs_vm_releasepage,
+	.invalidatepage		= xfs_vm_invalidatepage,
+	.write_begin		= xfs_vm_write_begin,
+	.write_end		= xfs_vm_write_end,
+	.bmap			= xfs_vm_bmap,
+	.direct_IO		= xfs_vm_direct_IO,
+	.migratepage		= buffer_migrate_page,
+	.is_partially_uptodate  = block_is_partially_uptodate,
+	.error_remove_page	= generic_error_remove_page,
+};
diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h
new file mode 100644
index 0000000..116dd5c
--- /dev/null
+++ b/fs/xfs/xfs_aops.h
@@ -0,0 +1,66 @@
+/*
+ * Copyright (c) 2005-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_AOPS_H__
+#define __XFS_AOPS_H__
+
+extern struct workqueue_struct *xfsdatad_workqueue;
+extern struct workqueue_struct *xfsconvertd_workqueue;
+extern mempool_t *xfs_ioend_pool;
+
+/*
+ * Types of I/O for bmap clustering and I/O completion tracking.
+ */
+enum {
+	IO_DIRECT = 0,	/* special case for direct I/O ioends */
+	IO_DELALLOC,	/* mapping covers delalloc region */
+	IO_UNWRITTEN,	/* mapping covers allocated but uninitialized data */
+	IO_OVERWRITE,	/* mapping covers already allocated extent */
+};
+
+#define XFS_IO_TYPES \
+	{ 0,			"" }, \
+	{ IO_DELALLOC,		"delalloc" }, \
+	{ IO_UNWRITTEN,		"unwritten" }, \
+	{ IO_OVERWRITE,		"overwrite" }
+
+/*
+ * xfs_ioend struct manages large extent writes for XFS.
+ * It can manage several multi-page bio's at once.
+ */
+typedef struct xfs_ioend {
+	struct xfs_ioend	*io_list;	/* next ioend in chain */
+	unsigned int		io_type;	/* delalloc / unwritten */
+	int			io_error;	/* I/O error code */
+	atomic_t		io_remaining;	/* hold count */
+	unsigned int		io_isasync : 1;	/* needs aio_complete */
+	struct inode		*io_inode;	/* file being written to */
+	struct buffer_head	*io_buffer_head;/* buffer linked list head */
+	struct buffer_head	*io_buffer_tail;/* buffer linked list tail */
+	size_t			io_size;	/* size of the extent */
+	xfs_off_t		io_offset;	/* offset in the file */
+	struct work_struct	io_work;	/* xfsdatad work queue */
+	struct kiocb		*io_iocb;
+	int			io_result;
+} xfs_ioend_t;
+
+extern const struct address_space_operations xfs_address_space_operations;
+extern int xfs_get_blocks(struct inode *, sector_t, struct buffer_head *, int);
+
+extern void xfs_count_page_state(struct page *, int *, int *);
+
+#endif /* __XFS_AOPS_H__ */
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
new file mode 100644
index 0000000..2f5a8f7
--- /dev/null
+++ b/fs/xfs/xfs_buf.c
@@ -0,0 +1,1838 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include <linux/stddef.h>
+#include <linux/errno.h>
+#include <linux/gfp.h>
+#include <linux/pagemap.h>
+#include <linux/init.h>
+#include <linux/vmalloc.h>
+#include <linux/bio.h>
+#include <linux/sysctl.h>
+#include <linux/proc_fs.h>
+#include <linux/workqueue.h>
+#include <linux/percpu.h>
+#include <linux/blkdev.h>
+#include <linux/hash.h>
+#include <linux/kthread.h>
+#include <linux/migrate.h>
+#include <linux/backing-dev.h>
+#include <linux/freezer.h>
+
+#include "xfs_sb.h"
+#include "xfs_inum.h"
+#include "xfs_log.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_trace.h"
+
+static kmem_zone_t *xfs_buf_zone;
+STATIC int xfsbufd(void *);
+
+static struct workqueue_struct *xfslogd_workqueue;
+struct workqueue_struct *xfsdatad_workqueue;
+struct workqueue_struct *xfsconvertd_workqueue;
+
+#ifdef XFS_BUF_LOCK_TRACKING
+# define XB_SET_OWNER(bp)	((bp)->b_last_holder = current->pid)
+# define XB_CLEAR_OWNER(bp)	((bp)->b_last_holder = -1)
+# define XB_GET_OWNER(bp)	((bp)->b_last_holder)
+#else
+# define XB_SET_OWNER(bp)	do { } while (0)
+# define XB_CLEAR_OWNER(bp)	do { } while (0)
+# define XB_GET_OWNER(bp)	do { } while (0)
+#endif
+
+#define xb_to_gfp(flags) \
+	((((flags) & XBF_READ_AHEAD) ? __GFP_NORETRY : \
+	  ((flags) & XBF_DONT_BLOCK) ? GFP_NOFS : GFP_KERNEL) | __GFP_NOWARN)
+
+#define xb_to_km(flags) \
+	 (((flags) & XBF_DONT_BLOCK) ? KM_NOFS : KM_SLEEP)
+
+
+static inline int
+xfs_buf_is_vmapped(
+	struct xfs_buf	*bp)
+{
+	/*
+	 * Return true if the buffer is vmapped.
+	 *
+	 * The XBF_MAPPED flag is set if the buffer should be mapped, but the
+	 * code is clever enough to know it doesn't have to map a single page,
+	 * so the check has to be both for XBF_MAPPED and bp->b_page_count > 1.
+	 */
+	return (bp->b_flags & XBF_MAPPED) && bp->b_page_count > 1;
+}
+
+static inline int
+xfs_buf_vmap_len(
+	struct xfs_buf	*bp)
+{
+	return (bp->b_page_count * PAGE_SIZE) - bp->b_offset;
+}
+
+/*
+ * xfs_buf_lru_add - add a buffer to the LRU.
+ *
+ * The LRU takes a new reference to the buffer so that it will only be freed
+ * once the shrinker takes the buffer off the LRU.
+ */
+STATIC void
+xfs_buf_lru_add(
+	struct xfs_buf	*bp)
+{
+	struct xfs_buftarg *btp = bp->b_target;
+
+	spin_lock(&btp->bt_lru_lock);
+	if (list_empty(&bp->b_lru)) {
+		atomic_inc(&bp->b_hold);
+		list_add_tail(&bp->b_lru, &btp->bt_lru);
+		btp->bt_lru_nr++;
+	}
+	spin_unlock(&btp->bt_lru_lock);
+}
+
+/*
+ * xfs_buf_lru_del - remove a buffer from the LRU
+ *
+ * The unlocked check is safe here because it only occurs when there are not
+ * b_lru_ref counts left on the inode under the pag->pag_buf_lock. it is there
+ * to optimise the shrinker removing the buffer from the LRU and calling
+ * xfs_buf_free(). i.e. it removes an unnecessary round trip on the
+ * bt_lru_lock.
+ */
+STATIC void
+xfs_buf_lru_del(
+	struct xfs_buf	*bp)
+{
+	struct xfs_buftarg *btp = bp->b_target;
+
+	if (list_empty(&bp->b_lru))
+		return;
+
+	spin_lock(&btp->bt_lru_lock);
+	if (!list_empty(&bp->b_lru)) {
+		list_del_init(&bp->b_lru);
+		btp->bt_lru_nr--;
+	}
+	spin_unlock(&btp->bt_lru_lock);
+}
+
+/*
+ * When we mark a buffer stale, we remove the buffer from the LRU and clear the
+ * b_lru_ref count so that the buffer is freed immediately when the buffer
+ * reference count falls to zero. If the buffer is already on the LRU, we need
+ * to remove the reference that LRU holds on the buffer.
+ *
+ * This prevents build-up of stale buffers on the LRU.
+ */
+void
+xfs_buf_stale(
+	struct xfs_buf	*bp)
+{
+	bp->b_flags |= XBF_STALE;
+	xfs_buf_delwri_dequeue(bp);
+	atomic_set(&(bp)->b_lru_ref, 0);
+	if (!list_empty(&bp->b_lru)) {
+		struct xfs_buftarg *btp = bp->b_target;
+
+		spin_lock(&btp->bt_lru_lock);
+		if (!list_empty(&bp->b_lru)) {
+			list_del_init(&bp->b_lru);
+			btp->bt_lru_nr--;
+			atomic_dec(&bp->b_hold);
+		}
+		spin_unlock(&btp->bt_lru_lock);
+	}
+	ASSERT(atomic_read(&bp->b_hold) >= 1);
+}
+
+struct xfs_buf *
+xfs_buf_alloc(
+	struct xfs_buftarg	*target,
+	xfs_off_t		range_base,
+	size_t			range_length,
+	xfs_buf_flags_t		flags)
+{
+	struct xfs_buf		*bp;
+
+	bp = kmem_zone_alloc(xfs_buf_zone, xb_to_km(flags));
+	if (unlikely(!bp))
+		return NULL;
+
+	/*
+	 * We don't want certain flags to appear in b_flags.
+	 */
+	flags &= ~(XBF_LOCK|XBF_MAPPED|XBF_DONT_BLOCK|XBF_READ_AHEAD);
+
+	memset(bp, 0, sizeof(xfs_buf_t));
+	atomic_set(&bp->b_hold, 1);
+	atomic_set(&bp->b_lru_ref, 1);
+	init_completion(&bp->b_iowait);
+	INIT_LIST_HEAD(&bp->b_lru);
+	INIT_LIST_HEAD(&bp->b_list);
+	RB_CLEAR_NODE(&bp->b_rbnode);
+	sema_init(&bp->b_sema, 0); /* held, no waiters */
+	XB_SET_OWNER(bp);
+	bp->b_target = target;
+	bp->b_file_offset = range_base;
+	/*
+	 * Set buffer_length and count_desired to the same value initially.
+	 * I/O routines should use count_desired, which will be the same in
+	 * most cases but may be reset (e.g. XFS recovery).
+	 */
+	bp->b_buffer_length = bp->b_count_desired = range_length;
+	bp->b_flags = flags;
+	bp->b_bn = XFS_BUF_DADDR_NULL;
+	atomic_set(&bp->b_pin_count, 0);
+	init_waitqueue_head(&bp->b_waiters);
+
+	XFS_STATS_INC(xb_create);
+	trace_xfs_buf_init(bp, _RET_IP_);
+
+	return bp;
+}
+
+/*
+ *	Allocate a page array capable of holding a specified number
+ *	of pages, and point the page buf at it.
+ */
+STATIC int
+_xfs_buf_get_pages(
+	xfs_buf_t		*bp,
+	int			page_count,
+	xfs_buf_flags_t		flags)
+{
+	/* Make sure that we have a page list */
+	if (bp->b_pages == NULL) {
+		bp->b_offset = xfs_buf_poff(bp->b_file_offset);
+		bp->b_page_count = page_count;
+		if (page_count <= XB_PAGES) {
+			bp->b_pages = bp->b_page_array;
+		} else {
+			bp->b_pages = kmem_alloc(sizeof(struct page *) *
+					page_count, xb_to_km(flags));
+			if (bp->b_pages == NULL)
+				return -ENOMEM;
+		}
+		memset(bp->b_pages, 0, sizeof(struct page *) * page_count);
+	}
+	return 0;
+}
+
+/*
+ *	Frees b_pages if it was allocated.
+ */
+STATIC void
+_xfs_buf_free_pages(
+	xfs_buf_t	*bp)
+{
+	if (bp->b_pages != bp->b_page_array) {
+		kmem_free(bp->b_pages);
+		bp->b_pages = NULL;
+	}
+}
+
+/*
+ *	Releases the specified buffer.
+ *
+ * 	The modification state of any associated pages is left unchanged.
+ * 	The buffer most not be on any hash - use xfs_buf_rele instead for
+ * 	hashed and refcounted buffers
+ */
+void
+xfs_buf_free(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_free(bp, _RET_IP_);
+
+	ASSERT(list_empty(&bp->b_lru));
+
+	if (bp->b_flags & _XBF_PAGES) {
+		uint		i;
+
+		if (xfs_buf_is_vmapped(bp))
+			vm_unmap_ram(bp->b_addr - bp->b_offset,
+					bp->b_page_count);
+
+		for (i = 0; i < bp->b_page_count; i++) {
+			struct page	*page = bp->b_pages[i];
+
+			__free_page(page);
+		}
+	} else if (bp->b_flags & _XBF_KMEM)
+		kmem_free(bp->b_addr);
+	_xfs_buf_free_pages(bp);
+	kmem_zone_free(xfs_buf_zone, bp);
+}
+
+/*
+ * Allocates all the pages for buffer in question and builds it's page list.
+ */
+STATIC int
+xfs_buf_allocate_memory(
+	xfs_buf_t		*bp,
+	uint			flags)
+{
+	size_t			size = bp->b_count_desired;
+	size_t			nbytes, offset;
+	gfp_t			gfp_mask = xb_to_gfp(flags);
+	unsigned short		page_count, i;
+	xfs_off_t		end;
+	int			error;
+
+	/*
+	 * for buffers that are contained within a single page, just allocate
+	 * the memory from the heap - there's no need for the complexity of
+	 * page arrays to keep allocation down to order 0.
+	 */
+	if (bp->b_buffer_length < PAGE_SIZE) {
+		bp->b_addr = kmem_alloc(bp->b_buffer_length, xb_to_km(flags));
+		if (!bp->b_addr) {
+			/* low memory - use alloc_page loop instead */
+			goto use_alloc_page;
+		}
+
+		if (((unsigned long)(bp->b_addr + bp->b_buffer_length - 1) &
+								PAGE_MASK) !=
+		    ((unsigned long)bp->b_addr & PAGE_MASK)) {
+			/* b_addr spans two pages - use alloc_page instead */
+			kmem_free(bp->b_addr);
+			bp->b_addr = NULL;
+			goto use_alloc_page;
+		}
+		bp->b_offset = offset_in_page(bp->b_addr);
+		bp->b_pages = bp->b_page_array;
+		bp->b_pages[0] = virt_to_page(bp->b_addr);
+		bp->b_page_count = 1;
+		bp->b_flags |= XBF_MAPPED | _XBF_KMEM;
+		return 0;
+	}
+
+use_alloc_page:
+	end = bp->b_file_offset + bp->b_buffer_length;
+	page_count = xfs_buf_btoc(end) - xfs_buf_btoct(bp->b_file_offset);
+	error = _xfs_buf_get_pages(bp, page_count, flags);
+	if (unlikely(error))
+		return error;
+
+	offset = bp->b_offset;
+	bp->b_flags |= _XBF_PAGES;
+
+	for (i = 0; i < bp->b_page_count; i++) {
+		struct page	*page;
+		uint		retries = 0;
+retry:
+		page = alloc_page(gfp_mask);
+		if (unlikely(page == NULL)) {
+			if (flags & XBF_READ_AHEAD) {
+				bp->b_page_count = i;
+				error = ENOMEM;
+				goto out_free_pages;
+			}
+
+			/*
+			 * This could deadlock.
+			 *
+			 * But until all the XFS lowlevel code is revamped to
+			 * handle buffer allocation failures we can't do much.
+			 */
+			if (!(++retries % 100))
+				xfs_err(NULL,
+		"possible memory allocation deadlock in %s (mode:0x%x)",
+					__func__, gfp_mask);
+
+			XFS_STATS_INC(xb_page_retries);
+			congestion_wait(BLK_RW_ASYNC, HZ/50);
+			goto retry;
+		}
+
+		XFS_STATS_INC(xb_page_found);
+
+		nbytes = min_t(size_t, size, PAGE_SIZE - offset);
+		size -= nbytes;
+		bp->b_pages[i] = page;
+		offset = 0;
+	}
+	return 0;
+
+out_free_pages:
+	for (i = 0; i < bp->b_page_count; i++)
+		__free_page(bp->b_pages[i]);
+	return error;
+}
+
+/*
+ *	Map buffer into kernel address-space if necessary.
+ */
+STATIC int
+_xfs_buf_map_pages(
+	xfs_buf_t		*bp,
+	uint			flags)
+{
+	ASSERT(bp->b_flags & _XBF_PAGES);
+	if (bp->b_page_count == 1) {
+		/* A single page buffer is always mappable */
+		bp->b_addr = page_address(bp->b_pages[0]) + bp->b_offset;
+		bp->b_flags |= XBF_MAPPED;
+	} else if (flags & XBF_MAPPED) {
+		int retried = 0;
+
+		do {
+			bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count,
+						-1, PAGE_KERNEL);
+			if (bp->b_addr)
+				break;
+			vm_unmap_aliases();
+		} while (retried++ <= 1);
+
+		if (!bp->b_addr)
+			return -ENOMEM;
+		bp->b_addr += bp->b_offset;
+		bp->b_flags |= XBF_MAPPED;
+	}
+
+	return 0;
+}
+
+/*
+ *	Finding and Reading Buffers
+ */
+
+/*
+ *	Look up, and creates if absent, a lockable buffer for
+ *	a given range of an inode.  The buffer is returned
+ *	locked.	No I/O is implied by this call.
+ */
+xfs_buf_t *
+_xfs_buf_find(
+	xfs_buftarg_t		*btp,	/* block device target		*/
+	xfs_off_t		ioff,	/* starting offset of range	*/
+	size_t			isize,	/* length of range		*/
+	xfs_buf_flags_t		flags,
+	xfs_buf_t		*new_bp)
+{
+	xfs_off_t		range_base;
+	size_t			range_length;
+	struct xfs_perag	*pag;
+	struct rb_node		**rbp;
+	struct rb_node		*parent;
+	xfs_buf_t		*bp;
+
+	range_base = (ioff << BBSHIFT);
+	range_length = (isize << BBSHIFT);
+
+	/* Check for IOs smaller than the sector size / not sector aligned */
+	ASSERT(!(range_length < (1 << btp->bt_sshift)));
+	ASSERT(!(range_base & (xfs_off_t)btp->bt_smask));
+
+	/* get tree root */
+	pag = xfs_perag_get(btp->bt_mount,
+				xfs_daddr_to_agno(btp->bt_mount, ioff));
+
+	/* walk tree */
+	spin_lock(&pag->pag_buf_lock);
+	rbp = &pag->pag_buf_tree.rb_node;
+	parent = NULL;
+	bp = NULL;
+	while (*rbp) {
+		parent = *rbp;
+		bp = rb_entry(parent, struct xfs_buf, b_rbnode);
+
+		if (range_base < bp->b_file_offset)
+			rbp = &(*rbp)->rb_left;
+		else if (range_base > bp->b_file_offset)
+			rbp = &(*rbp)->rb_right;
+		else {
+			/*
+			 * found a block offset match. If the range doesn't
+			 * match, the only way this is allowed is if the buffer
+			 * in the cache is stale and the transaction that made
+			 * it stale has not yet committed. i.e. we are
+			 * reallocating a busy extent. Skip this buffer and
+			 * continue searching to the right for an exact match.
+			 */
+			if (bp->b_buffer_length != range_length) {
+				ASSERT(bp->b_flags & XBF_STALE);
+				rbp = &(*rbp)->rb_right;
+				continue;
+			}
+			atomic_inc(&bp->b_hold);
+			goto found;
+		}
+	}
+
+	/* No match found */
+	if (new_bp) {
+		rb_link_node(&new_bp->b_rbnode, parent, rbp);
+		rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree);
+		/* the buffer keeps the perag reference until it is freed */
+		new_bp->b_pag = pag;
+		spin_unlock(&pag->pag_buf_lock);
+	} else {
+		XFS_STATS_INC(xb_miss_locked);
+		spin_unlock(&pag->pag_buf_lock);
+		xfs_perag_put(pag);
+	}
+	return new_bp;
+
+found:
+	spin_unlock(&pag->pag_buf_lock);
+	xfs_perag_put(pag);
+
+	if (!xfs_buf_trylock(bp)) {
+		if (flags & XBF_TRYLOCK) {
+			xfs_buf_rele(bp);
+			XFS_STATS_INC(xb_busy_locked);
+			return NULL;
+		}
+		xfs_buf_lock(bp);
+		XFS_STATS_INC(xb_get_locked_waited);
+	}
+
+	/*
+	 * if the buffer is stale, clear all the external state associated with
+	 * it. We need to keep flags such as how we allocated the buffer memory
+	 * intact here.
+	 */
+	if (bp->b_flags & XBF_STALE) {
+		ASSERT((bp->b_flags & _XBF_DELWRI_Q) == 0);
+		bp->b_flags &= XBF_MAPPED | _XBF_KMEM | _XBF_PAGES;
+	}
+
+	trace_xfs_buf_find(bp, flags, _RET_IP_);
+	XFS_STATS_INC(xb_get_locked);
+	return bp;
+}
+
+/*
+ * Assembles a buffer covering the specified range. The code is optimised for
+ * cache hits, as metadata intensive workloads will see 3 orders of magnitude
+ * more hits than misses.
+ */
+struct xfs_buf *
+xfs_buf_get(
+	xfs_buftarg_t		*target,/* target for buffer		*/
+	xfs_off_t		ioff,	/* starting offset of range	*/
+	size_t			isize,	/* length of range		*/
+	xfs_buf_flags_t		flags)
+{
+	struct xfs_buf		*bp;
+	struct xfs_buf		*new_bp;
+	int			error = 0;
+
+	bp = _xfs_buf_find(target, ioff, isize, flags, NULL);
+	if (likely(bp))
+		goto found;
+
+	new_bp = xfs_buf_alloc(target, ioff << BBSHIFT, isize << BBSHIFT,
+			       flags);
+	if (unlikely(!new_bp))
+		return NULL;
+
+	bp = _xfs_buf_find(target, ioff, isize, flags, new_bp);
+	if (!bp) {
+		kmem_zone_free(xfs_buf_zone, new_bp);
+		return NULL;
+	}
+
+	if (bp == new_bp) {
+		error = xfs_buf_allocate_memory(bp, flags);
+		if (error)
+			goto no_buffer;
+	} else
+		kmem_zone_free(xfs_buf_zone, new_bp);
+
+	/*
+	 * Now we have a workable buffer, fill in the block number so
+	 * that we can do IO on it.
+	 */
+	bp->b_bn = ioff;
+	bp->b_count_desired = bp->b_buffer_length;
+
+found:
+	if (!(bp->b_flags & XBF_MAPPED)) {
+		error = _xfs_buf_map_pages(bp, flags);
+		if (unlikely(error)) {
+			xfs_warn(target->bt_mount,
+				"%s: failed to map pages\n", __func__);
+			goto no_buffer;
+		}
+	}
+
+	XFS_STATS_INC(xb_get);
+	trace_xfs_buf_get(bp, flags, _RET_IP_);
+	return bp;
+
+no_buffer:
+	if (flags & (XBF_LOCK | XBF_TRYLOCK))
+		xfs_buf_unlock(bp);
+	xfs_buf_rele(bp);
+	return NULL;
+}
+
+STATIC int
+_xfs_buf_read(
+	xfs_buf_t		*bp,
+	xfs_buf_flags_t		flags)
+{
+	int			status;
+
+	ASSERT(!(flags & (XBF_DELWRI|XBF_WRITE)));
+	ASSERT(bp->b_bn != XFS_BUF_DADDR_NULL);
+
+	bp->b_flags &= ~(XBF_WRITE | XBF_ASYNC | XBF_DELWRI | XBF_READ_AHEAD);
+	bp->b_flags |= flags & (XBF_READ | XBF_ASYNC | XBF_READ_AHEAD);
+
+	status = xfs_buf_iorequest(bp);
+	if (status || bp->b_error || (flags & XBF_ASYNC))
+		return status;
+	return xfs_buf_iowait(bp);
+}
+
+xfs_buf_t *
+xfs_buf_read(
+	xfs_buftarg_t		*target,
+	xfs_off_t		ioff,
+	size_t			isize,
+	xfs_buf_flags_t		flags)
+{
+	xfs_buf_t		*bp;
+
+	flags |= XBF_READ;
+
+	bp = xfs_buf_get(target, ioff, isize, flags);
+	if (bp) {
+		trace_xfs_buf_read(bp, flags, _RET_IP_);
+
+		if (!XFS_BUF_ISDONE(bp)) {
+			XFS_STATS_INC(xb_get_read);
+			_xfs_buf_read(bp, flags);
+		} else if (flags & XBF_ASYNC) {
+			/*
+			 * Read ahead call which is already satisfied,
+			 * drop the buffer
+			 */
+			goto no_buffer;
+		} else {
+			/* We do not want read in the flags */
+			bp->b_flags &= ~XBF_READ;
+		}
+	}
+
+	return bp;
+
+ no_buffer:
+	if (flags & (XBF_LOCK | XBF_TRYLOCK))
+		xfs_buf_unlock(bp);
+	xfs_buf_rele(bp);
+	return NULL;
+}
+
+/*
+ *	If we are not low on memory then do the readahead in a deadlock
+ *	safe manner.
+ */
+void
+xfs_buf_readahead(
+	xfs_buftarg_t		*target,
+	xfs_off_t		ioff,
+	size_t			isize)
+{
+	if (bdi_read_congested(target->bt_bdi))
+		return;
+
+	xfs_buf_read(target, ioff, isize,
+		     XBF_TRYLOCK|XBF_ASYNC|XBF_READ_AHEAD|XBF_DONT_BLOCK);
+}
+
+/*
+ * Read an uncached buffer from disk. Allocates and returns a locked
+ * buffer containing the disk contents or nothing.
+ */
+struct xfs_buf *
+xfs_buf_read_uncached(
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*target,
+	xfs_daddr_t		daddr,
+	size_t			length,
+	int			flags)
+{
+	xfs_buf_t		*bp;
+	int			error;
+
+	bp = xfs_buf_get_uncached(target, length, flags);
+	if (!bp)
+		return NULL;
+
+	/* set up the buffer for a read IO */
+	XFS_BUF_SET_ADDR(bp, daddr);
+	XFS_BUF_READ(bp);
+
+	xfsbdstrat(mp, bp);
+	error = xfs_buf_iowait(bp);
+	if (error || bp->b_error) {
+		xfs_buf_relse(bp);
+		return NULL;
+	}
+	return bp;
+}
+
+/*
+ * Return a buffer allocated as an empty buffer and associated to external
+ * memory via xfs_buf_associate_memory() back to it's empty state.
+ */
+void
+xfs_buf_set_empty(
+	struct xfs_buf		*bp,
+	size_t			len)
+{
+	if (bp->b_pages)
+		_xfs_buf_free_pages(bp);
+
+	bp->b_pages = NULL;
+	bp->b_page_count = 0;
+	bp->b_addr = NULL;
+	bp->b_file_offset = 0;
+	bp->b_buffer_length = bp->b_count_desired = len;
+	bp->b_bn = XFS_BUF_DADDR_NULL;
+	bp->b_flags &= ~XBF_MAPPED;
+}
+
+static inline struct page *
+mem_to_page(
+	void			*addr)
+{
+	if ((!is_vmalloc_addr(addr))) {
+		return virt_to_page(addr);
+	} else {
+		return vmalloc_to_page(addr);
+	}
+}
+
+int
+xfs_buf_associate_memory(
+	xfs_buf_t		*bp,
+	void			*mem,
+	size_t			len)
+{
+	int			rval;
+	int			i = 0;
+	unsigned long		pageaddr;
+	unsigned long		offset;
+	size_t			buflen;
+	int			page_count;
+
+	pageaddr = (unsigned long)mem & PAGE_MASK;
+	offset = (unsigned long)mem - pageaddr;
+	buflen = PAGE_ALIGN(len + offset);
+	page_count = buflen >> PAGE_SHIFT;
+
+	/* Free any previous set of page pointers */
+	if (bp->b_pages)
+		_xfs_buf_free_pages(bp);
+
+	bp->b_pages = NULL;
+	bp->b_addr = mem;
+
+	rval = _xfs_buf_get_pages(bp, page_count, XBF_DONT_BLOCK);
+	if (rval)
+		return rval;
+
+	bp->b_offset = offset;
+
+	for (i = 0; i < bp->b_page_count; i++) {
+		bp->b_pages[i] = mem_to_page((void *)pageaddr);
+		pageaddr += PAGE_SIZE;
+	}
+
+	bp->b_count_desired = len;
+	bp->b_buffer_length = buflen;
+	bp->b_flags |= XBF_MAPPED;
+
+	return 0;
+}
+
+xfs_buf_t *
+xfs_buf_get_uncached(
+	struct xfs_buftarg	*target,
+	size_t			len,
+	int			flags)
+{
+	unsigned long		page_count = PAGE_ALIGN(len) >> PAGE_SHIFT;
+	int			error, i;
+	xfs_buf_t		*bp;
+
+	bp = xfs_buf_alloc(target, 0, len, 0);
+	if (unlikely(bp == NULL))
+		goto fail;
+
+	error = _xfs_buf_get_pages(bp, page_count, 0);
+	if (error)
+		goto fail_free_buf;
+
+	for (i = 0; i < page_count; i++) {
+		bp->b_pages[i] = alloc_page(xb_to_gfp(flags));
+		if (!bp->b_pages[i])
+			goto fail_free_mem;
+	}
+	bp->b_flags |= _XBF_PAGES;
+
+	error = _xfs_buf_map_pages(bp, XBF_MAPPED);
+	if (unlikely(error)) {
+		xfs_warn(target->bt_mount,
+			"%s: failed to map pages\n", __func__);
+		goto fail_free_mem;
+	}
+
+	trace_xfs_buf_get_uncached(bp, _RET_IP_);
+	return bp;
+
+ fail_free_mem:
+	while (--i >= 0)
+		__free_page(bp->b_pages[i]);
+	_xfs_buf_free_pages(bp);
+ fail_free_buf:
+	kmem_zone_free(xfs_buf_zone, bp);
+ fail:
+	return NULL;
+}
+
+/*
+ *	Increment reference count on buffer, to hold the buffer concurrently
+ *	with another thread which may release (free) the buffer asynchronously.
+ *	Must hold the buffer already to call this function.
+ */
+void
+xfs_buf_hold(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_hold(bp, _RET_IP_);
+	atomic_inc(&bp->b_hold);
+}
+
+/*
+ *	Releases a hold on the specified buffer.  If the
+ *	the hold count is 1, calls xfs_buf_free.
+ */
+void
+xfs_buf_rele(
+	xfs_buf_t		*bp)
+{
+	struct xfs_perag	*pag = bp->b_pag;
+
+	trace_xfs_buf_rele(bp, _RET_IP_);
+
+	if (!pag) {
+		ASSERT(list_empty(&bp->b_lru));
+		ASSERT(RB_EMPTY_NODE(&bp->b_rbnode));
+		if (atomic_dec_and_test(&bp->b_hold))
+			xfs_buf_free(bp);
+		return;
+	}
+
+	ASSERT(!RB_EMPTY_NODE(&bp->b_rbnode));
+
+	ASSERT(atomic_read(&bp->b_hold) > 0);
+	if (atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock)) {
+		if (!(bp->b_flags & XBF_STALE) &&
+			   atomic_read(&bp->b_lru_ref)) {
+			xfs_buf_lru_add(bp);
+			spin_unlock(&pag->pag_buf_lock);
+		} else {
+			xfs_buf_lru_del(bp);
+			ASSERT(!(bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)));
+			rb_erase(&bp->b_rbnode, &pag->pag_buf_tree);
+			spin_unlock(&pag->pag_buf_lock);
+			xfs_perag_put(pag);
+			xfs_buf_free(bp);
+		}
+	}
+}
+
+
+/*
+ *	Lock a buffer object, if it is not already locked.
+ *
+ *	If we come across a stale, pinned, locked buffer, we know that we are
+ *	being asked to lock a buffer that has been reallocated. Because it is
+ *	pinned, we know that the log has not been pushed to disk and hence it
+ *	will still be locked.  Rather than continuing to have trylock attempts
+ *	fail until someone else pushes the log, push it ourselves before
+ *	returning.  This means that the xfsaild will not get stuck trying
+ *	to push on stale inode buffers.
+ */
+int
+xfs_buf_trylock(
+	struct xfs_buf		*bp)
+{
+	int			locked;
+
+	locked = down_trylock(&bp->b_sema) == 0;
+	if (locked)
+		XB_SET_OWNER(bp);
+	else if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
+		xfs_log_force(bp->b_target->bt_mount, 0);
+
+	trace_xfs_buf_trylock(bp, _RET_IP_);
+	return locked;
+}
+
+/*
+ *	Lock a buffer object.
+ *
+ *	If we come across a stale, pinned, locked buffer, we know that we
+ *	are being asked to lock a buffer that has been reallocated. Because
+ *	it is pinned, we know that the log has not been pushed to disk and
+ *	hence it will still be locked. Rather than sleeping until someone
+ *	else pushes the log, push it ourselves before trying to get the lock.
+ */
+void
+xfs_buf_lock(
+	struct xfs_buf		*bp)
+{
+	trace_xfs_buf_lock(bp, _RET_IP_);
+
+	if (atomic_read(&bp->b_pin_count) && (bp->b_flags & XBF_STALE))
+		xfs_log_force(bp->b_target->bt_mount, 0);
+	down(&bp->b_sema);
+	XB_SET_OWNER(bp);
+
+	trace_xfs_buf_lock_done(bp, _RET_IP_);
+}
+
+/*
+ *	Releases the lock on the buffer object.
+ *	If the buffer is marked delwri but is not queued, do so before we
+ *	unlock the buffer as we need to set flags correctly.  We also need to
+ *	take a reference for the delwri queue because the unlocker is going to
+ *	drop their's and they don't know we just queued it.
+ */
+void
+xfs_buf_unlock(
+	struct xfs_buf		*bp)
+{
+	XB_CLEAR_OWNER(bp);
+	up(&bp->b_sema);
+
+	trace_xfs_buf_unlock(bp, _RET_IP_);
+}
+
+STATIC void
+xfs_buf_wait_unpin(
+	xfs_buf_t		*bp)
+{
+	DECLARE_WAITQUEUE	(wait, current);
+
+	if (atomic_read(&bp->b_pin_count) == 0)
+		return;
+
+	add_wait_queue(&bp->b_waiters, &wait);
+	for (;;) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (atomic_read(&bp->b_pin_count) == 0)
+			break;
+		io_schedule();
+	}
+	remove_wait_queue(&bp->b_waiters, &wait);
+	set_current_state(TASK_RUNNING);
+}
+
+/*
+ *	Buffer Utility Routines
+ */
+
+STATIC void
+xfs_buf_iodone_work(
+	struct work_struct	*work)
+{
+	xfs_buf_t		*bp =
+		container_of(work, xfs_buf_t, b_iodone_work);
+
+	if (bp->b_iodone)
+		(*(bp->b_iodone))(bp);
+	else if (bp->b_flags & XBF_ASYNC)
+		xfs_buf_relse(bp);
+}
+
+void
+xfs_buf_ioend(
+	xfs_buf_t		*bp,
+	int			schedule)
+{
+	trace_xfs_buf_iodone(bp, _RET_IP_);
+
+	bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_READ_AHEAD);
+	if (bp->b_error == 0)
+		bp->b_flags |= XBF_DONE;
+
+	if ((bp->b_iodone) || (bp->b_flags & XBF_ASYNC)) {
+		if (schedule) {
+			INIT_WORK(&bp->b_iodone_work, xfs_buf_iodone_work);
+			queue_work(xfslogd_workqueue, &bp->b_iodone_work);
+		} else {
+			xfs_buf_iodone_work(&bp->b_iodone_work);
+		}
+	} else {
+		complete(&bp->b_iowait);
+	}
+}
+
+void
+xfs_buf_ioerror(
+	xfs_buf_t		*bp,
+	int			error)
+{
+	ASSERT(error >= 0 && error <= 0xffff);
+	bp->b_error = (unsigned short)error;
+	trace_xfs_buf_ioerror(bp, error, _RET_IP_);
+}
+
+void
+xfs_buf_ioerror_alert(
+	struct xfs_buf		*bp,
+	const char		*func)
+{
+	xfs_alert(bp->b_target->bt_mount,
+"metadata I/O error: block 0x%llx (\"%s\") error %d buf count %zd",
+		(__uint64_t)XFS_BUF_ADDR(bp), func,
+		bp->b_error, XFS_BUF_COUNT(bp));
+}
+
+int
+xfs_bwrite(
+	struct xfs_buf		*bp)
+{
+	int			error;
+
+	bp->b_flags |= XBF_WRITE;
+	bp->b_flags &= ~(XBF_ASYNC | XBF_READ);
+
+	xfs_buf_delwri_dequeue(bp);
+	xfs_bdstrat_cb(bp);
+
+	error = xfs_buf_iowait(bp);
+	if (error) {
+		xfs_force_shutdown(bp->b_target->bt_mount,
+				   SHUTDOWN_META_IO_ERROR);
+	}
+	return error;
+}
+
+/*
+ * Called when we want to stop a buffer from getting written or read.
+ * We attach the EIO error, muck with its flags, and call xfs_buf_ioend
+ * so that the proper iodone callbacks get called.
+ */
+STATIC int
+xfs_bioerror(
+	xfs_buf_t *bp)
+{
+#ifdef XFSERRORDEBUG
+	ASSERT(XFS_BUF_ISREAD(bp) || bp->b_iodone);
+#endif
+
+	/*
+	 * No need to wait until the buffer is unpinned, we aren't flushing it.
+	 */
+	xfs_buf_ioerror(bp, EIO);
+
+	/*
+	 * We're calling xfs_buf_ioend, so delete XBF_DONE flag.
+	 */
+	XFS_BUF_UNREAD(bp);
+	XFS_BUF_UNDONE(bp);
+	xfs_buf_stale(bp);
+
+	xfs_buf_ioend(bp, 0);
+
+	return EIO;
+}
+
+/*
+ * Same as xfs_bioerror, except that we are releasing the buffer
+ * here ourselves, and avoiding the xfs_buf_ioend call.
+ * This is meant for userdata errors; metadata bufs come with
+ * iodone functions attached, so that we can track down errors.
+ */
+STATIC int
+xfs_bioerror_relse(
+	struct xfs_buf	*bp)
+{
+	int64_t		fl = bp->b_flags;
+	/*
+	 * No need to wait until the buffer is unpinned.
+	 * We aren't flushing it.
+	 *
+	 * chunkhold expects B_DONE to be set, whether
+	 * we actually finish the I/O or not. We don't want to
+	 * change that interface.
+	 */
+	XFS_BUF_UNREAD(bp);
+	XFS_BUF_DONE(bp);
+	xfs_buf_stale(bp);
+	bp->b_iodone = NULL;
+	if (!(fl & XBF_ASYNC)) {
+		/*
+		 * Mark b_error and B_ERROR _both_.
+		 * Lot's of chunkcache code assumes that.
+		 * There's no reason to mark error for
+		 * ASYNC buffers.
+		 */
+		xfs_buf_ioerror(bp, EIO);
+		complete(&bp->b_iowait);
+	} else {
+		xfs_buf_relse(bp);
+	}
+
+	return EIO;
+}
+
+
+/*
+ * All xfs metadata buffers except log state machine buffers
+ * get this attached as their b_bdstrat callback function.
+ * This is so that we can catch a buffer
+ * after prematurely unpinning it to forcibly shutdown the filesystem.
+ */
+int
+xfs_bdstrat_cb(
+	struct xfs_buf	*bp)
+{
+	if (XFS_FORCED_SHUTDOWN(bp->b_target->bt_mount)) {
+		trace_xfs_bdstrat_shut(bp, _RET_IP_);
+		/*
+		 * Metadata write that didn't get logged but
+		 * written delayed anyway. These aren't associated
+		 * with a transaction, and can be ignored.
+		 */
+		if (!bp->b_iodone && !XFS_BUF_ISREAD(bp))
+			return xfs_bioerror_relse(bp);
+		else
+			return xfs_bioerror(bp);
+	}
+
+	xfs_buf_iorequest(bp);
+	return 0;
+}
+
+/*
+ * Wrapper around bdstrat so that we can stop data from going to disk in case
+ * we are shutting down the filesystem.  Typically user data goes thru this
+ * path; one of the exceptions is the superblock.
+ */
+void
+xfsbdstrat(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp)
+{
+	if (XFS_FORCED_SHUTDOWN(mp)) {
+		trace_xfs_bdstrat_shut(bp, _RET_IP_);
+		xfs_bioerror_relse(bp);
+		return;
+	}
+
+	xfs_buf_iorequest(bp);
+}
+
+STATIC void
+_xfs_buf_ioend(
+	xfs_buf_t		*bp,
+	int			schedule)
+{
+	if (atomic_dec_and_test(&bp->b_io_remaining) == 1)
+		xfs_buf_ioend(bp, schedule);
+}
+
+STATIC void
+xfs_buf_bio_end_io(
+	struct bio		*bio,
+	int			error)
+{
+	xfs_buf_t		*bp = (xfs_buf_t *)bio->bi_private;
+
+	/*
+	 * don't overwrite existing errors - otherwise we can lose errors on
+	 * buffers that require multiple bios to complete.
+	 */
+	if (!bp->b_error)
+		xfs_buf_ioerror(bp, -error);
+
+	if (!bp->b_error && xfs_buf_is_vmapped(bp) && (bp->b_flags & XBF_READ))
+		invalidate_kernel_vmap_range(bp->b_addr, xfs_buf_vmap_len(bp));
+
+	_xfs_buf_ioend(bp, 1);
+	bio_put(bio);
+}
+
+STATIC void
+_xfs_buf_ioapply(
+	xfs_buf_t		*bp)
+{
+	int			rw, map_i, total_nr_pages, nr_pages;
+	struct bio		*bio;
+	int			offset = bp->b_offset;
+	int			size = bp->b_count_desired;
+	sector_t		sector = bp->b_bn;
+
+	total_nr_pages = bp->b_page_count;
+	map_i = 0;
+
+	if (bp->b_flags & XBF_WRITE) {
+		if (bp->b_flags & XBF_SYNCIO)
+			rw = WRITE_SYNC;
+		else
+			rw = WRITE;
+		if (bp->b_flags & XBF_FUA)
+			rw |= REQ_FUA;
+		if (bp->b_flags & XBF_FLUSH)
+			rw |= REQ_FLUSH;
+	} else if (bp->b_flags & XBF_READ_AHEAD) {
+		rw = READA;
+	} else {
+		rw = READ;
+	}
+
+	/* we only use the buffer cache for meta-data */
+	rw |= REQ_META;
+
+next_chunk:
+	atomic_inc(&bp->b_io_remaining);
+	nr_pages = BIO_MAX_SECTORS >> (PAGE_SHIFT - BBSHIFT);
+	if (nr_pages > total_nr_pages)
+		nr_pages = total_nr_pages;
+
+	bio = bio_alloc(GFP_NOIO, nr_pages);
+	bio->bi_bdev = bp->b_target->bt_bdev;
+	bio->bi_sector = sector;
+	bio->bi_end_io = xfs_buf_bio_end_io;
+	bio->bi_private = bp;
+
+
+	for (; size && nr_pages; nr_pages--, map_i++) {
+		int	rbytes, nbytes = PAGE_SIZE - offset;
+
+		if (nbytes > size)
+			nbytes = size;
+
+		rbytes = bio_add_page(bio, bp->b_pages[map_i], nbytes, offset);
+		if (rbytes < nbytes)
+			break;
+
+		offset = 0;
+		sector += nbytes >> BBSHIFT;
+		size -= nbytes;
+		total_nr_pages--;
+	}
+
+	if (likely(bio->bi_size)) {
+		if (xfs_buf_is_vmapped(bp)) {
+			flush_kernel_vmap_range(bp->b_addr,
+						xfs_buf_vmap_len(bp));
+		}
+		submit_bio(rw, bio);
+		if (size)
+			goto next_chunk;
+	} else {
+		/*
+		 * This is guaranteed not to be the last io reference count
+		 * because the caller (xfs_buf_iorequest) holds a count itself.
+		 */
+		atomic_dec(&bp->b_io_remaining);
+		xfs_buf_ioerror(bp, EIO);
+		bio_put(bio);
+	}
+}
+
+int
+xfs_buf_iorequest(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_iorequest(bp, _RET_IP_);
+
+	ASSERT(!(bp->b_flags & XBF_DELWRI));
+
+	if (bp->b_flags & XBF_WRITE)
+		xfs_buf_wait_unpin(bp);
+	xfs_buf_hold(bp);
+
+	/* Set the count to 1 initially, this will stop an I/O
+	 * completion callout which happens before we have started
+	 * all the I/O from calling xfs_buf_ioend too early.
+	 */
+	atomic_set(&bp->b_io_remaining, 1);
+	_xfs_buf_ioapply(bp);
+	_xfs_buf_ioend(bp, 0);
+
+	xfs_buf_rele(bp);
+	return 0;
+}
+
+/*
+ *	Waits for I/O to complete on the buffer supplied.
+ *	It returns immediately if no I/O is pending.
+ *	It returns the I/O error code, if any, or 0 if there was no error.
+ */
+int
+xfs_buf_iowait(
+	xfs_buf_t		*bp)
+{
+	trace_xfs_buf_iowait(bp, _RET_IP_);
+
+	wait_for_completion(&bp->b_iowait);
+
+	trace_xfs_buf_iowait_done(bp, _RET_IP_);
+	return bp->b_error;
+}
+
+xfs_caddr_t
+xfs_buf_offset(
+	xfs_buf_t		*bp,
+	size_t			offset)
+{
+	struct page		*page;
+
+	if (bp->b_flags & XBF_MAPPED)
+		return bp->b_addr + offset;
+
+	offset += bp->b_offset;
+	page = bp->b_pages[offset >> PAGE_SHIFT];
+	return (xfs_caddr_t)page_address(page) + (offset & (PAGE_SIZE-1));
+}
+
+/*
+ *	Move data into or out of a buffer.
+ */
+void
+xfs_buf_iomove(
+	xfs_buf_t		*bp,	/* buffer to process		*/
+	size_t			boff,	/* starting buffer offset	*/
+	size_t			bsize,	/* length to copy		*/
+	void			*data,	/* data address			*/
+	xfs_buf_rw_t		mode)	/* read/write/zero flag		*/
+{
+	size_t			bend, cpoff, csize;
+	struct page		*page;
+
+	bend = boff + bsize;
+	while (boff < bend) {
+		page = bp->b_pages[xfs_buf_btoct(boff + bp->b_offset)];
+		cpoff = xfs_buf_poff(boff + bp->b_offset);
+		csize = min_t(size_t,
+			      PAGE_SIZE-cpoff, bp->b_count_desired-boff);
+
+		ASSERT(((csize + cpoff) <= PAGE_SIZE));
+
+		switch (mode) {
+		case XBRW_ZERO:
+			memset(page_address(page) + cpoff, 0, csize);
+			break;
+		case XBRW_READ:
+			memcpy(data, page_address(page) + cpoff, csize);
+			break;
+		case XBRW_WRITE:
+			memcpy(page_address(page) + cpoff, data, csize);
+		}
+
+		boff += csize;
+		data += csize;
+	}
+}
+
+/*
+ *	Handling of buffer targets (buftargs).
+ */
+
+/*
+ * Wait for any bufs with callbacks that have been submitted but have not yet
+ * returned. These buffers will have an elevated hold count, so wait on those
+ * while freeing all the buffers only held by the LRU.
+ */
+void
+xfs_wait_buftarg(
+	struct xfs_buftarg	*btp)
+{
+	struct xfs_buf		*bp;
+
+restart:
+	spin_lock(&btp->bt_lru_lock);
+	while (!list_empty(&btp->bt_lru)) {
+		bp = list_first_entry(&btp->bt_lru, struct xfs_buf, b_lru);
+		if (atomic_read(&bp->b_hold) > 1) {
+			spin_unlock(&btp->bt_lru_lock);
+			delay(100);
+			goto restart;
+		}
+		/*
+		 * clear the LRU reference count so the bufer doesn't get
+		 * ignored in xfs_buf_rele().
+		 */
+		atomic_set(&bp->b_lru_ref, 0);
+		spin_unlock(&btp->bt_lru_lock);
+		xfs_buf_rele(bp);
+		spin_lock(&btp->bt_lru_lock);
+	}
+	spin_unlock(&btp->bt_lru_lock);
+}
+
+int
+xfs_buftarg_shrink(
+	struct shrinker		*shrink,
+	struct shrink_control	*sc)
+{
+	struct xfs_buftarg	*btp = container_of(shrink,
+					struct xfs_buftarg, bt_shrinker);
+	struct xfs_buf		*bp;
+	int nr_to_scan = sc->nr_to_scan;
+	LIST_HEAD(dispose);
+
+	if (!nr_to_scan)
+		return btp->bt_lru_nr;
+
+	spin_lock(&btp->bt_lru_lock);
+	while (!list_empty(&btp->bt_lru)) {
+		if (nr_to_scan-- <= 0)
+			break;
+
+		bp = list_first_entry(&btp->bt_lru, struct xfs_buf, b_lru);
+
+		/*
+		 * Decrement the b_lru_ref count unless the value is already
+		 * zero. If the value is already zero, we need to reclaim the
+		 * buffer, otherwise it gets another trip through the LRU.
+		 */
+		if (!atomic_add_unless(&bp->b_lru_ref, -1, 0)) {
+			list_move_tail(&bp->b_lru, &btp->bt_lru);
+			continue;
+		}
+
+		/*
+		 * remove the buffer from the LRU now to avoid needing another
+		 * lock round trip inside xfs_buf_rele().
+		 */
+		list_move(&bp->b_lru, &dispose);
+		btp->bt_lru_nr--;
+	}
+	spin_unlock(&btp->bt_lru_lock);
+
+	while (!list_empty(&dispose)) {
+		bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
+		list_del_init(&bp->b_lru);
+		xfs_buf_rele(bp);
+	}
+
+	return btp->bt_lru_nr;
+}
+
+void
+xfs_free_buftarg(
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*btp)
+{
+	unregister_shrinker(&btp->bt_shrinker);
+
+	xfs_flush_buftarg(btp, 1);
+	if (mp->m_flags & XFS_MOUNT_BARRIER)
+		xfs_blkdev_issue_flush(btp);
+
+	kthread_stop(btp->bt_task);
+	kmem_free(btp);
+}
+
+STATIC int
+xfs_setsize_buftarg_flags(
+	xfs_buftarg_t		*btp,
+	unsigned int		blocksize,
+	unsigned int		sectorsize,
+	int			verbose)
+{
+	btp->bt_bsize = blocksize;
+	btp->bt_sshift = ffs(sectorsize) - 1;
+	btp->bt_smask = sectorsize - 1;
+
+	if (set_blocksize(btp->bt_bdev, sectorsize)) {
+		char name[BDEVNAME_SIZE];
+
+		bdevname(btp->bt_bdev, name);
+
+		xfs_warn(btp->bt_mount,
+			"Cannot set_blocksize to %u on device %s\n",
+			sectorsize, name);
+		return EINVAL;
+	}
+
+	return 0;
+}
+
+/*
+ *	When allocating the initial buffer target we have not yet
+ *	read in the superblock, so don't know what sized sectors
+ *	are being used is 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);
+}
+
+STATIC int
+xfs_alloc_delwri_queue(
+	xfs_buftarg_t		*btp,
+	const char		*fsname)
+{
+	INIT_LIST_HEAD(&btp->bt_delwri_queue);
+	spin_lock_init(&btp->bt_delwri_lock);
+	btp->bt_flags = 0;
+	btp->bt_task = kthread_run(xfsbufd, btp, "xfsbufd/%s", fsname);
+	if (IS_ERR(btp->bt_task))
+		return PTR_ERR(btp->bt_task);
+	return 0;
+}
+
+xfs_buftarg_t *
+xfs_alloc_buftarg(
+	struct xfs_mount	*mp,
+	struct block_device	*bdev,
+	int			external,
+	const char		*fsname)
+{
+	xfs_buftarg_t		*btp;
+
+	btp = kmem_zalloc(sizeof(*btp), KM_SLEEP);
+
+	btp->bt_mount = mp;
+	btp->bt_dev =  bdev->bd_dev;
+	btp->bt_bdev = bdev;
+	btp->bt_bdi = blk_get_backing_dev_info(bdev);
+	if (!btp->bt_bdi)
+		goto error;
+
+	INIT_LIST_HEAD(&btp->bt_lru);
+	spin_lock_init(&btp->bt_lru_lock);
+	if (xfs_setsize_buftarg_early(btp, bdev))
+		goto error;
+	if (xfs_alloc_delwri_queue(btp, fsname))
+		goto error;
+	btp->bt_shrinker.shrink = xfs_buftarg_shrink;
+	btp->bt_shrinker.seeks = DEFAULT_SEEKS;
+	register_shrinker(&btp->bt_shrinker);
+	return btp;
+
+error:
+	kmem_free(btp);
+	return NULL;
+}
+
+
+/*
+ *	Delayed write buffer handling
+ */
+void
+xfs_buf_delwri_queue(
+	xfs_buf_t		*bp)
+{
+	struct xfs_buftarg	*btp = bp->b_target;
+
+	trace_xfs_buf_delwri_queue(bp, _RET_IP_);
+
+	ASSERT(!(bp->b_flags & XBF_READ));
+
+	spin_lock(&btp->bt_delwri_lock);
+	if (!list_empty(&bp->b_list)) {
+		/* if already in the queue, move it to the tail */
+		ASSERT(bp->b_flags & _XBF_DELWRI_Q);
+		list_move_tail(&bp->b_list, &btp->bt_delwri_queue);
+	} else {
+		/* start xfsbufd as it is about to have something to do */
+		if (list_empty(&btp->bt_delwri_queue))
+			wake_up_process(bp->b_target->bt_task);
+
+		atomic_inc(&bp->b_hold);
+		bp->b_flags |= XBF_DELWRI | _XBF_DELWRI_Q | XBF_ASYNC;
+		list_add_tail(&bp->b_list, &btp->bt_delwri_queue);
+	}
+	bp->b_queuetime = jiffies;
+	spin_unlock(&btp->bt_delwri_lock);
+}
+
+void
+xfs_buf_delwri_dequeue(
+	xfs_buf_t		*bp)
+{
+	int			dequeued = 0;
+
+	spin_lock(&bp->b_target->bt_delwri_lock);
+	if ((bp->b_flags & XBF_DELWRI) && !list_empty(&bp->b_list)) {
+		ASSERT(bp->b_flags & _XBF_DELWRI_Q);
+		list_del_init(&bp->b_list);
+		dequeued = 1;
+	}
+	bp->b_flags &= ~(XBF_DELWRI|_XBF_DELWRI_Q);
+	spin_unlock(&bp->b_target->bt_delwri_lock);
+
+	if (dequeued)
+		xfs_buf_rele(bp);
+
+	trace_xfs_buf_delwri_dequeue(bp, _RET_IP_);
+}
+
+/*
+ * If a delwri buffer needs to be pushed before it has aged out, then promote
+ * it to the head of the delwri queue so that it will be flushed on the next
+ * xfsbufd run. We do this by resetting the queuetime of the buffer to be older
+ * than the age currently needed to flush the buffer. Hence the next time the
+ * xfsbufd sees it is guaranteed to be considered old enough to flush.
+ */
+void
+xfs_buf_delwri_promote(
+	struct xfs_buf	*bp)
+{
+	struct xfs_buftarg *btp = bp->b_target;
+	long		age = xfs_buf_age_centisecs * msecs_to_jiffies(10) + 1;
+
+	ASSERT(bp->b_flags & XBF_DELWRI);
+	ASSERT(bp->b_flags & _XBF_DELWRI_Q);
+
+	/*
+	 * Check the buffer age before locking the delayed write queue as we
+	 * don't need to promote buffers that are already past the flush age.
+	 */
+	if (bp->b_queuetime < jiffies - age)
+		return;
+	bp->b_queuetime = jiffies - age;
+	spin_lock(&btp->bt_delwri_lock);
+	list_move(&bp->b_list, &btp->bt_delwri_queue);
+	spin_unlock(&btp->bt_delwri_lock);
+}
+
+/*
+ * Move as many buffers as specified to the supplied list
+ * idicating if we skipped any buffers to prevent deadlocks.
+ */
+STATIC int
+xfs_buf_delwri_split(
+	xfs_buftarg_t	*target,
+	struct list_head *list,
+	unsigned long	age)
+{
+	xfs_buf_t	*bp, *n;
+	int		skipped = 0;
+	int		force;
+
+	force = test_and_clear_bit(XBT_FORCE_FLUSH, &target->bt_flags);
+	INIT_LIST_HEAD(list);
+	spin_lock(&target->bt_delwri_lock);
+	list_for_each_entry_safe(bp, n, &target->bt_delwri_queue, b_list) {
+		ASSERT(bp->b_flags & XBF_DELWRI);
+
+		if (!xfs_buf_ispinned(bp) && xfs_buf_trylock(bp)) {
+			if (!force &&
+			    time_before(jiffies, bp->b_queuetime + age)) {
+				xfs_buf_unlock(bp);
+				break;
+			}
+
+			bp->b_flags &= ~(XBF_DELWRI | _XBF_DELWRI_Q);
+			bp->b_flags |= XBF_WRITE;
+			list_move_tail(&bp->b_list, list);
+			trace_xfs_buf_delwri_split(bp, _RET_IP_);
+		} else
+			skipped++;
+	}
+
+	spin_unlock(&target->bt_delwri_lock);
+	return skipped;
+}
+
+/*
+ * Compare function is more complex than it needs to be because
+ * the return value is only 32 bits and we are doing comparisons
+ * on 64 bit values
+ */
+static int
+xfs_buf_cmp(
+	void		*priv,
+	struct list_head *a,
+	struct list_head *b)
+{
+	struct xfs_buf	*ap = container_of(a, struct xfs_buf, b_list);
+	struct xfs_buf	*bp = container_of(b, struct xfs_buf, b_list);
+	xfs_daddr_t		diff;
+
+	diff = ap->b_bn - bp->b_bn;
+	if (diff < 0)
+		return -1;
+	if (diff > 0)
+		return 1;
+	return 0;
+}
+
+STATIC int
+xfsbufd(
+	void		*data)
+{
+	xfs_buftarg_t   *target = (xfs_buftarg_t *)data;
+
+	current->flags |= PF_MEMALLOC;
+
+	set_freezable();
+
+	do {
+		long	age = xfs_buf_age_centisecs * msecs_to_jiffies(10);
+		long	tout = xfs_buf_timer_centisecs * msecs_to_jiffies(10);
+		struct list_head tmp;
+		struct blk_plug plug;
+
+		if (unlikely(freezing(current))) {
+			set_bit(XBT_FORCE_SLEEP, &target->bt_flags);
+			refrigerator();
+		} else {
+			clear_bit(XBT_FORCE_SLEEP, &target->bt_flags);
+		}
+
+		/* sleep for a long time if there is nothing to do. */
+		if (list_empty(&target->bt_delwri_queue))
+			tout = MAX_SCHEDULE_TIMEOUT;
+		schedule_timeout_interruptible(tout);
+
+		xfs_buf_delwri_split(target, &tmp, age);
+		list_sort(NULL, &tmp, xfs_buf_cmp);
+
+		blk_start_plug(&plug);
+		while (!list_empty(&tmp)) {
+			struct xfs_buf *bp;
+			bp = list_first_entry(&tmp, struct xfs_buf, b_list);
+			list_del_init(&bp->b_list);
+			xfs_bdstrat_cb(bp);
+		}
+		blk_finish_plug(&plug);
+	} while (!kthread_should_stop());
+
+	return 0;
+}
+
+/*
+ *	Go through all incore buffers, and release buffers if they belong to
+ *	the given device. This is used in filesystem error handling to
+ *	preserve the consistency of its metadata.
+ */
+int
+xfs_flush_buftarg(
+	xfs_buftarg_t	*target,
+	int		wait)
+{
+	xfs_buf_t	*bp;
+	int		pincount = 0;
+	LIST_HEAD(tmp_list);
+	LIST_HEAD(wait_list);
+	struct blk_plug plug;
+
+	flush_workqueue(xfslogd_workqueue);
+
+	set_bit(XBT_FORCE_FLUSH, &target->bt_flags);
+	pincount = xfs_buf_delwri_split(target, &tmp_list, 0);
+
+	/*
+	 * Dropped the delayed write list lock, now walk the temporary list.
+	 * All I/O is issued async and then if we need to wait for completion
+	 * we do that after issuing all the IO.
+	 */
+	list_sort(NULL, &tmp_list, xfs_buf_cmp);
+
+	blk_start_plug(&plug);
+	while (!list_empty(&tmp_list)) {
+		bp = list_first_entry(&tmp_list, struct xfs_buf, b_list);
+		ASSERT(target == bp->b_target);
+		list_del_init(&bp->b_list);
+		if (wait) {
+			bp->b_flags &= ~XBF_ASYNC;
+			list_add(&bp->b_list, &wait_list);
+		}
+		xfs_bdstrat_cb(bp);
+	}
+	blk_finish_plug(&plug);
+
+	if (wait) {
+		/* Wait for IO to complete. */
+		while (!list_empty(&wait_list)) {
+			bp = list_first_entry(&wait_list, struct xfs_buf, b_list);
+
+			list_del_init(&bp->b_list);
+			xfs_buf_iowait(bp);
+			xfs_buf_relse(bp);
+		}
+	}
+
+	return pincount;
+}
+
+int __init
+xfs_buf_init(void)
+{
+	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",
+						KM_ZONE_HWALIGN, NULL);
+	if (!xfs_buf_zone)
+		goto out;
+
+	xfslogd_workqueue = alloc_workqueue("xfslogd",
+					WQ_MEM_RECLAIM | WQ_HIGHPRI, 1);
+	if (!xfslogd_workqueue)
+		goto out_free_buf_zone;
+
+	xfsdatad_workqueue = alloc_workqueue("xfsdatad", WQ_MEM_RECLAIM, 1);
+	if (!xfsdatad_workqueue)
+		goto out_destroy_xfslogd_workqueue;
+
+	xfsconvertd_workqueue = alloc_workqueue("xfsconvertd",
+						WQ_MEM_RECLAIM, 1);
+	if (!xfsconvertd_workqueue)
+		goto out_destroy_xfsdatad_workqueue;
+
+	return 0;
+
+ out_destroy_xfsdatad_workqueue:
+	destroy_workqueue(xfsdatad_workqueue);
+ out_destroy_xfslogd_workqueue:
+	destroy_workqueue(xfslogd_workqueue);
+ out_free_buf_zone:
+	kmem_zone_destroy(xfs_buf_zone);
+ out:
+	return -ENOMEM;
+}
+
+void
+xfs_buf_terminate(void)
+{
+	destroy_workqueue(xfsconvertd_workqueue);
+	destroy_workqueue(xfsdatad_workqueue);
+	destroy_workqueue(xfslogd_workqueue);
+	kmem_zone_destroy(xfs_buf_zone);
+}
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
new file mode 100644
index 0000000..5bab046
--- /dev/null
+++ b/fs/xfs/xfs_buf.h
@@ -0,0 +1,297 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_BUF_H__
+#define __XFS_BUF_H__
+
+#include <linux/list.h>
+#include <linux/types.h>
+#include <linux/spinlock.h>
+#include <asm/system.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+#include <linux/buffer_head.h>
+#include <linux/uio.h>
+
+/*
+ *	Base types
+ */
+
+#define XFS_BUF_DADDR_NULL	((xfs_daddr_t) (-1LL))
+
+#define xfs_buf_ctob(pp)	((pp) * PAGE_CACHE_SIZE)
+#define xfs_buf_btoc(dd)	(((dd) + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT)
+#define xfs_buf_btoct(dd)	((dd) >> PAGE_CACHE_SHIFT)
+#define xfs_buf_poff(aa)	((aa) & ~PAGE_CACHE_MASK)
+
+typedef enum {
+	XBRW_READ = 1,			/* transfer into target memory */
+	XBRW_WRITE = 2,			/* transfer from target memory */
+	XBRW_ZERO = 3,			/* Zero target memory */
+} xfs_buf_rw_t;
+
+#define XBF_READ	(1 << 0) /* buffer intended for reading from device */
+#define XBF_WRITE	(1 << 1) /* buffer intended for writing to device */
+#define XBF_READ_AHEAD	(1 << 2) /* asynchronous read-ahead */
+#define XBF_MAPPED	(1 << 3) /* buffer mapped (b_addr valid) */
+#define XBF_ASYNC	(1 << 4) /* initiator will not wait for completion */
+#define XBF_DONE	(1 << 5) /* all pages in the buffer uptodate */
+#define XBF_DELWRI	(1 << 6) /* buffer has dirty pages */
+#define XBF_STALE	(1 << 7) /* buffer has been staled, do not find it */
+
+/* I/O hints for the BIO layer */
+#define XBF_SYNCIO	(1 << 10)/* treat this buffer as synchronous I/O */
+#define XBF_FUA		(1 << 11)/* force cache write through mode */
+#define XBF_FLUSH	(1 << 12)/* flush the disk cache before a write */
+
+/* flags used only as arguments to access routines */
+#define XBF_LOCK	(1 << 15)/* lock requested */
+#define XBF_TRYLOCK	(1 << 16)/* lock requested, but do not wait */
+#define XBF_DONT_BLOCK	(1 << 17)/* do not block in current thread */
+
+/* flags used only internally */
+#define _XBF_PAGES	(1 << 20)/* backed by refcounted pages */
+#define _XBF_KMEM	(1 << 21)/* backed by heap memory */
+#define _XBF_DELWRI_Q	(1 << 22)/* buffer on delwri queue */
+
+typedef unsigned int xfs_buf_flags_t;
+
+#define XFS_BUF_FLAGS \
+	{ XBF_READ,		"READ" }, \
+	{ XBF_WRITE,		"WRITE" }, \
+	{ XBF_READ_AHEAD,	"READ_AHEAD" }, \
+	{ XBF_MAPPED,		"MAPPED" }, \
+	{ XBF_ASYNC,		"ASYNC" }, \
+	{ XBF_DONE,		"DONE" }, \
+	{ XBF_DELWRI,		"DELWRI" }, \
+	{ XBF_STALE,		"STALE" }, \
+	{ XBF_SYNCIO,		"SYNCIO" }, \
+	{ XBF_FUA,		"FUA" }, \
+	{ XBF_FLUSH,		"FLUSH" }, \
+	{ XBF_LOCK,		"LOCK" },  	/* should never be set */\
+	{ XBF_TRYLOCK,		"TRYLOCK" }, 	/* ditto */\
+	{ XBF_DONT_BLOCK,	"DONT_BLOCK" },	/* ditto */\
+	{ _XBF_PAGES,		"PAGES" }, \
+	{ _XBF_KMEM,		"KMEM" }, \
+	{ _XBF_DELWRI_Q,	"DELWRI_Q" }
+
+typedef enum {
+	XBT_FORCE_SLEEP = 0,
+	XBT_FORCE_FLUSH = 1,
+} xfs_buftarg_flags_t;
+
+typedef struct xfs_buftarg {
+	dev_t			bt_dev;
+	struct block_device	*bt_bdev;
+	struct backing_dev_info	*bt_bdi;
+	struct xfs_mount	*bt_mount;
+	unsigned int		bt_bsize;
+	unsigned int		bt_sshift;
+	size_t			bt_smask;
+
+	/* per device delwri queue */
+	struct task_struct	*bt_task;
+	struct list_head	bt_delwri_queue;
+	spinlock_t		bt_delwri_lock;
+	unsigned long		bt_flags;
+
+	/* LRU control structures */
+	struct shrinker		bt_shrinker;
+	struct list_head	bt_lru;
+	spinlock_t		bt_lru_lock;
+	unsigned int		bt_lru_nr;
+} xfs_buftarg_t;
+
+struct xfs_buf;
+typedef void (*xfs_buf_iodone_t)(struct xfs_buf *);
+
+#define XB_PAGES	2
+
+typedef struct xfs_buf {
+	/*
+	 * first cacheline holds all the fields needed for an uncontended cache
+	 * hit to be fully processed. The semaphore straddles the cacheline
+	 * boundary, but the counter and lock sits on the first cacheline,
+	 * which is the only bit that is touched if we hit the semaphore
+	 * fast-path on locking.
+	 */
+	struct rb_node		b_rbnode;	/* rbtree node */
+	xfs_off_t		b_file_offset;	/* offset in file */
+	size_t			b_buffer_length;/* size of buffer in bytes */
+	atomic_t		b_hold;		/* reference count */
+	atomic_t		b_lru_ref;	/* lru reclaim ref count */
+	xfs_buf_flags_t		b_flags;	/* status flags */
+	struct semaphore	b_sema;		/* semaphore for lockables */
+
+	struct list_head	b_lru;		/* lru list */
+	wait_queue_head_t	b_waiters;	/* unpin waiters */
+	struct list_head	b_list;
+	struct xfs_perag	*b_pag;		/* contains rbtree root */
+	xfs_buftarg_t		*b_target;	/* buffer target (device) */
+	xfs_daddr_t		b_bn;		/* block number for I/O */
+	size_t			b_count_desired;/* desired transfer size */
+	void			*b_addr;	/* virtual address of buffer */
+	struct work_struct	b_iodone_work;
+	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
+	struct completion	b_iowait;	/* queue for I/O waiters */
+	void			*b_fspriv;
+	struct xfs_trans	*b_transp;
+	struct page		**b_pages;	/* array of page pointers */
+	struct page		*b_page_array[XB_PAGES]; /* inline pages */
+	unsigned long		b_queuetime;	/* time buffer was queued */
+	atomic_t		b_pin_count;	/* pin count */
+	atomic_t		b_io_remaining;	/* #outstanding I/O requests */
+	unsigned int		b_page_count;	/* size of page array */
+	unsigned int		b_offset;	/* page offset in first page */
+	unsigned short		b_error;	/* error code on I/O */
+#ifdef XFS_BUF_LOCK_TRACKING
+	int			b_last_holder;
+#endif
+} xfs_buf_t;
+
+
+/* Finding and Reading Buffers */
+extern xfs_buf_t *_xfs_buf_find(xfs_buftarg_t *, xfs_off_t, size_t,
+				xfs_buf_flags_t, xfs_buf_t *);
+#define xfs_incore(buftarg,blkno,len,lockit) \
+	_xfs_buf_find(buftarg, blkno ,len, lockit, NULL)
+
+extern xfs_buf_t *xfs_buf_get(xfs_buftarg_t *, xfs_off_t, size_t,
+				xfs_buf_flags_t);
+extern xfs_buf_t *xfs_buf_read(xfs_buftarg_t *, xfs_off_t, size_t,
+				xfs_buf_flags_t);
+
+struct xfs_buf *xfs_buf_alloc(struct xfs_buftarg *, xfs_off_t, size_t,
+			      xfs_buf_flags_t);
+extern void xfs_buf_set_empty(struct xfs_buf *bp, size_t len);
+extern xfs_buf_t *xfs_buf_get_uncached(struct xfs_buftarg *, size_t, int);
+extern int xfs_buf_associate_memory(xfs_buf_t *, void *, size_t);
+extern void xfs_buf_hold(xfs_buf_t *);
+extern void xfs_buf_readahead(xfs_buftarg_t *, xfs_off_t, size_t);
+struct xfs_buf *xfs_buf_read_uncached(struct xfs_mount *mp,
+				struct xfs_buftarg *target,
+				xfs_daddr_t daddr, size_t length, int flags);
+
+/* Releasing Buffers */
+extern void xfs_buf_free(xfs_buf_t *);
+extern void xfs_buf_rele(xfs_buf_t *);
+
+/* Locking and Unlocking Buffers */
+extern int xfs_buf_trylock(xfs_buf_t *);
+extern void xfs_buf_lock(xfs_buf_t *);
+extern void xfs_buf_unlock(xfs_buf_t *);
+#define xfs_buf_islocked(bp) \
+	((bp)->b_sema.count <= 0)
+
+/* Buffer Read and Write Routines */
+extern int xfs_bwrite(struct xfs_buf *bp);
+
+extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *);
+extern int xfs_bdstrat_cb(struct xfs_buf *);
+
+extern void xfs_buf_ioend(xfs_buf_t *,	int);
+extern void xfs_buf_ioerror(xfs_buf_t *, int);
+extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func);
+extern int xfs_buf_iorequest(xfs_buf_t *);
+extern int xfs_buf_iowait(xfs_buf_t *);
+extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *,
+				xfs_buf_rw_t);
+#define xfs_buf_zero(bp, off, len) \
+	    xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
+
+static inline int xfs_buf_geterror(xfs_buf_t *bp)
+{
+	return bp ? bp->b_error : ENOMEM;
+}
+
+/* Buffer Utility Routines */
+extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t);
+
+/* Delayed Write Buffer Routines */
+extern void xfs_buf_delwri_queue(struct xfs_buf *);
+extern void xfs_buf_delwri_dequeue(struct xfs_buf *);
+extern void xfs_buf_delwri_promote(struct xfs_buf *);
+
+/* Buffer Daemon Setup Routines */
+extern int xfs_buf_init(void);
+extern void xfs_buf_terminate(void);
+
+#define XFS_BUF_ZEROFLAGS(bp) \
+	((bp)->b_flags &= ~(XBF_READ|XBF_WRITE|XBF_ASYNC|XBF_DELWRI| \
+			    XBF_SYNCIO|XBF_FUA|XBF_FLUSH))
+
+void xfs_buf_stale(struct xfs_buf *bp);
+#define XFS_BUF_UNSTALE(bp)	((bp)->b_flags &= ~XBF_STALE)
+#define XFS_BUF_ISSTALE(bp)	((bp)->b_flags & XBF_STALE)
+
+#define XFS_BUF_ISDELAYWRITE(bp)	((bp)->b_flags & XBF_DELWRI)
+
+#define XFS_BUF_DONE(bp)	((bp)->b_flags |= XBF_DONE)
+#define XFS_BUF_UNDONE(bp)	((bp)->b_flags &= ~XBF_DONE)
+#define XFS_BUF_ISDONE(bp)	((bp)->b_flags & XBF_DONE)
+
+#define XFS_BUF_ASYNC(bp)	((bp)->b_flags |= XBF_ASYNC)
+#define XFS_BUF_UNASYNC(bp)	((bp)->b_flags &= ~XBF_ASYNC)
+#define XFS_BUF_ISASYNC(bp)	((bp)->b_flags & XBF_ASYNC)
+
+#define XFS_BUF_READ(bp)	((bp)->b_flags |= XBF_READ)
+#define XFS_BUF_UNREAD(bp)	((bp)->b_flags &= ~XBF_READ)
+#define XFS_BUF_ISREAD(bp)	((bp)->b_flags & XBF_READ)
+
+#define XFS_BUF_WRITE(bp)	((bp)->b_flags |= XBF_WRITE)
+#define XFS_BUF_UNWRITE(bp)	((bp)->b_flags &= ~XBF_WRITE)
+#define XFS_BUF_ISWRITE(bp)	((bp)->b_flags & XBF_WRITE)
+
+#define XFS_BUF_ADDR(bp)		((bp)->b_bn)
+#define XFS_BUF_SET_ADDR(bp, bno)	((bp)->b_bn = (xfs_daddr_t)(bno))
+#define XFS_BUF_OFFSET(bp)		((bp)->b_file_offset)
+#define XFS_BUF_SET_OFFSET(bp, off)	((bp)->b_file_offset = (off))
+#define XFS_BUF_COUNT(bp)		((bp)->b_count_desired)
+#define XFS_BUF_SET_COUNT(bp, cnt)	((bp)->b_count_desired = (cnt))
+#define XFS_BUF_SIZE(bp)		((bp)->b_buffer_length)
+#define XFS_BUF_SET_SIZE(bp, cnt)	((bp)->b_buffer_length = (cnt))
+
+static inline void xfs_buf_set_ref(struct xfs_buf *bp, int lru_ref)
+{
+	atomic_set(&bp->b_lru_ref, lru_ref);
+}
+
+static inline int xfs_buf_ispinned(struct xfs_buf *bp)
+{
+	return atomic_read(&bp->b_pin_count);
+}
+
+static inline void xfs_buf_relse(xfs_buf_t *bp)
+{
+	xfs_buf_unlock(bp);
+	xfs_buf_rele(bp);
+}
+
+/*
+ *	Handling of buftargs.
+ */
+extern xfs_buftarg_t *xfs_alloc_buftarg(struct xfs_mount *,
+			struct block_device *, int, const char *);
+extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *);
+extern void xfs_wait_buftarg(xfs_buftarg_t *);
+extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int);
+extern int xfs_flush_buftarg(xfs_buftarg_t *, int);
+
+#define xfs_getsize_buftarg(buftarg)	block_size((buftarg)->bt_bdev)
+#define xfs_readonly_buftarg(buftarg)	bdev_read_only((buftarg)->bt_bdev)
+
+#endif	/* __XFS_BUF_H__ */
diff --git a/fs/xfs/xfs_dir2_format.h b/fs/xfs/xfs_dir2_format.h
new file mode 100644
index 0000000..0727098
--- /dev/null
+++ b/fs/xfs/xfs_dir2_format.h
@@ -0,0 +1,597 @@
+/*
+ * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_DIR2_FORMAT_H__
+#define __XFS_DIR2_FORMAT_H__
+
+/*
+ * Directory version 2.
+ *
+ * There are 4 possible formats:
+ *  - shortform - embedded into the inode
+ *  - single block - data with embedded leaf at the end
+ *  - multiple data blocks, single leaf+freeindex block
+ *  - data blocks, node and leaf blocks (btree), freeindex blocks
+ *
+ * Note: many node blocks structures and constants are shared with the attr
+ * code and defined in xfs_da_btree.h.
+ */
+
+#define	XFS_DIR2_BLOCK_MAGIC	0x58443242	/* XD2B: single block dirs */
+#define	XFS_DIR2_DATA_MAGIC	0x58443244	/* XD2D: multiblock dirs */
+#define	XFS_DIR2_FREE_MAGIC	0x58443246	/* XD2F: free index blocks */
+
+/*
+ * Byte offset in data block and shortform entry.
+ */
+typedef	__uint16_t	xfs_dir2_data_off_t;
+#define	NULLDATAOFF	0xffffU
+typedef uint		xfs_dir2_data_aoff_t;	/* argument form */
+
+/*
+ * Normalized offset (in a data block) of the entry, really xfs_dir2_data_off_t.
+ * Only need 16 bits, this is the byte offset into the single block form.
+ */
+typedef struct { __uint8_t i[2]; } __arch_pack xfs_dir2_sf_off_t;
+
+/*
+ * Offset in data space of a data entry.
+ */
+typedef	__uint32_t	xfs_dir2_dataptr_t;
+#define	XFS_DIR2_MAX_DATAPTR	((xfs_dir2_dataptr_t)0xffffffff)
+#define	XFS_DIR2_NULL_DATAPTR	((xfs_dir2_dataptr_t)0)
+
+/*
+ * Byte offset in a directory.
+ */
+typedef	xfs_off_t	xfs_dir2_off_t;
+
+/*
+ * Directory block number (logical dirblk in file)
+ */
+typedef	__uint32_t	xfs_dir2_db_t;
+
+/*
+ * Inode number stored as 8 8-bit values.
+ */
+typedef	struct { __uint8_t i[8]; } xfs_dir2_ino8_t;
+
+/*
+ * Inode number stored as 4 8-bit values.
+ * Works a lot of the time, when all the inode numbers in a directory
+ * fit in 32 bits.
+ */
+typedef struct { __uint8_t i[4]; } xfs_dir2_ino4_t;
+
+typedef union {
+	xfs_dir2_ino8_t	i8;
+	xfs_dir2_ino4_t	i4;
+} xfs_dir2_inou_t;
+#define	XFS_DIR2_MAX_SHORT_INUM	((xfs_ino_t)0xffffffffULL)
+
+/*
+ * Directory layout when stored internal to an inode.
+ *
+ * Small directories are packed as tightly as possible so as to fit into the
+ * literal area of the inode.  These "shortform" directories consist of a
+ * single xfs_dir2_sf_hdr header followed by zero or more xfs_dir2_sf_entry
+ * structures.  Due the different inode number storage size and the variable
+ * length name field in the xfs_dir2_sf_entry all these structure are
+ * variable length, and the accessors in this file should be used to iterate
+ * over them.
+ */
+typedef struct xfs_dir2_sf_hdr {
+	__uint8_t		count;		/* count of entries */
+	__uint8_t		i8count;	/* count of 8-byte inode #s */
+	xfs_dir2_inou_t		parent;		/* parent dir inode number */
+} __arch_pack xfs_dir2_sf_hdr_t;
+
+typedef struct xfs_dir2_sf_entry {
+	__u8			namelen;	/* actual name length */
+	xfs_dir2_sf_off_t	offset;		/* saved offset */
+	__u8			name[];		/* name, variable size */
+	/*
+	 * A xfs_dir2_ino8_t or xfs_dir2_ino4_t follows here, at a
+	 * variable offset after the name.
+	 */
+} __arch_pack xfs_dir2_sf_entry_t;
+
+static inline int xfs_dir2_sf_hdr_size(int i8count)
+{
+	return sizeof(struct xfs_dir2_sf_hdr) -
+		(i8count == 0) *
+		(sizeof(xfs_dir2_ino8_t) - sizeof(xfs_dir2_ino4_t));
+}
+
+static inline xfs_dir2_data_aoff_t
+xfs_dir2_sf_get_offset(xfs_dir2_sf_entry_t *sfep)
+{
+	return get_unaligned_be16(&sfep->offset.i);
+}
+
+static inline void
+xfs_dir2_sf_put_offset(xfs_dir2_sf_entry_t *sfep, xfs_dir2_data_aoff_t off)
+{
+	put_unaligned_be16(off, &sfep->offset.i);
+}
+
+static inline int
+xfs_dir2_sf_entsize(struct xfs_dir2_sf_hdr *hdr, int len)
+{
+	return sizeof(struct xfs_dir2_sf_entry) +	/* namelen + offset */
+		len +					/* name */
+		(hdr->i8count ?				/* ino */
+		 sizeof(xfs_dir2_ino8_t) :
+		 sizeof(xfs_dir2_ino4_t));
+}
+
+static inline struct xfs_dir2_sf_entry *
+xfs_dir2_sf_firstentry(struct xfs_dir2_sf_hdr *hdr)
+{
+	return (struct xfs_dir2_sf_entry *)
+		((char *)hdr + xfs_dir2_sf_hdr_size(hdr->i8count));
+}
+
+static inline struct xfs_dir2_sf_entry *
+xfs_dir2_sf_nextentry(struct xfs_dir2_sf_hdr *hdr,
+		struct xfs_dir2_sf_entry *sfep)
+{
+	return (struct xfs_dir2_sf_entry *)
+		((char *)sfep + xfs_dir2_sf_entsize(hdr, sfep->namelen));
+}
+
+
+/*
+ * Data block structures.
+ *
+ * A pure data block looks like the following drawing on disk:
+ *
+ *    +-------------------------------------------------+
+ *    | xfs_dir2_data_hdr_t                             |
+ *    +-------------------------------------------------+
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t |
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t |
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t |
+ *    | ...                                             |
+ *    +-------------------------------------------------+
+ *    | unused space                                    |
+ *    +-------------------------------------------------+
+ *
+ * As all the entries are variable size structures the accessors below should
+ * be used to iterate over them.
+ *
+ * In addition to the pure data blocks for the data and node formats,
+ * most structures are also used for the combined data/freespace "block"
+ * format below.
+ */
+
+#define	XFS_DIR2_DATA_ALIGN_LOG	3		/* i.e., 8 bytes */
+#define	XFS_DIR2_DATA_ALIGN	(1 << XFS_DIR2_DATA_ALIGN_LOG)
+#define	XFS_DIR2_DATA_FREE_TAG	0xffff
+#define	XFS_DIR2_DATA_FD_COUNT	3
+
+/*
+ * Directory address space divided into sections,
+ * spaces separated by 32GB.
+ */
+#define	XFS_DIR2_SPACE_SIZE	(1ULL << (32 + XFS_DIR2_DATA_ALIGN_LOG))
+#define	XFS_DIR2_DATA_SPACE	0
+#define	XFS_DIR2_DATA_OFFSET	(XFS_DIR2_DATA_SPACE * XFS_DIR2_SPACE_SIZE)
+#define	XFS_DIR2_DATA_FIRSTDB(mp)	\
+	xfs_dir2_byte_to_db(mp, XFS_DIR2_DATA_OFFSET)
+
+/*
+ * Offsets of . and .. in data space (always block 0)
+ */
+#define	XFS_DIR2_DATA_DOT_OFFSET	\
+	((xfs_dir2_data_aoff_t)sizeof(struct xfs_dir2_data_hdr))
+#define	XFS_DIR2_DATA_DOTDOT_OFFSET	\
+	(XFS_DIR2_DATA_DOT_OFFSET + xfs_dir2_data_entsize(1))
+#define	XFS_DIR2_DATA_FIRST_OFFSET		\
+	(XFS_DIR2_DATA_DOTDOT_OFFSET + xfs_dir2_data_entsize(2))
+
+/*
+ * Describe a free area in the data block.
+ *
+ * The freespace will be formatted as a xfs_dir2_data_unused_t.
+ */
+typedef struct xfs_dir2_data_free {
+	__be16			offset;		/* start of freespace */
+	__be16			length;		/* length of freespace */
+} xfs_dir2_data_free_t;
+
+/*
+ * Header for the data blocks.
+ *
+ * The code knows that XFS_DIR2_DATA_FD_COUNT is 3.
+ */
+typedef struct xfs_dir2_data_hdr {
+	__be32			magic;		/* XFS_DIR2_DATA_MAGIC or */
+						/* XFS_DIR2_BLOCK_MAGIC */
+	xfs_dir2_data_free_t	bestfree[XFS_DIR2_DATA_FD_COUNT];
+} xfs_dir2_data_hdr_t;
+
+/*
+ * Active entry in a data block.
+ *
+ * Aligned to 8 bytes.  After the variable length name field there is a
+ * 2 byte tag field, which can be accessed using xfs_dir2_data_entry_tag_p.
+ */
+typedef struct xfs_dir2_data_entry {
+	__be64			inumber;	/* inode number */
+	__u8			namelen;	/* name length */
+	__u8			name[];		/* name bytes, no null */
+     /*	__be16                  tag; */		/* starting offset of us */
+} xfs_dir2_data_entry_t;
+
+/*
+ * Unused entry in a data block.
+ *
+ * Aligned to 8 bytes.  Tag appears as the last 2 bytes and must be accessed
+ * using xfs_dir2_data_unused_tag_p.
+ */
+typedef struct xfs_dir2_data_unused {
+	__be16			freetag;	/* XFS_DIR2_DATA_FREE_TAG */
+	__be16			length;		/* total free length */
+						/* variable offset */
+	__be16			tag;		/* starting offset of us */
+} xfs_dir2_data_unused_t;
+
+/*
+ * Size of a data entry.
+ */
+static inline int xfs_dir2_data_entsize(int n)
+{
+	return (int)roundup(offsetof(struct xfs_dir2_data_entry, name[0]) + n +
+		 (uint)sizeof(xfs_dir2_data_off_t), XFS_DIR2_DATA_ALIGN);
+}
+
+/*
+ * Pointer to an entry's tag word.
+ */
+static inline __be16 *
+xfs_dir2_data_entry_tag_p(struct xfs_dir2_data_entry *dep)
+{
+	return (__be16 *)((char *)dep +
+		xfs_dir2_data_entsize(dep->namelen) - sizeof(__be16));
+}
+
+/*
+ * Pointer to a freespace's tag word.
+ */
+static inline __be16 *
+xfs_dir2_data_unused_tag_p(struct xfs_dir2_data_unused *dup)
+{
+	return (__be16 *)((char *)dup +
+			be16_to_cpu(dup->length) - sizeof(__be16));
+}
+
+/*
+ * Leaf block structures.
+ *
+ * A pure leaf block looks like the following drawing on disk:
+ *
+ *    +---------------------------+
+ *    | xfs_dir2_leaf_hdr_t       |
+ *    +---------------------------+
+ *    | xfs_dir2_leaf_entry_t     |
+ *    | xfs_dir2_leaf_entry_t     |
+ *    | xfs_dir2_leaf_entry_t     |
+ *    | xfs_dir2_leaf_entry_t     |
+ *    | ...                       |
+ *    +---------------------------+
+ *    | xfs_dir2_data_off_t       |
+ *    | xfs_dir2_data_off_t       |
+ *    | xfs_dir2_data_off_t       |
+ *    | ...                       |
+ *    +---------------------------+
+ *    | xfs_dir2_leaf_tail_t      |
+ *    +---------------------------+
+ *
+ * The xfs_dir2_data_off_t members (bests) and tail are at the end of the block
+ * for single-leaf (magic = XFS_DIR2_LEAF1_MAGIC) blocks only, but not present
+ * for directories with separate leaf nodes and free space blocks
+ * (magic = XFS_DIR2_LEAFN_MAGIC).
+ *
+ * As all the entries are variable size structures the accessors below should
+ * be used to iterate over them.
+ */
+
+/*
+ * Offset of the leaf/node space.  First block in this space
+ * is the btree root.
+ */
+#define	XFS_DIR2_LEAF_SPACE	1
+#define	XFS_DIR2_LEAF_OFFSET	(XFS_DIR2_LEAF_SPACE * XFS_DIR2_SPACE_SIZE)
+#define	XFS_DIR2_LEAF_FIRSTDB(mp)	\
+	xfs_dir2_byte_to_db(mp, XFS_DIR2_LEAF_OFFSET)
+
+/*
+ * Leaf block header.
+ */
+typedef struct xfs_dir2_leaf_hdr {
+	xfs_da_blkinfo_t	info;		/* header for da routines */
+	__be16			count;		/* count of entries */
+	__be16			stale;		/* count of stale entries */
+} xfs_dir2_leaf_hdr_t;
+
+/*
+ * Leaf block entry.
+ */
+typedef struct xfs_dir2_leaf_entry {
+	__be32			hashval;	/* hash value of name */
+	__be32			address;	/* address of data entry */
+} xfs_dir2_leaf_entry_t;
+
+/*
+ * Leaf block tail.
+ */
+typedef struct xfs_dir2_leaf_tail {
+	__be32			bestcount;
+} xfs_dir2_leaf_tail_t;
+
+/*
+ * Leaf block.
+ */
+typedef struct xfs_dir2_leaf {
+	xfs_dir2_leaf_hdr_t	hdr;		/* leaf header */
+	xfs_dir2_leaf_entry_t	ents[];		/* entries */
+} xfs_dir2_leaf_t;
+
+/*
+ * DB blocks here are logical directory block numbers, not filesystem blocks.
+ */
+
+static inline int xfs_dir2_max_leaf_ents(struct xfs_mount *mp)
+{
+	return (mp->m_dirblksize - (uint)sizeof(struct xfs_dir2_leaf_hdr)) /
+		(uint)sizeof(struct xfs_dir2_leaf_entry);
+}
+
+/*
+ * Get address of the bestcount field in the single-leaf block.
+ */
+static inline struct xfs_dir2_leaf_tail *
+xfs_dir2_leaf_tail_p(struct xfs_mount *mp, struct xfs_dir2_leaf *lp)
+{
+	return (struct xfs_dir2_leaf_tail *)
+		((char *)lp + mp->m_dirblksize -
+		  sizeof(struct xfs_dir2_leaf_tail));
+}
+
+/*
+ * Get address of the bests array in the single-leaf block.
+ */
+static inline __be16 *
+xfs_dir2_leaf_bests_p(struct xfs_dir2_leaf_tail *ltp)
+{
+	return (__be16 *)ltp - be32_to_cpu(ltp->bestcount);
+}
+
+/*
+ * Convert dataptr to byte in file space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_dataptr_to_byte(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
+{
+	return (xfs_dir2_off_t)dp << XFS_DIR2_DATA_ALIGN_LOG;
+}
+
+/*
+ * Convert byte in file space to dataptr.  It had better be aligned.
+ */
+static inline xfs_dir2_dataptr_t
+xfs_dir2_byte_to_dataptr(struct xfs_mount *mp, xfs_dir2_off_t by)
+{
+	return (xfs_dir2_dataptr_t)(by >> XFS_DIR2_DATA_ALIGN_LOG);
+}
+
+/*
+ * Convert byte in space to (DB) block
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_byte_to_db(struct xfs_mount *mp, xfs_dir2_off_t by)
+{
+	return (xfs_dir2_db_t)
+		(by >> (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog));
+}
+
+/*
+ * Convert dataptr to a block number
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_dataptr_to_db(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
+{
+	return xfs_dir2_byte_to_db(mp, xfs_dir2_dataptr_to_byte(mp, dp));
+}
+
+/*
+ * Convert byte in space to offset in a block
+ */
+static inline xfs_dir2_data_aoff_t
+xfs_dir2_byte_to_off(struct xfs_mount *mp, xfs_dir2_off_t by)
+{
+	return (xfs_dir2_data_aoff_t)(by &
+		((1 << (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) - 1));
+}
+
+/*
+ * Convert dataptr to a byte offset in a block
+ */
+static inline xfs_dir2_data_aoff_t
+xfs_dir2_dataptr_to_off(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
+{
+	return xfs_dir2_byte_to_off(mp, xfs_dir2_dataptr_to_byte(mp, dp));
+}
+
+/*
+ * Convert block and offset to byte in space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_db_off_to_byte(struct xfs_mount *mp, xfs_dir2_db_t db,
+			xfs_dir2_data_aoff_t o)
+{
+	return ((xfs_dir2_off_t)db <<
+		(mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) + o;
+}
+
+/*
+ * Convert block (DB) to block (dablk)
+ */
+static inline xfs_dablk_t
+xfs_dir2_db_to_da(struct xfs_mount *mp, xfs_dir2_db_t db)
+{
+	return (xfs_dablk_t)(db << mp->m_sb.sb_dirblklog);
+}
+
+/*
+ * Convert byte in space to (DA) block
+ */
+static inline xfs_dablk_t
+xfs_dir2_byte_to_da(struct xfs_mount *mp, xfs_dir2_off_t by)
+{
+	return xfs_dir2_db_to_da(mp, xfs_dir2_byte_to_db(mp, by));
+}
+
+/*
+ * Convert block and offset to dataptr
+ */
+static inline xfs_dir2_dataptr_t
+xfs_dir2_db_off_to_dataptr(struct xfs_mount *mp, xfs_dir2_db_t db,
+			   xfs_dir2_data_aoff_t o)
+{
+	return xfs_dir2_byte_to_dataptr(mp, xfs_dir2_db_off_to_byte(mp, db, o));
+}
+
+/*
+ * Convert block (dablk) to block (DB)
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_da_to_db(struct xfs_mount *mp, xfs_dablk_t da)
+{
+	return (xfs_dir2_db_t)(da >> mp->m_sb.sb_dirblklog);
+}
+
+/*
+ * Convert block (dablk) to byte offset in space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_da_to_byte(struct xfs_mount *mp, xfs_dablk_t da)
+{
+	return xfs_dir2_db_off_to_byte(mp, xfs_dir2_da_to_db(mp, da), 0);
+}
+
+/*
+ * Free space block defintions for the node format.
+ */
+
+/*
+ * Offset of the freespace index.
+ */
+#define	XFS_DIR2_FREE_SPACE	2
+#define	XFS_DIR2_FREE_OFFSET	(XFS_DIR2_FREE_SPACE * XFS_DIR2_SPACE_SIZE)
+#define	XFS_DIR2_FREE_FIRSTDB(mp)	\
+	xfs_dir2_byte_to_db(mp, XFS_DIR2_FREE_OFFSET)
+
+typedef	struct xfs_dir2_free_hdr {
+	__be32			magic;		/* XFS_DIR2_FREE_MAGIC */
+	__be32			firstdb;	/* db of first entry */
+	__be32			nvalid;		/* count of valid entries */
+	__be32			nused;		/* count of used entries */
+} xfs_dir2_free_hdr_t;
+
+typedef struct xfs_dir2_free {
+	xfs_dir2_free_hdr_t	hdr;		/* block header */
+	__be16			bests[];	/* best free counts */
+						/* unused entries are -1 */
+} xfs_dir2_free_t;
+
+static inline int xfs_dir2_free_max_bests(struct xfs_mount *mp)
+{
+	return (mp->m_dirblksize - sizeof(struct xfs_dir2_free_hdr)) /
+		sizeof(xfs_dir2_data_off_t);
+}
+
+/*
+ * Convert data space db to the corresponding free db.
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_db_to_fdb(struct xfs_mount *mp, xfs_dir2_db_t db)
+{
+	return XFS_DIR2_FREE_FIRSTDB(mp) + db / xfs_dir2_free_max_bests(mp);
+}
+
+/*
+ * Convert data space db to the corresponding index in a free db.
+ */
+static inline int
+xfs_dir2_db_to_fdindex(struct xfs_mount *mp, xfs_dir2_db_t db)
+{
+	return db % xfs_dir2_free_max_bests(mp);
+}
+
+/*
+ * Single block format.
+ *
+ * The single block format looks like the following drawing on disk:
+ *
+ *    +-------------------------------------------------+
+ *    | xfs_dir2_data_hdr_t                             |
+ *    +-------------------------------------------------+
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t |
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t |
+ *    | xfs_dir2_data_entry_t OR xfs_dir2_data_unused_t :
+ *    | ...                                             |
+ *    +-------------------------------------------------+
+ *    | unused space                                    |
+ *    +-------------------------------------------------+
+ *    | ...                                             |
+ *    | xfs_dir2_leaf_entry_t                           |
+ *    | xfs_dir2_leaf_entry_t                           |
+ *    +-------------------------------------------------+
+ *    | xfs_dir2_block_tail_t                           |
+ *    +-------------------------------------------------+
+ *
+ * As all the entries are variable size structures the accessors below should
+ * be used to iterate over them.
+ */
+
+typedef struct xfs_dir2_block_tail {
+	__be32		count;			/* count of leaf entries */
+	__be32		stale;			/* count of stale lf entries */
+} xfs_dir2_block_tail_t;
+
+/*
+ * Pointer to the leaf header embedded in a data block (1-block format)
+ */
+static inline struct xfs_dir2_block_tail *
+xfs_dir2_block_tail_p(struct xfs_mount *mp, struct xfs_dir2_data_hdr *hdr)
+{
+	return ((struct xfs_dir2_block_tail *)
+		((char *)hdr + mp->m_dirblksize)) - 1;
+}
+
+/*
+ * Pointer to the leaf entries embedded in a data block (1-block format)
+ */
+static inline struct xfs_dir2_leaf_entry *
+xfs_dir2_block_leaf_p(struct xfs_dir2_block_tail *btp)
+{
+	return ((struct xfs_dir2_leaf_entry *)btp) - be32_to_cpu(btp->count);
+}
+
+#endif /* __XFS_DIR2_FORMAT_H__ */
diff --git a/fs/xfs/xfs_dir2_priv.h b/fs/xfs/xfs_dir2_priv.h
new file mode 100644
index 0000000..067f403
--- /dev/null
+++ b/fs/xfs/xfs_dir2_priv.h
@@ -0,0 +1,135 @@
+/*
+ * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_DIR2_PRIV_H__
+#define __XFS_DIR2_PRIV_H__
+
+/* xfs_dir2.c */
+extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino);
+extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
+extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
+extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
+				xfs_dir2_db_t *dbp);
+extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
+				struct xfs_dabuf *bp);
+extern int xfs_dir_cilookup_result(struct xfs_da_args *args,
+				const unsigned char *name, int len);
+
+/* xfs_dir2_block.c */
+extern int xfs_dir2_block_addname(struct xfs_da_args *args);
+extern int xfs_dir2_block_getdents(struct xfs_inode *dp, void *dirent,
+		xfs_off_t *offset, filldir_t filldir);
+extern int xfs_dir2_block_lookup(struct xfs_da_args *args);
+extern int xfs_dir2_block_removename(struct xfs_da_args *args);
+extern int xfs_dir2_block_replace(struct xfs_da_args *args);
+extern int xfs_dir2_leaf_to_block(struct xfs_da_args *args,
+		struct xfs_dabuf *lbp, struct xfs_dabuf *dbp);
+
+/* xfs_dir2_data.c */
+#ifdef DEBUG
+extern void xfs_dir2_data_check(struct xfs_inode *dp, struct xfs_dabuf *bp);
+#else
+#define	xfs_dir2_data_check(dp,bp)
+#endif
+extern struct xfs_dir2_data_free *
+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,
+		struct xfs_dabuf **bpp);
+extern void xfs_dir2_data_log_entry(struct xfs_trans *tp, struct xfs_dabuf *bp,
+		struct xfs_dir2_data_entry *dep);
+extern void xfs_dir2_data_log_header(struct xfs_trans *tp,
+		struct xfs_dabuf *bp);
+extern void xfs_dir2_data_log_unused(struct xfs_trans *tp, struct xfs_dabuf *bp,
+		struct xfs_dir2_data_unused *dup);
+extern void xfs_dir2_data_make_free(struct xfs_trans *tp, struct xfs_dabuf *bp,
+		xfs_dir2_data_aoff_t offset, xfs_dir2_data_aoff_t len,
+		int *needlogp, int *needscanp);
+extern void xfs_dir2_data_use_free(struct xfs_trans *tp, struct xfs_dabuf *bp,
+		struct xfs_dir2_data_unused *dup, xfs_dir2_data_aoff_t offset,
+		xfs_dir2_data_aoff_t len, int *needlogp, int *needscanp);
+
+/* xfs_dir2_leaf.c */
+extern int xfs_dir2_block_to_leaf(struct xfs_da_args *args,
+		struct xfs_dabuf *dbp);
+extern int xfs_dir2_leaf_addname(struct xfs_da_args *args);
+extern void xfs_dir2_leaf_compact(struct xfs_da_args *args,
+		struct xfs_dabuf *bp);
+extern void xfs_dir2_leaf_compact_x1(struct xfs_dabuf *bp, int *indexp,
+		int *lowstalep, int *highstalep, int *lowlogp, int *highlogp);
+extern int xfs_dir2_leaf_getdents(struct xfs_inode *dp, void *dirent,
+		size_t bufsize, xfs_off_t *offset, filldir_t filldir);
+extern int xfs_dir2_leaf_init(struct xfs_da_args *args, xfs_dir2_db_t bno,
+		struct xfs_dabuf **bpp, int magic);
+extern void xfs_dir2_leaf_log_ents(struct xfs_trans *tp, struct xfs_dabuf *bp,
+		int first, int last);
+extern void xfs_dir2_leaf_log_header(struct xfs_trans *tp,
+		struct xfs_dabuf *bp);
+extern int xfs_dir2_leaf_lookup(struct xfs_da_args *args);
+extern int xfs_dir2_leaf_removename(struct xfs_da_args *args);
+extern int xfs_dir2_leaf_replace(struct xfs_da_args *args);
+extern int xfs_dir2_leaf_search_hash(struct xfs_da_args *args,
+		struct xfs_dabuf *lbp);
+extern int xfs_dir2_leaf_trim_data(struct xfs_da_args *args,
+		struct xfs_dabuf *lbp, xfs_dir2_db_t db);
+extern struct xfs_dir2_leaf_entry *
+xfs_dir2_leaf_find_entry(struct xfs_dir2_leaf *leaf, int index, int compact,
+		int lowstale, int highstale,
+		int *lfloglow, int *lfloghigh);
+extern int xfs_dir2_node_to_leaf(struct xfs_da_state *state);
+
+/* xfs_dir2_node.c */
+extern int xfs_dir2_leaf_to_node(struct xfs_da_args *args,
+		struct xfs_dabuf *lbp);
+extern xfs_dahash_t xfs_dir2_leafn_lasthash(struct xfs_dabuf *bp, int *count);
+extern int xfs_dir2_leafn_lookup_int(struct xfs_dabuf *bp,
+		struct xfs_da_args *args, int *indexp,
+		struct xfs_da_state *state);
+extern int xfs_dir2_leafn_order(struct xfs_dabuf *leaf1_bp,
+		struct xfs_dabuf *leaf2_bp);
+extern int xfs_dir2_leafn_split(struct xfs_da_state *state,
+	struct xfs_da_state_blk *oldblk, struct xfs_da_state_blk *newblk);
+extern int xfs_dir2_leafn_toosmall(struct xfs_da_state *state, int *action);
+extern void xfs_dir2_leafn_unbalance(struct xfs_da_state *state,
+		struct xfs_da_state_blk *drop_blk,
+		struct xfs_da_state_blk *save_blk);
+extern int xfs_dir2_node_addname(struct xfs_da_args *args);
+extern int xfs_dir2_node_lookup(struct xfs_da_args *args);
+extern int xfs_dir2_node_removename(struct xfs_da_args *args);
+extern int xfs_dir2_node_replace(struct xfs_da_args *args);
+extern int xfs_dir2_node_trim_free(struct xfs_da_args *args, xfs_fileoff_t fo,
+		int *rvalp);
+
+/* xfs_dir2_sf.c */
+extern xfs_ino_t xfs_dir2_sf_get_parent_ino(struct xfs_dir2_sf_hdr *sfp);
+extern xfs_ino_t xfs_dir2_sfe_get_ino(struct xfs_dir2_sf_hdr *sfp,
+		struct xfs_dir2_sf_entry *sfep);
+extern int xfs_dir2_block_sfsize(struct xfs_inode *dp,
+		struct xfs_dir2_data_hdr *block, struct xfs_dir2_sf_hdr *sfhp);
+extern int xfs_dir2_block_to_sf(struct xfs_da_args *args, struct xfs_dabuf *bp,
+		int size, xfs_dir2_sf_hdr_t *sfhp);
+extern int xfs_dir2_sf_addname(struct xfs_da_args *args);
+extern int xfs_dir2_sf_create(struct xfs_da_args *args, xfs_ino_t pino);
+extern int xfs_dir2_sf_getdents(struct xfs_inode *dp, void *dirent,
+		xfs_off_t *offset, filldir_t filldir);
+extern int xfs_dir2_sf_lookup(struct xfs_da_args *args);
+extern int xfs_dir2_sf_removename(struct xfs_da_args *args);
+extern int xfs_dir2_sf_replace(struct xfs_da_args *args);
+
+#endif /* __XFS_DIR2_PRIV_H__ */
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
new file mode 100644
index 0000000..286a051
--- /dev/null
+++ b/fs/xfs/xfs_discard.c
@@ -0,0 +1,222 @@
+/*
+ * Copyright (C) 2010 Red Hat, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_sb.h"
+#include "xfs_inum.h"
+#include "xfs_log.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_quota.h"
+#include "xfs_trans.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_btree.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_error.h"
+#include "xfs_discard.h"
+#include "xfs_trace.h"
+
+STATIC int
+xfs_trim_extents(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	xfs_fsblock_t		start,
+	xfs_fsblock_t		end,
+	xfs_fsblock_t		minlen,
+	__uint64_t		*blocks_trimmed)
+{
+	struct block_device	*bdev = mp->m_ddev_targp->bt_bdev;
+	struct xfs_btree_cur	*cur;
+	struct xfs_buf		*agbp;
+	struct xfs_perag	*pag;
+	int			error;
+	int			i;
+
+	pag = xfs_perag_get(mp, agno);
+
+	error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp);
+	if (error || !agbp)
+		goto out_put_perag;
+
+	cur = xfs_allocbt_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_CNT);
+
+	/*
+	 * Force out the log.  This means any transactions that might have freed
+	 * space before we took the AGF buffer lock are now on disk, and the
+	 * volatile disk cache is flushed.
+	 */
+	xfs_log_force(mp, XFS_LOG_SYNC);
+
+	/*
+	 * Look up the longest btree in the AGF and start with it.
+	 */
+	error = xfs_alloc_lookup_le(cur, 0,
+			    be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_longest), &i);
+	if (error)
+		goto out_del_cursor;
+
+	/*
+	 * Loop until we are done with all extents that are large
+	 * enough to be worth discarding.
+	 */
+	while (i) {
+		xfs_agblock_t fbno;
+		xfs_extlen_t flen;
+
+		error = xfs_alloc_get_rec(cur, &fbno, &flen, &i);
+		if (error)
+			goto out_del_cursor;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, out_del_cursor);
+		ASSERT(flen <= be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_longest));
+
+		/*
+		 * Too small?  Give up.
+		 */
+		if (flen < minlen) {
+			trace_xfs_discard_toosmall(mp, agno, fbno, flen);
+			goto out_del_cursor;
+		}
+
+		/*
+		 * If the extent is entirely outside of the range we are
+		 * supposed to discard skip it.  Do not bother to trim
+		 * down partially overlapping ranges for now.
+		 */
+		if (XFS_AGB_TO_FSB(mp, agno, fbno) + flen < start ||
+		    XFS_AGB_TO_FSB(mp, agno, fbno) > end) {
+			trace_xfs_discard_exclude(mp, agno, fbno, flen);
+			goto next_extent;
+		}
+
+		/*
+		 * If any blocks in the range are still busy, skip the
+		 * discard and try again the next time.
+		 */
+		if (xfs_alloc_busy_search(mp, agno, fbno, flen)) {
+			trace_xfs_discard_busy(mp, agno, fbno, flen);
+			goto next_extent;
+		}
+
+		trace_xfs_discard_extent(mp, agno, fbno, flen);
+		error = -blkdev_issue_discard(bdev,
+				XFS_AGB_TO_DADDR(mp, agno, fbno),
+				XFS_FSB_TO_BB(mp, flen),
+				GFP_NOFS, 0);
+		if (error)
+			goto out_del_cursor;
+		*blocks_trimmed += flen;
+
+next_extent:
+		error = xfs_btree_decrement(cur, 0, &i);
+		if (error)
+			goto out_del_cursor;
+	}
+
+out_del_cursor:
+	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+	xfs_buf_relse(agbp);
+out_put_perag:
+	xfs_perag_put(pag);
+	return error;
+}
+
+int
+xfs_ioc_trim(
+	struct xfs_mount		*mp,
+	struct fstrim_range __user	*urange)
+{
+	struct request_queue	*q = mp->m_ddev_targp->bt_bdev->bd_disk->queue;
+	unsigned int		granularity = q->limits.discard_granularity;
+	struct fstrim_range	range;
+	xfs_fsblock_t		start, end, minlen;
+	xfs_agnumber_t		start_agno, end_agno, agno;
+	__uint64_t		blocks_trimmed = 0;
+	int			error, last_error = 0;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (!blk_queue_discard(q))
+		return -XFS_ERROR(EOPNOTSUPP);
+	if (copy_from_user(&range, urange, sizeof(range)))
+		return -XFS_ERROR(EFAULT);
+
+	/*
+	 * Truncating down the len isn't actually quite correct, but using
+	 * XFS_B_TO_FSB would mean we trivially get overflows for values
+	 * of ULLONG_MAX or slightly lower.  And ULLONG_MAX is the default
+	 * used by the fstrim application.  In the end it really doesn't
+	 * matter as trimming blocks is an advisory interface.
+	 */
+	start = XFS_B_TO_FSBT(mp, range.start);
+	end = start + XFS_B_TO_FSBT(mp, range.len) - 1;
+	minlen = XFS_B_TO_FSB(mp, max_t(u64, granularity, range.minlen));
+
+	if (start >= mp->m_sb.sb_dblocks)
+		return -XFS_ERROR(EINVAL);
+	if (end > mp->m_sb.sb_dblocks - 1)
+		end = mp->m_sb.sb_dblocks - 1;
+
+	start_agno = XFS_FSB_TO_AGNO(mp, start);
+	end_agno = XFS_FSB_TO_AGNO(mp, end);
+
+	for (agno = start_agno; agno <= end_agno; agno++) {
+		error = -xfs_trim_extents(mp, agno, start, end, minlen,
+					  &blocks_trimmed);
+		if (error)
+			last_error = error;
+	}
+
+	if (last_error)
+		return last_error;
+
+	range.len = XFS_FSB_TO_B(mp, blocks_trimmed);
+	if (copy_to_user(urange, &range, sizeof(range)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+int
+xfs_discard_extents(
+	struct xfs_mount	*mp,
+	struct list_head	*list)
+{
+	struct xfs_busy_extent	*busyp;
+	int			error = 0;
+
+	list_for_each_entry(busyp, list, list) {
+		trace_xfs_discard_extent(mp, busyp->agno, busyp->bno,
+					 busyp->length);
+
+		error = -blkdev_issue_discard(mp->m_ddev_targp->bt_bdev,
+				XFS_AGB_TO_DADDR(mp, busyp->agno, busyp->bno),
+				XFS_FSB_TO_BB(mp, busyp->length),
+				GFP_NOFS, 0);
+		if (error && error != EOPNOTSUPP) {
+			xfs_info(mp,
+	 "discard failed for extent [0x%llu,%u], error %d",
+				 (unsigned long long)busyp->bno,
+				 busyp->length,
+				 error);
+			return error;
+		}
+	}
+
+	return 0;
+}
diff --git a/fs/xfs/xfs_discard.h b/fs/xfs/xfs_discard.h
new file mode 100644
index 0000000..344879a
--- /dev/null
+++ b/fs/xfs/xfs_discard.h
@@ -0,0 +1,10 @@
+#ifndef XFS_DISCARD_H
+#define XFS_DISCARD_H 1
+
+struct fstrim_range;
+struct list_head;
+
+extern int	xfs_ioc_trim(struct xfs_mount *, struct fstrim_range __user *);
+extern int	xfs_discard_extents(struct xfs_mount *, struct list_head *);
+
+#endif /* XFS_DISCARD_H */
diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
new file mode 100644
index 0000000..25d7280
--- /dev/null
+++ b/fs/xfs/xfs_dquot.c
@@ -0,0 +1,1456 @@
+/*
+ * Copyright (c) 2000-2003 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_itable.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_space.h"
+#include "xfs_trans_priv.h"
+#include "xfs_qm.h"
+#include "xfs_trace.h"
+
+
+/*
+   LOCK ORDER
+
+   inode lock		    (ilock)
+   dquot hash-chain lock    (hashlock)
+   xqm dquot freelist lock  (freelistlock
+   mount's dquot list lock  (mplistlock)
+   user dquot lock - lock ordering among dquots is based on the uid or gid
+   group dquot lock - similar to udquots. Between the two dquots, the udquot
+		      has to be locked first.
+   pin lock - the dquot lock must be held to take this lock.
+   flush lock - ditto.
+*/
+
+#ifdef DEBUG
+xfs_buftarg_t *xfs_dqerror_target;
+int xfs_do_dqerror;
+int xfs_dqreq_num;
+int xfs_dqerror_mod = 33;
+#endif
+
+static struct lock_class_key xfs_dquot_other_class;
+
+/*
+ * Allocate and initialize a dquot. We don't always allocate fresh memory;
+ * we try to reclaim a free dquot if the number of incore dquots are above
+ * a threshold.
+ * The only field inside the core that gets initialized at this point
+ * is the d_id field. The idea is to fill in the entire q_core
+ * when we read in the on disk dquot.
+ */
+STATIC xfs_dquot_t *
+xfs_qm_dqinit(
+	xfs_mount_t  *mp,
+	xfs_dqid_t   id,
+	uint	     type)
+{
+	xfs_dquot_t	*dqp;
+	boolean_t	brandnewdquot;
+
+	brandnewdquot = xfs_qm_dqalloc_incore(&dqp);
+	dqp->dq_flags = type;
+	dqp->q_core.d_id = cpu_to_be32(id);
+	dqp->q_mount = mp;
+
+	/*
+	 * No need to re-initialize these if this is a reclaimed dquot.
+	 */
+	if (brandnewdquot) {
+		INIT_LIST_HEAD(&dqp->q_freelist);
+		mutex_init(&dqp->q_qlock);
+		init_waitqueue_head(&dqp->q_pinwait);
+
+		/*
+		 * Because we want to use a counting completion, complete
+		 * the flush completion once to allow a single access to
+		 * the flush completion without blocking.
+		 */
+		init_completion(&dqp->q_flush);
+		complete(&dqp->q_flush);
+
+		trace_xfs_dqinit(dqp);
+	} else {
+		/*
+		 * Only the q_core portion was zeroed in dqreclaim_one().
+		 * So, we need to reset others.
+		 */
+		dqp->q_nrefs = 0;
+		dqp->q_blkno = 0;
+		INIT_LIST_HEAD(&dqp->q_mplist);
+		INIT_LIST_HEAD(&dqp->q_hashlist);
+		dqp->q_bufoffset = 0;
+		dqp->q_fileoffset = 0;
+		dqp->q_transp = NULL;
+		dqp->q_gdquot = NULL;
+		dqp->q_res_bcount = 0;
+		dqp->q_res_icount = 0;
+		dqp->q_res_rtbcount = 0;
+		atomic_set(&dqp->q_pincount, 0);
+		dqp->q_hash = NULL;
+		ASSERT(list_empty(&dqp->q_freelist));
+
+		trace_xfs_dqreuse(dqp);
+	}
+
+	/*
+	 * In either case we need to make sure group quotas have a different
+	 * lock class than user quotas, to make sure lockdep knows we can
+	 * locks of one of each at the same time.
+	 */
+	if (!(type & XFS_DQ_USER))
+		lockdep_set_class(&dqp->q_qlock, &xfs_dquot_other_class);
+
+	/*
+	 * log item gets initialized later
+	 */
+	return (dqp);
+}
+
+/*
+ * This is called to free all the memory associated with a dquot
+ */
+void
+xfs_qm_dqdestroy(
+	xfs_dquot_t	*dqp)
+{
+	ASSERT(list_empty(&dqp->q_freelist));
+
+	mutex_destroy(&dqp->q_qlock);
+	kmem_zone_free(xfs_Gqm->qm_dqzone, dqp);
+
+	atomic_dec(&xfs_Gqm->qm_totaldquots);
+}
+
+/*
+ * This is what a 'fresh' dquot inside a dquot chunk looks like on disk.
+ */
+STATIC void
+xfs_qm_dqinit_core(
+	xfs_dqid_t	id,
+	uint		type,
+	xfs_dqblk_t	*d)
+{
+	/*
+	 * Caller has zero'd the entire dquot 'chunk' already.
+	 */
+	d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
+	d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
+	d->dd_diskdq.d_id = cpu_to_be32(id);
+	d->dd_diskdq.d_flags = type;
+}
+
+/*
+ * If default limits are in force, push them into the dquot now.
+ * We overwrite the dquot limits only if they are zero and this
+ * is not the root dquot.
+ */
+void
+xfs_qm_adjust_dqlimits(
+	xfs_mount_t		*mp,
+	xfs_disk_dquot_t	*d)
+{
+	xfs_quotainfo_t		*q = mp->m_quotainfo;
+
+	ASSERT(d->d_id);
+
+	if (q->qi_bsoftlimit && !d->d_blk_softlimit)
+		d->d_blk_softlimit = cpu_to_be64(q->qi_bsoftlimit);
+	if (q->qi_bhardlimit && !d->d_blk_hardlimit)
+		d->d_blk_hardlimit = cpu_to_be64(q->qi_bhardlimit);
+	if (q->qi_isoftlimit && !d->d_ino_softlimit)
+		d->d_ino_softlimit = cpu_to_be64(q->qi_isoftlimit);
+	if (q->qi_ihardlimit && !d->d_ino_hardlimit)
+		d->d_ino_hardlimit = cpu_to_be64(q->qi_ihardlimit);
+	if (q->qi_rtbsoftlimit && !d->d_rtb_softlimit)
+		d->d_rtb_softlimit = cpu_to_be64(q->qi_rtbsoftlimit);
+	if (q->qi_rtbhardlimit && !d->d_rtb_hardlimit)
+		d->d_rtb_hardlimit = cpu_to_be64(q->qi_rtbhardlimit);
+}
+
+/*
+ * Check the limits and timers of a dquot and start or reset timers
+ * if necessary.
+ * This gets called even when quota enforcement is OFF, which makes our
+ * life a little less complicated. (We just don't reject any quota
+ * reservations in that case, when enforcement is off).
+ * We also return 0 as the values of the timers in Q_GETQUOTA calls, when
+ * enforcement's off.
+ * In contrast, warnings are a little different in that they don't
+ * 'automatically' get started when limits get exceeded.  They do
+ * get reset to zero, however, when we find the count to be under
+ * the soft limit (they are only ever set non-zero via userspace).
+ */
+void
+xfs_qm_adjust_dqtimers(
+	xfs_mount_t		*mp,
+	xfs_disk_dquot_t	*d)
+{
+	ASSERT(d->d_id);
+
+#ifdef DEBUG
+	if (d->d_blk_hardlimit)
+		ASSERT(be64_to_cpu(d->d_blk_softlimit) <=
+		       be64_to_cpu(d->d_blk_hardlimit));
+	if (d->d_ino_hardlimit)
+		ASSERT(be64_to_cpu(d->d_ino_softlimit) <=
+		       be64_to_cpu(d->d_ino_hardlimit));
+	if (d->d_rtb_hardlimit)
+		ASSERT(be64_to_cpu(d->d_rtb_softlimit) <=
+		       be64_to_cpu(d->d_rtb_hardlimit));
+#endif
+
+	if (!d->d_btimer) {
+		if ((d->d_blk_softlimit &&
+		     (be64_to_cpu(d->d_bcount) >=
+		      be64_to_cpu(d->d_blk_softlimit))) ||
+		    (d->d_blk_hardlimit &&
+		     (be64_to_cpu(d->d_bcount) >=
+		      be64_to_cpu(d->d_blk_hardlimit)))) {
+			d->d_btimer = cpu_to_be32(get_seconds() +
+					mp->m_quotainfo->qi_btimelimit);
+		} else {
+			d->d_bwarns = 0;
+		}
+	} else {
+		if ((!d->d_blk_softlimit ||
+		     (be64_to_cpu(d->d_bcount) <
+		      be64_to_cpu(d->d_blk_softlimit))) &&
+		    (!d->d_blk_hardlimit ||
+		    (be64_to_cpu(d->d_bcount) <
+		     be64_to_cpu(d->d_blk_hardlimit)))) {
+			d->d_btimer = 0;
+		}
+	}
+
+	if (!d->d_itimer) {
+		if ((d->d_ino_softlimit &&
+		     (be64_to_cpu(d->d_icount) >=
+		      be64_to_cpu(d->d_ino_softlimit))) ||
+		    (d->d_ino_hardlimit &&
+		     (be64_to_cpu(d->d_icount) >=
+		      be64_to_cpu(d->d_ino_hardlimit)))) {
+			d->d_itimer = cpu_to_be32(get_seconds() +
+					mp->m_quotainfo->qi_itimelimit);
+		} else {
+			d->d_iwarns = 0;
+		}
+	} else {
+		if ((!d->d_ino_softlimit ||
+		     (be64_to_cpu(d->d_icount) <
+		      be64_to_cpu(d->d_ino_softlimit)))  &&
+		    (!d->d_ino_hardlimit ||
+		     (be64_to_cpu(d->d_icount) <
+		      be64_to_cpu(d->d_ino_hardlimit)))) {
+			d->d_itimer = 0;
+		}
+	}
+
+	if (!d->d_rtbtimer) {
+		if ((d->d_rtb_softlimit &&
+		     (be64_to_cpu(d->d_rtbcount) >=
+		      be64_to_cpu(d->d_rtb_softlimit))) ||
+		    (d->d_rtb_hardlimit &&
+		     (be64_to_cpu(d->d_rtbcount) >=
+		      be64_to_cpu(d->d_rtb_hardlimit)))) {
+			d->d_rtbtimer = cpu_to_be32(get_seconds() +
+					mp->m_quotainfo->qi_rtbtimelimit);
+		} else {
+			d->d_rtbwarns = 0;
+		}
+	} else {
+		if ((!d->d_rtb_softlimit ||
+		     (be64_to_cpu(d->d_rtbcount) <
+		      be64_to_cpu(d->d_rtb_softlimit))) &&
+		    (!d->d_rtb_hardlimit ||
+		     (be64_to_cpu(d->d_rtbcount) <
+		      be64_to_cpu(d->d_rtb_hardlimit)))) {
+			d->d_rtbtimer = 0;
+		}
+	}
+}
+
+/*
+ * initialize a buffer full of dquots and log the whole thing
+ */
+STATIC void
+xfs_qm_init_dquot_blk(
+	xfs_trans_t	*tp,
+	xfs_mount_t	*mp,
+	xfs_dqid_t	id,
+	uint		type,
+	xfs_buf_t	*bp)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	xfs_dqblk_t	*d;
+	int		curid, i;
+
+	ASSERT(tp);
+	ASSERT(xfs_buf_islocked(bp));
+
+	d = bp->b_addr;
+
+	/*
+	 * ID of the first dquot in the block - id's are zero based.
+	 */
+	curid = id - (id % q->qi_dqperchunk);
+	ASSERT(curid >= 0);
+	memset(d, 0, BBTOB(q->qi_dqchunklen));
+	for (i = 0; i < q->qi_dqperchunk; i++, d++, curid++)
+		xfs_qm_dqinit_core(curid, type, d);
+	xfs_trans_dquot_buf(tp, bp,
+			    (type & XFS_DQ_USER ? XFS_BLF_UDQUOT_BUF :
+			    ((type & XFS_DQ_PROJ) ? XFS_BLF_PDQUOT_BUF :
+			     XFS_BLF_GDQUOT_BUF)));
+	xfs_trans_log_buf(tp, bp, 0, BBTOB(q->qi_dqchunklen) - 1);
+}
+
+
+
+/*
+ * Allocate a block and fill it with dquots.
+ * This is called when the bmapi finds a hole.
+ */
+STATIC int
+xfs_qm_dqalloc(
+	xfs_trans_t	**tpp,
+	xfs_mount_t	*mp,
+	xfs_dquot_t	*dqp,
+	xfs_inode_t	*quotip,
+	xfs_fileoff_t	offset_fsb,
+	xfs_buf_t	**O_bpp)
+{
+	xfs_fsblock_t	firstblock;
+	xfs_bmap_free_t flist;
+	xfs_bmbt_irec_t map;
+	int		nmaps, error, committed;
+	xfs_buf_t	*bp;
+	xfs_trans_t	*tp = *tpp;
+
+	ASSERT(tp != NULL);
+
+	trace_xfs_dqalloc(dqp);
+
+	/*
+	 * Initialize the bmap freelist prior to calling bmapi code.
+	 */
+	xfs_bmap_init(&flist, &firstblock);
+	xfs_ilock(quotip, XFS_ILOCK_EXCL);
+	/*
+	 * Return if this type of quotas is turned off while we didn't
+	 * have an inode lock
+	 */
+	if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
+		xfs_iunlock(quotip, XFS_ILOCK_EXCL);
+		return (ESRCH);
+	}
+
+	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
+	nmaps = 1;
+	error = xfs_bmapi_write(tp, quotip, offset_fsb,
+				XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA,
+				&firstblock, XFS_QM_DQALLOC_SPACE_RES(mp),
+				&map, &nmaps, &flist);
+	if (error)
+		goto error0;
+	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
+	ASSERT(nmaps == 1);
+	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
+	       (map.br_startblock != HOLESTARTBLOCK));
+
+	/*
+	 * Keep track of the blkno to save a lookup later
+	 */
+	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
+
+	/* now we can just get the buffer (there's nothing to read yet) */
+	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
+			       dqp->q_blkno,
+			       mp->m_quotainfo->qi_dqchunklen,
+			       0);
+
+	error = xfs_buf_geterror(bp);
+	if (error)
+		goto error1;
+
+	/*
+	 * Make a chunk of dquots out of this buffer and log
+	 * the entire thing.
+	 */
+	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
+			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);
+
+	/*
+	 * xfs_bmap_finish() may commit the current transaction and
+	 * start a second transaction if the freelist is not empty.
+	 *
+	 * Since we still want to modify this buffer, we need to
+	 * ensure that the buffer is not released on commit of
+	 * the first transaction and ensure the buffer is added to the
+	 * second transaction.
+	 *
+	 * If there is only one transaction then don't stop the buffer
+	 * from being released when it commits later on.
+	 */
+
+	xfs_trans_bhold(tp, bp);
+
+	if ((error = xfs_bmap_finish(tpp, &flist, &committed))) {
+		goto error1;
+	}
+
+	if (committed) {
+		tp = *tpp;
+		xfs_trans_bjoin(tp, bp);
+	} else {
+		xfs_trans_bhold_release(tp, bp);
+	}
+
+	*O_bpp = bp;
+	return 0;
+
+      error1:
+	xfs_bmap_cancel(&flist);
+      error0:
+	xfs_iunlock(quotip, XFS_ILOCK_EXCL);
+
+	return (error);
+}
+
+/*
+ * Maps a dquot to the buffer containing its on-disk version.
+ * This returns a ptr to the buffer containing the on-disk dquot
+ * in the bpp param, and a ptr to the on-disk dquot within that buffer
+ */
+STATIC int
+xfs_qm_dqtobp(
+	xfs_trans_t		**tpp,
+	xfs_dquot_t		*dqp,
+	xfs_disk_dquot_t	**O_ddpp,
+	xfs_buf_t		**O_bpp,
+	uint			flags)
+{
+	xfs_bmbt_irec_t map;
+	int		nmaps = 1, error;
+	xfs_buf_t	*bp;
+	xfs_inode_t	*quotip = XFS_DQ_TO_QIP(dqp);
+	xfs_mount_t	*mp = dqp->q_mount;
+	xfs_disk_dquot_t *ddq;
+	xfs_dqid_t	id = be32_to_cpu(dqp->q_core.d_id);
+	xfs_trans_t	*tp = (tpp ? *tpp : NULL);
+
+	dqp->q_fileoffset = (xfs_fileoff_t)id / mp->m_quotainfo->qi_dqperchunk;
+
+	xfs_ilock(quotip, XFS_ILOCK_SHARED);
+	if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
+		/*
+		 * Return if this type of quotas is turned off while we
+		 * didn't have the quota inode lock.
+		 */
+		xfs_iunlock(quotip, XFS_ILOCK_SHARED);
+		return ESRCH;
+	}
+
+	/*
+	 * Find the block map; no allocations yet
+	 */
+	error = xfs_bmapi_read(quotip, dqp->q_fileoffset,
+			       XFS_DQUOT_CLUSTER_SIZE_FSB, &map, &nmaps, 0);
+
+	xfs_iunlock(quotip, XFS_ILOCK_SHARED);
+	if (error)
+		return error;
+
+	ASSERT(nmaps == 1);
+	ASSERT(map.br_blockcount == 1);
+
+	/*
+	 * Offset of dquot in the (fixed sized) dquot chunk.
+	 */
+	dqp->q_bufoffset = (id % mp->m_quotainfo->qi_dqperchunk) *
+		sizeof(xfs_dqblk_t);
+
+	ASSERT(map.br_startblock != DELAYSTARTBLOCK);
+	if (map.br_startblock == HOLESTARTBLOCK) {
+		/*
+		 * We don't allocate unless we're asked to
+		 */
+		if (!(flags & XFS_QMOPT_DQALLOC))
+			return ENOENT;
+
+		ASSERT(tp);
+		error = xfs_qm_dqalloc(tpp, mp, dqp, quotip,
+					dqp->q_fileoffset, &bp);
+		if (error)
+			return error;
+		tp = *tpp;
+	} else {
+		trace_xfs_dqtobp_read(dqp);
+
+		/*
+		 * store the blkno etc so that we don't have to do the
+		 * mapping all the time
+		 */
+		dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
+
+		error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
+					   dqp->q_blkno,
+					   mp->m_quotainfo->qi_dqchunklen,
+					   0, &bp);
+		if (error || !bp)
+			return XFS_ERROR(error);
+	}
+
+	ASSERT(xfs_buf_islocked(bp));
+
+	/*
+	 * calculate the location of the dquot inside the buffer.
+	 */
+	ddq = bp->b_addr + dqp->q_bufoffset;
+
+	/*
+	 * A simple sanity check in case we got a corrupted dquot...
+	 */
+	error = xfs_qm_dqcheck(mp, ddq, id, dqp->dq_flags & XFS_DQ_ALLTYPES,
+			   flags & (XFS_QMOPT_DQREPAIR|XFS_QMOPT_DOWARN),
+			   "dqtobp");
+	if (error) {
+		if (!(flags & XFS_QMOPT_DQREPAIR)) {
+			xfs_trans_brelse(tp, bp);
+			return XFS_ERROR(EIO);
+		}
+	}
+
+	*O_bpp = bp;
+	*O_ddpp = ddq;
+
+	return (0);
+}
+
+
+/*
+ * Read in the ondisk dquot using dqtobp() then copy it to an incore version,
+ * and release the buffer immediately.
+ *
+ */
+/* ARGSUSED */
+STATIC int
+xfs_qm_dqread(
+	xfs_trans_t	**tpp,
+	xfs_dqid_t	id,
+	xfs_dquot_t	*dqp,	/* dquot to get filled in */
+	uint		flags)
+{
+	xfs_disk_dquot_t *ddqp;
+	xfs_buf_t	 *bp;
+	int		 error;
+	xfs_trans_t	 *tp;
+
+	ASSERT(tpp);
+
+	trace_xfs_dqread(dqp);
+
+	/*
+	 * get a pointer to the on-disk dquot and the buffer containing it
+	 * dqp already knows its own type (GROUP/USER).
+	 */
+	if ((error = xfs_qm_dqtobp(tpp, dqp, &ddqp, &bp, flags))) {
+		return (error);
+	}
+	tp = *tpp;
+
+	/* copy everything from disk dquot to the incore dquot */
+	memcpy(&dqp->q_core, ddqp, sizeof(xfs_disk_dquot_t));
+	ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);
+	xfs_qm_dquot_logitem_init(dqp);
+
+	/*
+	 * Reservation counters are defined as reservation plus current usage
+	 * to avoid having to add every time.
+	 */
+	dqp->q_res_bcount = be64_to_cpu(ddqp->d_bcount);
+	dqp->q_res_icount = be64_to_cpu(ddqp->d_icount);
+	dqp->q_res_rtbcount = be64_to_cpu(ddqp->d_rtbcount);
+
+	/* Mark the buf so that this will stay incore a little longer */
+	xfs_buf_set_ref(bp, XFS_DQUOT_REF);
+
+	/*
+	 * We got the buffer with a xfs_trans_read_buf() (in dqtobp())
+	 * So we need to release with xfs_trans_brelse().
+	 * The strategy here is identical to that of inodes; we lock
+	 * the dquot in xfs_qm_dqget() before making it accessible to
+	 * others. This is because dquots, like inodes, need a good level of
+	 * concurrency, and we don't want to take locks on the entire buffers
+	 * for dquot accesses.
+	 * Note also that the dquot buffer may even be dirty at this point, if
+	 * this particular dquot was repaired. We still aren't afraid to
+	 * brelse it because we have the changes incore.
+	 */
+	ASSERT(xfs_buf_islocked(bp));
+	xfs_trans_brelse(tp, bp);
+
+	return (error);
+}
+
+
+/*
+ * allocate an incore dquot from the kernel heap,
+ * and fill its core with quota information kept on disk.
+ * If XFS_QMOPT_DQALLOC is set, it'll allocate a dquot on disk
+ * if it wasn't already allocated.
+ */
+STATIC int
+xfs_qm_idtodq(
+	xfs_mount_t	*mp,
+	xfs_dqid_t	id,	 /* gid or uid, depending on type */
+	uint		type,	 /* UDQUOT or GDQUOT */
+	uint		flags,	 /* DQALLOC, DQREPAIR */
+	xfs_dquot_t	**O_dqpp)/* OUT : incore dquot, not locked */
+{
+	xfs_dquot_t	*dqp;
+	int		error;
+	xfs_trans_t	*tp;
+	int		cancelflags=0;
+
+	dqp = xfs_qm_dqinit(mp, id, type);
+	tp = NULL;
+	if (flags & XFS_QMOPT_DQALLOC) {
+		tp = xfs_trans_alloc(mp, XFS_TRANS_QM_DQALLOC);
+		error = xfs_trans_reserve(tp, XFS_QM_DQALLOC_SPACE_RES(mp),
+				XFS_WRITE_LOG_RES(mp) +
+				BBTOB(mp->m_quotainfo->qi_dqchunklen) - 1 +
+				128,
+				0,
+				XFS_TRANS_PERM_LOG_RES,
+				XFS_WRITE_LOG_COUNT);
+		if (error) {
+			cancelflags = 0;
+			goto error0;
+		}
+		cancelflags = XFS_TRANS_RELEASE_LOG_RES;
+	}
+
+	/*
+	 * Read it from disk; xfs_dqread() takes care of
+	 * all the necessary initialization of dquot's fields (locks, etc)
+	 */
+	if ((error = xfs_qm_dqread(&tp, id, dqp, flags))) {
+		/*
+		 * This can happen if quotas got turned off (ESRCH),
+		 * or if the dquot didn't exist on disk and we ask to
+		 * allocate (ENOENT).
+		 */
+		trace_xfs_dqread_fail(dqp);
+		cancelflags |= XFS_TRANS_ABORT;
+		goto error0;
+	}
+	if (tp) {
+		if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES)))
+			goto error1;
+	}
+
+	*O_dqpp = dqp;
+	return (0);
+
+ error0:
+	ASSERT(error);
+	if (tp)
+		xfs_trans_cancel(tp, cancelflags);
+ error1:
+	xfs_qm_dqdestroy(dqp);
+	*O_dqpp = NULL;
+	return (error);
+}
+
+/*
+ * Lookup a dquot in the incore dquot hashtable. We keep two separate
+ * hashtables for user and group dquots; and, these are global tables
+ * inside the XQM, not per-filesystem tables.
+ * The hash chain must be locked by caller, and it is left locked
+ * on return. Returning dquot is locked.
+ */
+STATIC int
+xfs_qm_dqlookup(
+	xfs_mount_t		*mp,
+	xfs_dqid_t		id,
+	xfs_dqhash_t		*qh,
+	xfs_dquot_t		**O_dqpp)
+{
+	xfs_dquot_t		*dqp;
+	uint			flist_locked;
+
+	ASSERT(mutex_is_locked(&qh->qh_lock));
+
+	flist_locked = B_FALSE;
+
+	/*
+	 * Traverse the hashchain looking for a match
+	 */
+	list_for_each_entry(dqp, &qh->qh_list, q_hashlist) {
+		/*
+		 * We already have the hashlock. We don't need the
+		 * dqlock to look at the id field of the dquot, since the
+		 * id can't be modified without the hashlock anyway.
+		 */
+		if (be32_to_cpu(dqp->q_core.d_id) == id && dqp->q_mount == mp) {
+			trace_xfs_dqlookup_found(dqp);
+
+			/*
+			 * All in core dquots must be on the dqlist of mp
+			 */
+			ASSERT(!list_empty(&dqp->q_mplist));
+
+			xfs_dqlock(dqp);
+			if (dqp->q_nrefs == 0) {
+				ASSERT(!list_empty(&dqp->q_freelist));
+				if (!mutex_trylock(&xfs_Gqm->qm_dqfrlist_lock)) {
+					trace_xfs_dqlookup_want(dqp);
+
+					/*
+					 * We may have raced with dqreclaim_one()
+					 * (and lost). So, flag that we don't
+					 * want the dquot to be reclaimed.
+					 */
+					dqp->dq_flags |= XFS_DQ_WANT;
+					xfs_dqunlock(dqp);
+					mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);
+					xfs_dqlock(dqp);
+					dqp->dq_flags &= ~(XFS_DQ_WANT);
+				}
+				flist_locked = B_TRUE;
+			}
+
+			/*
+			 * id couldn't have changed; we had the hashlock all
+			 * along
+			 */
+			ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);
+
+			if (flist_locked) {
+				if (dqp->q_nrefs != 0) {
+					mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+					flist_locked = B_FALSE;
+				} else {
+					/* take it off the freelist */
+					trace_xfs_dqlookup_freelist(dqp);
+					list_del_init(&dqp->q_freelist);
+					xfs_Gqm->qm_dqfrlist_cnt--;
+				}
+			}
+
+			XFS_DQHOLD(dqp);
+
+			if (flist_locked)
+				mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+			/*
+			 * move the dquot to the front of the hashchain
+			 */
+			ASSERT(mutex_is_locked(&qh->qh_lock));
+			list_move(&dqp->q_hashlist, &qh->qh_list);
+			trace_xfs_dqlookup_done(dqp);
+			*O_dqpp = dqp;
+			return 0;
+		}
+	}
+
+	*O_dqpp = NULL;
+	ASSERT(mutex_is_locked(&qh->qh_lock));
+	return (1);
+}
+
+/*
+ * Given the file system, inode OR id, and type (UDQUOT/GDQUOT), return a
+ * a locked dquot, doing an allocation (if requested) as needed.
+ * When both an inode and an id are given, the inode's id takes precedence.
+ * That is, if the id changes while we don't hold the ilock inside this
+ * function, the new dquot is returned, not necessarily the one requested
+ * in the id argument.
+ */
+int
+xfs_qm_dqget(
+	xfs_mount_t	*mp,
+	xfs_inode_t	*ip,	  /* locked inode (optional) */
+	xfs_dqid_t	id,	  /* uid/projid/gid depending on type */
+	uint		type,	  /* XFS_DQ_USER/XFS_DQ_PROJ/XFS_DQ_GROUP */
+	uint		flags,	  /* DQALLOC, DQSUSER, DQREPAIR, DOWARN */
+	xfs_dquot_t	**O_dqpp) /* OUT : locked incore dquot */
+{
+	xfs_dquot_t	*dqp;
+	xfs_dqhash_t	*h;
+	uint		version;
+	int		error;
+
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+	if ((! XFS_IS_UQUOTA_ON(mp) && type == XFS_DQ_USER) ||
+	    (! XFS_IS_PQUOTA_ON(mp) && type == XFS_DQ_PROJ) ||
+	    (! XFS_IS_GQUOTA_ON(mp) && type == XFS_DQ_GROUP)) {
+		return (ESRCH);
+	}
+	h = XFS_DQ_HASH(mp, id, type);
+
+#ifdef DEBUG
+	if (xfs_do_dqerror) {
+		if ((xfs_dqerror_target == mp->m_ddev_targp) &&
+		    (xfs_dqreq_num++ % xfs_dqerror_mod) == 0) {
+			xfs_debug(mp, "Returning error in dqget");
+			return (EIO);
+		}
+	}
+#endif
+
+ again:
+
+#ifdef DEBUG
+	ASSERT(type == XFS_DQ_USER ||
+	       type == XFS_DQ_PROJ ||
+	       type == XFS_DQ_GROUP);
+	if (ip) {
+		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+		if (type == XFS_DQ_USER)
+			ASSERT(ip->i_udquot == NULL);
+		else
+			ASSERT(ip->i_gdquot == NULL);
+	}
+#endif
+	mutex_lock(&h->qh_lock);
+
+	/*
+	 * Look in the cache (hashtable).
+	 * The chain is kept locked during lookup.
+	 */
+	if (xfs_qm_dqlookup(mp, id, h, O_dqpp) == 0) {
+		XQM_STATS_INC(xqmstats.xs_qm_dqcachehits);
+		/*
+		 * The dquot was found, moved to the front of the chain,
+		 * taken off the freelist if it was on it, and locked
+		 * at this point. Just unlock the hashchain and return.
+		 */
+		ASSERT(*O_dqpp);
+		ASSERT(XFS_DQ_IS_LOCKED(*O_dqpp));
+		mutex_unlock(&h->qh_lock);
+		trace_xfs_dqget_hit(*O_dqpp);
+		return (0);	/* success */
+	}
+	XQM_STATS_INC(xqmstats.xs_qm_dqcachemisses);
+
+	/*
+	 * Dquot cache miss. We don't want to keep the inode lock across
+	 * a (potential) disk read. Also we don't want to deal with the lock
+	 * ordering between quotainode and this inode. OTOH, dropping the inode
+	 * lock here means dealing with a chown that can happen before
+	 * we re-acquire the lock.
+	 */
+	if (ip)
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	/*
+	 * Save the hashchain version stamp, and unlock the chain, so that
+	 * we don't keep the lock across a disk read
+	 */
+	version = h->qh_version;
+	mutex_unlock(&h->qh_lock);
+
+	/*
+	 * Allocate the dquot on the kernel heap, and read the ondisk
+	 * portion off the disk. Also, do all the necessary initialization
+	 * This can return ENOENT if dquot didn't exist on disk and we didn't
+	 * ask it to allocate; ESRCH if quotas got turned off suddenly.
+	 */
+	if ((error = xfs_qm_idtodq(mp, id, type,
+				  flags & (XFS_QMOPT_DQALLOC|XFS_QMOPT_DQREPAIR|
+					   XFS_QMOPT_DOWARN),
+				  &dqp))) {
+		if (ip)
+			xfs_ilock(ip, XFS_ILOCK_EXCL);
+		return (error);
+	}
+
+	/*
+	 * See if this is mount code calling to look at the overall quota limits
+	 * which are stored in the id == 0 user or group's dquot.
+	 * Since we may not have done a quotacheck by this point, just return
+	 * the dquot without attaching it to any hashtables, lists, etc, or even
+	 * taking a reference.
+	 * The caller must dqdestroy this once done.
+	 */
+	if (flags & XFS_QMOPT_DQSUSER) {
+		ASSERT(id == 0);
+		ASSERT(! ip);
+		goto dqret;
+	}
+
+	/*
+	 * Dquot lock comes after hashlock in the lock ordering
+	 */
+	if (ip) {
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+		/*
+		 * A dquot could be attached to this inode by now, since
+		 * we had dropped the ilock.
+		 */
+		if (type == XFS_DQ_USER) {
+			if (!XFS_IS_UQUOTA_ON(mp)) {
+				/* inode stays locked on return */
+				xfs_qm_dqdestroy(dqp);
+				return XFS_ERROR(ESRCH);
+			}
+			if (ip->i_udquot) {
+				xfs_qm_dqdestroy(dqp);
+				dqp = ip->i_udquot;
+				xfs_dqlock(dqp);
+				goto dqret;
+			}
+		} else {
+			if (!XFS_IS_OQUOTA_ON(mp)) {
+				/* inode stays locked on return */
+				xfs_qm_dqdestroy(dqp);
+				return XFS_ERROR(ESRCH);
+			}
+			if (ip->i_gdquot) {
+				xfs_qm_dqdestroy(dqp);
+				dqp = ip->i_gdquot;
+				xfs_dqlock(dqp);
+				goto dqret;
+			}
+		}
+	}
+
+	/*
+	 * Hashlock comes after ilock in lock order
+	 */
+	mutex_lock(&h->qh_lock);
+	if (version != h->qh_version) {
+		xfs_dquot_t *tmpdqp;
+		/*
+		 * Now, see if somebody else put the dquot in the
+		 * hashtable before us. This can happen because we didn't
+		 * keep the hashchain lock. We don't have to worry about
+		 * lock order between the two dquots here since dqp isn't
+		 * on any findable lists yet.
+		 */
+		if (xfs_qm_dqlookup(mp, id, h, &tmpdqp) == 0) {
+			/*
+			 * Duplicate found. Just throw away the new dquot
+			 * and start over.
+			 */
+			xfs_qm_dqput(tmpdqp);
+			mutex_unlock(&h->qh_lock);
+			xfs_qm_dqdestroy(dqp);
+			XQM_STATS_INC(xqmstats.xs_qm_dquot_dups);
+			goto again;
+		}
+	}
+
+	/*
+	 * Put the dquot at the beginning of the hash-chain and mp's list
+	 * LOCK ORDER: hashlock, freelistlock, mplistlock, udqlock, gdqlock ..
+	 */
+	ASSERT(mutex_is_locked(&h->qh_lock));
+	dqp->q_hash = h;
+	list_add(&dqp->q_hashlist, &h->qh_list);
+	h->qh_version++;
+
+	/*
+	 * Attach this dquot to this filesystem's list of all dquots,
+	 * kept inside the mount structure in m_quotainfo field
+	 */
+	mutex_lock(&mp->m_quotainfo->qi_dqlist_lock);
+
+	/*
+	 * We return a locked dquot to the caller, with a reference taken
+	 */
+	xfs_dqlock(dqp);
+	dqp->q_nrefs = 1;
+
+	list_add(&dqp->q_mplist, &mp->m_quotainfo->qi_dqlist);
+	mp->m_quotainfo->qi_dquots++;
+	mutex_unlock(&mp->m_quotainfo->qi_dqlist_lock);
+	mutex_unlock(&h->qh_lock);
+ dqret:
+	ASSERT((ip == NULL) || xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	trace_xfs_dqget_miss(dqp);
+	*O_dqpp = dqp;
+	return (0);
+}
+
+
+/*
+ * Release a reference to the dquot (decrement ref-count)
+ * and unlock it. If there is a group quota attached to this
+ * dquot, carefully release that too without tripping over
+ * deadlocks'n'stuff.
+ */
+void
+xfs_qm_dqput(
+	xfs_dquot_t	*dqp)
+{
+	xfs_dquot_t	*gdqp;
+
+	ASSERT(dqp->q_nrefs > 0);
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+
+	trace_xfs_dqput(dqp);
+
+	if (dqp->q_nrefs != 1) {
+		dqp->q_nrefs--;
+		xfs_dqunlock(dqp);
+		return;
+	}
+
+	/*
+	 * drop the dqlock and acquire the freelist and dqlock
+	 * in the right order; but try to get it out-of-order first
+	 */
+	if (!mutex_trylock(&xfs_Gqm->qm_dqfrlist_lock)) {
+		trace_xfs_dqput_wait(dqp);
+		xfs_dqunlock(dqp);
+		mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);
+		xfs_dqlock(dqp);
+	}
+
+	while (1) {
+		gdqp = NULL;
+
+		/* We can't depend on nrefs being == 1 here */
+		if (--dqp->q_nrefs == 0) {
+			trace_xfs_dqput_free(dqp);
+
+			list_add_tail(&dqp->q_freelist, &xfs_Gqm->qm_dqfrlist);
+			xfs_Gqm->qm_dqfrlist_cnt++;
+
+			/*
+			 * If we just added a udquot to the freelist, then
+			 * we want to release the gdquot reference that
+			 * it (probably) has. Otherwise it'll keep the
+			 * gdquot from getting reclaimed.
+			 */
+			if ((gdqp = dqp->q_gdquot)) {
+				/*
+				 * Avoid a recursive dqput call
+				 */
+				xfs_dqlock(gdqp);
+				dqp->q_gdquot = NULL;
+			}
+		}
+		xfs_dqunlock(dqp);
+
+		/*
+		 * If we had a group quota inside the user quota as a hint,
+		 * release it now.
+		 */
+		if (! gdqp)
+			break;
+		dqp = gdqp;
+	}
+	mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+}
+
+/*
+ * Release a dquot. Flush it if dirty, then dqput() it.
+ * dquot must not be locked.
+ */
+void
+xfs_qm_dqrele(
+	xfs_dquot_t	*dqp)
+{
+	if (!dqp)
+		return;
+
+	trace_xfs_dqrele(dqp);
+
+	xfs_dqlock(dqp);
+	/*
+	 * We don't care to flush it if the dquot is dirty here.
+	 * That will create stutters that we want to avoid.
+	 * Instead we do a delayed write when we try to reclaim
+	 * a dirty dquot. Also xfs_sync will take part of the burden...
+	 */
+	xfs_qm_dqput(dqp);
+}
+
+/*
+ * This is the dquot flushing I/O completion routine.  It is called
+ * from interrupt level when the buffer containing the dquot is
+ * flushed to disk.  It is responsible for removing the dquot logitem
+ * from the AIL if it has not been re-logged, and unlocking the dquot's
+ * flush lock. This behavior is very similar to that of inodes..
+ */
+STATIC void
+xfs_qm_dqflush_done(
+	struct xfs_buf		*bp,
+	struct xfs_log_item	*lip)
+{
+	xfs_dq_logitem_t	*qip = (struct xfs_dq_logitem *)lip;
+	xfs_dquot_t		*dqp = qip->qli_dquot;
+	struct xfs_ail		*ailp = lip->li_ailp;
+
+	/*
+	 * We only want to pull the item from the AIL if its
+	 * location in the log has not changed since we started the flush.
+	 * Thus, we only bother if the dquot's lsn has
+	 * not changed. First we check the lsn outside the lock
+	 * since it's cheaper, and then we recheck while
+	 * holding the lock before removing the dquot from the AIL.
+	 */
+	if ((lip->li_flags & XFS_LI_IN_AIL) &&
+	    lip->li_lsn == qip->qli_flush_lsn) {
+
+		/* xfs_trans_ail_delete() drops the AIL lock. */
+		spin_lock(&ailp->xa_lock);
+		if (lip->li_lsn == qip->qli_flush_lsn)
+			xfs_trans_ail_delete(ailp, lip);
+		else
+			spin_unlock(&ailp->xa_lock);
+	}
+
+	/*
+	 * Release the dq's flush lock since we're done with it.
+	 */
+	xfs_dqfunlock(dqp);
+}
+
+/*
+ * Write a modified dquot to disk.
+ * The dquot must be locked and the flush lock too taken by caller.
+ * The flush lock will not be unlocked until the dquot reaches the disk,
+ * but the dquot is free to be unlocked and modified by the caller
+ * in the interim. Dquot is still locked on return. This behavior is
+ * identical to that of inodes.
+ */
+int
+xfs_qm_dqflush(
+	xfs_dquot_t		*dqp,
+	uint			flags)
+{
+	struct xfs_mount	*mp = dqp->q_mount;
+	struct xfs_buf		*bp;
+	struct xfs_disk_dquot	*ddqp;
+	int			error;
+
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+	ASSERT(!completion_done(&dqp->q_flush));
+
+	trace_xfs_dqflush(dqp);
+
+	/*
+	 * If not dirty, or it's pinned and we are not supposed to block, nada.
+	 */
+	if (!XFS_DQ_IS_DIRTY(dqp) ||
+	    (!(flags & SYNC_WAIT) && atomic_read(&dqp->q_pincount) > 0)) {
+		xfs_dqfunlock(dqp);
+		return 0;
+	}
+	xfs_qm_dqunpin_wait(dqp);
+
+	/*
+	 * This may have been unpinned because the filesystem is shutting
+	 * down forcibly. If that's the case we must not write this dquot
+	 * to disk, because the log record didn't make it to disk!
+	 */
+	if (XFS_FORCED_SHUTDOWN(mp)) {
+		dqp->dq_flags &= ~XFS_DQ_DIRTY;
+		xfs_dqfunlock(dqp);
+		return XFS_ERROR(EIO);
+	}
+
+	/*
+	 * Get the buffer containing the on-disk dquot
+	 */
+	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dqp->q_blkno,
+				   mp->m_quotainfo->qi_dqchunklen, 0, &bp);
+	if (error) {
+		ASSERT(error != ENOENT);
+		xfs_dqfunlock(dqp);
+		return error;
+	}
+
+	/*
+	 * Calculate the location of the dquot inside the buffer.
+	 */
+	ddqp = bp->b_addr + dqp->q_bufoffset;
+
+	/*
+	 * A simple sanity check in case we got a corrupted dquot..
+	 */
+	error = xfs_qm_dqcheck(mp, &dqp->q_core, be32_to_cpu(ddqp->d_id), 0,
+			   XFS_QMOPT_DOWARN, "dqflush (incore copy)");
+	if (error) {
+		xfs_buf_relse(bp);
+		xfs_dqfunlock(dqp);
+		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		return XFS_ERROR(EIO);
+	}
+
+	/* This is the only portion of data that needs to persist */
+	memcpy(ddqp, &dqp->q_core, sizeof(xfs_disk_dquot_t));
+
+	/*
+	 * Clear the dirty field and remember the flush lsn for later use.
+	 */
+	dqp->dq_flags &= ~XFS_DQ_DIRTY;
+
+	xfs_trans_ail_copy_lsn(mp->m_ail, &dqp->q_logitem.qli_flush_lsn,
+					&dqp->q_logitem.qli_item.li_lsn);
+
+	/*
+	 * Attach an iodone routine so that we can remove this dquot from the
+	 * AIL and release the flush lock once the dquot is synced to disk.
+	 */
+	xfs_buf_attach_iodone(bp, xfs_qm_dqflush_done,
+				  &dqp->q_logitem.qli_item);
+
+	/*
+	 * If the buffer is pinned then push on the log so we won't
+	 * get stuck waiting in the write for too long.
+	 */
+	if (xfs_buf_ispinned(bp)) {
+		trace_xfs_dqflush_force(dqp);
+		xfs_log_force(mp, 0);
+	}
+
+	if (flags & SYNC_WAIT)
+		error = xfs_bwrite(bp);
+	else
+		xfs_buf_delwri_queue(bp);
+
+	xfs_buf_relse(bp);
+
+	trace_xfs_dqflush_done(dqp);
+
+	/*
+	 * dqp is still locked, but caller is free to unlock it now.
+	 */
+	return error;
+
+}
+
+int
+xfs_qm_dqlock_nowait(
+	xfs_dquot_t *dqp)
+{
+	return mutex_trylock(&dqp->q_qlock);
+}
+
+void
+xfs_dqlock(
+	xfs_dquot_t *dqp)
+{
+	mutex_lock(&dqp->q_qlock);
+}
+
+void
+xfs_dqunlock(
+	xfs_dquot_t *dqp)
+{
+	mutex_unlock(&(dqp->q_qlock));
+	if (dqp->q_logitem.qli_dquot == dqp) {
+		/* Once was dqp->q_mount, but might just have been cleared */
+		xfs_trans_unlocked_item(dqp->q_logitem.qli_item.li_ailp,
+					(xfs_log_item_t*)&(dqp->q_logitem));
+	}
+}
+
+
+void
+xfs_dqunlock_nonotify(
+	xfs_dquot_t *dqp)
+{
+	mutex_unlock(&(dqp->q_qlock));
+}
+
+/*
+ * Lock two xfs_dquot structures.
+ *
+ * To avoid deadlocks we always lock the quota structure with
+ * the lowerd id first.
+ */
+void
+xfs_dqlock2(
+	xfs_dquot_t	*d1,
+	xfs_dquot_t	*d2)
+{
+	if (d1 && d2) {
+		ASSERT(d1 != d2);
+		if (be32_to_cpu(d1->q_core.d_id) >
+		    be32_to_cpu(d2->q_core.d_id)) {
+			mutex_lock(&d2->q_qlock);
+			mutex_lock_nested(&d1->q_qlock, XFS_QLOCK_NESTED);
+		} else {
+			mutex_lock(&d1->q_qlock);
+			mutex_lock_nested(&d2->q_qlock, XFS_QLOCK_NESTED);
+		}
+	} else if (d1) {
+		mutex_lock(&d1->q_qlock);
+	} else if (d2) {
+		mutex_lock(&d2->q_qlock);
+	}
+}
+
+
+/*
+ * Take a dquot out of the mount's dqlist as well as the hashlist.
+ * This is called via unmount as well as quotaoff, and the purge
+ * will always succeed unless there are soft (temp) references
+ * outstanding.
+ *
+ * This returns 0 if it was purged, 1 if it wasn't. It's not an error code
+ * that we're returning! XXXsup - not cool.
+ */
+/* ARGSUSED */
+int
+xfs_qm_dqpurge(
+	xfs_dquot_t	*dqp)
+{
+	xfs_dqhash_t	*qh = dqp->q_hash;
+	xfs_mount_t	*mp = dqp->q_mount;
+
+	ASSERT(mutex_is_locked(&mp->m_quotainfo->qi_dqlist_lock));
+	ASSERT(mutex_is_locked(&dqp->q_hash->qh_lock));
+
+	xfs_dqlock(dqp);
+	/*
+	 * We really can't afford to purge a dquot that is
+	 * referenced, because these are hard refs.
+	 * It shouldn't happen in general because we went thru _all_ inodes in
+	 * dqrele_all_inodes before calling this and didn't let the mountlock go.
+	 * However it is possible that we have dquots with temporary
+	 * references that are not attached to an inode. e.g. see xfs_setattr().
+	 */
+	if (dqp->q_nrefs != 0) {
+		xfs_dqunlock(dqp);
+		mutex_unlock(&dqp->q_hash->qh_lock);
+		return (1);
+	}
+
+	ASSERT(!list_empty(&dqp->q_freelist));
+
+	/*
+	 * If we're turning off quotas, we have to make sure that, for
+	 * example, we don't delete quota disk blocks while dquots are
+	 * in the process of getting written to those disk blocks.
+	 * This dquot might well be on AIL, and we can't leave it there
+	 * if we're turning off quotas. Basically, we need this flush
+	 * lock, and are willing to block on it.
+	 */
+	if (!xfs_dqflock_nowait(dqp)) {
+		/*
+		 * Block on the flush lock after nudging dquot buffer,
+		 * if it is incore.
+		 */
+		xfs_qm_dqflock_pushbuf_wait(dqp);
+	}
+
+	/*
+	 * XXXIf we're turning this type of quotas off, we don't care
+	 * about the dirty metadata sitting in this dquot. OTOH, if
+	 * we're unmounting, we do care, so we flush it and wait.
+	 */
+	if (XFS_DQ_IS_DIRTY(dqp)) {
+		int	error;
+
+		/* dqflush unlocks dqflock */
+		/*
+		 * Given that dqpurge is a very rare occurrence, it is OK
+		 * that we're holding the hashlist and mplist locks
+		 * across the disk write. But, ... XXXsup
+		 *
+		 * We don't care about getting disk errors here. We need
+		 * to purge this dquot anyway, so we go ahead regardless.
+		 */
+		error = xfs_qm_dqflush(dqp, SYNC_WAIT);
+		if (error)
+			xfs_warn(mp, "%s: dquot %p flush failed",
+				__func__, dqp);
+		xfs_dqflock(dqp);
+	}
+	ASSERT(atomic_read(&dqp->q_pincount) == 0);
+	ASSERT(XFS_FORCED_SHUTDOWN(mp) ||
+	       !(dqp->q_logitem.qli_item.li_flags & XFS_LI_IN_AIL));
+
+	list_del_init(&dqp->q_hashlist);
+	qh->qh_version++;
+	list_del_init(&dqp->q_mplist);
+	mp->m_quotainfo->qi_dqreclaims++;
+	mp->m_quotainfo->qi_dquots--;
+	/*
+	 * XXX Move this to the front of the freelist, if we can get the
+	 * freelist lock.
+	 */
+	ASSERT(!list_empty(&dqp->q_freelist));
+
+	dqp->q_mount = NULL;
+	dqp->q_hash = NULL;
+	dqp->dq_flags = XFS_DQ_INACTIVE;
+	memset(&dqp->q_core, 0, sizeof(dqp->q_core));
+	xfs_dqfunlock(dqp);
+	xfs_dqunlock(dqp);
+	mutex_unlock(&qh->qh_lock);
+	return (0);
+}
+
+
+/*
+ * Give the buffer a little push if it is incore and
+ * wait on the flush lock.
+ */
+void
+xfs_qm_dqflock_pushbuf_wait(
+	xfs_dquot_t	*dqp)
+{
+	xfs_mount_t	*mp = dqp->q_mount;
+	xfs_buf_t	*bp;
+
+	/*
+	 * Check to see if the dquot has been flushed delayed
+	 * write.  If so, grab its buffer and send it
+	 * out immediately.  We'll be able to acquire
+	 * the flush lock when the I/O completes.
+	 */
+	bp = xfs_incore(mp->m_ddev_targp, dqp->q_blkno,
+			mp->m_quotainfo->qi_dqchunklen, XBF_TRYLOCK);
+	if (!bp)
+		goto out_lock;
+
+	if (XFS_BUF_ISDELAYWRITE(bp)) {
+		if (xfs_buf_ispinned(bp))
+			xfs_log_force(mp, 0);
+		xfs_buf_delwri_promote(bp);
+		wake_up_process(bp->b_target->bt_task);
+	}
+	xfs_buf_relse(bp);
+out_lock:
+	xfs_dqflock(dqp);
+}
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
new file mode 100644
index 0000000..34b7e94
--- /dev/null
+++ b/fs/xfs/xfs_dquot.h
@@ -0,0 +1,137 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_DQUOT_H__
+#define __XFS_DQUOT_H__
+
+/*
+ * Dquots are structures that hold quota information about a user or a group,
+ * much like inodes are for files. In fact, dquots share many characteristics
+ * with inodes. However, dquots can also be a centralized resource, relative
+ * to a collection of inodes. In this respect, dquots share some characteristics
+ * of the superblock.
+ * XFS dquots exploit both those in its algorithms. They make every attempt
+ * to not be a bottleneck when quotas are on and have minimal impact, if any,
+ * when quotas are off.
+ */
+
+/*
+ * The hash chain headers (hash buckets)
+ */
+typedef struct xfs_dqhash {
+	struct list_head  qh_list;
+	struct mutex	  qh_lock;
+	uint		  qh_version;	/* ever increasing version */
+	uint		  qh_nelems;	/* number of dquots on the list */
+} xfs_dqhash_t;
+
+struct xfs_mount;
+struct xfs_trans;
+
+/*
+ * The incore dquot structure
+ */
+typedef struct xfs_dquot {
+	uint		 dq_flags;	/* various flags (XFS_DQ_*) */
+	struct list_head q_freelist;	/* global free list of dquots */
+	struct list_head q_mplist;	/* mount's list of dquots */
+	struct list_head q_hashlist;	/* gloabl hash list of dquots */
+	xfs_dqhash_t	*q_hash;	/* the hashchain header */
+	struct xfs_mount*q_mount;	/* filesystem this relates to */
+	struct xfs_trans*q_transp;	/* trans this belongs to currently */
+	uint		 q_nrefs;	/* # active refs from inodes */
+	xfs_daddr_t	 q_blkno;	/* blkno of dquot buffer */
+	int		 q_bufoffset;	/* off of dq in buffer (# dquots) */
+	xfs_fileoff_t	 q_fileoffset;	/* offset in quotas file */
+
+	struct xfs_dquot*q_gdquot;	/* group dquot, hint only */
+	xfs_disk_dquot_t q_core;	/* actual usage & quotas */
+	xfs_dq_logitem_t q_logitem;	/* dquot log item */
+	xfs_qcnt_t	 q_res_bcount;	/* total regular nblks used+reserved */
+	xfs_qcnt_t	 q_res_icount;	/* total inos allocd+reserved */
+	xfs_qcnt_t	 q_res_rtbcount;/* total realtime blks used+reserved */
+	struct mutex	 q_qlock;	/* quota lock */
+	struct completion q_flush;	/* flush completion queue */
+	atomic_t          q_pincount;	/* dquot pin count */
+	wait_queue_head_t q_pinwait;	/* dquot pinning wait queue */
+} xfs_dquot_t;
+
+/*
+ * Lock hierarchy for q_qlock:
+ *	XFS_QLOCK_NORMAL is the implicit default,
+ * 	XFS_QLOCK_NESTED is the dquot with the higher id in xfs_dqlock2
+ */
+enum {
+	XFS_QLOCK_NORMAL = 0,
+	XFS_QLOCK_NESTED,
+};
+
+#define XFS_DQHOLD(dqp)		((dqp)->q_nrefs++)
+
+/*
+ * Manage the q_flush completion queue embedded in the dquot.  This completion
+ * queue synchronizes processes attempting to flush the in-core dquot back to
+ * disk.
+ */
+static inline void xfs_dqflock(xfs_dquot_t *dqp)
+{
+	wait_for_completion(&dqp->q_flush);
+}
+
+static inline int xfs_dqflock_nowait(xfs_dquot_t *dqp)
+{
+	return try_wait_for_completion(&dqp->q_flush);
+}
+
+static inline void xfs_dqfunlock(xfs_dquot_t *dqp)
+{
+	complete(&dqp->q_flush);
+}
+
+#define XFS_DQ_IS_LOCKED(dqp)	(mutex_is_locked(&((dqp)->q_qlock)))
+#define XFS_DQ_IS_DIRTY(dqp)	((dqp)->dq_flags & XFS_DQ_DIRTY)
+#define XFS_QM_ISUDQ(dqp)	((dqp)->dq_flags & XFS_DQ_USER)
+#define XFS_QM_ISPDQ(dqp)	((dqp)->dq_flags & XFS_DQ_PROJ)
+#define XFS_QM_ISGDQ(dqp)	((dqp)->dq_flags & XFS_DQ_GROUP)
+#define XFS_DQ_TO_QINF(dqp)	((dqp)->q_mount->m_quotainfo)
+#define XFS_DQ_TO_QIP(dqp)	(XFS_QM_ISUDQ(dqp) ? \
+				 XFS_DQ_TO_QINF(dqp)->qi_uquotaip : \
+				 XFS_DQ_TO_QINF(dqp)->qi_gquotaip)
+
+#define XFS_IS_THIS_QUOTA_OFF(d) (! (XFS_QM_ISUDQ(d) ? \
+				     (XFS_IS_UQUOTA_ON((d)->q_mount)) : \
+				     (XFS_IS_OQUOTA_ON((d)->q_mount))))
+
+extern void		xfs_qm_dqdestroy(xfs_dquot_t *);
+extern int		xfs_qm_dqflush(xfs_dquot_t *, uint);
+extern int		xfs_qm_dqpurge(xfs_dquot_t *);
+extern void		xfs_qm_dqunpin_wait(xfs_dquot_t *);
+extern int		xfs_qm_dqlock_nowait(xfs_dquot_t *);
+extern void		xfs_qm_dqflock_pushbuf_wait(xfs_dquot_t *dqp);
+extern void		xfs_qm_adjust_dqtimers(xfs_mount_t *,
+					xfs_disk_dquot_t *);
+extern void		xfs_qm_adjust_dqlimits(xfs_mount_t *,
+					xfs_disk_dquot_t *);
+extern int		xfs_qm_dqget(xfs_mount_t *, xfs_inode_t *,
+					xfs_dqid_t, uint, uint, xfs_dquot_t **);
+extern void		xfs_qm_dqput(xfs_dquot_t *);
+extern void		xfs_dqlock(xfs_dquot_t *);
+extern void		xfs_dqlock2(xfs_dquot_t *, xfs_dquot_t *);
+extern void		xfs_dqunlock(xfs_dquot_t *);
+extern void		xfs_dqunlock_nonotify(xfs_dquot_t *);
+
+#endif /* __XFS_DQUOT_H__ */
diff --git a/fs/xfs/xfs_dquot_item.c b/fs/xfs/xfs_dquot_item.c
new file mode 100644
index 0000000..0dee0b7
--- /dev/null
+++ b/fs/xfs/xfs_dquot_item.c
@@ -0,0 +1,533 @@
+/*
+ * Copyright (c) 2000-2003 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_itable.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_priv.h"
+#include "xfs_qm.h"
+
+static inline struct xfs_dq_logitem *DQUOT_ITEM(struct xfs_log_item *lip)
+{
+	return container_of(lip, struct xfs_dq_logitem, qli_item);
+}
+
+/*
+ * returns the number of iovecs needed to log the given dquot item.
+ */
+STATIC uint
+xfs_qm_dquot_logitem_size(
+	struct xfs_log_item	*lip)
+{
+	/*
+	 * we need only two iovecs, one for the format, one for the real thing
+	 */
+	return 2;
+}
+
+/*
+ * fills in the vector of log iovecs for the given dquot log item.
+ */
+STATIC void
+xfs_qm_dquot_logitem_format(
+	struct xfs_log_item	*lip,
+	struct xfs_log_iovec	*logvec)
+{
+	struct xfs_dq_logitem	*qlip = DQUOT_ITEM(lip);
+
+	logvec->i_addr = &qlip->qli_format;
+	logvec->i_len  = sizeof(xfs_dq_logformat_t);
+	logvec->i_type = XLOG_REG_TYPE_QFORMAT;
+	logvec++;
+	logvec->i_addr = &qlip->qli_dquot->q_core;
+	logvec->i_len  = sizeof(xfs_disk_dquot_t);
+	logvec->i_type = XLOG_REG_TYPE_DQUOT;
+
+	ASSERT(2 == lip->li_desc->lid_size);
+	qlip->qli_format.qlf_size = 2;
+
+}
+
+/*
+ * Increment the pin count of the given dquot.
+ */
+STATIC void
+xfs_qm_dquot_logitem_pin(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
+
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+	atomic_inc(&dqp->q_pincount);
+}
+
+/*
+ * Decrement the pin count of the given dquot, and wake up
+ * anyone in xfs_dqwait_unpin() if the count goes to 0.	 The
+ * dquot must have been previously pinned with a call to
+ * xfs_qm_dquot_logitem_pin().
+ */
+STATIC void
+xfs_qm_dquot_logitem_unpin(
+	struct xfs_log_item	*lip,
+	int			remove)
+{
+	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
+
+	ASSERT(atomic_read(&dqp->q_pincount) > 0);
+	if (atomic_dec_and_test(&dqp->q_pincount))
+		wake_up(&dqp->q_pinwait);
+}
+
+/*
+ * Given the logitem, this writes the corresponding dquot entry to disk
+ * asynchronously. This is called with the dquot entry securely locked;
+ * we simply get xfs_qm_dqflush() to do the work, and unlock the dquot
+ * at the end.
+ */
+STATIC void
+xfs_qm_dquot_logitem_push(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
+	int			error;
+
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+	ASSERT(!completion_done(&dqp->q_flush));
+
+	/*
+	 * Since we were able to lock the dquot's flush lock and
+	 * we found it on the AIL, the dquot must be dirty.  This
+	 * is because the dquot is removed from the AIL while still
+	 * holding the flush lock in xfs_dqflush_done().  Thus, if
+	 * we found it in the AIL and were able to obtain the flush
+	 * lock without sleeping, then there must not have been
+	 * anyone in the process of flushing the dquot.
+	 */
+	error = xfs_qm_dqflush(dqp, 0);
+	if (error)
+		xfs_warn(dqp->q_mount, "%s: push error %d on dqp %p",
+			__func__, error, dqp);
+	xfs_dqunlock(dqp);
+}
+
+STATIC xfs_lsn_t
+xfs_qm_dquot_logitem_committed(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		lsn)
+{
+	/*
+	 * We always re-log the entire dquot when it becomes dirty,
+	 * so, the latest copy _is_ the only one that matters.
+	 */
+	return lsn;
+}
+
+/*
+ * This is called to wait for the given dquot to be unpinned.
+ * Most of these pin/unpin routines are plagiarized from inode code.
+ */
+void
+xfs_qm_dqunpin_wait(
+	struct xfs_dquot	*dqp)
+{
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+	if (atomic_read(&dqp->q_pincount) == 0)
+		return;
+
+	/*
+	 * Give the log a push so we don't wait here too long.
+	 */
+	xfs_log_force(dqp->q_mount, 0);
+	wait_event(dqp->q_pinwait, (atomic_read(&dqp->q_pincount) == 0));
+}
+
+/*
+ * This is called when IOP_TRYLOCK returns XFS_ITEM_PUSHBUF to indicate that
+ * the dquot is locked by us, but the flush lock isn't. So, here we are
+ * going to see if the relevant dquot buffer is incore, waiting on DELWRI.
+ * If so, we want to push it out to help us take this item off the AIL as soon
+ * as possible.
+ *
+ * We must not be holding the AIL lock at this point. Calling incore() to
+ * search the buffer cache can be a time consuming thing, and AIL lock is a
+ * spinlock.
+ */
+STATIC bool
+xfs_qm_dquot_logitem_pushbuf(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_dq_logitem	*qlip = DQUOT_ITEM(lip);
+	struct xfs_dquot	*dqp = qlip->qli_dquot;
+	struct xfs_buf		*bp;
+	bool			ret = true;
+
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+
+	/*
+	 * If flushlock isn't locked anymore, chances are that the
+	 * inode flush completed and the inode was taken off the AIL.
+	 * So, just get out.
+	 */
+	if (completion_done(&dqp->q_flush) ||
+	    !(lip->li_flags & XFS_LI_IN_AIL)) {
+		xfs_dqunlock(dqp);
+		return true;
+	}
+
+	bp = xfs_incore(dqp->q_mount->m_ddev_targp, qlip->qli_format.qlf_blkno,
+			dqp->q_mount->m_quotainfo->qi_dqchunklen, XBF_TRYLOCK);
+	xfs_dqunlock(dqp);
+	if (!bp)
+		return true;
+	if (XFS_BUF_ISDELAYWRITE(bp))
+		xfs_buf_delwri_promote(bp);
+	if (xfs_buf_ispinned(bp))
+		ret = false;
+	xfs_buf_relse(bp);
+	return ret;
+}
+
+/*
+ * This is called to attempt to lock the dquot associated with this
+ * dquot log item.  Don't sleep on the dquot lock or the flush lock.
+ * If the flush lock is already held, indicating that the dquot has
+ * been or is in the process of being flushed, then see if we can
+ * find the dquot's buffer in the buffer cache without sleeping.  If
+ * we can and it is marked delayed write, then we want to send it out.
+ * We delay doing so until the push routine, though, to avoid sleeping
+ * in any device strategy routines.
+ */
+STATIC uint
+xfs_qm_dquot_logitem_trylock(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
+
+	if (atomic_read(&dqp->q_pincount) > 0)
+		return XFS_ITEM_PINNED;
+
+	if (!xfs_qm_dqlock_nowait(dqp))
+		return XFS_ITEM_LOCKED;
+
+	if (!xfs_dqflock_nowait(dqp)) {
+		/*
+		 * dquot has already been flushed to the backing buffer,
+		 * leave it locked, pushbuf routine will unlock it.
+		 */
+		return XFS_ITEM_PUSHBUF;
+	}
+
+	ASSERT(lip->li_flags & XFS_LI_IN_AIL);
+	return XFS_ITEM_SUCCESS;
+}
+
+/*
+ * Unlock the dquot associated with the log item.
+ * Clear the fields of the dquot and dquot log item that
+ * are specific to the current transaction.  If the
+ * hold flags is set, do not unlock the dquot.
+ */
+STATIC void
+xfs_qm_dquot_logitem_unlock(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
+
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+
+	/*
+	 * Clear the transaction pointer in the dquot
+	 */
+	dqp->q_transp = NULL;
+
+	/*
+	 * dquots are never 'held' from getting unlocked at the end of
+	 * a transaction.  Their locking and unlocking is hidden inside the
+	 * transaction layer, within trans_commit. Hence, no LI_HOLD flag
+	 * for the logitem.
+	 */
+	xfs_dqunlock(dqp);
+}
+
+/*
+ * this needs to stamp an lsn into the dquot, I think.
+ * rpc's that look at user dquot's would then have to
+ * push on the dependency recorded in the dquot
+ */
+STATIC void
+xfs_qm_dquot_logitem_committing(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		lsn)
+{
+}
+
+/*
+ * This is the ops vector for dquots
+ */
+static const struct xfs_item_ops xfs_dquot_item_ops = {
+	.iop_size	= xfs_qm_dquot_logitem_size,
+	.iop_format	= xfs_qm_dquot_logitem_format,
+	.iop_pin	= xfs_qm_dquot_logitem_pin,
+	.iop_unpin	= xfs_qm_dquot_logitem_unpin,
+	.iop_trylock	= xfs_qm_dquot_logitem_trylock,
+	.iop_unlock	= xfs_qm_dquot_logitem_unlock,
+	.iop_committed	= xfs_qm_dquot_logitem_committed,
+	.iop_push	= xfs_qm_dquot_logitem_push,
+	.iop_pushbuf	= xfs_qm_dquot_logitem_pushbuf,
+	.iop_committing = xfs_qm_dquot_logitem_committing
+};
+
+/*
+ * Initialize the dquot log item for a newly allocated dquot.
+ * The dquot isn't locked at this point, but it isn't on any of the lists
+ * either, so we don't care.
+ */
+void
+xfs_qm_dquot_logitem_init(
+	struct xfs_dquot	*dqp)
+{
+	struct xfs_dq_logitem	*lp = &dqp->q_logitem;
+
+	xfs_log_item_init(dqp->q_mount, &lp->qli_item, XFS_LI_DQUOT,
+					&xfs_dquot_item_ops);
+	lp->qli_dquot = dqp;
+	lp->qli_format.qlf_type = XFS_LI_DQUOT;
+	lp->qli_format.qlf_id = be32_to_cpu(dqp->q_core.d_id);
+	lp->qli_format.qlf_blkno = dqp->q_blkno;
+	lp->qli_format.qlf_len = 1;
+	/*
+	 * This is just the offset of this dquot within its buffer
+	 * (which is currently 1 FSB and probably won't change).
+	 * Hence 32 bits for this offset should be just fine.
+	 * Alternatively, we can store (bufoffset / sizeof(xfs_dqblk_t))
+	 * here, and recompute it at recovery time.
+	 */
+	lp->qli_format.qlf_boffset = (__uint32_t)dqp->q_bufoffset;
+}
+
+/*------------------  QUOTAOFF LOG ITEMS  -------------------*/
+
+static inline struct xfs_qoff_logitem *QOFF_ITEM(struct xfs_log_item *lip)
+{
+	return container_of(lip, struct xfs_qoff_logitem, qql_item);
+}
+
+
+/*
+ * This returns the number of iovecs needed to log the given quotaoff item.
+ * We only need 1 iovec for an quotaoff item.  It just logs the
+ * quotaoff_log_format structure.
+ */
+STATIC uint
+xfs_qm_qoff_logitem_size(
+	struct xfs_log_item	*lip)
+{
+	return 1;
+}
+
+/*
+ * This is called to fill in the vector of log iovecs for the
+ * given quotaoff log item. We use only 1 iovec, and we point that
+ * at the quotaoff_log_format structure embedded in the quotaoff item.
+ * It is at this point that we assert that all of the extent
+ * slots in the quotaoff item have been filled.
+ */
+STATIC void
+xfs_qm_qoff_logitem_format(
+	struct xfs_log_item	*lip,
+	struct xfs_log_iovec	*log_vector)
+{
+	struct xfs_qoff_logitem	*qflip = QOFF_ITEM(lip);
+
+	ASSERT(qflip->qql_format.qf_type == XFS_LI_QUOTAOFF);
+
+	log_vector->i_addr = &qflip->qql_format;
+	log_vector->i_len = sizeof(xfs_qoff_logitem_t);
+	log_vector->i_type = XLOG_REG_TYPE_QUOTAOFF;
+	qflip->qql_format.qf_size = 1;
+}
+
+/*
+ * Pinning has no meaning for an quotaoff item, so just return.
+ */
+STATIC void
+xfs_qm_qoff_logitem_pin(
+	struct xfs_log_item	*lip)
+{
+}
+
+/*
+ * Since pinning has no meaning for an quotaoff item, unpinning does
+ * not either.
+ */
+STATIC void
+xfs_qm_qoff_logitem_unpin(
+	struct xfs_log_item	*lip,
+	int			remove)
+{
+}
+
+/*
+ * Quotaoff items have no locking, so just return success.
+ */
+STATIC uint
+xfs_qm_qoff_logitem_trylock(
+	struct xfs_log_item	*lip)
+{
+	return XFS_ITEM_LOCKED;
+}
+
+/*
+ * Quotaoff items have no locking or pushing, so return failure
+ * so that the caller doesn't bother with us.
+ */
+STATIC void
+xfs_qm_qoff_logitem_unlock(
+	struct xfs_log_item	*lip)
+{
+}
+
+/*
+ * The quotaoff-start-item is logged only once and cannot be moved in the log,
+ * so simply return the lsn at which it's been logged.
+ */
+STATIC xfs_lsn_t
+xfs_qm_qoff_logitem_committed(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		lsn)
+{
+	return lsn;
+}
+
+/*
+ * There isn't much you can do to push on an quotaoff item.  It is simply
+ * stuck waiting for the log to be flushed to disk.
+ */
+STATIC void
+xfs_qm_qoff_logitem_push(
+	struct xfs_log_item	*lip)
+{
+}
+
+
+STATIC xfs_lsn_t
+xfs_qm_qoffend_logitem_committed(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		lsn)
+{
+	struct xfs_qoff_logitem	*qfe = QOFF_ITEM(lip);
+	struct xfs_qoff_logitem	*qfs = qfe->qql_start_lip;
+	struct xfs_ail		*ailp = qfs->qql_item.li_ailp;
+
+	/*
+	 * Delete the qoff-start logitem from the AIL.
+	 * xfs_trans_ail_delete() drops the AIL lock.
+	 */
+	spin_lock(&ailp->xa_lock);
+	xfs_trans_ail_delete(ailp, (xfs_log_item_t *)qfs);
+
+	kmem_free(qfs);
+	kmem_free(qfe);
+	return (xfs_lsn_t)-1;
+}
+
+/*
+ * XXX rcc - don't know quite what to do with this.  I think we can
+ * just ignore it.  The only time that isn't the case is if we allow
+ * the client to somehow see that quotas have been turned off in which
+ * we can't allow that to get back until the quotaoff hits the disk.
+ * So how would that happen?  Also, do we need different routines for
+ * quotaoff start and quotaoff end?  I suspect the answer is yes but
+ * to be sure, I need to look at the recovery code and see how quota off
+ * recovery is handled (do we roll forward or back or do something else).
+ * If we roll forwards or backwards, then we need two separate routines,
+ * one that does nothing and one that stamps in the lsn that matters
+ * (truly makes the quotaoff irrevocable).  If we do something else,
+ * then maybe we don't need two.
+ */
+STATIC void
+xfs_qm_qoff_logitem_committing(
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		commit_lsn)
+{
+}
+
+static const struct xfs_item_ops xfs_qm_qoffend_logitem_ops = {
+	.iop_size	= xfs_qm_qoff_logitem_size,
+	.iop_format	= xfs_qm_qoff_logitem_format,
+	.iop_pin	= xfs_qm_qoff_logitem_pin,
+	.iop_unpin	= xfs_qm_qoff_logitem_unpin,
+	.iop_trylock	= xfs_qm_qoff_logitem_trylock,
+	.iop_unlock	= xfs_qm_qoff_logitem_unlock,
+	.iop_committed	= xfs_qm_qoffend_logitem_committed,
+	.iop_push	= xfs_qm_qoff_logitem_push,
+	.iop_committing = xfs_qm_qoff_logitem_committing
+};
+
+/*
+ * This is the ops vector shared by all quotaoff-start log items.
+ */
+static const struct xfs_item_ops xfs_qm_qoff_logitem_ops = {
+	.iop_size	= xfs_qm_qoff_logitem_size,
+	.iop_format	= xfs_qm_qoff_logitem_format,
+	.iop_pin	= xfs_qm_qoff_logitem_pin,
+	.iop_unpin	= xfs_qm_qoff_logitem_unpin,
+	.iop_trylock	= xfs_qm_qoff_logitem_trylock,
+	.iop_unlock	= xfs_qm_qoff_logitem_unlock,
+	.iop_committed	= xfs_qm_qoff_logitem_committed,
+	.iop_push	= xfs_qm_qoff_logitem_push,
+	.iop_committing = xfs_qm_qoff_logitem_committing
+};
+
+/*
+ * Allocate and initialize an quotaoff item of the correct quota type(s).
+ */
+struct xfs_qoff_logitem *
+xfs_qm_qoff_logitem_init(
+	struct xfs_mount	*mp,
+	struct xfs_qoff_logitem	*start,
+	uint			flags)
+{
+	struct xfs_qoff_logitem	*qf;
+
+	qf = kmem_zalloc(sizeof(struct xfs_qoff_logitem), KM_SLEEP);
+
+	xfs_log_item_init(mp, &qf->qql_item, XFS_LI_QUOTAOFF, start ?
+			&xfs_qm_qoffend_logitem_ops : &xfs_qm_qoff_logitem_ops);
+	qf->qql_item.li_mountp = mp;
+	qf->qql_format.qf_type = XFS_LI_QUOTAOFF;
+	qf->qql_format.qf_flags = flags;
+	qf->qql_start_lip = start;
+	return qf;
+}
diff --git a/fs/xfs/xfs_dquot_item.h b/fs/xfs/xfs_dquot_item.h
new file mode 100644
index 0000000..5acae2a
--- /dev/null
+++ b/fs/xfs/xfs_dquot_item.h
@@ -0,0 +1,48 @@
+/*
+ * Copyright (c) 2000-2003 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_DQUOT_ITEM_H__
+#define __XFS_DQUOT_ITEM_H__
+
+struct xfs_dquot;
+struct xfs_trans;
+struct xfs_mount;
+struct xfs_qoff_logitem;
+
+typedef struct xfs_dq_logitem {
+	xfs_log_item_t		 qli_item;	   /* common portion */
+	struct xfs_dquot	*qli_dquot;	   /* dquot ptr */
+	xfs_lsn_t		 qli_flush_lsn;	   /* lsn at last flush */
+	xfs_dq_logformat_t	 qli_format;	   /* logged structure */
+} xfs_dq_logitem_t;
+
+typedef struct xfs_qoff_logitem {
+	xfs_log_item_t		 qql_item;	/* common portion */
+	struct xfs_qoff_logitem *qql_start_lip; /* qoff-start logitem, if any */
+	xfs_qoff_logformat_t	 qql_format;	/* logged structure */
+} xfs_qoff_logitem_t;
+
+
+extern void		   xfs_qm_dquot_logitem_init(struct xfs_dquot *);
+extern xfs_qoff_logitem_t *xfs_qm_qoff_logitem_init(struct xfs_mount *,
+					struct xfs_qoff_logitem *, uint);
+extern xfs_qoff_logitem_t *xfs_trans_get_qoff_item(struct xfs_trans *,
+					struct xfs_qoff_logitem *, uint);
+extern void		   xfs_trans_log_quotaoff_item(struct xfs_trans *,
+					struct xfs_qoff_logitem *);
+
+#endif	/* __XFS_DQUOT_ITEM_H__ */
diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c
new file mode 100644
index 0000000..5703fb8
--- /dev/null
+++ b/fs/xfs/xfs_export.c
@@ -0,0 +1,253 @@
+/*
+ * Copyright (c) 2004-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_types.h"
+#include "xfs_inum.h"
+#include "xfs_log.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_mount.h"
+#include "xfs_export.h"
+#include "xfs_vnodeops.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_inode_item.h"
+#include "xfs_trace.h"
+
+/*
+ * Note that we only accept fileids which are long enough rather than allow
+ * the parent generation number to default to zero.  XFS considers zero a
+ * valid generation number not an invalid/wildcard value.
+ */
+static int xfs_fileid_length(int fileid_type)
+{
+	switch (fileid_type) {
+	case FILEID_INO32_GEN:
+		return 2;
+	case FILEID_INO32_GEN_PARENT:
+		return 4;
+	case FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG:
+		return 3;
+	case FILEID_INO32_GEN_PARENT | XFS_FILEID_TYPE_64FLAG:
+		return 6;
+	}
+	return 255; /* invalid */
+}
+
+STATIC int
+xfs_fs_encode_fh(
+	struct dentry		*dentry,
+	__u32			*fh,
+	int			*max_len,
+	int			connectable)
+{
+	struct fid		*fid = (struct fid *)fh;
+	struct xfs_fid64	*fid64 = (struct xfs_fid64 *)fh;
+	struct inode		*inode = dentry->d_inode;
+	int			fileid_type;
+	int			len;
+
+	/* Directories don't need their parent encoded, they have ".." */
+	if (S_ISDIR(inode->i_mode) || !connectable)
+		fileid_type = FILEID_INO32_GEN;
+	else
+		fileid_type = FILEID_INO32_GEN_PARENT;
+
+	/*
+	 * If the the filesystem may contain 64bit inode numbers, we need
+	 * to use larger file handles that can represent them.
+	 *
+	 * While we only allocate inodes that do not fit into 32 bits any
+	 * large enough filesystem may contain them, thus the slightly
+	 * confusing looking conditional below.
+	 */
+	if (!(XFS_M(inode->i_sb)->m_flags & XFS_MOUNT_SMALL_INUMS) ||
+	    (XFS_M(inode->i_sb)->m_flags & XFS_MOUNT_32BITINODES))
+		fileid_type |= XFS_FILEID_TYPE_64FLAG;
+
+	/*
+	 * Only encode if there is enough space given.  In practice
+	 * this means we can't export a filesystem with 64bit inodes
+	 * over NFSv2 with the subtree_check export option; the other
+	 * seven combinations work.  The real answer is "don't use v2".
+	 */
+	len = xfs_fileid_length(fileid_type);
+	if (*max_len < len) {
+		*max_len = len;
+		return 255;
+	}
+	*max_len = len;
+
+	switch (fileid_type) {
+	case FILEID_INO32_GEN_PARENT:
+		spin_lock(&dentry->d_lock);
+		fid->i32.parent_ino = XFS_I(dentry->d_parent->d_inode)->i_ino;
+		fid->i32.parent_gen = dentry->d_parent->d_inode->i_generation;
+		spin_unlock(&dentry->d_lock);
+		/*FALLTHRU*/
+	case FILEID_INO32_GEN:
+		fid->i32.ino = XFS_I(inode)->i_ino;
+		fid->i32.gen = inode->i_generation;
+		break;
+	case FILEID_INO32_GEN_PARENT | XFS_FILEID_TYPE_64FLAG:
+		spin_lock(&dentry->d_lock);
+		fid64->parent_ino = XFS_I(dentry->d_parent->d_inode)->i_ino;
+		fid64->parent_gen = dentry->d_parent->d_inode->i_generation;
+		spin_unlock(&dentry->d_lock);
+		/*FALLTHRU*/
+	case FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG:
+		fid64->ino = XFS_I(inode)->i_ino;
+		fid64->gen = inode->i_generation;
+		break;
+	}
+
+	return fileid_type;
+}
+
+STATIC struct inode *
+xfs_nfs_get_inode(
+	struct super_block	*sb,
+	u64			ino,
+	u32			generation)
+ {
+ 	xfs_mount_t		*mp = XFS_M(sb);
+	xfs_inode_t		*ip;
+	int			error;
+
+	/*
+	 * NFS can sometimes send requests for ino 0.  Fail them gracefully.
+	 */
+	if (ino == 0)
+		return ERR_PTR(-ESTALE);
+
+	/*
+	 * The XFS_IGET_UNTRUSTED means that an invalid inode number is just
+	 * fine and not an indication of a corrupted filesystem as clients can
+	 * send invalid file handles and we have to handle it gracefully..
+	 */
+	error = xfs_iget(mp, NULL, ino, XFS_IGET_UNTRUSTED, 0, &ip);
+	if (error) {
+		/*
+		 * EINVAL means the inode cluster doesn't exist anymore.
+		 * This implies the filehandle is stale, so we should
+		 * translate it here.
+		 * We don't use ESTALE directly down the chain to not
+		 * confuse applications using bulkstat that expect EINVAL.
+		 */
+		if (error == EINVAL || error == ENOENT)
+			error = ESTALE;
+		return ERR_PTR(-error);
+	}
+
+	if (ip->i_d.di_gen != generation) {
+		IRELE(ip);
+		return ERR_PTR(-ESTALE);
+	}
+
+	return VFS_I(ip);
+}
+
+STATIC struct dentry *
+xfs_fs_fh_to_dentry(struct super_block *sb, struct fid *fid,
+		 int fh_len, int fileid_type)
+{
+	struct xfs_fid64	*fid64 = (struct xfs_fid64 *)fid;
+	struct inode		*inode = NULL;
+
+	if (fh_len < xfs_fileid_length(fileid_type))
+		return NULL;
+
+	switch (fileid_type) {
+	case FILEID_INO32_GEN_PARENT:
+	case FILEID_INO32_GEN:
+		inode = xfs_nfs_get_inode(sb, fid->i32.ino, fid->i32.gen);
+		break;
+	case FILEID_INO32_GEN_PARENT | XFS_FILEID_TYPE_64FLAG:
+	case FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG:
+		inode = xfs_nfs_get_inode(sb, fid64->ino, fid64->gen);
+		break;
+	}
+
+	return d_obtain_alias(inode);
+}
+
+STATIC struct dentry *
+xfs_fs_fh_to_parent(struct super_block *sb, struct fid *fid,
+		 int fh_len, int fileid_type)
+{
+	struct xfs_fid64	*fid64 = (struct xfs_fid64 *)fid;
+	struct inode		*inode = NULL;
+
+	if (fh_len < xfs_fileid_length(fileid_type))
+		return NULL;
+
+	switch (fileid_type) {
+	case FILEID_INO32_GEN_PARENT:
+		inode = xfs_nfs_get_inode(sb, fid->i32.parent_ino,
+					      fid->i32.parent_gen);
+		break;
+	case FILEID_INO32_GEN_PARENT | XFS_FILEID_TYPE_64FLAG:
+		inode = xfs_nfs_get_inode(sb, fid64->parent_ino,
+					      fid64->parent_gen);
+		break;
+	}
+
+	return d_obtain_alias(inode);
+}
+
+STATIC struct dentry *
+xfs_fs_get_parent(
+	struct dentry		*child)
+{
+	int			error;
+	struct xfs_inode	*cip;
+
+	error = xfs_lookup(XFS_I(child->d_inode), &xfs_name_dotdot, &cip, NULL);
+	if (unlikely(error))
+		return ERR_PTR(-error);
+
+	return d_obtain_alias(VFS_I(cip));
+}
+
+STATIC int
+xfs_fs_nfs_commit_metadata(
+	struct inode		*inode)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_lsn_t		lsn = 0;
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	if (xfs_ipincount(ip))
+		lsn = ip->i_itemp->ili_last_lsn;
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (!lsn)
+		return 0;
+	return _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL);
+}
+
+const struct export_operations xfs_export_operations = {
+	.encode_fh		= xfs_fs_encode_fh,
+	.fh_to_dentry		= xfs_fs_fh_to_dentry,
+	.fh_to_parent		= xfs_fs_fh_to_parent,
+	.get_parent		= xfs_fs_get_parent,
+	.commit_metadata	= xfs_fs_nfs_commit_metadata,
+};
diff --git a/fs/xfs/xfs_export.h b/fs/xfs/xfs_export.h
new file mode 100644
index 0000000..3272b6a
--- /dev/null
+++ b/fs/xfs/xfs_export.h
@@ -0,0 +1,72 @@
+/*
+ * Copyright (c) 2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_EXPORT_H__
+#define __XFS_EXPORT_H__
+
+/*
+ * Common defines for code related to exporting XFS filesystems over NFS.
+ *
+ * The NFS fileid goes out on the wire as an array of
+ * 32bit unsigned ints in host order.  There are 5 possible
+ * formats.
+ *
+ * (1)	fileid_type=0x00
+ *	(no fileid data; handled by the generic code)
+ *
+ * (2)	fileid_type=0x01
+ *	inode-num
+ *	generation
+ *
+ * (3)	fileid_type=0x02
+ *	inode-num
+ *	generation
+ *	parent-inode-num
+ *	parent-generation
+ *
+ * (4)	fileid_type=0x81
+ *	inode-num-lo32
+ *	inode-num-hi32
+ *	generation
+ *
+ * (5)	fileid_type=0x82
+ *	inode-num-lo32
+ *	inode-num-hi32
+ *	generation
+ *	parent-inode-num-lo32
+ *	parent-inode-num-hi32
+ *	parent-generation
+ *
+ * Note, the NFS filehandle also includes an fsid portion which
+ * may have an inode number in it.  That number is hardcoded to
+ * 32bits and there is no way for XFS to intercept it.  In
+ * practice this means when exporting an XFS filesystem with 64bit
+ * inodes you should either export the mountpoint (rather than
+ * a subdirectory) or use the "fsid" export option.
+ */
+
+struct xfs_fid64 {
+	u64 ino;
+	u32 gen;
+	u64 parent_ino;
+	u32 parent_gen;
+} __attribute__((packed));
+
+/* This flag goes on the wire.  Don't play with it. */
+#define XFS_FILEID_TYPE_64FLAG	0x80	/* NFS fileid has 64bit inodes */
+
+#endif	/* __XFS_EXPORT_H__ */
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
new file mode 100644
index 0000000..8ae937a
--- /dev/null
+++ b/fs/xfs/xfs_file.c
@@ -0,0 +1,1184 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_trans.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_alloc.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_inode_item.h"
+#include "xfs_bmap.h"
+#include "xfs_error.h"
+#include "xfs_vnodeops.h"
+#include "xfs_da_btree.h"
+#include "xfs_ioctl.h"
+#include "xfs_trace.h"
+
+#include <linux/dcache.h>
+#include <linux/falloc.h>
+
+static const struct vm_operations_struct xfs_file_vm_ops;
+
+/*
+ * Locking primitives for read and write IO paths to ensure we consistently use
+ * and order the inode->i_mutex, ip->i_lock and ip->i_iolock.
+ */
+static inline void
+xfs_rw_ilock(
+	struct xfs_inode	*ip,
+	int			type)
+{
+	if (type & XFS_IOLOCK_EXCL)
+		mutex_lock(&VFS_I(ip)->i_mutex);
+	xfs_ilock(ip, type);
+}
+
+static inline void
+xfs_rw_iunlock(
+	struct xfs_inode	*ip,
+	int			type)
+{
+	xfs_iunlock(ip, type);
+	if (type & XFS_IOLOCK_EXCL)
+		mutex_unlock(&VFS_I(ip)->i_mutex);
+}
+
+static inline void
+xfs_rw_ilock_demote(
+	struct xfs_inode	*ip,
+	int			type)
+{
+	xfs_ilock_demote(ip, type);
+	if (type & XFS_IOLOCK_EXCL)
+		mutex_unlock(&VFS_I(ip)->i_mutex);
+}
+
+/*
+ *	xfs_iozero
+ *
+ *	xfs_iozero clears the specified range of buffer supplied,
+ *	and marks all the affected blocks as valid and modified.  If
+ *	an affected block is not allocated, it will be allocated.  If
+ *	an affected block is not completely overwritten, and is not
+ *	valid before the operation, it will be read from disk before
+ *	being partially zeroed.
+ */
+STATIC int
+xfs_iozero(
+	struct xfs_inode	*ip,	/* inode			*/
+	loff_t			pos,	/* offset in file		*/
+	size_t			count)	/* size of data to zero		*/
+{
+	struct page		*page;
+	struct address_space	*mapping;
+	int			status;
+
+	mapping = VFS_I(ip)->i_mapping;
+	do {
+		unsigned offset, bytes;
+		void *fsdata;
+
+		offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */
+		bytes = PAGE_CACHE_SIZE - offset;
+		if (bytes > count)
+			bytes = count;
+
+		status = pagecache_write_begin(NULL, mapping, pos, bytes,
+					AOP_FLAG_UNINTERRUPTIBLE,
+					&page, &fsdata);
+		if (status)
+			break;
+
+		zero_user(page, offset, bytes);
+
+		status = pagecache_write_end(NULL, mapping, pos, bytes, bytes,
+					page, fsdata);
+		WARN_ON(status <= 0); /* can't return less than zero! */
+		pos += bytes;
+		count -= bytes;
+		status = 0;
+	} while (count);
+
+	return (-status);
+}
+
+/*
+ * Fsync operations on directories are much simpler than on regular files,
+ * as there is no file data to flush, and thus also no need for explicit
+ * cache flush operations, and there are no non-transaction metadata updates
+ * on directories either.
+ */
+STATIC int
+xfs_dir_fsync(
+	struct file		*file,
+	loff_t			start,
+	loff_t			end,
+	int			datasync)
+{
+	struct xfs_inode	*ip = XFS_I(file->f_mapping->host);
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_lsn_t		lsn = 0;
+
+	trace_xfs_dir_fsync(ip);
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	if (xfs_ipincount(ip))
+		lsn = ip->i_itemp->ili_last_lsn;
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (!lsn)
+		return 0;
+	return _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL);
+}
+
+STATIC int
+xfs_file_fsync(
+	struct file		*file,
+	loff_t			start,
+	loff_t			end,
+	int			datasync)
+{
+	struct inode		*inode = file->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	int			error = 0;
+	int			log_flushed = 0;
+	xfs_lsn_t		lsn = 0;
+
+	trace_xfs_file_fsync(ip);
+
+	error = filemap_write_and_wait_range(inode->i_mapping, start, end);
+	if (error)
+		return error;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+
+	if (mp->m_flags & XFS_MOUNT_BARRIER) {
+		/*
+		 * If we have an RT and/or log subvolume we need to make sure
+		 * to flush the write cache the device used for file data
+		 * first.  This is to ensure newly written file data make
+		 * it to disk before logging the new inode size in case of
+		 * an extending write.
+		 */
+		if (XFS_IS_REALTIME_INODE(ip))
+			xfs_blkdev_issue_flush(mp->m_rtdev_targp);
+		else if (mp->m_logdev_targp != mp->m_ddev_targp)
+			xfs_blkdev_issue_flush(mp->m_ddev_targp);
+	}
+
+	/*
+	 * We always need to make sure that the required inode state is safe on
+	 * disk.  The inode might be clean but we still might need to force the
+	 * log because of committed transactions that haven't hit the disk yet.
+	 * Likewise, there could be unflushed non-transactional changes to the
+	 * inode core that have to go to disk and this requires us to issue
+	 * a synchronous transaction to capture these changes correctly.
+	 *
+	 * This code relies on the assumption that if the i_update_core field
+	 * of the inode is clear and the inode is unpinned then it is clean
+	 * and no action is required.
+	 */
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+
+	/*
+	 * First check if the VFS inode is marked dirty.  All the dirtying
+	 * of non-transactional updates no goes through mark_inode_dirty*,
+	 * which allows us to distinguish beteeen pure timestamp updates
+	 * and i_size updates which need to be caught for fdatasync.
+	 * After that also theck for the dirty state in the XFS inode, which
+	 * might gets cleared when the inode gets written out via the AIL
+	 * or xfs_iflush_cluster.
+	 */
+	if (((inode->i_state & I_DIRTY_DATASYNC) ||
+	    ((inode->i_state & I_DIRTY_SYNC) && !datasync)) &&
+	    ip->i_update_core) {
+		/*
+		 * Kick off a transaction to log the inode core to get the
+		 * updates.  The sync transaction will also force the log.
+		 */
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+		tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS);
+		error = xfs_trans_reserve(tp, 0,
+				XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0);
+		if (error) {
+			xfs_trans_cancel(tp, 0);
+			return -error;
+		}
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+		/*
+		 * Note - it's possible that we might have pushed ourselves out
+		 * of the way during trans_reserve which would flush the inode.
+		 * But there's no guarantee that the inode buffer has actually
+		 * gone out yet (it's delwri).	Plus the buffer could be pinned
+		 * anyway if it's part of an inode in another recent
+		 * transaction.	 So we play it safe and fire off the
+		 * transaction anyway.
+		 */
+		xfs_trans_ijoin(tp, ip, 0);
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+		error = xfs_trans_commit(tp, 0);
+
+		lsn = ip->i_itemp->ili_last_lsn;
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	} else {
+		/*
+		 * Timestamps/size haven't changed since last inode flush or
+		 * inode transaction commit.  That means either nothing got
+		 * written or a transaction committed which caught the updates.
+		 * If the latter happened and the transaction hasn't hit the
+		 * disk yet, the inode will be still be pinned.  If it is,
+		 * force the log.
+		 */
+		if (xfs_ipincount(ip))
+			lsn = ip->i_itemp->ili_last_lsn;
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	}
+
+	if (!error && lsn)
+		error = _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, &log_flushed);
+
+	/*
+	 * If we only have a single device, and the log force about was
+	 * a no-op we might have to flush the data device cache here.
+	 * This can only happen for fdatasync/O_DSYNC if we were overwriting
+	 * an already allocated file and thus do not have any metadata to
+	 * commit.
+	 */
+	if ((mp->m_flags & XFS_MOUNT_BARRIER) &&
+	    mp->m_logdev_targp == mp->m_ddev_targp &&
+	    !XFS_IS_REALTIME_INODE(ip) &&
+	    !log_flushed)
+		xfs_blkdev_issue_flush(mp->m_ddev_targp);
+
+	return -error;
+}
+
+STATIC ssize_t
+xfs_file_aio_read(
+	struct kiocb		*iocb,
+	const struct iovec	*iovp,
+	unsigned long		nr_segs,
+	loff_t			pos)
+{
+	struct file		*file = iocb->ki_filp;
+	struct inode		*inode = file->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	size_t			size = 0;
+	ssize_t			ret = 0;
+	int			ioflags = 0;
+	xfs_fsize_t		n;
+	unsigned long		seg;
+
+	XFS_STATS_INC(xs_read_calls);
+
+	BUG_ON(iocb->ki_pos != pos);
+
+	if (unlikely(file->f_flags & O_DIRECT))
+		ioflags |= IO_ISDIRECT;
+	if (file->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	/* START copy & waste from filemap.c */
+	for (seg = 0; seg < nr_segs; seg++) {
+		const struct iovec *iv = &iovp[seg];
+
+		/*
+		 * If any segment has a negative length, or the cumulative
+		 * length ever wraps negative then return -EINVAL.
+		 */
+		size += iv->iov_len;
+		if (unlikely((ssize_t)(size|iv->iov_len) < 0))
+			return XFS_ERROR(-EINVAL);
+	}
+	/* END copy & waste from filemap.c */
+
+	if (unlikely(ioflags & IO_ISDIRECT)) {
+		xfs_buftarg_t	*target =
+			XFS_IS_REALTIME_INODE(ip) ?
+				mp->m_rtdev_targp : mp->m_ddev_targp;
+		if ((iocb->ki_pos & target->bt_smask) ||
+		    (size & target->bt_smask)) {
+			if (iocb->ki_pos == ip->i_size)
+				return 0;
+			return -XFS_ERROR(EINVAL);
+		}
+	}
+
+	n = XFS_MAXIOFFSET(mp) - iocb->ki_pos;
+	if (n <= 0 || size == 0)
+		return 0;
+
+	if (n < size)
+		size = n;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -EIO;
+
+	/*
+	 * Locking is a bit tricky here. If we take an exclusive lock
+	 * for direct IO, we effectively serialise all new concurrent
+	 * read IO to this file and block it behind IO that is currently in
+	 * progress because IO in progress holds the IO lock shared. We only
+	 * need to hold the lock exclusive to blow away the page cache, so
+	 * only take lock exclusively if the page cache needs invalidation.
+	 * This allows the normal direct IO case of no page cache pages to
+	 * proceeed concurrently without serialisation.
+	 */
+	xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
+	if ((ioflags & IO_ISDIRECT) && inode->i_mapping->nrpages) {
+		xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
+		xfs_rw_ilock(ip, XFS_IOLOCK_EXCL);
+
+		if (inode->i_mapping->nrpages) {
+			ret = -xfs_flushinval_pages(ip,
+					(iocb->ki_pos & PAGE_CACHE_MASK),
+					-1, FI_REMAPF_LOCKED);
+			if (ret) {
+				xfs_rw_iunlock(ip, XFS_IOLOCK_EXCL);
+				return ret;
+			}
+		}
+		xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL);
+	}
+
+	trace_xfs_file_read(ip, size, iocb->ki_pos, ioflags);
+
+	ret = generic_file_aio_read(iocb, iovp, nr_segs, iocb->ki_pos);
+	if (ret > 0)
+		XFS_STATS_ADD(xs_read_bytes, ret);
+
+	xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
+	return ret;
+}
+
+STATIC ssize_t
+xfs_file_splice_read(
+	struct file		*infilp,
+	loff_t			*ppos,
+	struct pipe_inode_info	*pipe,
+	size_t			count,
+	unsigned int		flags)
+{
+	struct xfs_inode	*ip = XFS_I(infilp->f_mapping->host);
+	int			ioflags = 0;
+	ssize_t			ret;
+
+	XFS_STATS_INC(xs_read_calls);
+
+	if (infilp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return -EIO;
+
+	xfs_rw_ilock(ip, XFS_IOLOCK_SHARED);
+
+	trace_xfs_file_splice_read(ip, count, *ppos, ioflags);
+
+	ret = generic_file_splice_read(infilp, ppos, pipe, count, flags);
+	if (ret > 0)
+		XFS_STATS_ADD(xs_read_bytes, ret);
+
+	xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED);
+	return ret;
+}
+
+STATIC void
+xfs_aio_write_isize_update(
+	struct inode	*inode,
+	loff_t		*ppos,
+	ssize_t		bytes_written)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	xfs_fsize_t		isize = i_size_read(inode);
+
+	if (bytes_written > 0)
+		XFS_STATS_ADD(xs_write_bytes, bytes_written);
+
+	if (unlikely(bytes_written < 0 && bytes_written != -EFAULT &&
+					*ppos > isize))
+		*ppos = isize;
+
+	if (*ppos > ip->i_size) {
+		xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
+		if (*ppos > ip->i_size)
+			ip->i_size = *ppos;
+		xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
+	}
+}
+
+/*
+ * If this was a direct or synchronous I/O that failed (such as ENOSPC) then
+ * part of the I/O may have been written to disk before the error occurred.  In
+ * this case the on-disk file size may have been adjusted beyond the in-memory
+ * file size and now needs to be truncated back.
+ */
+STATIC void
+xfs_aio_write_newsize_update(
+	struct xfs_inode	*ip,
+	xfs_fsize_t		new_size)
+{
+	if (new_size == ip->i_new_size) {
+		xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
+		if (new_size == ip->i_new_size)
+			ip->i_new_size = 0;
+		if (ip->i_d.di_size > ip->i_size)
+			ip->i_d.di_size = ip->i_size;
+		xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
+	}
+}
+
+/*
+ * xfs_file_splice_write() does not use xfs_rw_ilock() because
+ * generic_file_splice_write() takes the i_mutex itself. This, in theory,
+ * couuld cause lock inversions between the aio_write path and the splice path
+ * if someone is doing concurrent splice(2) based writes and write(2) based
+ * writes to the same inode. The only real way to fix this is to re-implement
+ * the generic code here with correct locking orders.
+ */
+STATIC ssize_t
+xfs_file_splice_write(
+	struct pipe_inode_info	*pipe,
+	struct file		*outfilp,
+	loff_t			*ppos,
+	size_t			count,
+	unsigned int		flags)
+{
+	struct inode		*inode = outfilp->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	xfs_fsize_t		new_size;
+	int			ioflags = 0;
+	ssize_t			ret;
+
+	XFS_STATS_INC(xs_write_calls);
+
+	if (outfilp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return -EIO;
+
+	xfs_ilock(ip, XFS_IOLOCK_EXCL);
+
+	new_size = *ppos + count;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	if (new_size > ip->i_size)
+		ip->i_new_size = new_size;
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	trace_xfs_file_splice_write(ip, count, *ppos, ioflags);
+
+	ret = generic_file_splice_write(pipe, outfilp, ppos, count, flags);
+
+	xfs_aio_write_isize_update(inode, ppos, ret);
+	xfs_aio_write_newsize_update(ip, new_size);
+	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+	return ret;
+}
+
+/*
+ * This routine is called to handle zeroing any space in the last
+ * block of the file that is beyond the EOF.  We do this since the
+ * size is being increased without writing anything to that block
+ * and we don't want anyone to read the garbage on the disk.
+ */
+STATIC int				/* error (positive) */
+xfs_zero_last_block(
+	xfs_inode_t	*ip,
+	xfs_fsize_t	offset,
+	xfs_fsize_t	isize,
+	bool		*did_zeroing)
+{
+	xfs_fileoff_t	last_fsb;
+	xfs_mount_t	*mp = ip->i_mount;
+	int		nimaps;
+	int		zero_offset;
+	int		zero_len;
+	int		error = 0;
+	xfs_bmbt_irec_t	imap;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	zero_offset = XFS_B_FSB_OFFSET(mp, isize);
+	if (zero_offset == 0) {
+		/*
+		 * There are no extra bytes in the last block on disk to
+		 * zero, so return.
+		 */
+		return 0;
+	}
+
+	last_fsb = XFS_B_TO_FSBT(mp, isize);
+	nimaps = 1;
+	error = xfs_bmapi_read(ip, last_fsb, 1, &imap, &nimaps, 0);
+	if (error)
+		return error;
+	ASSERT(nimaps > 0);
+	/*
+	 * If the block underlying isize is just a hole, then there
+	 * is nothing to zero.
+	 */
+	if (imap.br_startblock == HOLESTARTBLOCK) {
+		return 0;
+	}
+	/*
+	 * Zero the part of the last block beyond the EOF, and write it
+	 * out sync.  We need to drop the ilock while we do this so we
+	 * don't deadlock when the buffer cache calls back to us.
+	 */
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	zero_len = mp->m_sb.sb_blocksize - zero_offset;
+	if (isize + zero_len > offset)
+		zero_len = offset - isize;
+	*did_zeroing = true;
+	error = xfs_iozero(ip, isize, zero_len);
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	ASSERT(error >= 0);
+	return error;
+}
+
+/*
+ * Zero any on disk space between the current EOF and the new,
+ * larger EOF.  This handles the normal case of zeroing the remainder
+ * of the last block in the file and the unusual case of zeroing blocks
+ * out beyond the size of the file.  This second case only happens
+ * with fixed size extents and when the system crashes before the inode
+ * size was updated but after blocks were allocated.  If fill is set,
+ * then any holes in the range are filled and zeroed.  If not, the holes
+ * are left alone as holes.
+ */
+
+int					/* error (positive) */
+xfs_zero_eof(
+	xfs_inode_t	*ip,
+	xfs_off_t	offset,		/* starting I/O offset */
+	xfs_fsize_t	isize,		/* current inode size */
+	bool		*did_zeroing)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	xfs_fileoff_t	start_zero_fsb;
+	xfs_fileoff_t	end_zero_fsb;
+	xfs_fileoff_t	zero_count_fsb;
+	xfs_fileoff_t	last_fsb;
+	xfs_fileoff_t	zero_off;
+	xfs_fsize_t	zero_len;
+	int		nimaps;
+	int		error = 0;
+	xfs_bmbt_irec_t	imap;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
+	ASSERT(offset > isize);
+
+	/*
+	 * First handle zeroing the block on which isize resides.
+	 * We only zero a part of that block so it is handled specially.
+	 */
+	error = xfs_zero_last_block(ip, offset, isize, did_zeroing);
+	if (error) {
+		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
+		return error;
+	}
+
+	/*
+	 * Calculate the range between the new size and the old
+	 * where blocks needing to be zeroed may exist.  To get the
+	 * block where the last byte in the file currently resides,
+	 * we need to subtract one from the size and truncate back
+	 * to a block boundary.  We subtract 1 in case the size is
+	 * exactly on a block boundary.
+	 */
+	last_fsb = isize ? XFS_B_TO_FSBT(mp, isize - 1) : (xfs_fileoff_t)-1;
+	start_zero_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)isize);
+	end_zero_fsb = XFS_B_TO_FSBT(mp, offset - 1);
+	ASSERT((xfs_sfiloff_t)last_fsb < (xfs_sfiloff_t)start_zero_fsb);
+	if (last_fsb == end_zero_fsb) {
+		/*
+		 * The size was only incremented on its last block.
+		 * We took care of that above, so just return.
+		 */
+		return 0;
+	}
+
+	ASSERT(start_zero_fsb <= end_zero_fsb);
+	while (start_zero_fsb <= end_zero_fsb) {
+		nimaps = 1;
+		zero_count_fsb = end_zero_fsb - start_zero_fsb + 1;
+		error = xfs_bmapi_read(ip, start_zero_fsb, zero_count_fsb,
+					  &imap, &nimaps, 0);
+		if (error) {
+			ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
+			return error;
+		}
+		ASSERT(nimaps > 0);
+
+		if (imap.br_state == XFS_EXT_UNWRITTEN ||
+		    imap.br_startblock == HOLESTARTBLOCK) {
+			/*
+			 * This loop handles initializing pages that were
+			 * partially initialized by the code below this
+			 * loop. It basically zeroes the part of the page
+			 * that sits on a hole and sets the page as P_HOLE
+			 * and calls remapf if it is a mapped file.
+			 */
+			start_zero_fsb = imap.br_startoff + imap.br_blockcount;
+			ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
+			continue;
+		}
+
+		/*
+		 * There are blocks we need to zero.
+		 * Drop the inode lock while we're doing the I/O.
+		 * We'll still have the iolock to protect us.
+		 */
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+		zero_off = XFS_FSB_TO_B(mp, start_zero_fsb);
+		zero_len = XFS_FSB_TO_B(mp, imap.br_blockcount);
+
+		if ((zero_off + zero_len) > offset)
+			zero_len = offset - zero_off;
+
+		error = xfs_iozero(ip, zero_off, zero_len);
+		if (error) {
+			goto out_lock;
+		}
+
+		*did_zeroing = true;
+		start_zero_fsb = imap.br_startoff + imap.br_blockcount;
+		ASSERT(start_zero_fsb <= (end_zero_fsb + 1));
+
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+	}
+
+	return 0;
+
+out_lock:
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	ASSERT(error >= 0);
+	return error;
+}
+
+/*
+ * Common pre-write limit and setup checks.
+ *
+ * Returns with iolock held according to @iolock.
+ */
+STATIC ssize_t
+xfs_file_aio_write_checks(
+	struct file		*file,
+	loff_t			*pos,
+	size_t			*count,
+	xfs_fsize_t		*new_sizep,
+	int			*iolock)
+{
+	struct inode		*inode = file->f_mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	xfs_fsize_t		new_size;
+	int			error = 0;
+
+	xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
+	*new_sizep = 0;
+restart:
+	error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
+	if (error) {
+		xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
+		*iolock = 0;
+		return error;
+	}
+
+	if (likely(!(file->f_mode & FMODE_NOCMTIME)))
+		file_update_time(file);
+
+	/*
+	 * If the offset is beyond the size of the file, we need to zero any
+	 * blocks that fall between the existing EOF and the start of this
+	 * write. There is no need to issue zeroing if another in-flght IO ends
+	 * at or before this one If zeronig is needed and we are currently
+	 * holding the iolock shared, we need to update it to exclusive which
+	 * involves dropping all locks and relocking to maintain correct locking
+	 * order. If we do this, restart the function to ensure all checks and
+	 * values are still valid.
+	 */
+	if ((ip->i_new_size && *pos > ip->i_new_size) ||
+	    (!ip->i_new_size && *pos > ip->i_size)) {
+		bool	zero = false;
+
+		if (*iolock == XFS_IOLOCK_SHARED) {
+			xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
+			*iolock = XFS_IOLOCK_EXCL;
+			xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
+			goto restart;
+		}
+		error = -xfs_zero_eof(ip, *pos, ip->i_size, &zero);
+	}
+
+	/*
+	 * If this IO extends beyond EOF, we may need to update ip->i_new_size.
+	 * We have already zeroed space beyond EOF (if necessary).  Only update
+	 * ip->i_new_size if this IO ends beyond any other in-flight writes.
+	 */
+	new_size = *pos + *count;
+	if (new_size > ip->i_size) {
+		if (new_size > ip->i_new_size)
+			ip->i_new_size = new_size;
+		*new_sizep = new_size;
+	}
+
+	xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
+	if (error)
+		return error;
+
+	/*
+	 * If we're writing the file then make sure to clear the setuid and
+	 * setgid bits if the process is not being run by root.  This keeps
+	 * people from modifying setuid and setgid binaries.
+	 */
+	return file_remove_suid(file);
+
+}
+
+/*
+ * xfs_file_dio_aio_write - handle direct IO writes
+ *
+ * Lock the inode appropriately to prepare for and issue a direct IO write.
+ * By separating it from the buffered write path we remove all the tricky to
+ * follow locking changes and looping.
+ *
+ * If there are cached pages or we're extending the file, we need IOLOCK_EXCL
+ * until we're sure the bytes at the new EOF have been zeroed and/or the cached
+ * pages are flushed out.
+ *
+ * In most cases the direct IO writes will be done holding IOLOCK_SHARED
+ * allowing them to be done in parallel with reads and other direct IO writes.
+ * However, if the IO is not aligned to filesystem blocks, the direct IO layer
+ * needs to do sub-block zeroing and that requires serialisation against other
+ * direct IOs to the same block. In this case we need to serialise the
+ * submission of the unaligned IOs so that we don't get racing block zeroing in
+ * the dio layer.  To avoid the problem with aio, we also need to wait for
+ * outstanding IOs to complete so that unwritten extent conversion is completed
+ * before we try to map the overlapping block. This is currently implemented by
+ * hitting it with a big hammer (i.e. inode_dio_wait()).
+ *
+ * Returns with locks held indicated by @iolock and errors indicated by
+ * negative return values.
+ */
+STATIC ssize_t
+xfs_file_dio_aio_write(
+	struct kiocb		*iocb,
+	const struct iovec	*iovp,
+	unsigned long		nr_segs,
+	loff_t			pos,
+	size_t			ocount,
+	xfs_fsize_t		*new_size,
+	int			*iolock)
+{
+	struct file		*file = iocb->ki_filp;
+	struct address_space	*mapping = file->f_mapping;
+	struct inode		*inode = mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	ssize_t			ret = 0;
+	size_t			count = ocount;
+	int			unaligned_io = 0;
+	struct xfs_buftarg	*target = XFS_IS_REALTIME_INODE(ip) ?
+					mp->m_rtdev_targp : mp->m_ddev_targp;
+
+	*iolock = 0;
+	if ((pos & target->bt_smask) || (count & target->bt_smask))
+		return -XFS_ERROR(EINVAL);
+
+	if ((pos & mp->m_blockmask) || ((pos + count) & mp->m_blockmask))
+		unaligned_io = 1;
+
+	/*
+	 * We don't need to take an exclusive lock unless there page cache needs
+	 * to be invalidated or unaligned IO is being executed. We don't need to
+	 * consider the EOF extension case here because
+	 * xfs_file_aio_write_checks() will relock the inode as necessary for
+	 * EOF zeroing cases and fill out the new inode size as appropriate.
+	 */
+	if (unaligned_io || mapping->nrpages)
+		*iolock = XFS_IOLOCK_EXCL;
+	else
+		*iolock = XFS_IOLOCK_SHARED;
+	xfs_rw_ilock(ip, *iolock);
+
+	/*
+	 * Recheck if there are cached pages that need invalidate after we got
+	 * the iolock to protect against other threads adding new pages while
+	 * we were waiting for the iolock.
+	 */
+	if (mapping->nrpages && *iolock == XFS_IOLOCK_SHARED) {
+		xfs_rw_iunlock(ip, *iolock);
+		*iolock = XFS_IOLOCK_EXCL;
+		xfs_rw_ilock(ip, *iolock);
+	}
+
+	ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock);
+	if (ret)
+		return ret;
+
+	if (mapping->nrpages) {
+		ret = -xfs_flushinval_pages(ip, (pos & PAGE_CACHE_MASK), -1,
+							FI_REMAPF_LOCKED);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * If we are doing unaligned IO, wait for all other IO to drain,
+	 * otherwise demote the lock if we had to flush cached pages
+	 */
+	if (unaligned_io)
+		inode_dio_wait(inode);
+	else if (*iolock == XFS_IOLOCK_EXCL) {
+		xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL);
+		*iolock = XFS_IOLOCK_SHARED;
+	}
+
+	trace_xfs_file_direct_write(ip, count, iocb->ki_pos, 0);
+	ret = generic_file_direct_write(iocb, iovp,
+			&nr_segs, pos, &iocb->ki_pos, count, ocount);
+
+	/* No fallback to buffered IO on errors for XFS. */
+	ASSERT(ret < 0 || ret == count);
+	return ret;
+}
+
+STATIC ssize_t
+xfs_file_buffered_aio_write(
+	struct kiocb		*iocb,
+	const struct iovec	*iovp,
+	unsigned long		nr_segs,
+	loff_t			pos,
+	size_t			ocount,
+	xfs_fsize_t		*new_size,
+	int			*iolock)
+{
+	struct file		*file = iocb->ki_filp;
+	struct address_space	*mapping = file->f_mapping;
+	struct inode		*inode = mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	ssize_t			ret;
+	int			enospc = 0;
+	size_t			count = ocount;
+
+	*iolock = XFS_IOLOCK_EXCL;
+	xfs_rw_ilock(ip, *iolock);
+
+	ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock);
+	if (ret)
+		return ret;
+
+	/* We can write back this queue in page reclaim */
+	current->backing_dev_info = mapping->backing_dev_info;
+
+write_retry:
+	trace_xfs_file_buffered_write(ip, count, iocb->ki_pos, 0);
+	ret = generic_file_buffered_write(iocb, iovp, nr_segs,
+			pos, &iocb->ki_pos, count, ret);
+	/*
+	 * if we just got an ENOSPC, flush the inode now we aren't holding any
+	 * page locks and retry *once*
+	 */
+	if (ret == -ENOSPC && !enospc) {
+		ret = -xfs_flush_pages(ip, 0, -1, 0, FI_NONE);
+		if (ret)
+			return ret;
+		enospc = 1;
+		goto write_retry;
+	}
+	current->backing_dev_info = NULL;
+	return ret;
+}
+
+STATIC ssize_t
+xfs_file_aio_write(
+	struct kiocb		*iocb,
+	const struct iovec	*iovp,
+	unsigned long		nr_segs,
+	loff_t			pos)
+{
+	struct file		*file = iocb->ki_filp;
+	struct address_space	*mapping = file->f_mapping;
+	struct inode		*inode = mapping->host;
+	struct xfs_inode	*ip = XFS_I(inode);
+	ssize_t			ret;
+	int			iolock;
+	size_t			ocount = 0;
+	xfs_fsize_t		new_size = 0;
+
+	XFS_STATS_INC(xs_write_calls);
+
+	BUG_ON(iocb->ki_pos != pos);
+
+	ret = generic_segment_checks(iovp, &nr_segs, &ocount, VERIFY_READ);
+	if (ret)
+		return ret;
+
+	if (ocount == 0)
+		return 0;
+
+	xfs_wait_for_freeze(ip->i_mount, SB_FREEZE_WRITE);
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return -EIO;
+
+	if (unlikely(file->f_flags & O_DIRECT))
+		ret = xfs_file_dio_aio_write(iocb, iovp, nr_segs, pos,
+						ocount, &new_size, &iolock);
+	else
+		ret = xfs_file_buffered_aio_write(iocb, iovp, nr_segs, pos,
+						ocount, &new_size, &iolock);
+
+	xfs_aio_write_isize_update(inode, &iocb->ki_pos, ret);
+
+	if (ret <= 0)
+		goto out_unlock;
+
+	/* Handle various SYNC-type writes */
+	if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) {
+		loff_t end = pos + ret - 1;
+		int error;
+
+		xfs_rw_iunlock(ip, iolock);
+		error = xfs_file_fsync(file, pos, end,
+				      (file->f_flags & __O_SYNC) ? 0 : 1);
+		xfs_rw_ilock(ip, iolock);
+		if (error)
+			ret = error;
+	}
+
+out_unlock:
+	xfs_aio_write_newsize_update(ip, new_size);
+	xfs_rw_iunlock(ip, iolock);
+	return ret;
+}
+
+STATIC long
+xfs_file_fallocate(
+	struct file	*file,
+	int		mode,
+	loff_t		offset,
+	loff_t		len)
+{
+	struct inode	*inode = file->f_path.dentry->d_inode;
+	long		error;
+	loff_t		new_size = 0;
+	xfs_flock64_t	bf;
+	xfs_inode_t	*ip = XFS_I(inode);
+	int		cmd = XFS_IOC_RESVSP;
+	int		attr_flags = XFS_ATTR_NOLOCK;
+
+	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
+		return -EOPNOTSUPP;
+
+	bf.l_whence = 0;
+	bf.l_start = offset;
+	bf.l_len = len;
+
+	xfs_ilock(ip, XFS_IOLOCK_EXCL);
+
+	if (mode & FALLOC_FL_PUNCH_HOLE)
+		cmd = XFS_IOC_UNRESVSP;
+
+	/* check the new inode size is valid before allocating */
+	if (!(mode & FALLOC_FL_KEEP_SIZE) &&
+	    offset + len > i_size_read(inode)) {
+		new_size = offset + len;
+		error = inode_newsize_ok(inode, new_size);
+		if (error)
+			goto out_unlock;
+	}
+
+	if (file->f_flags & O_DSYNC)
+		attr_flags |= XFS_ATTR_SYNC;
+
+	error = -xfs_change_file_space(ip, cmd, &bf, 0, attr_flags);
+	if (error)
+		goto out_unlock;
+
+	/* Change file size if needed */
+	if (new_size) {
+		struct iattr iattr;
+
+		iattr.ia_valid = ATTR_SIZE;
+		iattr.ia_size = new_size;
+		error = -xfs_setattr_size(ip, &iattr, XFS_ATTR_NOLOCK);
+	}
+
+out_unlock:
+	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+	return error;
+}
+
+
+STATIC int
+xfs_file_open(
+	struct inode	*inode,
+	struct file	*file)
+{
+	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
+		return -EFBIG;
+	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
+		return -EIO;
+	return 0;
+}
+
+STATIC int
+xfs_dir_open(
+	struct inode	*inode,
+	struct file	*file)
+{
+	struct xfs_inode *ip = XFS_I(inode);
+	int		mode;
+	int		error;
+
+	error = xfs_file_open(inode, file);
+	if (error)
+		return error;
+
+	/*
+	 * If there are any blocks, read-ahead block 0 as we're almost
+	 * certain to have the next operation be a read there.
+	 */
+	mode = xfs_ilock_map_shared(ip);
+	if (ip->i_d.di_nextents > 0)
+		xfs_da_reada_buf(NULL, ip, 0, XFS_DATA_FORK);
+	xfs_iunlock(ip, mode);
+	return 0;
+}
+
+STATIC int
+xfs_file_release(
+	struct inode	*inode,
+	struct file	*filp)
+{
+	return -xfs_release(XFS_I(inode));
+}
+
+STATIC int
+xfs_file_readdir(
+	struct file	*filp,
+	void		*dirent,
+	filldir_t	filldir)
+{
+	struct inode	*inode = filp->f_path.dentry->d_inode;
+	xfs_inode_t	*ip = XFS_I(inode);
+	int		error;
+	size_t		bufsize;
+
+	/*
+	 * The Linux API doesn't pass down the total size of the buffer
+	 * we read into down to the filesystem.  With the filldir concept
+	 * it's not needed for correct information, but the XFS dir2 leaf
+	 * code wants an estimate of the buffer size to calculate it's
+	 * readahead window and size the buffers used for mapping to
+	 * physical blocks.
+	 *
+	 * Try to give it an estimate that's good enough, maybe at some
+	 * point we can change the ->readdir prototype to include the
+	 * buffer size.  For now we use the current glibc buffer size.
+	 */
+	bufsize = (size_t)min_t(loff_t, 32768, ip->i_d.di_size);
+
+	error = xfs_readdir(ip, dirent, bufsize,
+				(xfs_off_t *)&filp->f_pos, filldir);
+	if (error)
+		return -error;
+	return 0;
+}
+
+STATIC int
+xfs_file_mmap(
+	struct file	*filp,
+	struct vm_area_struct *vma)
+{
+	vma->vm_ops = &xfs_file_vm_ops;
+	vma->vm_flags |= VM_CAN_NONLINEAR;
+
+	file_accessed(filp);
+	return 0;
+}
+
+/*
+ * mmap()d file has taken write protection fault and is being made
+ * writable. We can set the page state up correctly for a writable
+ * page, which means we can do correct delalloc accounting (ENOSPC
+ * checking!) and unwritten extent mapping.
+ */
+STATIC int
+xfs_vm_page_mkwrite(
+	struct vm_area_struct	*vma,
+	struct vm_fault		*vmf)
+{
+	return block_page_mkwrite(vma, vmf, xfs_get_blocks);
+}
+
+const struct file_operations xfs_file_operations = {
+	.llseek		= generic_file_llseek,
+	.read		= do_sync_read,
+	.write		= do_sync_write,
+	.aio_read	= xfs_file_aio_read,
+	.aio_write	= xfs_file_aio_write,
+	.splice_read	= xfs_file_splice_read,
+	.splice_write	= xfs_file_splice_write,
+	.unlocked_ioctl	= xfs_file_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= xfs_file_compat_ioctl,
+#endif
+	.mmap		= xfs_file_mmap,
+	.open		= xfs_file_open,
+	.release	= xfs_file_release,
+	.fsync		= xfs_file_fsync,
+	.fallocate	= xfs_file_fallocate,
+};
+
+const struct file_operations xfs_dir_file_operations = {
+	.open		= xfs_dir_open,
+	.read		= generic_read_dir,
+	.readdir	= xfs_file_readdir,
+	.llseek		= generic_file_llseek,
+	.unlocked_ioctl	= xfs_file_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= xfs_file_compat_ioctl,
+#endif
+	.fsync		= xfs_dir_fsync,
+};
+
+static const struct vm_operations_struct xfs_file_vm_ops = {
+	.fault		= filemap_fault,
+	.page_mkwrite	= xfs_vm_page_mkwrite,
+};
diff --git a/fs/xfs/xfs_fs_subr.c b/fs/xfs/xfs_fs_subr.c
new file mode 100644
index 0000000..ed88ed1
--- /dev/null
+++ b/fs/xfs/xfs_fs_subr.c
@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2000-2002,2005-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_vnodeops.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_trace.h"
+
+/*
+ * note: all filemap functions return negative error codes. These
+ * need to be inverted before returning to the xfs core functions.
+ */
+void
+xfs_tosspages(
+	xfs_inode_t	*ip,
+	xfs_off_t	first,
+	xfs_off_t	last,
+	int		fiopt)
+{
+	/* can't toss partial tail pages, so mask them out */
+	last &= ~(PAGE_SIZE - 1);
+	truncate_inode_pages_range(VFS_I(ip)->i_mapping, first, last - 1);
+}
+
+int
+xfs_flushinval_pages(
+	xfs_inode_t	*ip,
+	xfs_off_t	first,
+	xfs_off_t	last,
+	int		fiopt)
+{
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
+	int		ret = 0;
+
+	trace_xfs_pagecache_inval(ip, first, last);
+
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	ret = filemap_write_and_wait_range(mapping, first,
+				last == -1 ? LLONG_MAX : last);
+	if (!ret)
+		truncate_inode_pages_range(mapping, first, last);
+	return -ret;
+}
+
+int
+xfs_flush_pages(
+	xfs_inode_t	*ip,
+	xfs_off_t	first,
+	xfs_off_t	last,
+	uint64_t	flags,
+	int		fiopt)
+{
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
+	int		ret = 0;
+	int		ret2;
+
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	ret = -filemap_fdatawrite_range(mapping, first,
+				last == -1 ? LLONG_MAX : last);
+	if (flags & XBF_ASYNC)
+		return ret;
+	ret2 = xfs_wait_on_pages(ip, first, last);
+	if (!ret)
+		ret = ret2;
+	return ret;
+}
+
+int
+xfs_wait_on_pages(
+	xfs_inode_t	*ip,
+	xfs_off_t	first,
+	xfs_off_t	last)
+{
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
+
+	if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+		return -filemap_fdatawait_range(mapping, first,
+					last == -1 ? ip->i_size - 1 : last);
+	}
+	return 0;
+}
diff --git a/fs/xfs/xfs_globals.c b/fs/xfs/xfs_globals.c
new file mode 100644
index 0000000..76e81cf
--- /dev/null
+++ b/fs/xfs/xfs_globals.c
@@ -0,0 +1,43 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_sysctl.h"
+
+/*
+ * Tunable XFS parameters.  xfs_params is required even when CONFIG_SYSCTL=n,
+ * other XFS code uses these values.  Times are measured in centisecs (i.e.
+ * 100ths of a second).
+ */
+xfs_param_t xfs_params = {
+			  /*	MIN		DFLT		MAX	*/
+	.sgid_inherit	= {	0,		0,		1	},
+	.symlink_mode	= {	0,		0,		1	},
+	.panic_mask	= {	0,		0,		255	},
+	.error_level	= {	0,		3,		11	},
+	.syncd_timer	= {	1*100,		30*100,		7200*100},
+	.stats_clear	= {	0,		0,		1	},
+	.inherit_sync	= {	0,		1,		1	},
+	.inherit_nodump	= {	0,		1,		1	},
+	.inherit_noatim = {	0,		1,		1	},
+	.xfs_buf_timer	= {	100/2,		1*100,		30*100	},
+	.xfs_buf_age	= {	1*100,		15*100,		7200*100},
+	.inherit_nosym	= {	0,		0,		1	},
+	.rotorstep	= {	1,		1,		255	},
+	.inherit_nodfrg	= {	0,		1,		1	},
+	.fstrm_timer	= {	1,		30*100,		3600*100},
+};
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
new file mode 100644
index 0000000..eb519de
--- /dev/null
+++ b/fs/xfs/xfs_ioctl.c
@@ -0,0 +1,1557 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_ioctl.h"
+#include "xfs_rtalloc.h"
+#include "xfs_itable.h"
+#include "xfs_error.h"
+#include "xfs_attr.h"
+#include "xfs_bmap.h"
+#include "xfs_buf_item.h"
+#include "xfs_utils.h"
+#include "xfs_dfrag.h"
+#include "xfs_fsops.h"
+#include "xfs_vnodeops.h"
+#include "xfs_discard.h"
+#include "xfs_quota.h"
+#include "xfs_inode_item.h"
+#include "xfs_export.h"
+#include "xfs_trace.h"
+
+#include <linux/capability.h>
+#include <linux/dcache.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/pagemap.h>
+#include <linux/slab.h>
+#include <linux/exportfs.h>
+
+/*
+ * xfs_find_handle maps from userspace xfs_fsop_handlereq structure to
+ * a file or fs handle.
+ *
+ * XFS_IOC_PATH_TO_FSHANDLE
+ *    returns fs handle for a mount point or path within that mount point
+ * XFS_IOC_FD_TO_HANDLE
+ *    returns full handle for a FD opened in user space
+ * XFS_IOC_PATH_TO_HANDLE
+ *    returns full handle for a path
+ */
+int
+xfs_find_handle(
+	unsigned int		cmd,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	int			hsize;
+	xfs_handle_t		handle;
+	struct inode		*inode;
+	struct file		*file = NULL;
+	struct path		path;
+	int			error;
+	struct xfs_inode	*ip;
+
+	if (cmd == XFS_IOC_FD_TO_HANDLE) {
+		file = fget(hreq->fd);
+		if (!file)
+			return -EBADF;
+		inode = file->f_path.dentry->d_inode;
+	} else {
+		error = user_lpath((const char __user *)hreq->path, &path);
+		if (error)
+			return error;
+		inode = path.dentry->d_inode;
+	}
+	ip = XFS_I(inode);
+
+	/*
+	 * We can only generate handles for inodes residing on a XFS filesystem,
+	 * and only for regular files, directories or symbolic links.
+	 */
+	error = -EINVAL;
+	if (inode->i_sb->s_magic != XFS_SB_MAGIC)
+		goto out_put;
+
+	error = -EBADF;
+	if (!S_ISREG(inode->i_mode) &&
+	    !S_ISDIR(inode->i_mode) &&
+	    !S_ISLNK(inode->i_mode))
+		goto out_put;
+
+
+	memcpy(&handle.ha_fsid, ip->i_mount->m_fixedfsid, sizeof(xfs_fsid_t));
+
+	if (cmd == XFS_IOC_PATH_TO_FSHANDLE) {
+		/*
+		 * This handle only contains an fsid, zero the rest.
+		 */
+		memset(&handle.ha_fid, 0, sizeof(handle.ha_fid));
+		hsize = sizeof(xfs_fsid_t);
+	} else {
+		int		lock_mode;
+
+		lock_mode = xfs_ilock_map_shared(ip);
+		handle.ha_fid.fid_len = sizeof(xfs_fid_t) -
+					sizeof(handle.ha_fid.fid_len);
+		handle.ha_fid.fid_pad = 0;
+		handle.ha_fid.fid_gen = ip->i_d.di_gen;
+		handle.ha_fid.fid_ino = ip->i_ino;
+		xfs_iunlock_map_shared(ip, lock_mode);
+
+		hsize = XFS_HSIZE(handle);
+	}
+
+	error = -EFAULT;
+	if (copy_to_user(hreq->ohandle, &handle, hsize) ||
+	    copy_to_user(hreq->ohandlen, &hsize, sizeof(__s32)))
+		goto out_put;
+
+	error = 0;
+
+ out_put:
+	if (cmd == XFS_IOC_FD_TO_HANDLE)
+		fput(file);
+	else
+		path_put(&path);
+	return error;
+}
+
+/*
+ * No need to do permission checks on the various pathname components
+ * as the handle operations are privileged.
+ */
+STATIC int
+xfs_handle_acceptable(
+	void			*context,
+	struct dentry		*dentry)
+{
+	return 1;
+}
+
+/*
+ * Convert userspace handle data into a dentry.
+ */
+struct dentry *
+xfs_handle_to_dentry(
+	struct file		*parfilp,
+	void __user		*uhandle,
+	u32			hlen)
+{
+	xfs_handle_t		handle;
+	struct xfs_fid64	fid;
+
+	/*
+	 * Only allow handle opens under a directory.
+	 */
+	if (!S_ISDIR(parfilp->f_path.dentry->d_inode->i_mode))
+		return ERR_PTR(-ENOTDIR);
+
+	if (hlen != sizeof(xfs_handle_t))
+		return ERR_PTR(-EINVAL);
+	if (copy_from_user(&handle, uhandle, hlen))
+		return ERR_PTR(-EFAULT);
+	if (handle.ha_fid.fid_len !=
+	    sizeof(handle.ha_fid) - sizeof(handle.ha_fid.fid_len))
+		return ERR_PTR(-EINVAL);
+
+	memset(&fid, 0, sizeof(struct fid));
+	fid.ino = handle.ha_fid.fid_ino;
+	fid.gen = handle.ha_fid.fid_gen;
+
+	return exportfs_decode_fh(parfilp->f_path.mnt, (struct fid *)&fid, 3,
+			FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG,
+			xfs_handle_acceptable, NULL);
+}
+
+STATIC struct dentry *
+xfs_handlereq_to_dentry(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	return xfs_handle_to_dentry(parfilp, hreq->ihandle, hreq->ihandlen);
+}
+
+int
+xfs_open_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	const struct cred	*cred = current_cred();
+	int			error;
+	int			fd;
+	int			permflag;
+	struct file		*filp;
+	struct inode		*inode;
+	struct dentry		*dentry;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+
+	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+	inode = dentry->d_inode;
+
+	/* Restrict xfs_open_by_handle to directories & regular files. */
+	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))) {
+		error = -XFS_ERROR(EPERM);
+		goto out_dput;
+	}
+
+#if BITS_PER_LONG != 32
+	hreq->oflags |= O_LARGEFILE;
+#endif
+
+	/* Put open permission in namei format. */
+	permflag = hreq->oflags;
+	if ((permflag+1) & O_ACCMODE)
+		permflag++;
+	if (permflag & O_TRUNC)
+		permflag |= 2;
+
+	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
+	    (permflag & FMODE_WRITE) && IS_APPEND(inode)) {
+		error = -XFS_ERROR(EPERM);
+		goto out_dput;
+	}
+
+	if ((permflag & FMODE_WRITE) && IS_IMMUTABLE(inode)) {
+		error = -XFS_ERROR(EACCES);
+		goto out_dput;
+	}
+
+	/* Can't write directories. */
+	if (S_ISDIR(inode->i_mode) && (permflag & FMODE_WRITE)) {
+		error = -XFS_ERROR(EISDIR);
+		goto out_dput;
+	}
+
+	fd = get_unused_fd();
+	if (fd < 0) {
+		error = fd;
+		goto out_dput;
+	}
+
+	filp = dentry_open(dentry, mntget(parfilp->f_path.mnt),
+			   hreq->oflags, cred);
+	if (IS_ERR(filp)) {
+		put_unused_fd(fd);
+		return PTR_ERR(filp);
+	}
+
+	if (S_ISREG(inode->i_mode)) {
+		filp->f_flags |= O_NOATIME;
+		filp->f_mode |= FMODE_NOCMTIME;
+	}
+
+	fd_install(fd, filp);
+	return fd;
+
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+/*
+ * This is a copy from fs/namei.c:vfs_readlink(), except for removing it's
+ * unused first argument.
+ */
+STATIC int
+do_readlink(
+	char __user		*buffer,
+	int			buflen,
+	const char		*link)
+{
+        int len;
+
+	len = PTR_ERR(link);
+	if (IS_ERR(link))
+		goto out;
+
+	len = strlen(link);
+	if (len > (unsigned) buflen)
+		len = buflen;
+	if (copy_to_user(buffer, link, len))
+		len = -EFAULT;
+ out:
+	return len;
+}
+
+
+int
+xfs_readlink_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq)
+{
+	struct dentry		*dentry;
+	__u32			olen;
+	void			*link;
+	int			error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+
+	dentry = xfs_handlereq_to_dentry(parfilp, hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	/* Restrict this handle operation to symlinks only. */
+	if (!S_ISLNK(dentry->d_inode->i_mode)) {
+		error = -XFS_ERROR(EINVAL);
+		goto out_dput;
+	}
+
+	if (copy_from_user(&olen, hreq->ohandlen, sizeof(__u32))) {
+		error = -XFS_ERROR(EFAULT);
+		goto out_dput;
+	}
+
+	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
+	if (!link) {
+		error = -XFS_ERROR(ENOMEM);
+		goto out_dput;
+	}
+
+	error = -xfs_readlink(XFS_I(dentry->d_inode), link);
+	if (error)
+		goto out_kfree;
+	error = do_readlink(hreq->ohandle, olen, link);
+	if (error)
+		goto out_kfree;
+
+ out_kfree:
+	kfree(link);
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+STATIC int
+xfs_fssetdm_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error;
+	struct fsdmidata	fsd;
+	xfs_fsop_setdm_handlereq_t dmhreq;
+	struct dentry		*dentry;
+
+	if (!capable(CAP_MKNOD))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&dmhreq, arg, sizeof(xfs_fsop_setdm_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &dmhreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	if (IS_IMMUTABLE(dentry->d_inode) || IS_APPEND(dentry->d_inode)) {
+		error = -XFS_ERROR(EPERM);
+		goto out;
+	}
+
+	if (copy_from_user(&fsd, dmhreq.data, sizeof(fsd))) {
+		error = -XFS_ERROR(EFAULT);
+		goto out;
+	}
+
+	error = -xfs_set_dmattrs(XFS_I(dentry->d_inode), fsd.fsd_dmevmask,
+				 fsd.fsd_dmstate);
+
+ out:
+	dput(dentry);
+	return error;
+}
+
+STATIC int
+xfs_attrlist_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error = -ENOMEM;
+	attrlist_cursor_kern_t	*cursor;
+	xfs_fsop_attrlist_handlereq_t al_hreq;
+	struct dentry		*dentry;
+	char			*kbuf;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&al_hreq, arg, sizeof(xfs_fsop_attrlist_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+	if (al_hreq.buflen < sizeof(struct attrlist) ||
+	    al_hreq.buflen > XATTR_LIST_MAX)
+		return -XFS_ERROR(EINVAL);
+
+	/*
+	 * Reject flags, only allow namespaces.
+	 */
+	if (al_hreq.flags & ~(ATTR_ROOT | ATTR_SECURE))
+		return -XFS_ERROR(EINVAL);
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &al_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	kbuf = kzalloc(al_hreq.buflen, GFP_KERNEL);
+	if (!kbuf)
+		goto out_dput;
+
+	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
+	error = -xfs_attr_list(XFS_I(dentry->d_inode), kbuf, al_hreq.buflen,
+					al_hreq.flags, cursor);
+	if (error)
+		goto out_kfree;
+
+	if (copy_to_user(al_hreq.buffer, kbuf, al_hreq.buflen))
+		error = -EFAULT;
+
+ out_kfree:
+	kfree(kbuf);
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+int
+xfs_attrmulti_attr_get(
+	struct inode		*inode,
+	unsigned char		*name,
+	unsigned char		__user *ubuf,
+	__uint32_t		*len,
+	__uint32_t		flags)
+{
+	unsigned char		*kbuf;
+	int			error = EFAULT;
+
+	if (*len > XATTR_SIZE_MAX)
+		return EINVAL;
+	kbuf = kmalloc(*len, GFP_KERNEL);
+	if (!kbuf)
+		return ENOMEM;
+
+	error = xfs_attr_get(XFS_I(inode), name, kbuf, (int *)len, flags);
+	if (error)
+		goto out_kfree;
+
+	if (copy_to_user(ubuf, kbuf, *len))
+		error = EFAULT;
+
+ out_kfree:
+	kfree(kbuf);
+	return error;
+}
+
+int
+xfs_attrmulti_attr_set(
+	struct inode		*inode,
+	unsigned char		*name,
+	const unsigned char	__user *ubuf,
+	__uint32_t		len,
+	__uint32_t		flags)
+{
+	unsigned char		*kbuf;
+	int			error = EFAULT;
+
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
+		return EPERM;
+	if (len > XATTR_SIZE_MAX)
+		return EINVAL;
+
+	kbuf = memdup_user(ubuf, len);
+	if (IS_ERR(kbuf))
+		return PTR_ERR(kbuf);
+
+	error = xfs_attr_set(XFS_I(inode), name, kbuf, len, flags);
+
+	return error;
+}
+
+int
+xfs_attrmulti_attr_remove(
+	struct inode		*inode,
+	unsigned char		*name,
+	__uint32_t		flags)
+{
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
+		return EPERM;
+	return xfs_attr_remove(XFS_I(inode), name, flags);
+}
+
+STATIC int
+xfs_attrmulti_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error;
+	xfs_attr_multiop_t	*ops;
+	xfs_fsop_attrmulti_handlereq_t am_hreq;
+	struct dentry		*dentry;
+	unsigned int		i, size;
+	unsigned char		*attr_name;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&am_hreq, arg, sizeof(xfs_fsop_attrmulti_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	/* overflow check */
+	if (am_hreq.opcount >= INT_MAX / sizeof(xfs_attr_multiop_t))
+		return -E2BIG;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &am_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = E2BIG;
+	size = am_hreq.opcount * sizeof(xfs_attr_multiop_t);
+	if (!size || size > 16 * PAGE_SIZE)
+		goto out_dput;
+
+	ops = memdup_user(am_hreq.ops, size);
+	if (IS_ERR(ops)) {
+		error = PTR_ERR(ops);
+		goto out_dput;
+	}
+
+	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
+	if (!attr_name)
+		goto out_kfree_ops;
+
+	error = 0;
+	for (i = 0; i < am_hreq.opcount; i++) {
+		ops[i].am_error = strncpy_from_user((char *)attr_name,
+				ops[i].am_attrname, MAXNAMELEN);
+		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
+			error = -ERANGE;
+		if (ops[i].am_error < 0)
+			break;
+
+		switch (ops[i].am_opcode) {
+		case ATTR_OP_GET:
+			ops[i].am_error = xfs_attrmulti_attr_get(
+					dentry->d_inode, attr_name,
+					ops[i].am_attrvalue, &ops[i].am_length,
+					ops[i].am_flags);
+			break;
+		case ATTR_OP_SET:
+			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
+			if (ops[i].am_error)
+				break;
+			ops[i].am_error = xfs_attrmulti_attr_set(
+					dentry->d_inode, attr_name,
+					ops[i].am_attrvalue, ops[i].am_length,
+					ops[i].am_flags);
+			mnt_drop_write(parfilp->f_path.mnt);
+			break;
+		case ATTR_OP_REMOVE:
+			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
+			if (ops[i].am_error)
+				break;
+			ops[i].am_error = xfs_attrmulti_attr_remove(
+					dentry->d_inode, attr_name,
+					ops[i].am_flags);
+			mnt_drop_write(parfilp->f_path.mnt);
+			break;
+		default:
+			ops[i].am_error = EINVAL;
+		}
+	}
+
+	if (copy_to_user(am_hreq.ops, ops, size))
+		error = XFS_ERROR(EFAULT);
+
+	kfree(attr_name);
+ out_kfree_ops:
+	kfree(ops);
+ out_dput:
+	dput(dentry);
+	return -error;
+}
+
+int
+xfs_ioc_space(
+	struct xfs_inode	*ip,
+	struct inode		*inode,
+	struct file		*filp,
+	int			ioflags,
+	unsigned int		cmd,
+	xfs_flock64_t		*bf)
+{
+	int			attr_flags = 0;
+	int			error;
+
+	/*
+	 * Only allow the sys admin to reserve space unless
+	 * unwritten extents are enabled.
+	 */
+	if (!xfs_sb_version_hasextflgbit(&ip->i_mount->m_sb) &&
+	    !capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+
+	if (inode->i_flags & (S_IMMUTABLE|S_APPEND))
+		return -XFS_ERROR(EPERM);
+
+	if (!(filp->f_mode & FMODE_WRITE))
+		return -XFS_ERROR(EBADF);
+
+	if (!S_ISREG(inode->i_mode))
+		return -XFS_ERROR(EINVAL);
+
+	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
+		attr_flags |= XFS_ATTR_NONBLOCK;
+
+	if (filp->f_flags & O_DSYNC)
+		attr_flags |= XFS_ATTR_SYNC;
+
+	if (ioflags & IO_INVIS)
+		attr_flags |= XFS_ATTR_DMI;
+
+	error = xfs_change_file_space(ip, cmd, bf, filp->f_pos, attr_flags);
+	return -error;
+}
+
+STATIC int
+xfs_ioc_bulkstat(
+	xfs_mount_t		*mp,
+	unsigned int		cmd,
+	void			__user *arg)
+{
+	xfs_fsop_bulkreq_t	bulkreq;
+	int			count;	/* # of records returned */
+	xfs_ino_t		inlast;	/* last inode number */
+	int			done;
+	int			error;
+
+	/* done = 1 if there are more stats to get and if bulkstat */
+	/* should be called again (unused here, but used in dmapi) */
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	if (copy_from_user(&bulkreq, arg, sizeof(xfs_fsop_bulkreq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	if (copy_from_user(&inlast, bulkreq.lastip, sizeof(__s64)))
+		return -XFS_ERROR(EFAULT);
+
+	if ((count = bulkreq.icount) <= 0)
+		return -XFS_ERROR(EINVAL);
+
+	if (bulkreq.ubuffer == NULL)
+		return -XFS_ERROR(EINVAL);
+
+	if (cmd == XFS_IOC_FSINUMBERS)
+		error = xfs_inumbers(mp, &inlast, &count,
+					bulkreq.ubuffer, xfs_inumbers_fmt);
+	else if (cmd == XFS_IOC_FSBULKSTAT_SINGLE)
+		error = xfs_bulkstat_single(mp, &inlast,
+						bulkreq.ubuffer, &done);
+	else	/* XFS_IOC_FSBULKSTAT */
+		error = xfs_bulkstat(mp, &inlast, &count, xfs_bulkstat_one,
+				     sizeof(xfs_bstat_t), bulkreq.ubuffer,
+				     &done);
+
+	if (error)
+		return -error;
+
+	if (bulkreq.ocount != NULL) {
+		if (copy_to_user(bulkreq.lastip, &inlast,
+						sizeof(xfs_ino_t)))
+			return -XFS_ERROR(EFAULT);
+
+		if (copy_to_user(bulkreq.ocount, &count, sizeof(count)))
+			return -XFS_ERROR(EFAULT);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_ioc_fsgeometry_v1(
+	xfs_mount_t		*mp,
+	void			__user *arg)
+{
+	xfs_fsop_geom_t         fsgeo;
+	int			error;
+
+	error = xfs_fs_geometry(mp, &fsgeo, 3);
+	if (error)
+		return -error;
+
+	/*
+	 * Caller should have passed an argument of type
+	 * xfs_fsop_geom_v1_t.  This is a proper subset of the
+	 * xfs_fsop_geom_t that xfs_fs_geometry() fills in.
+	 */
+	if (copy_to_user(arg, &fsgeo, sizeof(xfs_fsop_geom_v1_t)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_ioc_fsgeometry(
+	xfs_mount_t		*mp,
+	void			__user *arg)
+{
+	xfs_fsop_geom_t		fsgeo;
+	int			error;
+
+	error = xfs_fs_geometry(mp, &fsgeo, 4);
+	if (error)
+		return -error;
+
+	if (copy_to_user(arg, &fsgeo, sizeof(fsgeo)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+/*
+ * Linux extended inode flags interface.
+ */
+
+STATIC unsigned int
+xfs_merge_ioc_xflags(
+	unsigned int	flags,
+	unsigned int	start)
+{
+	unsigned int	xflags = start;
+
+	if (flags & FS_IMMUTABLE_FL)
+		xflags |= XFS_XFLAG_IMMUTABLE;
+	else
+		xflags &= ~XFS_XFLAG_IMMUTABLE;
+	if (flags & FS_APPEND_FL)
+		xflags |= XFS_XFLAG_APPEND;
+	else
+		xflags &= ~XFS_XFLAG_APPEND;
+	if (flags & FS_SYNC_FL)
+		xflags |= XFS_XFLAG_SYNC;
+	else
+		xflags &= ~XFS_XFLAG_SYNC;
+	if (flags & FS_NOATIME_FL)
+		xflags |= XFS_XFLAG_NOATIME;
+	else
+		xflags &= ~XFS_XFLAG_NOATIME;
+	if (flags & FS_NODUMP_FL)
+		xflags |= XFS_XFLAG_NODUMP;
+	else
+		xflags &= ~XFS_XFLAG_NODUMP;
+
+	return xflags;
+}
+
+STATIC unsigned int
+xfs_di2lxflags(
+	__uint16_t	di_flags)
+{
+	unsigned int	flags = 0;
+
+	if (di_flags & XFS_DIFLAG_IMMUTABLE)
+		flags |= FS_IMMUTABLE_FL;
+	if (di_flags & XFS_DIFLAG_APPEND)
+		flags |= FS_APPEND_FL;
+	if (di_flags & XFS_DIFLAG_SYNC)
+		flags |= FS_SYNC_FL;
+	if (di_flags & XFS_DIFLAG_NOATIME)
+		flags |= FS_NOATIME_FL;
+	if (di_flags & XFS_DIFLAG_NODUMP)
+		flags |= FS_NODUMP_FL;
+	return flags;
+}
+
+STATIC int
+xfs_ioc_fsgetxattr(
+	xfs_inode_t		*ip,
+	int			attr,
+	void			__user *arg)
+{
+	struct fsxattr		fa;
+
+	memset(&fa, 0, sizeof(struct fsxattr));
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	fa.fsx_xflags = xfs_ip2xflags(ip);
+	fa.fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog;
+	fa.fsx_projid = xfs_get_projid(ip);
+
+	if (attr) {
+		if (ip->i_afp) {
+			if (ip->i_afp->if_flags & XFS_IFEXTENTS)
+				fa.fsx_nextents = ip->i_afp->if_bytes /
+							sizeof(xfs_bmbt_rec_t);
+			else
+				fa.fsx_nextents = ip->i_d.di_anextents;
+		} else
+			fa.fsx_nextents = 0;
+	} else {
+		if (ip->i_df.if_flags & XFS_IFEXTENTS)
+			fa.fsx_nextents = ip->i_df.if_bytes /
+						sizeof(xfs_bmbt_rec_t);
+		else
+			fa.fsx_nextents = ip->i_d.di_nextents;
+	}
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (copy_to_user(arg, &fa, sizeof(fa)))
+		return -EFAULT;
+	return 0;
+}
+
+STATIC void
+xfs_set_diflags(
+	struct xfs_inode	*ip,
+	unsigned int		xflags)
+{
+	unsigned int		di_flags;
+
+	/* can't set PREALLOC this way, just preserve it */
+	di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
+	if (xflags & XFS_XFLAG_IMMUTABLE)
+		di_flags |= XFS_DIFLAG_IMMUTABLE;
+	if (xflags & XFS_XFLAG_APPEND)
+		di_flags |= XFS_DIFLAG_APPEND;
+	if (xflags & XFS_XFLAG_SYNC)
+		di_flags |= XFS_DIFLAG_SYNC;
+	if (xflags & XFS_XFLAG_NOATIME)
+		di_flags |= XFS_DIFLAG_NOATIME;
+	if (xflags & XFS_XFLAG_NODUMP)
+		di_flags |= XFS_DIFLAG_NODUMP;
+	if (xflags & XFS_XFLAG_PROJINHERIT)
+		di_flags |= XFS_DIFLAG_PROJINHERIT;
+	if (xflags & XFS_XFLAG_NODEFRAG)
+		di_flags |= XFS_DIFLAG_NODEFRAG;
+	if (xflags & XFS_XFLAG_FILESTREAM)
+		di_flags |= XFS_DIFLAG_FILESTREAM;
+	if (S_ISDIR(ip->i_d.di_mode)) {
+		if (xflags & XFS_XFLAG_RTINHERIT)
+			di_flags |= XFS_DIFLAG_RTINHERIT;
+		if (xflags & XFS_XFLAG_NOSYMLINKS)
+			di_flags |= XFS_DIFLAG_NOSYMLINKS;
+		if (xflags & XFS_XFLAG_EXTSZINHERIT)
+			di_flags |= XFS_DIFLAG_EXTSZINHERIT;
+	} else if (S_ISREG(ip->i_d.di_mode)) {
+		if (xflags & XFS_XFLAG_REALTIME)
+			di_flags |= XFS_DIFLAG_REALTIME;
+		if (xflags & XFS_XFLAG_EXTSIZE)
+			di_flags |= XFS_DIFLAG_EXTSIZE;
+	}
+
+	ip->i_d.di_flags = di_flags;
+}
+
+STATIC void
+xfs_diflags_to_linux(
+	struct xfs_inode	*ip)
+{
+	struct inode		*inode = VFS_I(ip);
+	unsigned int		xflags = xfs_ip2xflags(ip);
+
+	if (xflags & XFS_XFLAG_IMMUTABLE)
+		inode->i_flags |= S_IMMUTABLE;
+	else
+		inode->i_flags &= ~S_IMMUTABLE;
+	if (xflags & XFS_XFLAG_APPEND)
+		inode->i_flags |= S_APPEND;
+	else
+		inode->i_flags &= ~S_APPEND;
+	if (xflags & XFS_XFLAG_SYNC)
+		inode->i_flags |= S_SYNC;
+	else
+		inode->i_flags &= ~S_SYNC;
+	if (xflags & XFS_XFLAG_NOATIME)
+		inode->i_flags |= S_NOATIME;
+	else
+		inode->i_flags &= ~S_NOATIME;
+}
+
+#define FSX_PROJID	1
+#define FSX_EXTSIZE	2
+#define FSX_XFLAGS	4
+#define FSX_NONBLOCK	8
+
+STATIC int
+xfs_ioctl_setattr(
+	xfs_inode_t		*ip,
+	struct fsxattr		*fa,
+	int			mask)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	unsigned int		lock_flags = 0;
+	struct xfs_dquot	*udqp = NULL;
+	struct xfs_dquot	*gdqp = NULL;
+	struct xfs_dquot	*olddquot = NULL;
+	int			code;
+
+	trace_xfs_ioctl_setattr(ip);
+
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
+		return XFS_ERROR(EROFS);
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	/*
+	 * Disallow 32bit project ids when projid32bit feature is not enabled.
+	 */
+	if ((mask & FSX_PROJID) && (fa->fsx_projid > (__uint16_t)-1) &&
+			!xfs_sb_version_hasprojid32bit(&ip->i_mount->m_sb))
+		return XFS_ERROR(EINVAL);
+
+	/*
+	 * If disk quotas is on, we make sure that the dquots do exist on disk,
+	 * before we start any other transactions. Trying to do this later
+	 * is messy. We don't care to take a readlock to look at the ids
+	 * in inode here, because we can't hold it across the trans_reserve.
+	 * If the IDs do change before we take the ilock, we're covered
+	 * because the i_*dquot fields will get updated anyway.
+	 */
+	if (XFS_IS_QUOTA_ON(mp) && (mask & FSX_PROJID)) {
+		code = xfs_qm_vop_dqalloc(ip, ip->i_d.di_uid,
+					 ip->i_d.di_gid, fa->fsx_projid,
+					 XFS_QMOPT_PQUOTA, &udqp, &gdqp);
+		if (code)
+			return code;
+	}
+
+	/*
+	 * For the other attributes, we acquire the inode lock and
+	 * first do an error checking pass.
+	 */
+	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
+	code = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
+	if (code)
+		goto error_return;
+
+	lock_flags = XFS_ILOCK_EXCL;
+	xfs_ilock(ip, lock_flags);
+
+	/*
+	 * CAP_FOWNER overrides the following restrictions:
+	 *
+	 * The user ID of the calling process must be equal
+	 * to the file owner ID, except in cases where the
+	 * CAP_FSETID capability is applicable.
+	 */
+	if (current_fsuid() != ip->i_d.di_uid && !capable(CAP_FOWNER)) {
+		code = XFS_ERROR(EPERM);
+		goto error_return;
+	}
+
+	/*
+	 * Do a quota reservation only if projid is actually going to change.
+	 */
+	if (mask & FSX_PROJID) {
+		if (XFS_IS_QUOTA_RUNNING(mp) &&
+		    XFS_IS_PQUOTA_ON(mp) &&
+		    xfs_get_projid(ip) != fa->fsx_projid) {
+			ASSERT(tp);
+			code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
+						capable(CAP_FOWNER) ?
+						XFS_QMOPT_FORCE_RES : 0);
+			if (code)	/* out of quota */
+				goto error_return;
+		}
+	}
+
+	if (mask & FSX_EXTSIZE) {
+		/*
+		 * Can't change extent size if any extents are allocated.
+		 */
+		if (ip->i_d.di_nextents &&
+		    ((ip->i_d.di_extsize << mp->m_sb.sb_blocklog) !=
+		     fa->fsx_extsize)) {
+			code = XFS_ERROR(EINVAL);	/* EFBIG? */
+			goto error_return;
+		}
+
+		/*
+		 * Extent size must be a multiple of the appropriate block
+		 * size, if set at all. It must also be smaller than the
+		 * maximum extent size supported by the filesystem.
+		 *
+		 * Also, for non-realtime files, limit the extent size hint to
+		 * half the size of the AGs in the filesystem so alignment
+		 * doesn't result in extents larger than an AG.
+		 */
+		if (fa->fsx_extsize != 0) {
+			xfs_extlen_t    size;
+			xfs_fsblock_t   extsize_fsb;
+
+			extsize_fsb = XFS_B_TO_FSB(mp, fa->fsx_extsize);
+			if (extsize_fsb > MAXEXTLEN) {
+				code = XFS_ERROR(EINVAL);
+				goto error_return;
+			}
+
+			if (XFS_IS_REALTIME_INODE(ip) ||
+			    ((mask & FSX_XFLAGS) &&
+			    (fa->fsx_xflags & XFS_XFLAG_REALTIME))) {
+				size = mp->m_sb.sb_rextsize <<
+				       mp->m_sb.sb_blocklog;
+			} else {
+				size = mp->m_sb.sb_blocksize;
+				if (extsize_fsb > mp->m_sb.sb_agblocks / 2) {
+					code = XFS_ERROR(EINVAL);
+					goto error_return;
+				}
+			}
+
+			if (fa->fsx_extsize % size) {
+				code = XFS_ERROR(EINVAL);
+				goto error_return;
+			}
+		}
+	}
+
+
+	if (mask & FSX_XFLAGS) {
+		/*
+		 * Can't change realtime flag if any extents are allocated.
+		 */
+		if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
+		    (XFS_IS_REALTIME_INODE(ip)) !=
+		    (fa->fsx_xflags & XFS_XFLAG_REALTIME)) {
+			code = XFS_ERROR(EINVAL);	/* EFBIG? */
+			goto error_return;
+		}
+
+		/*
+		 * If realtime flag is set then must have realtime data.
+		 */
+		if ((fa->fsx_xflags & XFS_XFLAG_REALTIME)) {
+			if ((mp->m_sb.sb_rblocks == 0) ||
+			    (mp->m_sb.sb_rextsize == 0) ||
+			    (ip->i_d.di_extsize % mp->m_sb.sb_rextsize)) {
+				code = XFS_ERROR(EINVAL);
+				goto error_return;
+			}
+		}
+
+		/*
+		 * Can't modify an immutable/append-only file unless
+		 * we have appropriate permission.
+		 */
+		if ((ip->i_d.di_flags &
+				(XFS_DIFLAG_IMMUTABLE|XFS_DIFLAG_APPEND) ||
+		     (fa->fsx_xflags &
+				(XFS_XFLAG_IMMUTABLE | XFS_XFLAG_APPEND))) &&
+		    !capable(CAP_LINUX_IMMUTABLE)) {
+			code = XFS_ERROR(EPERM);
+			goto error_return;
+		}
+	}
+
+	xfs_trans_ijoin(tp, ip, 0);
+
+	/*
+	 * Change file ownership.  Must be the owner or privileged.
+	 */
+	if (mask & FSX_PROJID) {
+		/*
+		 * CAP_FSETID overrides the following restrictions:
+		 *
+		 * The set-user-ID and set-group-ID bits of a file will be
+		 * cleared upon successful return from chown()
+		 */
+		if ((ip->i_d.di_mode & (S_ISUID|S_ISGID)) &&
+		    !capable(CAP_FSETID))
+			ip->i_d.di_mode &= ~(S_ISUID|S_ISGID);
+
+		/*
+		 * Change the ownerships and register quota modifications
+		 * in the transaction.
+		 */
+		if (xfs_get_projid(ip) != fa->fsx_projid) {
+			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_PQUOTA_ON(mp)) {
+				olddquot = xfs_qm_vop_chown(tp, ip,
+							&ip->i_gdquot, gdqp);
+			}
+			xfs_set_projid(ip, fa->fsx_projid);
+
+			/*
+			 * We may have to rev the inode as well as
+			 * the superblock version number since projids didn't
+			 * exist before DINODE_VERSION_2 and SB_VERSION_NLINK.
+			 */
+			if (ip->i_d.di_version == 1)
+				xfs_bump_ino_vers2(tp, ip);
+		}
+
+	}
+
+	if (mask & FSX_EXTSIZE)
+		ip->i_d.di_extsize = fa->fsx_extsize >> mp->m_sb.sb_blocklog;
+	if (mask & FSX_XFLAGS) {
+		xfs_set_diflags(ip, fa->fsx_xflags);
+		xfs_diflags_to_linux(ip);
+	}
+
+	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	XFS_STATS_INC(xs_ig_attrchg);
+
+	/*
+	 * If this is a synchronous mount, make sure that the
+	 * transaction goes to disk before returning to the user.
+	 * This is slightly sub-optimal in that truncates require
+	 * two sync transactions instead of one for wsync filesystems.
+	 * One for the truncate and one for the timestamps since we
+	 * don't want to change the timestamps unless we're sure the
+	 * truncate worked.  Truncates are less than 1% of the laddis
+	 * mix so this probably isn't worth the trouble to optimize.
+	 */
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
+		xfs_trans_set_sync(tp);
+	code = xfs_trans_commit(tp, 0);
+	xfs_iunlock(ip, lock_flags);
+
+	/*
+	 * Release any dquot(s) the inode had kept before chown.
+	 */
+	xfs_qm_dqrele(olddquot);
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+
+	return code;
+
+ error_return:
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+	xfs_trans_cancel(tp, 0);
+	if (lock_flags)
+		xfs_iunlock(ip, lock_flags);
+	return code;
+}
+
+STATIC int
+xfs_ioc_fssetxattr(
+	xfs_inode_t		*ip,
+	struct file		*filp,
+	void			__user *arg)
+{
+	struct fsxattr		fa;
+	unsigned int		mask;
+
+	if (copy_from_user(&fa, arg, sizeof(fa)))
+		return -EFAULT;
+
+	mask = FSX_XFLAGS | FSX_EXTSIZE | FSX_PROJID;
+	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
+		mask |= FSX_NONBLOCK;
+
+	return -xfs_ioctl_setattr(ip, &fa, mask);
+}
+
+STATIC int
+xfs_ioc_getxflags(
+	xfs_inode_t		*ip,
+	void			__user *arg)
+{
+	unsigned int		flags;
+
+	flags = xfs_di2lxflags(ip->i_d.di_flags);
+	if (copy_to_user(arg, &flags, sizeof(flags)))
+		return -EFAULT;
+	return 0;
+}
+
+STATIC int
+xfs_ioc_setxflags(
+	xfs_inode_t		*ip,
+	struct file		*filp,
+	void			__user *arg)
+{
+	struct fsxattr		fa;
+	unsigned int		flags;
+	unsigned int		mask;
+
+	if (copy_from_user(&flags, arg, sizeof(flags)))
+		return -EFAULT;
+
+	if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
+		      FS_NOATIME_FL | FS_NODUMP_FL | \
+		      FS_SYNC_FL))
+		return -EOPNOTSUPP;
+
+	mask = FSX_XFLAGS;
+	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
+		mask |= FSX_NONBLOCK;
+	fa.fsx_xflags = xfs_merge_ioc_xflags(flags, xfs_ip2xflags(ip));
+
+	return -xfs_ioctl_setattr(ip, &fa, mask);
+}
+
+STATIC int
+xfs_getbmap_format(void **ap, struct getbmapx *bmv, int *full)
+{
+	struct getbmap __user	*base = *ap;
+
+	/* copy only getbmap portion (not getbmapx) */
+	if (copy_to_user(base, bmv, sizeof(struct getbmap)))
+		return XFS_ERROR(EFAULT);
+
+	*ap += sizeof(struct getbmap);
+	return 0;
+}
+
+STATIC int
+xfs_ioc_getbmap(
+	struct xfs_inode	*ip,
+	int			ioflags,
+	unsigned int		cmd,
+	void			__user *arg)
+{
+	struct getbmapx		bmx;
+	int			error;
+
+	if (copy_from_user(&bmx, arg, sizeof(struct getbmapx)))
+		return -XFS_ERROR(EFAULT);
+
+	if (bmx.bmv_count < 2)
+		return -XFS_ERROR(EINVAL);
+
+	bmx.bmv_iflags = (cmd == XFS_IOC_GETBMAPA ? BMV_IF_ATTRFORK : 0);
+	if (ioflags & IO_INVIS)
+		bmx.bmv_iflags |= BMV_IF_NO_DMAPI_READ;
+
+	error = xfs_getbmap(ip, &bmx, xfs_getbmap_format,
+			    (struct getbmap *)arg+1);
+	if (error)
+		return -error;
+
+	/* copy back header - only size of getbmap */
+	if (copy_to_user(arg, &bmx, sizeof(struct getbmap)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_getbmapx_format(void **ap, struct getbmapx *bmv, int *full)
+{
+	struct getbmapx __user	*base = *ap;
+
+	if (copy_to_user(base, bmv, sizeof(struct getbmapx)))
+		return XFS_ERROR(EFAULT);
+
+	*ap += sizeof(struct getbmapx);
+	return 0;
+}
+
+STATIC int
+xfs_ioc_getbmapx(
+	struct xfs_inode	*ip,
+	void			__user *arg)
+{
+	struct getbmapx		bmx;
+	int			error;
+
+	if (copy_from_user(&bmx, arg, sizeof(bmx)))
+		return -XFS_ERROR(EFAULT);
+
+	if (bmx.bmv_count < 2)
+		return -XFS_ERROR(EINVAL);
+
+	if (bmx.bmv_iflags & (~BMV_IF_VALID))
+		return -XFS_ERROR(EINVAL);
+
+	error = xfs_getbmap(ip, &bmx, xfs_getbmapx_format,
+			    (struct getbmapx *)arg+1);
+	if (error)
+		return -error;
+
+	/* copy back header */
+	if (copy_to_user(arg, &bmx, sizeof(struct getbmapx)))
+		return -XFS_ERROR(EFAULT);
+
+	return 0;
+}
+
+/*
+ * Note: some of the ioctl's return positive numbers as a
+ * byte count indicating success, such as readlink_by_handle.
+ * So we don't "sign flip" like most other routines.  This means
+ * true errors need to be returned as a negative value.
+ */
+long
+xfs_file_ioctl(
+	struct file		*filp,
+	unsigned int		cmd,
+	unsigned long		p)
+{
+	struct inode		*inode = filp->f_path.dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	void			__user *arg = (void __user *)p;
+	int			ioflags = 0;
+	int			error;
+
+	if (filp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	trace_xfs_file_ioctl(ip);
+
+	switch (cmd) {
+	case FITRIM:
+		return xfs_ioc_trim(mp, arg);
+	case XFS_IOC_ALLOCSP:
+	case XFS_IOC_FREESP:
+	case XFS_IOC_RESVSP:
+	case XFS_IOC_UNRESVSP:
+	case XFS_IOC_ALLOCSP64:
+	case XFS_IOC_FREESP64:
+	case XFS_IOC_RESVSP64:
+	case XFS_IOC_UNRESVSP64:
+	case XFS_IOC_ZERO_RANGE: {
+		xfs_flock64_t		bf;
+
+		if (copy_from_user(&bf, arg, sizeof(bf)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, &bf);
+	}
+	case XFS_IOC_DIOINFO: {
+		struct dioattr	da;
+		xfs_buftarg_t	*target =
+			XFS_IS_REALTIME_INODE(ip) ?
+			mp->m_rtdev_targp : mp->m_ddev_targp;
+
+		da.d_mem = da.d_miniosz = 1 << target->bt_sshift;
+		da.d_maxiosz = INT_MAX & ~(da.d_miniosz - 1);
+
+		if (copy_to_user(arg, &da, sizeof(da)))
+			return -XFS_ERROR(EFAULT);
+		return 0;
+	}
+
+	case XFS_IOC_FSBULKSTAT_SINGLE:
+	case XFS_IOC_FSBULKSTAT:
+	case XFS_IOC_FSINUMBERS:
+		return xfs_ioc_bulkstat(mp, cmd, arg);
+
+	case XFS_IOC_FSGEOMETRY_V1:
+		return xfs_ioc_fsgeometry_v1(mp, arg);
+
+	case XFS_IOC_FSGEOMETRY:
+		return xfs_ioc_fsgeometry(mp, arg);
+
+	case XFS_IOC_GETVERSION:
+		return put_user(inode->i_generation, (int __user *)arg);
+
+	case XFS_IOC_FSGETXATTR:
+		return xfs_ioc_fsgetxattr(ip, 0, arg);
+	case XFS_IOC_FSGETXATTRA:
+		return xfs_ioc_fsgetxattr(ip, 1, arg);
+	case XFS_IOC_FSSETXATTR:
+		return xfs_ioc_fssetxattr(ip, filp, arg);
+	case XFS_IOC_GETXFLAGS:
+		return xfs_ioc_getxflags(ip, arg);
+	case XFS_IOC_SETXFLAGS:
+		return xfs_ioc_setxflags(ip, filp, arg);
+
+	case XFS_IOC_FSSETDM: {
+		struct fsdmidata	dmi;
+
+		if (copy_from_user(&dmi, arg, sizeof(dmi)))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_set_dmattrs(ip, dmi.fsd_dmevmask,
+				dmi.fsd_dmstate);
+		return -error;
+	}
+
+	case XFS_IOC_GETBMAP:
+	case XFS_IOC_GETBMAPA:
+		return xfs_ioc_getbmap(ip, ioflags, cmd, arg);
+
+	case XFS_IOC_GETBMAPX:
+		return xfs_ioc_getbmapx(ip, arg);
+
+	case XFS_IOC_FD_TO_HANDLE:
+	case XFS_IOC_PATH_TO_HANDLE:
+	case XFS_IOC_PATH_TO_FSHANDLE: {
+		xfs_fsop_handlereq_t	hreq;
+
+		if (copy_from_user(&hreq, arg, sizeof(hreq)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_find_handle(cmd, &hreq);
+	}
+	case XFS_IOC_OPEN_BY_HANDLE: {
+		xfs_fsop_handlereq_t	hreq;
+
+		if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_open_by_handle(filp, &hreq);
+	}
+	case XFS_IOC_FSSETDM_BY_HANDLE:
+		return xfs_fssetdm_by_handle(filp, arg);
+
+	case XFS_IOC_READLINK_BY_HANDLE: {
+		xfs_fsop_handlereq_t	hreq;
+
+		if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_readlink_by_handle(filp, &hreq);
+	}
+	case XFS_IOC_ATTRLIST_BY_HANDLE:
+		return xfs_attrlist_by_handle(filp, arg);
+
+	case XFS_IOC_ATTRMULTI_BY_HANDLE:
+		return xfs_attrmulti_by_handle(filp, arg);
+
+	case XFS_IOC_SWAPEXT: {
+		struct xfs_swapext	sxp;
+
+		if (copy_from_user(&sxp, arg, sizeof(xfs_swapext_t)))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_swapext(&sxp);
+		return -error;
+	}
+
+	case XFS_IOC_FSCOUNTS: {
+		xfs_fsop_counts_t out;
+
+		error = xfs_fs_counts(mp, &out);
+		if (error)
+			return -error;
+
+		if (copy_to_user(arg, &out, sizeof(out)))
+			return -XFS_ERROR(EFAULT);
+		return 0;
+	}
+
+	case XFS_IOC_SET_RESBLKS: {
+		xfs_fsop_resblks_t inout;
+		__uint64_t	   in;
+
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		if (mp->m_flags & XFS_MOUNT_RDONLY)
+			return -XFS_ERROR(EROFS);
+
+		if (copy_from_user(&inout, arg, sizeof(inout)))
+			return -XFS_ERROR(EFAULT);
+
+		/* input parameter is passed in resblks field of structure */
+		in = inout.resblks;
+		error = xfs_reserve_blocks(mp, &in, &inout);
+		if (error)
+			return -error;
+
+		if (copy_to_user(arg, &inout, sizeof(inout)))
+			return -XFS_ERROR(EFAULT);
+		return 0;
+	}
+
+	case XFS_IOC_GET_RESBLKS: {
+		xfs_fsop_resblks_t out;
+
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		error = xfs_reserve_blocks(mp, NULL, &out);
+		if (error)
+			return -error;
+
+		if (copy_to_user(arg, &out, sizeof(out)))
+			return -XFS_ERROR(EFAULT);
+
+		return 0;
+	}
+
+	case XFS_IOC_FSGROWFSDATA: {
+		xfs_growfs_data_t in;
+
+		if (copy_from_user(&in, arg, sizeof(in)))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_growfs_data(mp, &in);
+		return -error;
+	}
+
+	case XFS_IOC_FSGROWFSLOG: {
+		xfs_growfs_log_t in;
+
+		if (copy_from_user(&in, arg, sizeof(in)))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_growfs_log(mp, &in);
+		return -error;
+	}
+
+	case XFS_IOC_FSGROWFSRT: {
+		xfs_growfs_rt_t in;
+
+		if (copy_from_user(&in, arg, sizeof(in)))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_growfs_rt(mp, &in);
+		return -error;
+	}
+
+	case XFS_IOC_GOINGDOWN: {
+		__uint32_t in;
+
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		if (get_user(in, (__uint32_t __user *)arg))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_fs_goingdown(mp, in);
+		return -error;
+	}
+
+	case XFS_IOC_ERROR_INJECTION: {
+		xfs_error_injection_t in;
+
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		if (copy_from_user(&in, arg, sizeof(in)))
+			return -XFS_ERROR(EFAULT);
+
+		error = xfs_errortag_add(in.errtag, mp);
+		return -error;
+	}
+
+	case XFS_IOC_ERROR_CLEARALL:
+		if (!capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		error = xfs_errortag_clearall(mp, 1);
+		return -error;
+
+	default:
+		return -ENOTTY;
+	}
+}
diff --git a/fs/xfs/xfs_ioctl.h b/fs/xfs/xfs_ioctl.h
new file mode 100644
index 0000000..d56173b
--- /dev/null
+++ b/fs/xfs/xfs_ioctl.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright (c) 2008 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_IOCTL_H__
+#define __XFS_IOCTL_H__
+
+extern int
+xfs_ioc_space(
+	struct xfs_inode	*ip,
+	struct inode		*inode,
+	struct file		*filp,
+	int			ioflags,
+	unsigned int		cmd,
+	xfs_flock64_t		*bf);
+
+extern int
+xfs_find_handle(
+	unsigned int		cmd,
+	xfs_fsop_handlereq_t	*hreq);
+
+extern int
+xfs_open_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq);
+
+extern int
+xfs_readlink_by_handle(
+	struct file		*parfilp,
+	xfs_fsop_handlereq_t	*hreq);
+
+extern int
+xfs_attrmulti_attr_get(
+	struct inode		*inode,
+	unsigned char		*name,
+	unsigned char		__user *ubuf,
+	__uint32_t		*len,
+	__uint32_t		flags);
+
+extern int
+xfs_attrmulti_attr_set(
+	struct inode		*inode,
+	unsigned char		*name,
+	const unsigned char	__user *ubuf,
+	__uint32_t		len,
+	__uint32_t		flags);
+
+extern int
+xfs_attrmulti_attr_remove(
+	struct inode		*inode,
+	unsigned char		*name,
+	__uint32_t		flags);
+
+extern struct dentry *
+xfs_handle_to_dentry(
+	struct file		*parfilp,
+	void __user		*uhandle,
+	u32			hlen);
+
+extern long
+xfs_file_ioctl(
+	struct file		*filp,
+	unsigned int		cmd,
+	unsigned long		p);
+
+extern long
+xfs_file_compat_ioctl(
+	struct file		*file,
+	unsigned int		cmd,
+	unsigned long		arg);
+
+#endif
diff --git a/fs/xfs/xfs_ioctl32.c b/fs/xfs/xfs_ioctl32.c
new file mode 100644
index 0000000..0d685b3
--- /dev/null
+++ b/fs/xfs/xfs_ioctl32.c
@@ -0,0 +1,673 @@
+/*
+ * Copyright (c) 2004-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include <linux/compat.h>
+#include <linux/ioctl.h>
+#include <linux/mount.h>
+#include <linux/slab.h>
+#include <asm/uaccess.h>
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_vnode.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_itable.h"
+#include "xfs_error.h"
+#include "xfs_dfrag.h"
+#include "xfs_vnodeops.h"
+#include "xfs_fsops.h"
+#include "xfs_alloc.h"
+#include "xfs_rtalloc.h"
+#include "xfs_attr.h"
+#include "xfs_ioctl.h"
+#include "xfs_ioctl32.h"
+#include "xfs_trace.h"
+
+#define  _NATIVE_IOC(cmd, type) \
+	  _IOC(_IOC_DIR(cmd), _IOC_TYPE(cmd), _IOC_NR(cmd), sizeof(type))
+
+#ifdef BROKEN_X86_ALIGNMENT
+STATIC int
+xfs_compat_flock64_copyin(
+	xfs_flock64_t		*bf,
+	compat_xfs_flock64_t	__user *arg32)
+{
+	if (get_user(bf->l_type,	&arg32->l_type) ||
+	    get_user(bf->l_whence,	&arg32->l_whence) ||
+	    get_user(bf->l_start,	&arg32->l_start) ||
+	    get_user(bf->l_len,		&arg32->l_len) ||
+	    get_user(bf->l_sysid,	&arg32->l_sysid) ||
+	    get_user(bf->l_pid,		&arg32->l_pid) ||
+	    copy_from_user(bf->l_pad,	&arg32->l_pad,	4*sizeof(u32)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_compat_ioc_fsgeometry_v1(
+	struct xfs_mount	  *mp,
+	compat_xfs_fsop_geom_v1_t __user *arg32)
+{
+	xfs_fsop_geom_t		  fsgeo;
+	int			  error;
+
+	error = xfs_fs_geometry(mp, &fsgeo, 3);
+	if (error)
+		return -error;
+	/* The 32-bit variant simply has some padding at the end */
+	if (copy_to_user(arg32, &fsgeo, sizeof(struct compat_xfs_fsop_geom_v1)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_compat_growfs_data_copyin(
+	struct xfs_growfs_data	 *in,
+	compat_xfs_growfs_data_t __user *arg32)
+{
+	if (get_user(in->newblocks, &arg32->newblocks) ||
+	    get_user(in->imaxpct,   &arg32->imaxpct))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_compat_growfs_rt_copyin(
+	struct xfs_growfs_rt	 *in,
+	compat_xfs_growfs_rt_t	__user *arg32)
+{
+	if (get_user(in->newblocks, &arg32->newblocks) ||
+	    get_user(in->extsize,   &arg32->extsize))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_inumbers_fmt_compat(
+	void			__user *ubuffer,
+	const xfs_inogrp_t	*buffer,
+	long			count,
+	long			*written)
+{
+	compat_xfs_inogrp_t	__user *p32 = ubuffer;
+	long			i;
+
+	for (i = 0; i < count; i++) {
+		if (put_user(buffer[i].xi_startino,   &p32[i].xi_startino) ||
+		    put_user(buffer[i].xi_alloccount, &p32[i].xi_alloccount) ||
+		    put_user(buffer[i].xi_allocmask,  &p32[i].xi_allocmask))
+			return -XFS_ERROR(EFAULT);
+	}
+	*written = count * sizeof(*p32);
+	return 0;
+}
+
+#else
+#define xfs_inumbers_fmt_compat xfs_inumbers_fmt
+#endif	/* BROKEN_X86_ALIGNMENT */
+
+STATIC int
+xfs_ioctl32_bstime_copyin(
+	xfs_bstime_t		*bstime,
+	compat_xfs_bstime_t	__user *bstime32)
+{
+	compat_time_t		sec32;	/* tv_sec differs on 64 vs. 32 */
+
+	if (get_user(sec32,		&bstime32->tv_sec)	||
+	    get_user(bstime->tv_nsec,	&bstime32->tv_nsec))
+		return -XFS_ERROR(EFAULT);
+	bstime->tv_sec = sec32;
+	return 0;
+}
+
+/* xfs_bstat_t has differing alignment on intel, & bstime_t sizes everywhere */
+STATIC int
+xfs_ioctl32_bstat_copyin(
+	xfs_bstat_t		*bstat,
+	compat_xfs_bstat_t	__user *bstat32)
+{
+	if (get_user(bstat->bs_ino,	&bstat32->bs_ino)	||
+	    get_user(bstat->bs_mode,	&bstat32->bs_mode)	||
+	    get_user(bstat->bs_nlink,	&bstat32->bs_nlink)	||
+	    get_user(bstat->bs_uid,	&bstat32->bs_uid)	||
+	    get_user(bstat->bs_gid,	&bstat32->bs_gid)	||
+	    get_user(bstat->bs_rdev,	&bstat32->bs_rdev)	||
+	    get_user(bstat->bs_blksize,	&bstat32->bs_blksize)	||
+	    get_user(bstat->bs_size,	&bstat32->bs_size)	||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_atime, &bstat32->bs_atime) ||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_mtime, &bstat32->bs_mtime) ||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_ctime, &bstat32->bs_ctime) ||
+	    get_user(bstat->bs_blocks,	&bstat32->bs_size)	||
+	    get_user(bstat->bs_xflags,	&bstat32->bs_size)	||
+	    get_user(bstat->bs_extsize,	&bstat32->bs_extsize)	||
+	    get_user(bstat->bs_extents,	&bstat32->bs_extents)	||
+	    get_user(bstat->bs_gen,	&bstat32->bs_gen)	||
+	    get_user(bstat->bs_projid_lo, &bstat32->bs_projid_lo) ||
+	    get_user(bstat->bs_projid_hi, &bstat32->bs_projid_hi) ||
+	    get_user(bstat->bs_dmevmask, &bstat32->bs_dmevmask)	||
+	    get_user(bstat->bs_dmstate,	&bstat32->bs_dmstate)	||
+	    get_user(bstat->bs_aextents, &bstat32->bs_aextents))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+/* XFS_IOC_FSBULKSTAT and friends */
+
+STATIC int
+xfs_bstime_store_compat(
+	compat_xfs_bstime_t	__user *p32,
+	const xfs_bstime_t	*p)
+{
+	__s32			sec32;
+
+	sec32 = p->tv_sec;
+	if (put_user(sec32, &p32->tv_sec) ||
+	    put_user(p->tv_nsec, &p32->tv_nsec))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+/* Return 0 on success or positive error (to xfs_bulkstat()) */
+STATIC int
+xfs_bulkstat_one_fmt_compat(
+	void			__user *ubuffer,
+	int			ubsize,
+	int			*ubused,
+	const xfs_bstat_t	*buffer)
+{
+	compat_xfs_bstat_t	__user *p32 = ubuffer;
+
+	if (ubsize < sizeof(*p32))
+		return XFS_ERROR(ENOMEM);
+
+	if (put_user(buffer->bs_ino,	  &p32->bs_ino)		||
+	    put_user(buffer->bs_mode,	  &p32->bs_mode)	||
+	    put_user(buffer->bs_nlink,	  &p32->bs_nlink)	||
+	    put_user(buffer->bs_uid,	  &p32->bs_uid)		||
+	    put_user(buffer->bs_gid,	  &p32->bs_gid)		||
+	    put_user(buffer->bs_rdev,	  &p32->bs_rdev)	||
+	    put_user(buffer->bs_blksize,  &p32->bs_blksize)	||
+	    put_user(buffer->bs_size,	  &p32->bs_size)	||
+	    xfs_bstime_store_compat(&p32->bs_atime, &buffer->bs_atime) ||
+	    xfs_bstime_store_compat(&p32->bs_mtime, &buffer->bs_mtime) ||
+	    xfs_bstime_store_compat(&p32->bs_ctime, &buffer->bs_ctime) ||
+	    put_user(buffer->bs_blocks,	  &p32->bs_blocks)	||
+	    put_user(buffer->bs_xflags,	  &p32->bs_xflags)	||
+	    put_user(buffer->bs_extsize,  &p32->bs_extsize)	||
+	    put_user(buffer->bs_extents,  &p32->bs_extents)	||
+	    put_user(buffer->bs_gen,	  &p32->bs_gen)		||
+	    put_user(buffer->bs_projid,	  &p32->bs_projid)	||
+	    put_user(buffer->bs_projid_hi,	&p32->bs_projid_hi)	||
+	    put_user(buffer->bs_dmevmask, &p32->bs_dmevmask)	||
+	    put_user(buffer->bs_dmstate,  &p32->bs_dmstate)	||
+	    put_user(buffer->bs_aextents, &p32->bs_aextents))
+		return XFS_ERROR(EFAULT);
+	if (ubused)
+		*ubused = sizeof(*p32);
+	return 0;
+}
+
+STATIC int
+xfs_bulkstat_one_compat(
+	xfs_mount_t	*mp,		/* mount point for filesystem */
+	xfs_ino_t	ino,		/* inode number to get data for */
+	void		__user *buffer,	/* buffer to place output in */
+	int		ubsize,		/* size of buffer */
+	int		*ubused,	/* bytes used by me */
+	int		*stat)		/* BULKSTAT_RV_... */
+{
+	return xfs_bulkstat_one_int(mp, ino, buffer, ubsize,
+				    xfs_bulkstat_one_fmt_compat,
+				    ubused, stat);
+}
+
+/* copied from xfs_ioctl.c */
+STATIC int
+xfs_compat_ioc_bulkstat(
+	xfs_mount_t		  *mp,
+	unsigned int		  cmd,
+	compat_xfs_fsop_bulkreq_t __user *p32)
+{
+	u32			addr;
+	xfs_fsop_bulkreq_t	bulkreq;
+	int			count;	/* # of records returned */
+	xfs_ino_t		inlast;	/* last inode number */
+	int			done;
+	int			error;
+
+	/* done = 1 if there are more stats to get and if bulkstat */
+	/* should be called again (unused here, but used in dmapi) */
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	if (get_user(addr, &p32->lastip))
+		return -XFS_ERROR(EFAULT);
+	bulkreq.lastip = compat_ptr(addr);
+	if (get_user(bulkreq.icount, &p32->icount) ||
+	    get_user(addr, &p32->ubuffer))
+		return -XFS_ERROR(EFAULT);
+	bulkreq.ubuffer = compat_ptr(addr);
+	if (get_user(addr, &p32->ocount))
+		return -XFS_ERROR(EFAULT);
+	bulkreq.ocount = compat_ptr(addr);
+
+	if (copy_from_user(&inlast, bulkreq.lastip, sizeof(__s64)))
+		return -XFS_ERROR(EFAULT);
+
+	if ((count = bulkreq.icount) <= 0)
+		return -XFS_ERROR(EINVAL);
+
+	if (bulkreq.ubuffer == NULL)
+		return -XFS_ERROR(EINVAL);
+
+	if (cmd == XFS_IOC_FSINUMBERS_32) {
+		error = xfs_inumbers(mp, &inlast, &count,
+				bulkreq.ubuffer, xfs_inumbers_fmt_compat);
+	} else if (cmd == XFS_IOC_FSBULKSTAT_SINGLE_32) {
+		int res;
+
+		error = xfs_bulkstat_one_compat(mp, inlast, bulkreq.ubuffer,
+				sizeof(compat_xfs_bstat_t), 0, &res);
+	} else if (cmd == XFS_IOC_FSBULKSTAT_32) {
+		error = xfs_bulkstat(mp, &inlast, &count,
+			xfs_bulkstat_one_compat, sizeof(compat_xfs_bstat_t),
+			bulkreq.ubuffer, &done);
+	} else
+		error = XFS_ERROR(EINVAL);
+	if (error)
+		return -error;
+
+	if (bulkreq.ocount != NULL) {
+		if (copy_to_user(bulkreq.lastip, &inlast,
+						sizeof(xfs_ino_t)))
+			return -XFS_ERROR(EFAULT);
+
+		if (copy_to_user(bulkreq.ocount, &count, sizeof(count)))
+			return -XFS_ERROR(EFAULT);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_compat_handlereq_copyin(
+	xfs_fsop_handlereq_t		*hreq,
+	compat_xfs_fsop_handlereq_t	__user *arg32)
+{
+	compat_xfs_fsop_handlereq_t	hreq32;
+
+	if (copy_from_user(&hreq32, arg32, sizeof(compat_xfs_fsop_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	hreq->fd = hreq32.fd;
+	hreq->path = compat_ptr(hreq32.path);
+	hreq->oflags = hreq32.oflags;
+	hreq->ihandle = compat_ptr(hreq32.ihandle);
+	hreq->ihandlen = hreq32.ihandlen;
+	hreq->ohandle = compat_ptr(hreq32.ohandle);
+	hreq->ohandlen = compat_ptr(hreq32.ohandlen);
+
+	return 0;
+}
+
+STATIC struct dentry *
+xfs_compat_handlereq_to_dentry(
+	struct file		*parfilp,
+	compat_xfs_fsop_handlereq_t *hreq)
+{
+	return xfs_handle_to_dentry(parfilp,
+			compat_ptr(hreq->ihandle), hreq->ihandlen);
+}
+
+STATIC int
+xfs_compat_attrlist_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error;
+	attrlist_cursor_kern_t	*cursor;
+	compat_xfs_fsop_attrlist_handlereq_t al_hreq;
+	struct dentry		*dentry;
+	char			*kbuf;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&al_hreq, arg,
+			   sizeof(compat_xfs_fsop_attrlist_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+	if (al_hreq.buflen < sizeof(struct attrlist) ||
+	    al_hreq.buflen > XATTR_LIST_MAX)
+		return -XFS_ERROR(EINVAL);
+
+	/*
+	 * Reject flags, only allow namespaces.
+	 */
+	if (al_hreq.flags & ~(ATTR_ROOT | ATTR_SECURE))
+		return -XFS_ERROR(EINVAL);
+
+	dentry = xfs_compat_handlereq_to_dentry(parfilp, &al_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = -ENOMEM;
+	kbuf = kmalloc(al_hreq.buflen, GFP_KERNEL);
+	if (!kbuf)
+		goto out_dput;
+
+	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
+	error = -xfs_attr_list(XFS_I(dentry->d_inode), kbuf, al_hreq.buflen,
+					al_hreq.flags, cursor);
+	if (error)
+		goto out_kfree;
+
+	if (copy_to_user(compat_ptr(al_hreq.buffer), kbuf, al_hreq.buflen))
+		error = -EFAULT;
+
+ out_kfree:
+	kfree(kbuf);
+ out_dput:
+	dput(dentry);
+	return error;
+}
+
+STATIC int
+xfs_compat_attrmulti_by_handle(
+	struct file				*parfilp,
+	void					__user *arg)
+{
+	int					error;
+	compat_xfs_attr_multiop_t		*ops;
+	compat_xfs_fsop_attrmulti_handlereq_t	am_hreq;
+	struct dentry				*dentry;
+	unsigned int				i, size;
+	unsigned char				*attr_name;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&am_hreq, arg,
+			   sizeof(compat_xfs_fsop_attrmulti_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	/* overflow check */
+	if (am_hreq.opcount >= INT_MAX / sizeof(compat_xfs_attr_multiop_t))
+		return -E2BIG;
+
+	dentry = xfs_compat_handlereq_to_dentry(parfilp, &am_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = E2BIG;
+	size = am_hreq.opcount * sizeof(compat_xfs_attr_multiop_t);
+	if (!size || size > 16 * PAGE_SIZE)
+		goto out_dput;
+
+	ops = memdup_user(compat_ptr(am_hreq.ops), size);
+	if (IS_ERR(ops)) {
+		error = PTR_ERR(ops);
+		goto out_dput;
+	}
+
+	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
+	if (!attr_name)
+		goto out_kfree_ops;
+
+	error = 0;
+	for (i = 0; i < am_hreq.opcount; i++) {
+		ops[i].am_error = strncpy_from_user((char *)attr_name,
+				compat_ptr(ops[i].am_attrname),
+				MAXNAMELEN);
+		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
+			error = -ERANGE;
+		if (ops[i].am_error < 0)
+			break;
+
+		switch (ops[i].am_opcode) {
+		case ATTR_OP_GET:
+			ops[i].am_error = xfs_attrmulti_attr_get(
+					dentry->d_inode, attr_name,
+					compat_ptr(ops[i].am_attrvalue),
+					&ops[i].am_length, ops[i].am_flags);
+			break;
+		case ATTR_OP_SET:
+			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
+			if (ops[i].am_error)
+				break;
+			ops[i].am_error = xfs_attrmulti_attr_set(
+					dentry->d_inode, attr_name,
+					compat_ptr(ops[i].am_attrvalue),
+					ops[i].am_length, ops[i].am_flags);
+			mnt_drop_write(parfilp->f_path.mnt);
+			break;
+		case ATTR_OP_REMOVE:
+			ops[i].am_error = mnt_want_write(parfilp->f_path.mnt);
+			if (ops[i].am_error)
+				break;
+			ops[i].am_error = xfs_attrmulti_attr_remove(
+					dentry->d_inode, attr_name,
+					ops[i].am_flags);
+			mnt_drop_write(parfilp->f_path.mnt);
+			break;
+		default:
+			ops[i].am_error = EINVAL;
+		}
+	}
+
+	if (copy_to_user(compat_ptr(am_hreq.ops), ops, size))
+		error = XFS_ERROR(EFAULT);
+
+	kfree(attr_name);
+ out_kfree_ops:
+	kfree(ops);
+ out_dput:
+	dput(dentry);
+	return -error;
+}
+
+STATIC int
+xfs_compat_fssetdm_by_handle(
+	struct file		*parfilp,
+	void			__user *arg)
+{
+	int			error;
+	struct fsdmidata	fsd;
+	compat_xfs_fsop_setdm_handlereq_t dmhreq;
+	struct dentry		*dentry;
+
+	if (!capable(CAP_MKNOD))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&dmhreq, arg,
+			   sizeof(compat_xfs_fsop_setdm_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	dentry = xfs_compat_handlereq_to_dentry(parfilp, &dmhreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	if (IS_IMMUTABLE(dentry->d_inode) || IS_APPEND(dentry->d_inode)) {
+		error = -XFS_ERROR(EPERM);
+		goto out;
+	}
+
+	if (copy_from_user(&fsd, compat_ptr(dmhreq.data), sizeof(fsd))) {
+		error = -XFS_ERROR(EFAULT);
+		goto out;
+	}
+
+	error = -xfs_set_dmattrs(XFS_I(dentry->d_inode), fsd.fsd_dmevmask,
+				 fsd.fsd_dmstate);
+
+out:
+	dput(dentry);
+	return error;
+}
+
+long
+xfs_file_compat_ioctl(
+	struct file		*filp,
+	unsigned		cmd,
+	unsigned long		p)
+{
+	struct inode		*inode = filp->f_path.dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	void			__user *arg = (void __user *)p;
+	int			ioflags = 0;
+	int			error;
+
+	if (filp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	trace_xfs_file_compat_ioctl(ip);
+
+	switch (cmd) {
+	/* No size or alignment issues on any arch */
+	case XFS_IOC_DIOINFO:
+	case XFS_IOC_FSGEOMETRY:
+	case XFS_IOC_FSGETXATTR:
+	case XFS_IOC_FSSETXATTR:
+	case XFS_IOC_FSGETXATTRA:
+	case XFS_IOC_FSSETDM:
+	case XFS_IOC_GETBMAP:
+	case XFS_IOC_GETBMAPA:
+	case XFS_IOC_GETBMAPX:
+	case XFS_IOC_FSCOUNTS:
+	case XFS_IOC_SET_RESBLKS:
+	case XFS_IOC_GET_RESBLKS:
+	case XFS_IOC_FSGROWFSLOG:
+	case XFS_IOC_GOINGDOWN:
+	case XFS_IOC_ERROR_INJECTION:
+	case XFS_IOC_ERROR_CLEARALL:
+		return xfs_file_ioctl(filp, cmd, p);
+#ifndef BROKEN_X86_ALIGNMENT
+	/* These are handled fine if no alignment issues */
+	case XFS_IOC_ALLOCSP:
+	case XFS_IOC_FREESP:
+	case XFS_IOC_RESVSP:
+	case XFS_IOC_UNRESVSP:
+	case XFS_IOC_ALLOCSP64:
+	case XFS_IOC_FREESP64:
+	case XFS_IOC_RESVSP64:
+	case XFS_IOC_UNRESVSP64:
+	case XFS_IOC_FSGEOMETRY_V1:
+	case XFS_IOC_FSGROWFSDATA:
+	case XFS_IOC_FSGROWFSRT:
+	case XFS_IOC_ZERO_RANGE:
+		return xfs_file_ioctl(filp, cmd, p);
+#else
+	case XFS_IOC_ALLOCSP_32:
+	case XFS_IOC_FREESP_32:
+	case XFS_IOC_ALLOCSP64_32:
+	case XFS_IOC_FREESP64_32:
+	case XFS_IOC_RESVSP_32:
+	case XFS_IOC_UNRESVSP_32:
+	case XFS_IOC_RESVSP64_32:
+	case XFS_IOC_UNRESVSP64_32:
+	case XFS_IOC_ZERO_RANGE_32: {
+		struct xfs_flock64	bf;
+
+		if (xfs_compat_flock64_copyin(&bf, arg))
+			return -XFS_ERROR(EFAULT);
+		cmd = _NATIVE_IOC(cmd, struct xfs_flock64);
+		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, &bf);
+	}
+	case XFS_IOC_FSGEOMETRY_V1_32:
+		return xfs_compat_ioc_fsgeometry_v1(mp, arg);
+	case XFS_IOC_FSGROWFSDATA_32: {
+		struct xfs_growfs_data	in;
+
+		if (xfs_compat_growfs_data_copyin(&in, arg))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_growfs_data(mp, &in);
+		return -error;
+	}
+	case XFS_IOC_FSGROWFSRT_32: {
+		struct xfs_growfs_rt	in;
+
+		if (xfs_compat_growfs_rt_copyin(&in, arg))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_growfs_rt(mp, &in);
+		return -error;
+	}
+#endif
+	/* long changes size, but xfs only copiese out 32 bits */
+	case XFS_IOC_GETXFLAGS_32:
+	case XFS_IOC_SETXFLAGS_32:
+	case XFS_IOC_GETVERSION_32:
+		cmd = _NATIVE_IOC(cmd, long);
+		return xfs_file_ioctl(filp, cmd, p);
+	case XFS_IOC_SWAPEXT_32: {
+		struct xfs_swapext	  sxp;
+		struct compat_xfs_swapext __user *sxu = arg;
+
+		/* Bulk copy in up to the sx_stat field, then copy bstat */
+		if (copy_from_user(&sxp, sxu,
+				   offsetof(struct xfs_swapext, sx_stat)) ||
+		    xfs_ioctl32_bstat_copyin(&sxp.sx_stat, &sxu->sx_stat))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_swapext(&sxp);
+		return -error;
+	}
+	case XFS_IOC_FSBULKSTAT_32:
+	case XFS_IOC_FSBULKSTAT_SINGLE_32:
+	case XFS_IOC_FSINUMBERS_32:
+		return xfs_compat_ioc_bulkstat(mp, cmd, arg);
+	case XFS_IOC_FD_TO_HANDLE_32:
+	case XFS_IOC_PATH_TO_HANDLE_32:
+	case XFS_IOC_PATH_TO_FSHANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
+
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
+		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_handlereq);
+		return xfs_find_handle(cmd, &hreq);
+	}
+	case XFS_IOC_OPEN_BY_HANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
+
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
+		return xfs_open_by_handle(filp, &hreq);
+	}
+	case XFS_IOC_READLINK_BY_HANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
+
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
+		return xfs_readlink_by_handle(filp, &hreq);
+	}
+	case XFS_IOC_ATTRLIST_BY_HANDLE_32:
+		return xfs_compat_attrlist_by_handle(filp, arg);
+	case XFS_IOC_ATTRMULTI_BY_HANDLE_32:
+		return xfs_compat_attrmulti_by_handle(filp, arg);
+	case XFS_IOC_FSSETDM_BY_HANDLE_32:
+		return xfs_compat_fssetdm_by_handle(filp, arg);
+	default:
+		return -XFS_ERROR(ENOIOCTLCMD);
+	}
+}
diff --git a/fs/xfs/xfs_ioctl32.h b/fs/xfs/xfs_ioctl32.h
new file mode 100644
index 0000000..80f4060
--- /dev/null
+++ b/fs/xfs/xfs_ioctl32.h
@@ -0,0 +1,237 @@
+/*
+ * Copyright (c) 2004-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_IOCTL32_H__
+#define __XFS_IOCTL32_H__
+
+#include <linux/compat.h>
+
+/*
+ * on 32-bit arches, ioctl argument structures may have different sizes
+ * and/or alignment.  We define compat structures which match the
+ * 32-bit sizes/alignments here, and their associated ioctl numbers.
+ *
+ * xfs_ioctl32.c contains routines to copy these structures in and out.
+ */
+
+/* stock kernel-level ioctls we support */
+#define XFS_IOC_GETXFLAGS_32	FS_IOC32_GETFLAGS
+#define XFS_IOC_SETXFLAGS_32	FS_IOC32_SETFLAGS
+#define XFS_IOC_GETVERSION_32	FS_IOC32_GETVERSION
+
+/*
+ * On intel, even if sizes match, alignment and/or padding may differ.
+ */
+#if defined(CONFIG_IA64) || defined(CONFIG_X86_64)
+#define BROKEN_X86_ALIGNMENT
+#define __compat_packed __attribute__((packed))
+#else
+#define __compat_packed
+#endif
+
+typedef struct compat_xfs_bstime {
+	compat_time_t	tv_sec;		/* seconds		*/
+	__s32		tv_nsec;	/* and nanoseconds	*/
+} compat_xfs_bstime_t;
+
+typedef struct compat_xfs_bstat {
+	__u64		bs_ino;		/* inode number			*/
+	__u16		bs_mode;	/* type and mode		*/
+	__u16		bs_nlink;	/* number of links		*/
+	__u32		bs_uid;		/* user id			*/
+	__u32		bs_gid;		/* group id			*/
+	__u32		bs_rdev;	/* device value			*/
+	__s32		bs_blksize;	/* block size			*/
+	__s64		bs_size;	/* file size			*/
+	compat_xfs_bstime_t bs_atime;	/* access time			*/
+	compat_xfs_bstime_t bs_mtime;	/* modify time			*/
+	compat_xfs_bstime_t bs_ctime;	/* inode change time		*/
+	int64_t		bs_blocks;	/* number of blocks		*/
+	__u32		bs_xflags;	/* extended flags		*/
+	__s32		bs_extsize;	/* extent size			*/
+	__s32		bs_extents;	/* number of extents		*/
+	__u32		bs_gen;		/* generation count		*/
+	__u16		bs_projid_lo;	/* lower part of project id	*/
+#define	bs_projid	bs_projid_lo	/* (previously just bs_projid)	*/
+	__u16		bs_projid_hi;	/* high part of project id	*/
+	unsigned char	bs_pad[12];	/* pad space, unused		*/
+	__u32		bs_dmevmask;	/* DMIG event mask		*/
+	__u16		bs_dmstate;	/* DMIG state info		*/
+	__u16		bs_aextents;	/* attribute number of extents	*/
+} __compat_packed compat_xfs_bstat_t;
+
+typedef struct compat_xfs_fsop_bulkreq {
+	compat_uptr_t	lastip;		/* last inode # pointer		*/
+	__s32		icount;		/* count of entries in buffer	*/
+	compat_uptr_t	ubuffer;	/* user buffer for inode desc.	*/
+	compat_uptr_t	ocount;		/* output count pointer		*/
+} compat_xfs_fsop_bulkreq_t;
+
+#define XFS_IOC_FSBULKSTAT_32 \
+	_IOWR('X', 101, struct compat_xfs_fsop_bulkreq)
+#define XFS_IOC_FSBULKSTAT_SINGLE_32 \
+	_IOWR('X', 102, struct compat_xfs_fsop_bulkreq)
+#define XFS_IOC_FSINUMBERS_32 \
+	_IOWR('X', 103, struct compat_xfs_fsop_bulkreq)
+
+typedef struct compat_xfs_fsop_handlereq {
+	__u32		fd;		/* fd for FD_TO_HANDLE		*/
+	compat_uptr_t	path;		/* user pathname		*/
+	__u32		oflags;		/* open flags			*/
+	compat_uptr_t	ihandle;	/* user supplied handle		*/
+	__u32		ihandlen;	/* user supplied length		*/
+	compat_uptr_t	ohandle;	/* user buffer for handle	*/
+	compat_uptr_t	ohandlen;	/* user buffer length		*/
+} compat_xfs_fsop_handlereq_t;
+
+#define XFS_IOC_PATH_TO_FSHANDLE_32 \
+	_IOWR('X', 104, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_PATH_TO_HANDLE_32 \
+	_IOWR('X', 105, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_FD_TO_HANDLE_32 \
+	_IOWR('X', 106, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_OPEN_BY_HANDLE_32 \
+	_IOWR('X', 107, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_READLINK_BY_HANDLE_32 \
+	_IOWR('X', 108, struct compat_xfs_fsop_handlereq)
+
+/* The bstat field in the swapext struct needs translation */
+typedef struct compat_xfs_swapext {
+	__int64_t		sx_version;	/* version */
+	__int64_t		sx_fdtarget;	/* fd of target file */
+	__int64_t		sx_fdtmp;	/* fd of tmp file */
+	xfs_off_t		sx_offset;	/* offset into file */
+	xfs_off_t		sx_length;	/* leng from offset */
+	char			sx_pad[16];	/* pad space, unused */
+	compat_xfs_bstat_t	sx_stat;	/* stat of target b4 copy */
+} __compat_packed compat_xfs_swapext_t;
+
+#define XFS_IOC_SWAPEXT_32	_IOWR('X', 109, struct compat_xfs_swapext)
+
+typedef struct compat_xfs_fsop_attrlist_handlereq {
+	struct compat_xfs_fsop_handlereq hreq; /* handle interface structure */
+	struct xfs_attrlist_cursor	pos; /* opaque cookie, list offset */
+	__u32				flags;	/* which namespace to use */
+	__u32				buflen;	/* length of buffer supplied */
+	compat_uptr_t			buffer;	/* returned names */
+} __compat_packed compat_xfs_fsop_attrlist_handlereq_t;
+
+/* Note: actually this is read/write */
+#define XFS_IOC_ATTRLIST_BY_HANDLE_32 \
+	_IOW('X', 122, struct compat_xfs_fsop_attrlist_handlereq)
+
+/* am_opcodes defined in xfs_fs.h */
+typedef struct compat_xfs_attr_multiop {
+	__u32		am_opcode;
+	__s32		am_error;
+	compat_uptr_t	am_attrname;
+	compat_uptr_t	am_attrvalue;
+	__u32		am_length;
+	__u32		am_flags;
+} compat_xfs_attr_multiop_t;
+
+typedef struct compat_xfs_fsop_attrmulti_handlereq {
+	struct compat_xfs_fsop_handlereq hreq; /* handle interface structure */
+	__u32				opcount;/* count of following multiop */
+	/* ptr to compat_xfs_attr_multiop */
+	compat_uptr_t			ops; /* attr_multi data */
+} compat_xfs_fsop_attrmulti_handlereq_t;
+
+#define XFS_IOC_ATTRMULTI_BY_HANDLE_32 \
+	_IOW('X', 123, struct compat_xfs_fsop_attrmulti_handlereq)
+
+typedef struct compat_xfs_fsop_setdm_handlereq {
+	struct compat_xfs_fsop_handlereq hreq;	/* handle information   */
+	/* ptr to struct fsdmidata */
+	compat_uptr_t			data;	/* DMAPI data   */
+} compat_xfs_fsop_setdm_handlereq_t;
+
+#define XFS_IOC_FSSETDM_BY_HANDLE_32 \
+	_IOW('X', 121, struct compat_xfs_fsop_setdm_handlereq)
+
+#ifdef BROKEN_X86_ALIGNMENT
+/* on ia32 l_start is on a 32-bit boundary */
+typedef struct compat_xfs_flock64 {
+	__s16		l_type;
+	__s16		l_whence;
+	__s64		l_start	__attribute__((packed));
+			/* len == 0 means until end of file */
+	__s64		l_len __attribute__((packed));
+	__s32		l_sysid;
+	__u32		l_pid;
+	__s32		l_pad[4];	/* reserve area */
+} compat_xfs_flock64_t;
+
+#define XFS_IOC_ALLOCSP_32	_IOW('X', 10, struct compat_xfs_flock64)
+#define XFS_IOC_FREESP_32	_IOW('X', 11, struct compat_xfs_flock64)
+#define XFS_IOC_ALLOCSP64_32	_IOW('X', 36, struct compat_xfs_flock64)
+#define XFS_IOC_FREESP64_32	_IOW('X', 37, struct compat_xfs_flock64)
+#define XFS_IOC_RESVSP_32	_IOW('X', 40, struct compat_xfs_flock64)
+#define XFS_IOC_UNRESVSP_32	_IOW('X', 41, struct compat_xfs_flock64)
+#define XFS_IOC_RESVSP64_32	_IOW('X', 42, struct compat_xfs_flock64)
+#define XFS_IOC_UNRESVSP64_32	_IOW('X', 43, struct compat_xfs_flock64)
+#define XFS_IOC_ZERO_RANGE_32	_IOW('X', 57, struct compat_xfs_flock64)
+
+typedef struct compat_xfs_fsop_geom_v1 {
+	__u32		blocksize;	/* filesystem (data) block size */
+	__u32		rtextsize;	/* realtime extent size		*/
+	__u32		agblocks;	/* fsblocks in an AG		*/
+	__u32		agcount;	/* number of allocation groups	*/
+	__u32		logblocks;	/* fsblocks in the log		*/
+	__u32		sectsize;	/* (data) sector size, bytes	*/
+	__u32		inodesize;	/* inode size in bytes		*/
+	__u32		imaxpct;	/* max allowed inode space(%)	*/
+	__u64		datablocks;	/* fsblocks in data subvolume	*/
+	__u64		rtblocks;	/* fsblocks in realtime subvol	*/
+	__u64		rtextents;	/* rt extents in realtime subvol*/
+	__u64		logstart;	/* starting fsblock of the log	*/
+	unsigned char	uuid[16];	/* unique id of the filesystem	*/
+	__u32		sunit;		/* stripe unit, fsblocks	*/
+	__u32		swidth;		/* stripe width, fsblocks	*/
+	__s32		version;	/* structure version		*/
+	__u32		flags;		/* superblock version flags	*/
+	__u32		logsectsize;	/* log sector size, bytes	*/
+	__u32		rtsectsize;	/* realtime sector size, bytes	*/
+	__u32		dirblocksize;	/* directory block size, bytes	*/
+} __attribute__((packed)) compat_xfs_fsop_geom_v1_t;
+
+#define XFS_IOC_FSGEOMETRY_V1_32  \
+	_IOR('X', 100, struct compat_xfs_fsop_geom_v1)
+
+typedef struct compat_xfs_inogrp {
+	__u64		xi_startino;	/* starting inode number	*/
+	__s32		xi_alloccount;	/* # bits set in allocmask	*/
+	__u64		xi_allocmask;	/* mask of allocated inodes	*/
+} __attribute__((packed)) compat_xfs_inogrp_t;
+
+/* These growfs input structures have padding on the end, so must translate */
+typedef struct compat_xfs_growfs_data {
+	__u64		newblocks;	/* new data subvol size, fsblocks */
+	__u32		imaxpct;	/* new inode space percentage limit */
+} __attribute__((packed)) compat_xfs_growfs_data_t;
+
+typedef struct compat_xfs_growfs_rt {
+	__u64		newblocks;	/* new realtime size, fsblocks */
+	__u32		extsize;	/* new realtime extent size, fsblocks */
+} __attribute__((packed)) compat_xfs_growfs_rt_t;
+
+#define XFS_IOC_FSGROWFSDATA_32 _IOW('X', 110, struct compat_xfs_growfs_data)
+#define XFS_IOC_FSGROWFSRT_32   _IOW('X', 112, struct compat_xfs_growfs_rt)
+
+#endif /* BROKEN_X86_ALIGNMENT */
+
+#endif /* __XFS_IOCTL32_H__ */
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
new file mode 100644
index 0000000..1c01f04
--- /dev/null
+++ b/fs/xfs/xfs_iops.c
@@ -0,0 +1,1228 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_acl.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_itable.h"
+#include "xfs_rw.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_utils.h"
+#include "xfs_vnodeops.h"
+#include "xfs_inode_item.h"
+#include "xfs_trace.h"
+
+#include <linux/capability.h>
+#include <linux/xattr.h>
+#include <linux/namei.h>
+#include <linux/posix_acl.h>
+#include <linux/security.h>
+#include <linux/fiemap.h>
+#include <linux/slab.h>
+
+/*
+ * Bring the timestamps in the XFS inode uptodate.
+ *
+ * Used before writing the inode to disk.
+ */
+void
+xfs_synchronize_times(
+	xfs_inode_t	*ip)
+{
+	struct inode	*inode = VFS_I(ip);
+
+	ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec;
+	ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec;
+	ip->i_d.di_ctime.t_sec = (__int32_t)inode->i_ctime.tv_sec;
+	ip->i_d.di_ctime.t_nsec = (__int32_t)inode->i_ctime.tv_nsec;
+	ip->i_d.di_mtime.t_sec = (__int32_t)inode->i_mtime.tv_sec;
+	ip->i_d.di_mtime.t_nsec = (__int32_t)inode->i_mtime.tv_nsec;
+}
+
+/*
+ * If the linux inode is valid, mark it dirty, else mark the dirty state
+ * in the XFS inode to make sure we pick it up when reclaiming the inode.
+ */
+void
+xfs_mark_inode_dirty_sync(
+	xfs_inode_t	*ip)
+{
+	struct inode	*inode = VFS_I(ip);
+
+	if (!(inode->i_state & (I_WILL_FREE|I_FREEING)))
+		mark_inode_dirty_sync(inode);
+	else {
+		barrier();
+		ip->i_update_core = 1;
+	}
+}
+
+void
+xfs_mark_inode_dirty(
+	xfs_inode_t	*ip)
+{
+	struct inode	*inode = VFS_I(ip);
+
+	if (!(inode->i_state & (I_WILL_FREE|I_FREEING)))
+		mark_inode_dirty(inode);
+	else {
+		barrier();
+		ip->i_update_core = 1;
+	}
+
+}
+
+
+int xfs_initxattrs(struct inode *inode, const struct xattr *xattr_array,
+		   void *fs_info)
+{
+	const struct xattr *xattr;
+	struct xfs_inode *ip = XFS_I(inode);
+	int error = 0;
+
+	for (xattr = xattr_array; xattr->name != NULL; xattr++) {
+		error = xfs_attr_set(ip, xattr->name, xattr->value,
+				     xattr->value_len, ATTR_SECURE);
+		if (error < 0)
+			break;
+	}
+	return error;
+}
+
+/*
+ * Hook in SELinux.  This is not quite correct yet, what we really need
+ * here (as we do for default ACLs) is a mechanism by which creation of
+ * these attrs can be journalled at inode creation time (along with the
+ * inode, of course, such that log replay can't cause these to be lost).
+ */
+
+STATIC int
+xfs_init_security(
+	struct inode	*inode,
+	struct inode	*dir,
+	const struct qstr *qstr)
+{
+	return security_inode_init_security(inode, dir, qstr,
+					    &xfs_initxattrs, NULL);
+}
+
+static void
+xfs_dentry_to_name(
+	struct xfs_name	*namep,
+	struct dentry	*dentry)
+{
+	namep->name = dentry->d_name.name;
+	namep->len = dentry->d_name.len;
+}
+
+STATIC void
+xfs_cleanup_inode(
+	struct inode	*dir,
+	struct inode	*inode,
+	struct dentry	*dentry)
+{
+	struct xfs_name	teardown;
+
+	/* Oh, the horror.
+	 * If we can't add the ACL or we fail in
+	 * xfs_init_security we must back out.
+	 * ENOSPC can hit here, among other things.
+	 */
+	xfs_dentry_to_name(&teardown, dentry);
+
+	xfs_remove(XFS_I(dir), &teardown, XFS_I(inode));
+	iput(inode);
+}
+
+STATIC int
+xfs_vn_mknod(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	int		mode,
+	dev_t		rdev)
+{
+	struct inode	*inode;
+	struct xfs_inode *ip = NULL;
+	struct posix_acl *default_acl = NULL;
+	struct xfs_name	name;
+	int		error;
+
+	/*
+	 * Irix uses Missed'em'V split, but doesn't want to see
+	 * the upper 5 bits of (14bit) major.
+	 */
+	if (S_ISCHR(mode) || S_ISBLK(mode)) {
+		if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff))
+			return -EINVAL;
+		rdev = sysv_encode_dev(rdev);
+	} else {
+		rdev = 0;
+	}
+
+	if (IS_POSIXACL(dir)) {
+		default_acl = xfs_get_acl(dir, ACL_TYPE_DEFAULT);
+		if (IS_ERR(default_acl))
+			return PTR_ERR(default_acl);
+
+		if (!default_acl)
+			mode &= ~current_umask();
+	}
+
+	xfs_dentry_to_name(&name, dentry);
+	error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip);
+	if (unlikely(error))
+		goto out_free_acl;
+
+	inode = VFS_I(ip);
+
+	error = xfs_init_security(inode, dir, &dentry->d_name);
+	if (unlikely(error))
+		goto out_cleanup_inode;
+
+	if (default_acl) {
+		error = -xfs_inherit_acl(inode, default_acl);
+		default_acl = NULL;
+		if (unlikely(error))
+			goto out_cleanup_inode;
+	}
+
+
+	d_instantiate(dentry, inode);
+	return -error;
+
+ out_cleanup_inode:
+	xfs_cleanup_inode(dir, inode, dentry);
+ out_free_acl:
+	posix_acl_release(default_acl);
+	return -error;
+}
+
+STATIC int
+xfs_vn_create(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	int		mode,
+	struct nameidata *nd)
+{
+	return xfs_vn_mknod(dir, dentry, mode, 0);
+}
+
+STATIC int
+xfs_vn_mkdir(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	int		mode)
+{
+	return xfs_vn_mknod(dir, dentry, mode|S_IFDIR, 0);
+}
+
+STATIC struct dentry *
+xfs_vn_lookup(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	struct nameidata *nd)
+{
+	struct xfs_inode *cip;
+	struct xfs_name	name;
+	int		error;
+
+	if (dentry->d_name.len >= MAXNAMELEN)
+		return ERR_PTR(-ENAMETOOLONG);
+
+	xfs_dentry_to_name(&name, dentry);
+	error = xfs_lookup(XFS_I(dir), &name, &cip, NULL);
+	if (unlikely(error)) {
+		if (unlikely(error != ENOENT))
+			return ERR_PTR(-error);
+		d_add(dentry, NULL);
+		return NULL;
+	}
+
+	return d_splice_alias(VFS_I(cip), dentry);
+}
+
+STATIC struct dentry *
+xfs_vn_ci_lookup(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	struct nameidata *nd)
+{
+	struct xfs_inode *ip;
+	struct xfs_name	xname;
+	struct xfs_name ci_name;
+	struct qstr	dname;
+	int		error;
+
+	if (dentry->d_name.len >= MAXNAMELEN)
+		return ERR_PTR(-ENAMETOOLONG);
+
+	xfs_dentry_to_name(&xname, dentry);
+	error = xfs_lookup(XFS_I(dir), &xname, &ip, &ci_name);
+	if (unlikely(error)) {
+		if (unlikely(error != ENOENT))
+			return ERR_PTR(-error);
+		/*
+		 * call d_add(dentry, NULL) here when d_drop_negative_children
+		 * is called in xfs_vn_mknod (ie. allow negative dentries
+		 * with CI filesystems).
+		 */
+		return NULL;
+	}
+
+	/* if exact match, just splice and exit */
+	if (!ci_name.name)
+		return d_splice_alias(VFS_I(ip), dentry);
+
+	/* else case-insensitive match... */
+	dname.name = ci_name.name;
+	dname.len = ci_name.len;
+	dentry = d_add_ci(dentry, VFS_I(ip), &dname);
+	kmem_free(ci_name.name);
+	return dentry;
+}
+
+STATIC int
+xfs_vn_link(
+	struct dentry	*old_dentry,
+	struct inode	*dir,
+	struct dentry	*dentry)
+{
+	struct inode	*inode = old_dentry->d_inode;
+	struct xfs_name	name;
+	int		error;
+
+	xfs_dentry_to_name(&name, dentry);
+
+	error = xfs_link(XFS_I(dir), XFS_I(inode), &name);
+	if (unlikely(error))
+		return -error;
+
+	ihold(inode);
+	d_instantiate(dentry, inode);
+	return 0;
+}
+
+STATIC int
+xfs_vn_unlink(
+	struct inode	*dir,
+	struct dentry	*dentry)
+{
+	struct xfs_name	name;
+	int		error;
+
+	xfs_dentry_to_name(&name, dentry);
+
+	error = -xfs_remove(XFS_I(dir), &name, XFS_I(dentry->d_inode));
+	if (error)
+		return error;
+
+	/*
+	 * With unlink, the VFS makes the dentry "negative": no inode,
+	 * but still hashed. This is incompatible with case-insensitive
+	 * mode, so invalidate (unhash) the dentry in CI-mode.
+	 */
+	if (xfs_sb_version_hasasciici(&XFS_M(dir->i_sb)->m_sb))
+		d_invalidate(dentry);
+	return 0;
+}
+
+STATIC int
+xfs_vn_symlink(
+	struct inode	*dir,
+	struct dentry	*dentry,
+	const char	*symname)
+{
+	struct inode	*inode;
+	struct xfs_inode *cip = NULL;
+	struct xfs_name	name;
+	int		error;
+	mode_t		mode;
+
+	mode = S_IFLNK |
+		(irix_symlink_mode ? 0777 & ~current_umask() : S_IRWXUGO);
+	xfs_dentry_to_name(&name, dentry);
+
+	error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip);
+	if (unlikely(error))
+		goto out;
+
+	inode = VFS_I(cip);
+
+	error = xfs_init_security(inode, dir, &dentry->d_name);
+	if (unlikely(error))
+		goto out_cleanup_inode;
+
+	d_instantiate(dentry, inode);
+	return 0;
+
+ out_cleanup_inode:
+	xfs_cleanup_inode(dir, inode, dentry);
+ out:
+	return -error;
+}
+
+STATIC int
+xfs_vn_rename(
+	struct inode	*odir,
+	struct dentry	*odentry,
+	struct inode	*ndir,
+	struct dentry	*ndentry)
+{
+	struct inode	*new_inode = ndentry->d_inode;
+	struct xfs_name	oname;
+	struct xfs_name	nname;
+
+	xfs_dentry_to_name(&oname, odentry);
+	xfs_dentry_to_name(&nname, ndentry);
+
+	return -xfs_rename(XFS_I(odir), &oname, XFS_I(odentry->d_inode),
+			   XFS_I(ndir), &nname, new_inode ?
+			   			XFS_I(new_inode) : NULL);
+}
+
+/*
+ * careful here - this function can get called recursively, so
+ * we need to be very careful about how much stack we use.
+ * uio is kmalloced for this reason...
+ */
+STATIC void *
+xfs_vn_follow_link(
+	struct dentry		*dentry,
+	struct nameidata	*nd)
+{
+	char			*link;
+	int			error = -ENOMEM;
+
+	link = kmalloc(MAXPATHLEN+1, GFP_KERNEL);
+	if (!link)
+		goto out_err;
+
+	error = -xfs_readlink(XFS_I(dentry->d_inode), link);
+	if (unlikely(error))
+		goto out_kfree;
+
+	nd_set_link(nd, link);
+	return NULL;
+
+ out_kfree:
+	kfree(link);
+ out_err:
+	nd_set_link(nd, ERR_PTR(error));
+	return NULL;
+}
+
+STATIC void
+xfs_vn_put_link(
+	struct dentry	*dentry,
+	struct nameidata *nd,
+	void		*p)
+{
+	char		*s = nd_get_link(nd);
+
+	if (!IS_ERR(s))
+		kfree(s);
+}
+
+STATIC int
+xfs_vn_getattr(
+	struct vfsmount		*mnt,
+	struct dentry		*dentry,
+	struct kstat		*stat)
+{
+	struct inode		*inode = dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+
+	trace_xfs_getattr(ip);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	stat->size = XFS_ISIZE(ip);
+	stat->dev = inode->i_sb->s_dev;
+	stat->mode = ip->i_d.di_mode;
+	stat->nlink = ip->i_d.di_nlink;
+	stat->uid = ip->i_d.di_uid;
+	stat->gid = ip->i_d.di_gid;
+	stat->ino = ip->i_ino;
+	stat->atime = inode->i_atime;
+	stat->mtime = inode->i_mtime;
+	stat->ctime = inode->i_ctime;
+	stat->blocks =
+		XFS_FSB_TO_BB(mp, ip->i_d.di_nblocks + ip->i_delayed_blks);
+
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFBLK:
+	case S_IFCHR:
+		stat->blksize = BLKDEV_IOSIZE;
+		stat->rdev = MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
+				   sysv_minor(ip->i_df.if_u2.if_rdev));
+		break;
+	default:
+		if (XFS_IS_REALTIME_INODE(ip)) {
+			/*
+			 * If the file blocks are being allocated from a
+			 * realtime volume, then return the inode's realtime
+			 * extent size or the realtime volume's extent size.
+			 */
+			stat->blksize =
+				xfs_get_extsz_hint(ip) << mp->m_sb.sb_blocklog;
+		} else
+			stat->blksize = xfs_preferred_iosize(mp);
+		stat->rdev = 0;
+		break;
+	}
+
+	return 0;
+}
+
+static void
+xfs_setattr_mode(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	struct iattr		*iattr)
+{
+	struct inode	*inode = VFS_I(ip);
+	umode_t		mode = iattr->ia_mode;
+
+	ASSERT(tp);
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
+		mode &= ~S_ISGID;
+
+	ip->i_d.di_mode &= S_IFMT;
+	ip->i_d.di_mode |= mode & ~S_IFMT;
+
+	inode->i_mode &= S_IFMT;
+	inode->i_mode |= mode & ~S_IFMT;
+}
+
+int
+xfs_setattr_nonsize(
+	struct xfs_inode	*ip,
+	struct iattr		*iattr,
+	int			flags)
+{
+	xfs_mount_t		*mp = ip->i_mount;
+	struct inode		*inode = VFS_I(ip);
+	int			mask = iattr->ia_valid;
+	xfs_trans_t		*tp;
+	int			error;
+	uid_t			uid = 0, iuid = 0;
+	gid_t			gid = 0, igid = 0;
+	struct xfs_dquot	*udqp = NULL, *gdqp = NULL;
+	struct xfs_dquot	*olddquot1 = NULL, *olddquot2 = NULL;
+
+	trace_xfs_setattr(ip);
+
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
+		return XFS_ERROR(EROFS);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	error = -inode_change_ok(inode, iattr);
+	if (error)
+		return XFS_ERROR(error);
+
+	ASSERT((mask & ATTR_SIZE) == 0);
+
+	/*
+	 * If disk quotas is on, we make sure that the dquots do exist on disk,
+	 * before we start any other transactions. Trying to do this later
+	 * is messy. We don't care to take a readlock to look at the ids
+	 * in inode here, because we can't hold it across the trans_reserve.
+	 * If the IDs do change before we take the ilock, we're covered
+	 * because the i_*dquot fields will get updated anyway.
+	 */
+	if (XFS_IS_QUOTA_ON(mp) && (mask & (ATTR_UID|ATTR_GID))) {
+		uint	qflags = 0;
+
+		if ((mask & ATTR_UID) && XFS_IS_UQUOTA_ON(mp)) {
+			uid = iattr->ia_uid;
+			qflags |= XFS_QMOPT_UQUOTA;
+		} else {
+			uid = ip->i_d.di_uid;
+		}
+		if ((mask & ATTR_GID) && XFS_IS_GQUOTA_ON(mp)) {
+			gid = iattr->ia_gid;
+			qflags |= XFS_QMOPT_GQUOTA;
+		}  else {
+			gid = ip->i_d.di_gid;
+		}
+
+		/*
+		 * We take a reference when we initialize udqp and gdqp,
+		 * so it is important that we never blindly double trip on
+		 * the same variable. See xfs_create() for an example.
+		 */
+		ASSERT(udqp == NULL);
+		ASSERT(gdqp == NULL);
+		error = xfs_qm_vop_dqalloc(ip, uid, gid, xfs_get_projid(ip),
+					 qflags, &udqp, &gdqp);
+		if (error)
+			return error;
+	}
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
+	error = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
+	if (error)
+		goto out_dqrele;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+	/*
+	 * Change file ownership.  Must be the owner or privileged.
+	 */
+	if (mask & (ATTR_UID|ATTR_GID)) {
+		/*
+		 * These IDs could have changed since we last looked at them.
+		 * But, we're assured that if the ownership did change
+		 * while we didn't have the inode locked, inode's dquot(s)
+		 * would have changed also.
+		 */
+		iuid = ip->i_d.di_uid;
+		igid = ip->i_d.di_gid;
+		gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
+		uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
+
+		/*
+		 * Do a quota reservation only if uid/gid is actually
+		 * going to change.
+		 */
+		if (XFS_IS_QUOTA_RUNNING(mp) &&
+		    ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
+		     (XFS_IS_GQUOTA_ON(mp) && igid != gid))) {
+			ASSERT(tp);
+			error = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
+						capable(CAP_FOWNER) ?
+						XFS_QMOPT_FORCE_RES : 0);
+			if (error)	/* out of quota */
+				goto out_trans_cancel;
+		}
+	}
+
+	xfs_trans_ijoin(tp, ip, 0);
+
+	/*
+	 * Change file ownership.  Must be the owner or privileged.
+	 */
+	if (mask & (ATTR_UID|ATTR_GID)) {
+		/*
+		 * CAP_FSETID overrides the following restrictions:
+		 *
+		 * The set-user-ID and set-group-ID bits of a file will be
+		 * cleared upon successful return from chown()
+		 */
+		if ((ip->i_d.di_mode & (S_ISUID|S_ISGID)) &&
+		    !capable(CAP_FSETID))
+			ip->i_d.di_mode &= ~(S_ISUID|S_ISGID);
+
+		/*
+		 * Change the ownerships and register quota modifications
+		 * in the transaction.
+		 */
+		if (iuid != uid) {
+			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
+				ASSERT(mask & ATTR_UID);
+				ASSERT(udqp);
+				olddquot1 = xfs_qm_vop_chown(tp, ip,
+							&ip->i_udquot, udqp);
+			}
+			ip->i_d.di_uid = uid;
+			inode->i_uid = uid;
+		}
+		if (igid != gid) {
+			if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_GQUOTA_ON(mp)) {
+				ASSERT(!XFS_IS_PQUOTA_ON(mp));
+				ASSERT(mask & ATTR_GID);
+				ASSERT(gdqp);
+				olddquot2 = xfs_qm_vop_chown(tp, ip,
+							&ip->i_gdquot, gdqp);
+			}
+			ip->i_d.di_gid = gid;
+			inode->i_gid = gid;
+		}
+	}
+
+	/*
+	 * Change file access modes.
+	 */
+	if (mask & ATTR_MODE)
+		xfs_setattr_mode(tp, ip, iattr);
+
+	/*
+	 * Change file access or modified times.
+	 */
+	if (mask & ATTR_ATIME) {
+		inode->i_atime = iattr->ia_atime;
+		ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
+		ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
+		ip->i_update_core = 1;
+	}
+	if (mask & ATTR_CTIME) {
+		inode->i_ctime = iattr->ia_ctime;
+		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
+		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
+		ip->i_update_core = 1;
+	}
+	if (mask & ATTR_MTIME) {
+		inode->i_mtime = iattr->ia_mtime;
+		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
+		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
+		ip->i_update_core = 1;
+	}
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	XFS_STATS_INC(xs_ig_attrchg);
+
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
+		xfs_trans_set_sync(tp);
+	error = xfs_trans_commit(tp, 0);
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	/*
+	 * Release any dquot(s) the inode had kept before chown.
+	 */
+	xfs_qm_dqrele(olddquot1);
+	xfs_qm_dqrele(olddquot2);
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+
+	if (error)
+		return XFS_ERROR(error);
+
+	/*
+	 * XXX(hch): Updating the ACL entries is not atomic vs the i_mode
+	 * 	     update.  We could avoid this with linked transactions
+	 * 	     and passing down the transaction pointer all the way
+	 *	     to attr_set.  No previous user of the generic
+	 * 	     Posix ACL code seems to care about this issue either.
+	 */
+	if ((mask & ATTR_MODE) && !(flags & XFS_ATTR_NOACL)) {
+		error = -xfs_acl_chmod(inode);
+		if (error)
+			return XFS_ERROR(error);
+	}
+
+	return 0;
+
+out_trans_cancel:
+	xfs_trans_cancel(tp, 0);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+out_dqrele:
+	xfs_qm_dqrele(udqp);
+	xfs_qm_dqrele(gdqp);
+	return error;
+}
+
+/*
+ * Truncate file.  Must have write permission and not be a directory.
+ */
+int
+xfs_setattr_size(
+	struct xfs_inode	*ip,
+	struct iattr		*iattr,
+	int			flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct inode		*inode = VFS_I(ip);
+	int			mask = iattr->ia_valid;
+	struct xfs_trans	*tp;
+	int			error;
+	uint			lock_flags;
+	uint			commit_flags = 0;
+	bool			did_zeroing = false;
+
+	trace_xfs_setattr(ip);
+
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
+		return XFS_ERROR(EROFS);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return XFS_ERROR(EIO);
+
+	error = -inode_change_ok(inode, iattr);
+	if (error)
+		return XFS_ERROR(error);
+
+	ASSERT(S_ISREG(ip->i_d.di_mode));
+	ASSERT((mask & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
+			ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
+
+	lock_flags = XFS_ILOCK_EXCL;
+	if (!(flags & XFS_ATTR_NOLOCK))
+		lock_flags |= XFS_IOLOCK_EXCL;
+	xfs_ilock(ip, lock_flags);
+
+	/*
+	 * Short circuit the truncate case for zero length files.
+	 */
+	if (iattr->ia_size == 0 &&
+	    ip->i_size == 0 && ip->i_d.di_nextents == 0) {
+		if (!(mask & (ATTR_CTIME|ATTR_MTIME)))
+			goto out_unlock;
+
+		/*
+		 * Use the regular setattr path to update the timestamps.
+		 */
+		xfs_iunlock(ip, lock_flags);
+		iattr->ia_valid &= ~ATTR_SIZE;
+		return xfs_setattr_nonsize(ip, iattr, 0);
+	}
+
+	/*
+	 * Make sure that the dquots are attached to the inode.
+	 */
+	error = xfs_qm_dqattach_locked(ip, 0);
+	if (error)
+		goto out_unlock;
+
+	/*
+	 * File data changes must be complete before we start the transaction to
+	 * modify the inode.  This needs to be done before joining the inode to
+	 * the transaction because the inode cannot be unlocked once it is a
+	 * part of the transaction.
+	 *
+	 * Start with zeroing any data block beyond EOF that we may expose on
+	 * file extension.
+	 */
+	if (iattr->ia_size > ip->i_size) {
+		error = xfs_zero_eof(ip, iattr->ia_size, ip->i_size, &did_zeroing);
+		if (error)
+			goto out_unlock;
+	}
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	lock_flags &= ~XFS_ILOCK_EXCL;
+
+	/*
+	 * We are going to log the inode size change in this transaction so
+	 * any previous writes that are beyond the on disk EOF and the new
+	 * EOF that have not been written out need to be written here.  If we
+	 * do not write the data out, we expose ourselves to the null files
+	 * problem. Note that this includes any block zeroing we did above;
+	 * otherwise those blocks may not be zeroed after a crash.
+	 */
+	if (iattr->ia_size > ip->i_d.di_size &&
+	    (ip->i_size != ip->i_d.di_size || did_zeroing)) {
+		error = xfs_flush_pages(ip, ip->i_d.di_size, iattr->ia_size, 0,
+					FI_NONE);
+		if (error)
+			goto out_unlock;
+	}
+
+	/* Now wait for all direct I/O to complete. */
+	inode_dio_wait(inode);
+
+	error = -block_truncate_page(inode->i_mapping, iattr->ia_size,
+				     xfs_get_blocks);
+	if (error)
+		goto out_unlock;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
+	error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0,
+				 XFS_TRANS_PERM_LOG_RES,
+				 XFS_ITRUNCATE_LOG_COUNT);
+	if (error)
+		goto out_trans_cancel;
+
+	truncate_setsize(inode, iattr->ia_size);
+
+	commit_flags = XFS_TRANS_RELEASE_LOG_RES;
+	lock_flags |= XFS_ILOCK_EXCL;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+	xfs_trans_ijoin(tp, ip, 0);
+
+	/*
+	 * Only change the c/mtime if we are changing the size or we are
+	 * explicitly asked to change it.  This handles the semantic difference
+	 * between truncate() and ftruncate() as implemented in the VFS.
+	 *
+	 * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
+	 * special case where we need to update the times despite not having
+	 * these flags set.  For all other operations the VFS set these flags
+	 * explicitly if it wants a timestamp update.
+	 */
+	if (iattr->ia_size != ip->i_size &&
+	    (!(mask & (ATTR_CTIME | ATTR_MTIME)))) {
+		iattr->ia_ctime = iattr->ia_mtime =
+			current_fs_time(inode->i_sb);
+		mask |= ATTR_CTIME | ATTR_MTIME;
+	}
+
+	if (iattr->ia_size > ip->i_size) {
+		ip->i_d.di_size = iattr->ia_size;
+		ip->i_size = iattr->ia_size;
+	} else if (iattr->ia_size <= ip->i_size ||
+		   (iattr->ia_size == 0 && ip->i_d.di_nextents)) {
+		error = xfs_itruncate_data(&tp, ip, iattr->ia_size);
+		if (error)
+			goto out_trans_abort;
+
+		/*
+		 * Truncated "down", so we're removing references to old data
+		 * here - if we delay flushing for a long time, we expose
+		 * ourselves unduly to the notorious NULL files problem.  So,
+		 * we mark this inode and flush it when the file is closed,
+		 * and do not wait the usual (long) time for writeout.
+		 */
+		xfs_iflags_set(ip, XFS_ITRUNCATED);
+	}
+
+	/*
+	 * Change file access modes.
+	 */
+	if (mask & ATTR_MODE)
+		xfs_setattr_mode(tp, ip, iattr);
+
+	if (mask & ATTR_CTIME) {
+		inode->i_ctime = iattr->ia_ctime;
+		ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
+		ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
+		ip->i_update_core = 1;
+	}
+	if (mask & ATTR_MTIME) {
+		inode->i_mtime = iattr->ia_mtime;
+		ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
+		ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
+		ip->i_update_core = 1;
+	}
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+	XFS_STATS_INC(xs_ig_attrchg);
+
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
+		xfs_trans_set_sync(tp);
+
+	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+out_unlock:
+	if (lock_flags)
+		xfs_iunlock(ip, lock_flags);
+	return error;
+
+out_trans_abort:
+	commit_flags |= XFS_TRANS_ABORT;
+out_trans_cancel:
+	xfs_trans_cancel(tp, commit_flags);
+	goto out_unlock;
+}
+
+STATIC int
+xfs_vn_setattr(
+	struct dentry	*dentry,
+	struct iattr	*iattr)
+{
+	if (iattr->ia_valid & ATTR_SIZE)
+		return -xfs_setattr_size(XFS_I(dentry->d_inode), iattr, 0);
+	return -xfs_setattr_nonsize(XFS_I(dentry->d_inode), iattr, 0);
+}
+
+#define XFS_FIEMAP_FLAGS	(FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
+
+/*
+ * Call fiemap helper to fill in user data.
+ * Returns positive errors to xfs_getbmap.
+ */
+STATIC int
+xfs_fiemap_format(
+	void			**arg,
+	struct getbmapx		*bmv,
+	int			*full)
+{
+	int			error;
+	struct fiemap_extent_info *fieinfo = *arg;
+	u32			fiemap_flags = 0;
+	u64			logical, physical, length;
+
+	/* Do nothing for a hole */
+	if (bmv->bmv_block == -1LL)
+		return 0;
+
+	logical = BBTOB(bmv->bmv_offset);
+	physical = BBTOB(bmv->bmv_block);
+	length = BBTOB(bmv->bmv_length);
+
+	if (bmv->bmv_oflags & BMV_OF_PREALLOC)
+		fiemap_flags |= FIEMAP_EXTENT_UNWRITTEN;
+	else if (bmv->bmv_oflags & BMV_OF_DELALLOC) {
+		fiemap_flags |= FIEMAP_EXTENT_DELALLOC;
+		physical = 0;   /* no block yet */
+	}
+	if (bmv->bmv_oflags & BMV_OF_LAST)
+		fiemap_flags |= FIEMAP_EXTENT_LAST;
+
+	error = fiemap_fill_next_extent(fieinfo, logical, physical,
+					length, fiemap_flags);
+	if (error > 0) {
+		error = 0;
+		*full = 1;	/* user array now full */
+	}
+
+	return -error;
+}
+
+STATIC int
+xfs_vn_fiemap(
+	struct inode		*inode,
+	struct fiemap_extent_info *fieinfo,
+	u64			start,
+	u64			length)
+{
+	xfs_inode_t		*ip = XFS_I(inode);
+	struct getbmapx		bm;
+	int			error;
+
+	error = fiemap_check_flags(fieinfo, XFS_FIEMAP_FLAGS);
+	if (error)
+		return error;
+
+	/* Set up bmap header for xfs internal routine */
+	bm.bmv_offset = BTOBB(start);
+	/* Special case for whole file */
+	if (length == FIEMAP_MAX_OFFSET)
+		bm.bmv_length = -1LL;
+	else
+		bm.bmv_length = BTOBB(length);
+
+	/* We add one because in getbmap world count includes the header */
+	bm.bmv_count = !fieinfo->fi_extents_max ? MAXEXTNUM :
+					fieinfo->fi_extents_max + 1;
+	bm.bmv_count = min_t(__s32, bm.bmv_count,
+			     (PAGE_SIZE * 16 / sizeof(struct getbmapx)));
+	bm.bmv_iflags = BMV_IF_PREALLOC | BMV_IF_NO_HOLES;
+	if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR)
+		bm.bmv_iflags |= BMV_IF_ATTRFORK;
+	if (!(fieinfo->fi_flags & FIEMAP_FLAG_SYNC))
+		bm.bmv_iflags |= BMV_IF_DELALLOC;
+
+	error = xfs_getbmap(ip, &bm, xfs_fiemap_format, fieinfo);
+	if (error)
+		return -error;
+
+	return 0;
+}
+
+static const struct inode_operations xfs_inode_operations = {
+	.get_acl		= xfs_get_acl,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= generic_setxattr,
+	.getxattr		= generic_getxattr,
+	.removexattr		= generic_removexattr,
+	.listxattr		= xfs_vn_listxattr,
+	.fiemap			= xfs_vn_fiemap,
+};
+
+static const struct inode_operations xfs_dir_inode_operations = {
+	.create			= xfs_vn_create,
+	.lookup			= xfs_vn_lookup,
+	.link			= xfs_vn_link,
+	.unlink			= xfs_vn_unlink,
+	.symlink		= xfs_vn_symlink,
+	.mkdir			= xfs_vn_mkdir,
+	/*
+	 * Yes, XFS uses the same method for rmdir and unlink.
+	 *
+	 * There are some subtile differences deeper in the code,
+	 * but we use S_ISDIR to check for those.
+	 */
+	.rmdir			= xfs_vn_unlink,
+	.mknod			= xfs_vn_mknod,
+	.rename			= xfs_vn_rename,
+	.get_acl		= xfs_get_acl,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= generic_setxattr,
+	.getxattr		= generic_getxattr,
+	.removexattr		= generic_removexattr,
+	.listxattr		= xfs_vn_listxattr,
+};
+
+static const struct inode_operations xfs_dir_ci_inode_operations = {
+	.create			= xfs_vn_create,
+	.lookup			= xfs_vn_ci_lookup,
+	.link			= xfs_vn_link,
+	.unlink			= xfs_vn_unlink,
+	.symlink		= xfs_vn_symlink,
+	.mkdir			= xfs_vn_mkdir,
+	/*
+	 * Yes, XFS uses the same method for rmdir and unlink.
+	 *
+	 * There are some subtile differences deeper in the code,
+	 * but we use S_ISDIR to check for those.
+	 */
+	.rmdir			= xfs_vn_unlink,
+	.mknod			= xfs_vn_mknod,
+	.rename			= xfs_vn_rename,
+	.get_acl		= xfs_get_acl,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= generic_setxattr,
+	.getxattr		= generic_getxattr,
+	.removexattr		= generic_removexattr,
+	.listxattr		= xfs_vn_listxattr,
+};
+
+static const struct inode_operations xfs_symlink_inode_operations = {
+	.readlink		= generic_readlink,
+	.follow_link		= xfs_vn_follow_link,
+	.put_link		= xfs_vn_put_link,
+	.get_acl		= xfs_get_acl,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= generic_setxattr,
+	.getxattr		= generic_getxattr,
+	.removexattr		= generic_removexattr,
+	.listxattr		= xfs_vn_listxattr,
+};
+
+STATIC void
+xfs_diflags_to_iflags(
+	struct inode		*inode,
+	struct xfs_inode	*ip)
+{
+	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
+		inode->i_flags |= S_IMMUTABLE;
+	else
+		inode->i_flags &= ~S_IMMUTABLE;
+	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
+		inode->i_flags |= S_APPEND;
+	else
+		inode->i_flags &= ~S_APPEND;
+	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
+		inode->i_flags |= S_SYNC;
+	else
+		inode->i_flags &= ~S_SYNC;
+	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
+		inode->i_flags |= S_NOATIME;
+	else
+		inode->i_flags &= ~S_NOATIME;
+}
+
+/*
+ * Initialize the Linux inode, set up the operation vectors and
+ * unlock the inode.
+ *
+ * When reading existing inodes from disk this is called directly
+ * from xfs_iget, when creating a new inode it is called from
+ * xfs_ialloc after setting up the inode.
+ *
+ * We are always called with an uninitialised linux inode here.
+ * We need to initialise the necessary fields and take a reference
+ * on it.
+ */
+void
+xfs_setup_inode(
+	struct xfs_inode	*ip)
+{
+	struct inode		*inode = &ip->i_vnode;
+
+	inode->i_ino = ip->i_ino;
+	inode->i_state = I_NEW;
+
+	inode_sb_list_add(inode);
+	/* make the inode look hashed for the writeback code */
+	hlist_add_fake(&inode->i_hash);
+
+	inode->i_mode	= ip->i_d.di_mode;
+	set_nlink(inode, ip->i_d.di_nlink);
+	inode->i_uid	= ip->i_d.di_uid;
+	inode->i_gid	= ip->i_d.di_gid;
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFBLK:
+	case S_IFCHR:
+		inode->i_rdev =
+			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
+			      sysv_minor(ip->i_df.if_u2.if_rdev));
+		break;
+	default:
+		inode->i_rdev = 0;
+		break;
+	}
+
+	inode->i_generation = ip->i_d.di_gen;
+	i_size_write(inode, ip->i_d.di_size);
+	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
+	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
+	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
+	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
+	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
+	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
+	xfs_diflags_to_iflags(inode, ip);
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFREG:
+		inode->i_op = &xfs_inode_operations;
+		inode->i_fop = &xfs_file_operations;
+		inode->i_mapping->a_ops = &xfs_address_space_operations;
+		break;
+	case S_IFDIR:
+		if (xfs_sb_version_hasasciici(&XFS_M(inode->i_sb)->m_sb))
+			inode->i_op = &xfs_dir_ci_inode_operations;
+		else
+			inode->i_op = &xfs_dir_inode_operations;
+		inode->i_fop = &xfs_dir_file_operations;
+		break;
+	case S_IFLNK:
+		inode->i_op = &xfs_symlink_inode_operations;
+		if (!(ip->i_df.if_flags & XFS_IFINLINE))
+			inode->i_mapping->a_ops = &xfs_address_space_operations;
+		break;
+	default:
+		inode->i_op = &xfs_inode_operations;
+		init_special_inode(inode, inode->i_mode, inode->i_rdev);
+		break;
+	}
+
+	/*
+	 * If there is no attribute fork no ACL can exist on this inode,
+	 * and it can't have any file capabilities attached to it either.
+	 */
+	if (!XFS_IFORK_Q(ip)) {
+		inode_has_no_xattr(inode);
+		cache_no_acl(inode);
+	}
+
+	xfs_iflags_clear(ip, XFS_INEW);
+	barrier();
+
+	unlock_new_inode(inode);
+}
diff --git a/fs/xfs/xfs_iops.h b/fs/xfs/xfs_iops.h
new file mode 100644
index 0000000..ef41c92
--- /dev/null
+++ b/fs/xfs/xfs_iops.h
@@ -0,0 +1,30 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_IOPS_H__
+#define __XFS_IOPS_H__
+
+struct xfs_inode;
+
+extern const struct file_operations xfs_file_operations;
+extern const struct file_operations xfs_dir_file_operations;
+
+extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);
+
+extern void xfs_setup_inode(struct xfs_inode *);
+
+#endif /* __XFS_IOPS_H__ */
diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h
new file mode 100644
index 0000000..828662f
--- /dev/null
+++ b/fs/xfs/xfs_linux.h
@@ -0,0 +1,311 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_LINUX__
+#define __XFS_LINUX__
+
+#include <linux/types.h>
+
+/*
+ * XFS_BIG_BLKNOS needs block layer disk addresses to be 64 bits.
+ * XFS_BIG_INUMS requires XFS_BIG_BLKNOS to be set.
+ */
+#if defined(CONFIG_LBDAF) || (BITS_PER_LONG == 64)
+# define XFS_BIG_BLKNOS	1
+# define XFS_BIG_INUMS	1
+#else
+# define XFS_BIG_BLKNOS	0
+# define XFS_BIG_INUMS	0
+#endif
+
+#include "xfs_types.h"
+
+#include "kmem.h"
+#include "mrlock.h"
+#include "time.h"
+#include "uuid.h"
+
+#include <linux/semaphore.h>
+#include <linux/mm.h>
+#include <linux/kernel.h>
+#include <linux/blkdev.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/file.h>
+#include <linux/swap.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/bitops.h>
+#include <linux/major.h>
+#include <linux/pagemap.h>
+#include <linux/vfs.h>
+#include <linux/seq_file.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/proc_fs.h>
+#include <linux/sort.h>
+#include <linux/cpu.h>
+#include <linux/notifier.h>
+#include <linux/delay.h>
+#include <linux/log2.h>
+#include <linux/spinlock.h>
+#include <linux/random.h>
+#include <linux/ctype.h>
+#include <linux/writeback.h>
+#include <linux/capability.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
+#include <linux/list_sort.h>
+
+#include <asm/page.h>
+#include <asm/div64.h>
+#include <asm/param.h>
+#include <asm/uaccess.h>
+#include <asm/byteorder.h>
+#include <asm/unaligned.h>
+
+#include "xfs_vnode.h"
+#include "xfs_stats.h"
+#include "xfs_sysctl.h"
+#include "xfs_iops.h"
+#include "xfs_aops.h"
+#include "xfs_super.h"
+#include "xfs_buf.h"
+#include "xfs_message.h"
+
+#ifdef __BIG_ENDIAN
+#define XFS_NATIVE_HOST 1
+#else
+#undef XFS_NATIVE_HOST
+#endif
+
+/*
+ * Feature macros (disable/enable)
+ */
+#ifdef CONFIG_SMP
+#define HAVE_PERCPU_SB	/* per cpu superblock counters are a 2.6 feature */
+#else
+#undef  HAVE_PERCPU_SB	/* per cpu superblock counters are a 2.6 feature */
+#endif
+
+#define irix_sgid_inherit	xfs_params.sgid_inherit.val
+#define irix_symlink_mode	xfs_params.symlink_mode.val
+#define xfs_panic_mask		xfs_params.panic_mask.val
+#define xfs_error_level		xfs_params.error_level.val
+#define xfs_syncd_centisecs	xfs_params.syncd_timer.val
+#define xfs_stats_clear		xfs_params.stats_clear.val
+#define xfs_inherit_sync	xfs_params.inherit_sync.val
+#define xfs_inherit_nodump	xfs_params.inherit_nodump.val
+#define xfs_inherit_noatime	xfs_params.inherit_noatim.val
+#define xfs_buf_timer_centisecs	xfs_params.xfs_buf_timer.val
+#define xfs_buf_age_centisecs	xfs_params.xfs_buf_age.val
+#define xfs_inherit_nosymlinks	xfs_params.inherit_nosym.val
+#define xfs_rotorstep		xfs_params.rotorstep.val
+#define xfs_inherit_nodefrag	xfs_params.inherit_nodfrg.val
+#define xfs_fstrm_centisecs	xfs_params.fstrm_timer.val
+
+#define current_cpu()		(raw_smp_processor_id())
+#define current_pid()		(current->pid)
+#define current_test_flags(f)	(current->flags & (f))
+#define current_set_flags_nested(sp, f)		\
+		(*(sp) = current->flags, current->flags |= (f))
+#define current_clear_flags_nested(sp, f)	\
+		(*(sp) = current->flags, current->flags &= ~(f))
+#define current_restore_flags_nested(sp, f)	\
+		(current->flags = ((current->flags & ~(f)) | (*(sp) & (f))))
+
+#define spinlock_destroy(lock)
+
+#define NBBY		8		/* number of bits per byte */
+
+/*
+ * Size of block device i/o is parameterized here.
+ * Currently the system supports page-sized i/o.
+ */
+#define	BLKDEV_IOSHIFT		PAGE_CACHE_SHIFT
+#define	BLKDEV_IOSIZE		(1<<BLKDEV_IOSHIFT)
+/* number of BB's per block device block */
+#define	BLKDEV_BB		BTOBB(BLKDEV_IOSIZE)
+
+#define ENOATTR		ENODATA		/* Attribute not found */
+#define EWRONGFS	EINVAL		/* Mount with wrong filesystem type */
+#define EFSCORRUPTED	EUCLEAN		/* Filesystem is corrupted */
+
+#define SYNCHRONIZE()	barrier()
+#define __return_address __builtin_return_address(0)
+
+#define XFS_PROJID_DEFAULT	0
+#define MAXPATHLEN	1024
+
+#define MIN(a,b)	(min(a,b))
+#define MAX(a,b)	(max(a,b))
+#define howmany(x, y)	(((x)+((y)-1))/(y))
+
+/*
+ * Various platform dependent calls that don't fit anywhere else
+ */
+#define xfs_sort(a,n,s,fn)	sort(a,n,s,fn,NULL)
+#define xfs_stack_trace()	dump_stack()
+
+
+/* Move the kernel do_div definition off to one side */
+
+#if defined __i386__
+/* For ia32 we need to pull some tricks to get past various versions
+ * of the compiler which do not like us using do_div in the middle
+ * of large functions.
+ */
+static inline __u32 xfs_do_div(void *a, __u32 b, int n)
+{
+	__u32	mod;
+
+	switch (n) {
+		case 4:
+			mod = *(__u32 *)a % b;
+			*(__u32 *)a = *(__u32 *)a / b;
+			return mod;
+		case 8:
+			{
+			unsigned long __upper, __low, __high, __mod;
+			__u64	c = *(__u64 *)a;
+			__upper = __high = c >> 32;
+			__low = c;
+			if (__high) {
+				__upper = __high % (b);
+				__high = __high / (b);
+			}
+			asm("divl %2":"=a" (__low), "=d" (__mod):"rm" (b), "0" (__low), "1" (__upper));
+			asm("":"=A" (c):"a" (__low),"d" (__high));
+			*(__u64 *)a = c;
+			return __mod;
+			}
+	}
+
+	/* NOTREACHED */
+	return 0;
+}
+
+/* Side effect free 64 bit mod operation */
+static inline __u32 xfs_do_mod(void *a, __u32 b, int n)
+{
+	switch (n) {
+		case 4:
+			return *(__u32 *)a % b;
+		case 8:
+			{
+			unsigned long __upper, __low, __high, __mod;
+			__u64	c = *(__u64 *)a;
+			__upper = __high = c >> 32;
+			__low = c;
+			if (__high) {
+				__upper = __high % (b);
+				__high = __high / (b);
+			}
+			asm("divl %2":"=a" (__low), "=d" (__mod):"rm" (b), "0" (__low), "1" (__upper));
+			asm("":"=A" (c):"a" (__low),"d" (__high));
+			return __mod;
+			}
+	}
+
+	/* NOTREACHED */
+	return 0;
+}
+#else
+static inline __u32 xfs_do_div(void *a, __u32 b, int n)
+{
+	__u32	mod;
+
+	switch (n) {
+		case 4:
+			mod = *(__u32 *)a % b;
+			*(__u32 *)a = *(__u32 *)a / b;
+			return mod;
+		case 8:
+			mod = do_div(*(__u64 *)a, b);
+			return mod;
+	}
+
+	/* NOTREACHED */
+	return 0;
+}
+
+/* Side effect free 64 bit mod operation */
+static inline __u32 xfs_do_mod(void *a, __u32 b, int n)
+{
+	switch (n) {
+		case 4:
+			return *(__u32 *)a % b;
+		case 8:
+			{
+			__u64	c = *(__u64 *)a;
+			return do_div(c, b);
+			}
+	}
+
+	/* NOTREACHED */
+	return 0;
+}
+#endif
+
+#undef do_div
+#define do_div(a, b)	xfs_do_div(&(a), (b), sizeof(a))
+#define do_mod(a, b)	xfs_do_mod(&(a), (b), sizeof(a))
+
+static inline __uint64_t roundup_64(__uint64_t x, __uint32_t y)
+{
+	x += y - 1;
+	do_div(x, y);
+	return(x * y);
+}
+
+static inline __uint64_t howmany_64(__uint64_t x, __uint32_t y)
+{
+	x += y - 1;
+	do_div(x, y);
+	return x;
+}
+
+/* ARM old ABI has some weird alignment/padding */
+#if defined(__arm__) && !defined(__ARM_EABI__)
+#define __arch_pack __attribute__((packed))
+#else
+#define __arch_pack
+#endif
+
+#define ASSERT_ALWAYS(expr)	\
+	(unlikely(expr) ? (void)0 : assfail(#expr, __FILE__, __LINE__))
+
+#ifndef DEBUG
+#define ASSERT(expr)	((void)0)
+
+#ifndef STATIC
+# define STATIC static noinline
+#endif
+
+#else /* DEBUG */
+
+#define ASSERT(expr)	\
+	(unlikely(expr) ? (void)0 : assfail(#expr, __FILE__, __LINE__))
+
+#ifndef STATIC
+# define STATIC noinline
+#endif
+
+#endif /* DEBUG */
+
+#endif /* __XFS_LINUX__ */
diff --git a/fs/xfs/xfs_message.c b/fs/xfs/xfs_message.c
new file mode 100644
index 0000000..bd672de
--- /dev/null
+++ b/fs/xfs/xfs_message.c
@@ -0,0 +1,108 @@
+/*
+ * Copyright (c) 2011 Red Hat, Inc.  All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_types.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+
+/*
+ * XFS logging functions
+ */
+static void
+__xfs_printk(
+	const char		*level,
+	const struct xfs_mount	*mp,
+	struct va_format	*vaf)
+{
+	if (mp && mp->m_fsname) {
+		printk("%sXFS (%s): %pV\n", level, mp->m_fsname, vaf);
+		return;
+	}
+	printk("%sXFS: %pV\n", level, vaf);
+}
+
+#define define_xfs_printk_level(func, kern_level)		\
+void func(const struct xfs_mount *mp, const char *fmt, ...)	\
+{								\
+	struct va_format	vaf;				\
+	va_list			args;				\
+								\
+	va_start(args, fmt);					\
+								\
+	vaf.fmt = fmt;						\
+	vaf.va = &args;						\
+								\
+	__xfs_printk(kern_level, mp, &vaf);			\
+	va_end(args);						\
+}								\
+
+define_xfs_printk_level(xfs_emerg, KERN_EMERG);
+define_xfs_printk_level(xfs_alert, KERN_ALERT);
+define_xfs_printk_level(xfs_crit, KERN_CRIT);
+define_xfs_printk_level(xfs_err, KERN_ERR);
+define_xfs_printk_level(xfs_warn, KERN_WARNING);
+define_xfs_printk_level(xfs_notice, KERN_NOTICE);
+define_xfs_printk_level(xfs_info, KERN_INFO);
+#ifdef DEBUG
+define_xfs_printk_level(xfs_debug, KERN_DEBUG);
+#endif
+
+void
+xfs_alert_tag(
+	const struct xfs_mount	*mp,
+	int			panic_tag,
+	const char		*fmt, ...)
+{
+	struct va_format	vaf;
+	va_list			args;
+	int			do_panic = 0;
+
+	if (xfs_panic_mask && (xfs_panic_mask & panic_tag)) {
+		xfs_alert(mp, "Transforming an alert into a BUG.");
+		do_panic = 1;
+	}
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	__xfs_printk(KERN_ALERT, mp, &vaf);
+	va_end(args);
+
+	BUG_ON(do_panic);
+}
+
+void
+assfail(char *expr, char *file, int line)
+{
+	xfs_emerg(NULL, "Assertion failed: %s, file: %s, line: %d",
+		expr, file, line);
+	BUG();
+}
+
+void
+xfs_hex_dump(void *p, int length)
+{
+	print_hex_dump(KERN_ALERT, "", DUMP_PREFIX_ADDRESS, 16, 1, p, length, 1);
+}
diff --git a/fs/xfs/xfs_message.h b/fs/xfs/xfs_message.h
new file mode 100644
index 0000000..56dc0c1
--- /dev/null
+++ b/fs/xfs/xfs_message.h
@@ -0,0 +1,37 @@
+#ifndef __XFS_MESSAGE_H
+#define __XFS_MESSAGE_H 1
+
+struct xfs_mount;
+
+extern __printf(2, 3)
+void xfs_emerg(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_alert(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(3, 4)
+void xfs_alert_tag(const struct xfs_mount *mp, int tag, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_crit(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_err(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_warn(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_notice(const struct xfs_mount *mp, const char *fmt, ...);
+extern __printf(2, 3)
+void xfs_info(const struct xfs_mount *mp, const char *fmt, ...);
+
+#ifdef DEBUG
+extern __printf(2, 3)
+void xfs_debug(const struct xfs_mount *mp, const char *fmt, ...);
+#else
+static inline __printf(2, 3)
+void xfs_debug(const struct xfs_mount *mp, const char *fmt, ...)
+{
+}
+#endif
+
+extern void assfail(char *expr, char *f, int l);
+
+extern void xfs_hex_dump(void *p, int length);
+
+#endif	/* __XFS_MESSAGE_H */
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
new file mode 100644
index 0000000..0bbb1a4
--- /dev/null
+++ b/fs/xfs/xfs_qm.c
@@ -0,0 +1,2415 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_ialloc.h"
+#include "xfs_itable.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_bmap.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_space.h"
+#include "xfs_utils.h"
+#include "xfs_qm.h"
+#include "xfs_trace.h"
+
+/*
+ * The global quota manager. There is only one of these for the entire
+ * system, _not_ one per file system. XQM keeps track of the overall
+ * quota functionality, including maintaining the freelist and hash
+ * tables of dquots.
+ */
+struct mutex	xfs_Gqm_lock;
+struct xfs_qm	*xfs_Gqm;
+uint		ndquot;
+
+kmem_zone_t	*qm_dqzone;
+kmem_zone_t	*qm_dqtrxzone;
+
+STATIC void	xfs_qm_list_init(xfs_dqlist_t *, char *, int);
+STATIC void	xfs_qm_list_destroy(xfs_dqlist_t *);
+
+STATIC int	xfs_qm_init_quotainos(xfs_mount_t *);
+STATIC int	xfs_qm_init_quotainfo(xfs_mount_t *);
+STATIC int	xfs_qm_shake(struct shrinker *, struct shrink_control *);
+
+static struct shrinker xfs_qm_shaker = {
+	.shrink = xfs_qm_shake,
+	.seeks = DEFAULT_SEEKS,
+};
+
+/*
+ * Initialize the XQM structure.
+ * Note that there is not one quota manager per file system.
+ */
+STATIC struct xfs_qm *
+xfs_Gqm_init(void)
+{
+	xfs_dqhash_t	*udqhash, *gdqhash;
+	xfs_qm_t	*xqm;
+	size_t		hsize;
+	uint		i;
+
+	/*
+	 * Initialize the dquot hash tables.
+	 */
+	udqhash = kmem_zalloc_greedy(&hsize,
+				     XFS_QM_HASHSIZE_LOW * sizeof(xfs_dqhash_t),
+				     XFS_QM_HASHSIZE_HIGH * sizeof(xfs_dqhash_t));
+	if (!udqhash)
+		goto out;
+
+	gdqhash = kmem_zalloc_large(hsize);
+	if (!gdqhash)
+		goto out_free_udqhash;
+
+	hsize /= sizeof(xfs_dqhash_t);
+	ndquot = hsize << 8;
+
+	xqm = kmem_zalloc(sizeof(xfs_qm_t), KM_SLEEP);
+	xqm->qm_dqhashmask = hsize - 1;
+	xqm->qm_usr_dqhtable = udqhash;
+	xqm->qm_grp_dqhtable = gdqhash;
+	ASSERT(xqm->qm_usr_dqhtable != NULL);
+	ASSERT(xqm->qm_grp_dqhtable != NULL);
+
+	for (i = 0; i < hsize; i++) {
+		xfs_qm_list_init(&(xqm->qm_usr_dqhtable[i]), "uxdqh", i);
+		xfs_qm_list_init(&(xqm->qm_grp_dqhtable[i]), "gxdqh", i);
+	}
+
+	/*
+	 * Freelist of all dquots of all file systems
+	 */
+	INIT_LIST_HEAD(&xqm->qm_dqfrlist);
+	xqm->qm_dqfrlist_cnt = 0;
+	mutex_init(&xqm->qm_dqfrlist_lock);
+
+	/*
+	 * dquot zone. we register our own low-memory callback.
+	 */
+	if (!qm_dqzone) {
+		xqm->qm_dqzone = kmem_zone_init(sizeof(xfs_dquot_t),
+						"xfs_dquots");
+		qm_dqzone = xqm->qm_dqzone;
+	} else
+		xqm->qm_dqzone = qm_dqzone;
+
+	register_shrinker(&xfs_qm_shaker);
+
+	/*
+	 * The t_dqinfo portion of transactions.
+	 */
+	if (!qm_dqtrxzone) {
+		xqm->qm_dqtrxzone = kmem_zone_init(sizeof(xfs_dquot_acct_t),
+						   "xfs_dqtrx");
+		qm_dqtrxzone = xqm->qm_dqtrxzone;
+	} else
+		xqm->qm_dqtrxzone = qm_dqtrxzone;
+
+	atomic_set(&xqm->qm_totaldquots, 0);
+	xqm->qm_dqfree_ratio = XFS_QM_DQFREE_RATIO;
+	xqm->qm_nrefs = 0;
+	return xqm;
+
+ out_free_udqhash:
+	kmem_free_large(udqhash);
+ out:
+	return NULL;
+}
+
+/*
+ * Destroy the global quota manager when its reference count goes to zero.
+ */
+STATIC void
+xfs_qm_destroy(
+	struct xfs_qm	*xqm)
+{
+	struct xfs_dquot *dqp, *n;
+	int		hsize, i;
+
+	ASSERT(xqm != NULL);
+	ASSERT(xqm->qm_nrefs == 0);
+	unregister_shrinker(&xfs_qm_shaker);
+	hsize = xqm->qm_dqhashmask + 1;
+	for (i = 0; i < hsize; i++) {
+		xfs_qm_list_destroy(&(xqm->qm_usr_dqhtable[i]));
+		xfs_qm_list_destroy(&(xqm->qm_grp_dqhtable[i]));
+	}
+	kmem_free_large(xqm->qm_usr_dqhtable);
+	kmem_free_large(xqm->qm_grp_dqhtable);
+	xqm->qm_usr_dqhtable = NULL;
+	xqm->qm_grp_dqhtable = NULL;
+	xqm->qm_dqhashmask = 0;
+
+	/* frlist cleanup */
+	mutex_lock(&xqm->qm_dqfrlist_lock);
+	list_for_each_entry_safe(dqp, n, &xqm->qm_dqfrlist, q_freelist) {
+		xfs_dqlock(dqp);
+		list_del_init(&dqp->q_freelist);
+		xfs_Gqm->qm_dqfrlist_cnt--;
+		xfs_dqunlock(dqp);
+		xfs_qm_dqdestroy(dqp);
+	}
+	mutex_unlock(&xqm->qm_dqfrlist_lock);
+	mutex_destroy(&xqm->qm_dqfrlist_lock);
+	kmem_free(xqm);
+}
+
+/*
+ * Called at mount time to let XQM know that another file system is
+ * starting quotas. This isn't crucial information as the individual mount
+ * structures are pretty independent, but it helps the XQM keep a
+ * global view of what's going on.
+ */
+/* ARGSUSED */
+STATIC int
+xfs_qm_hold_quotafs_ref(
+	struct xfs_mount *mp)
+{
+	/*
+	 * Need to lock the xfs_Gqm structure for things like this. For example,
+	 * the structure could disappear between the entry to this routine and
+	 * a HOLD operation if not locked.
+	 */
+	mutex_lock(&xfs_Gqm_lock);
+
+	if (!xfs_Gqm) {
+		xfs_Gqm = xfs_Gqm_init();
+		if (!xfs_Gqm) {
+			mutex_unlock(&xfs_Gqm_lock);
+			return ENOMEM;
+		}
+	}
+
+	/*
+	 * We can keep a list of all filesystems with quotas mounted for
+	 * debugging and statistical purposes, but ...
+	 * Just take a reference and get out.
+	 */
+	xfs_Gqm->qm_nrefs++;
+	mutex_unlock(&xfs_Gqm_lock);
+
+	return 0;
+}
+
+
+/*
+ * Release the reference that a filesystem took at mount time,
+ * so that we know when we need to destroy the entire quota manager.
+ */
+/* ARGSUSED */
+STATIC void
+xfs_qm_rele_quotafs_ref(
+	struct xfs_mount *mp)
+{
+	xfs_dquot_t	*dqp, *n;
+
+	ASSERT(xfs_Gqm);
+	ASSERT(xfs_Gqm->qm_nrefs > 0);
+
+	/*
+	 * Go thru the freelist and destroy all inactive dquots.
+	 */
+	mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);
+
+	list_for_each_entry_safe(dqp, n, &xfs_Gqm->qm_dqfrlist, q_freelist) {
+		xfs_dqlock(dqp);
+		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
+			ASSERT(dqp->q_mount == NULL);
+			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
+			ASSERT(list_empty(&dqp->q_hashlist));
+			ASSERT(list_empty(&dqp->q_mplist));
+			list_del_init(&dqp->q_freelist);
+			xfs_Gqm->qm_dqfrlist_cnt--;
+			xfs_dqunlock(dqp);
+			xfs_qm_dqdestroy(dqp);
+		} else {
+			xfs_dqunlock(dqp);
+		}
+	}
+	mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+
+	/*
+	 * Destroy the entire XQM. If somebody mounts with quotaon, this'll
+	 * be restarted.
+	 */
+	mutex_lock(&xfs_Gqm_lock);
+	if (--xfs_Gqm->qm_nrefs == 0) {
+		xfs_qm_destroy(xfs_Gqm);
+		xfs_Gqm = NULL;
+	}
+	mutex_unlock(&xfs_Gqm_lock);
+}
+
+/*
+ * Just destroy the quotainfo structure.
+ */
+void
+xfs_qm_unmount(
+	struct xfs_mount	*mp)
+{
+	if (mp->m_quotainfo) {
+		xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL);
+		xfs_qm_destroy_quotainfo(mp);
+	}
+}
+
+
+/*
+ * This is called from xfs_mountfs to start quotas and initialize all
+ * necessary data structures like quotainfo.  This is also responsible for
+ * running a quotacheck as necessary.  We are guaranteed that the superblock
+ * is consistently read in at this point.
+ *
+ * If we fail here, the mount will continue with quota turned off. We don't
+ * need to inidicate success or failure at all.
+ */
+void
+xfs_qm_mount_quotas(
+	xfs_mount_t	*mp)
+{
+	int		error = 0;
+	uint		sbf;
+
+	/*
+	 * If quotas on realtime volumes is not supported, we disable
+	 * quotas immediately.
+	 */
+	if (mp->m_sb.sb_rextents) {
+		xfs_notice(mp, "Cannot turn on quotas for realtime filesystem");
+		mp->m_qflags = 0;
+		goto write_changes;
+	}
+
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	/*
+	 * Allocate the quotainfo structure inside the mount struct, and
+	 * create quotainode(s), and change/rev superblock if necessary.
+	 */
+	error = xfs_qm_init_quotainfo(mp);
+	if (error) {
+		/*
+		 * We must turn off quotas.
+		 */
+		ASSERT(mp->m_quotainfo == NULL);
+		mp->m_qflags = 0;
+		goto write_changes;
+	}
+	/*
+	 * If any of the quotas are not consistent, do a quotacheck.
+	 */
+	if (XFS_QM_NEED_QUOTACHECK(mp)) {
+		error = xfs_qm_quotacheck(mp);
+		if (error) {
+			/* Quotacheck failed and disabled quotas. */
+			return;
+		}
+	}
+	/* 
+	 * If one type of quotas is off, then it will lose its
+	 * quotachecked status, since we won't be doing accounting for
+	 * that type anymore.
+	 */
+	if (!XFS_IS_UQUOTA_ON(mp))
+		mp->m_qflags &= ~XFS_UQUOTA_CHKD;
+	if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp)))
+		mp->m_qflags &= ~XFS_OQUOTA_CHKD;
+
+ write_changes:
+	/*
+	 * We actually don't have to acquire the m_sb_lock at all.
+	 * This can only be called from mount, and that's single threaded. XXX
+	 */
+	spin_lock(&mp->m_sb_lock);
+	sbf = mp->m_sb.sb_qflags;
+	mp->m_sb.sb_qflags = mp->m_qflags & XFS_MOUNT_QUOTA_ALL;
+	spin_unlock(&mp->m_sb_lock);
+
+	if (sbf != (mp->m_qflags & XFS_MOUNT_QUOTA_ALL)) {
+		if (xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS)) {
+			/*
+			 * We could only have been turning quotas off.
+			 * We aren't in very good shape actually because
+			 * the incore structures are convinced that quotas are
+			 * off, but the on disk superblock doesn't know that !
+			 */
+			ASSERT(!(XFS_IS_QUOTA_RUNNING(mp)));
+			xfs_alert(mp, "%s: Superblock update failed!",
+				__func__);
+		}
+	}
+
+	if (error) {
+		xfs_warn(mp, "Failed to initialize disk quotas.");
+		return;
+	}
+}
+
+/*
+ * Called from the vfsops layer.
+ */
+void
+xfs_qm_unmount_quotas(
+	xfs_mount_t	*mp)
+{
+	/*
+	 * Release the dquots that root inode, et al might be holding,
+	 * before we flush quotas and blow away the quotainfo structure.
+	 */
+	ASSERT(mp->m_rootip);
+	xfs_qm_dqdetach(mp->m_rootip);
+	if (mp->m_rbmip)
+		xfs_qm_dqdetach(mp->m_rbmip);
+	if (mp->m_rsumip)
+		xfs_qm_dqdetach(mp->m_rsumip);
+
+	/*
+	 * Release the quota inodes.
+	 */
+	if (mp->m_quotainfo) {
+		if (mp->m_quotainfo->qi_uquotaip) {
+			IRELE(mp->m_quotainfo->qi_uquotaip);
+			mp->m_quotainfo->qi_uquotaip = NULL;
+		}
+		if (mp->m_quotainfo->qi_gquotaip) {
+			IRELE(mp->m_quotainfo->qi_gquotaip);
+			mp->m_quotainfo->qi_gquotaip = NULL;
+		}
+	}
+}
+
+/*
+ * Flush all dquots of the given file system to disk. The dquots are
+ * _not_ purged from memory here, just their data written to disk.
+ */
+STATIC int
+xfs_qm_dqflush_all(
+	struct xfs_mount	*mp,
+	int			sync_mode)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	int			recl;
+	struct xfs_dquot	*dqp;
+	int			error;
+
+	if (!q)
+		return 0;
+again:
+	mutex_lock(&q->qi_dqlist_lock);
+	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
+		xfs_dqlock(dqp);
+		if (! XFS_DQ_IS_DIRTY(dqp)) {
+			xfs_dqunlock(dqp);
+			continue;
+		}
+
+		/* XXX a sentinel would be better */
+		recl = q->qi_dqreclaims;
+		if (!xfs_dqflock_nowait(dqp)) {
+			/*
+			 * If we can't grab the flush lock then check
+			 * to see if the dquot has been flushed delayed
+			 * write.  If so, grab its buffer and send it
+			 * out immediately.  We'll be able to acquire
+			 * the flush lock when the I/O completes.
+			 */
+			xfs_qm_dqflock_pushbuf_wait(dqp);
+		}
+		/*
+		 * Let go of the mplist lock. We don't want to hold it
+		 * across a disk write.
+		 */
+		mutex_unlock(&q->qi_dqlist_lock);
+		error = xfs_qm_dqflush(dqp, sync_mode);
+		xfs_dqunlock(dqp);
+		if (error)
+			return error;
+
+		mutex_lock(&q->qi_dqlist_lock);
+		if (recl != q->qi_dqreclaims) {
+			mutex_unlock(&q->qi_dqlist_lock);
+			/* XXX restart limit */
+			goto again;
+		}
+	}
+
+	mutex_unlock(&q->qi_dqlist_lock);
+	/* return ! busy */
+	return 0;
+}
+/*
+ * Release the group dquot pointers the user dquots may be
+ * carrying around as a hint. mplist is locked on entry and exit.
+ */
+STATIC void
+xfs_qm_detach_gdquots(
+	struct xfs_mount	*mp)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	struct xfs_dquot	*dqp, *gdqp;
+	int			nrecl;
+
+ again:
+	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
+	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
+		xfs_dqlock(dqp);
+		if ((gdqp = dqp->q_gdquot)) {
+			xfs_dqlock(gdqp);
+			dqp->q_gdquot = NULL;
+		}
+		xfs_dqunlock(dqp);
+
+		if (gdqp) {
+			/*
+			 * Can't hold the mplist lock across a dqput.
+			 * XXXmust convert to marker based iterations here.
+			 */
+			nrecl = q->qi_dqreclaims;
+			mutex_unlock(&q->qi_dqlist_lock);
+			xfs_qm_dqput(gdqp);
+
+			mutex_lock(&q->qi_dqlist_lock);
+			if (nrecl != q->qi_dqreclaims)
+				goto again;
+		}
+	}
+}
+
+/*
+ * Go through all the incore dquots of this file system and take them
+ * off the mplist and hashlist, if the dquot type matches the dqtype
+ * parameter. This is used when turning off quota accounting for
+ * users and/or groups, as well as when the filesystem is unmounting.
+ */
+STATIC int
+xfs_qm_dqpurge_int(
+	struct xfs_mount	*mp,
+	uint			flags)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	struct xfs_dquot	*dqp, *n;
+	uint			dqtype;
+	int			nrecl;
+	int			nmisses;
+
+	if (!q)
+		return 0;
+
+	dqtype = (flags & XFS_QMOPT_UQUOTA) ? XFS_DQ_USER : 0;
+	dqtype |= (flags & XFS_QMOPT_PQUOTA) ? XFS_DQ_PROJ : 0;
+	dqtype |= (flags & XFS_QMOPT_GQUOTA) ? XFS_DQ_GROUP : 0;
+
+	mutex_lock(&q->qi_dqlist_lock);
+
+	/*
+	 * In the first pass through all incore dquots of this filesystem,
+	 * we release the group dquot pointers the user dquots may be
+	 * carrying around as a hint. We need to do this irrespective of
+	 * what's being turned off.
+	 */
+	xfs_qm_detach_gdquots(mp);
+
+      again:
+	nmisses = 0;
+	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
+	/*
+	 * Try to get rid of all of the unwanted dquots. The idea is to
+	 * get them off mplist and hashlist, but leave them on freelist.
+	 */
+	list_for_each_entry_safe(dqp, n, &q->qi_dqlist, q_mplist) {
+		/*
+		 * It's OK to look at the type without taking dqlock here.
+		 * We're holding the mplist lock here, and that's needed for
+		 * a dqreclaim.
+		 */
+		if ((dqp->dq_flags & dqtype) == 0)
+			continue;
+
+		if (!mutex_trylock(&dqp->q_hash->qh_lock)) {
+			nrecl = q->qi_dqreclaims;
+			mutex_unlock(&q->qi_dqlist_lock);
+			mutex_lock(&dqp->q_hash->qh_lock);
+			mutex_lock(&q->qi_dqlist_lock);
+
+			/*
+			 * XXXTheoretically, we can get into a very long
+			 * ping pong game here.
+			 * No one can be adding dquots to the mplist at
+			 * this point, but somebody might be taking things off.
+			 */
+			if (nrecl != q->qi_dqreclaims) {
+				mutex_unlock(&dqp->q_hash->qh_lock);
+				goto again;
+			}
+		}
+
+		/*
+		 * Take the dquot off the mplist and hashlist. It may remain on
+		 * freelist in INACTIVE state.
+		 */
+		nmisses += xfs_qm_dqpurge(dqp);
+	}
+	mutex_unlock(&q->qi_dqlist_lock);
+	return nmisses;
+}
+
+int
+xfs_qm_dqpurge_all(
+	xfs_mount_t	*mp,
+	uint		flags)
+{
+	int		ndquots;
+
+	/*
+	 * Purge the dquot cache.
+	 * None of the dquots should really be busy at this point.
+	 */
+	if (mp->m_quotainfo) {
+		while ((ndquots = xfs_qm_dqpurge_int(mp, flags))) {
+			delay(ndquots * 10);
+		}
+	}
+	return 0;
+}
+
+STATIC int
+xfs_qm_dqattach_one(
+	xfs_inode_t	*ip,
+	xfs_dqid_t	id,
+	uint		type,
+	uint		doalloc,
+	xfs_dquot_t	*udqhint, /* hint */
+	xfs_dquot_t	**IO_idqpp)
+{
+	xfs_dquot_t	*dqp;
+	int		error;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	error = 0;
+
+	/*
+	 * See if we already have it in the inode itself. IO_idqpp is
+	 * &i_udquot or &i_gdquot. This made the code look weird, but
+	 * made the logic a lot simpler.
+	 */
+	dqp = *IO_idqpp;
+	if (dqp) {
+		trace_xfs_dqattach_found(dqp);
+		return 0;
+	}
+
+	/*
+	 * udqhint is the i_udquot field in inode, and is non-NULL only
+	 * when the type arg is group/project. Its purpose is to save a
+	 * lookup by dqid (xfs_qm_dqget) by caching a group dquot inside
+	 * the user dquot.
+	 */
+	if (udqhint) {
+		ASSERT(type == XFS_DQ_GROUP || type == XFS_DQ_PROJ);
+		xfs_dqlock(udqhint);
+
+		/*
+		 * No need to take dqlock to look at the id.
+		 *
+		 * The ID can't change until it gets reclaimed, and it won't
+		 * be reclaimed as long as we have a ref from inode and we
+		 * hold the ilock.
+		 */
+		dqp = udqhint->q_gdquot;
+		if (dqp && be32_to_cpu(dqp->q_core.d_id) == id) {
+			xfs_dqlock(dqp);
+			XFS_DQHOLD(dqp);
+			ASSERT(*IO_idqpp == NULL);
+			*IO_idqpp = dqp;
+
+			xfs_dqunlock(dqp);
+			xfs_dqunlock(udqhint);
+			return 0;
+		}
+
+		/*
+		 * We can't hold a dquot lock when we call the dqget code.
+		 * We'll deadlock in no time, because of (not conforming to)
+		 * lock ordering - the inodelock comes before any dquot lock,
+		 * and we may drop and reacquire the ilock in xfs_qm_dqget().
+		 */
+		xfs_dqunlock(udqhint);
+	}
+
+	/*
+	 * Find the dquot from somewhere. This bumps the
+	 * reference count of dquot and returns it locked.
+	 * This can return ENOENT if dquot didn't exist on
+	 * disk and we didn't ask it to allocate;
+	 * ESRCH if quotas got turned off suddenly.
+	 */
+	error = xfs_qm_dqget(ip->i_mount, ip, id, type,
+			     doalloc | XFS_QMOPT_DOWARN, &dqp);
+	if (error)
+		return error;
+
+	trace_xfs_dqattach_get(dqp);
+
+	/*
+	 * dqget may have dropped and re-acquired the ilock, but it guarantees
+	 * that the dquot returned is the one that should go in the inode.
+	 */
+	*IO_idqpp = dqp;
+	xfs_dqunlock(dqp);
+	return 0;
+}
+
+
+/*
+ * Given a udquot and gdquot, attach a ptr to the group dquot in the
+ * udquot as a hint for future lookups. The idea sounds simple, but the
+ * execution isn't, because the udquot might have a group dquot attached
+ * already and getting rid of that gets us into lock ordering constraints.
+ * The process is complicated more by the fact that the dquots may or may not
+ * be locked on entry.
+ */
+STATIC void
+xfs_qm_dqattach_grouphint(
+	xfs_dquot_t	*udq,
+	xfs_dquot_t	*gdq)
+{
+	xfs_dquot_t	*tmp;
+
+	xfs_dqlock(udq);
+
+	if ((tmp = udq->q_gdquot)) {
+		if (tmp == gdq) {
+			xfs_dqunlock(udq);
+			return;
+		}
+
+		udq->q_gdquot = NULL;
+		/*
+		 * We can't keep any dqlocks when calling dqrele,
+		 * because the freelist lock comes before dqlocks.
+		 */
+		xfs_dqunlock(udq);
+		/*
+		 * we took a hard reference once upon a time in dqget,
+		 * so give it back when the udquot no longer points at it
+		 * dqput() does the unlocking of the dquot.
+		 */
+		xfs_qm_dqrele(tmp);
+
+		xfs_dqlock(udq);
+		xfs_dqlock(gdq);
+
+	} else {
+		ASSERT(XFS_DQ_IS_LOCKED(udq));
+		xfs_dqlock(gdq);
+	}
+
+	ASSERT(XFS_DQ_IS_LOCKED(udq));
+	ASSERT(XFS_DQ_IS_LOCKED(gdq));
+	/*
+	 * Somebody could have attached a gdquot here,
+	 * when we dropped the uqlock. If so, just do nothing.
+	 */
+	if (udq->q_gdquot == NULL) {
+		XFS_DQHOLD(gdq);
+		udq->q_gdquot = gdq;
+	}
+
+	xfs_dqunlock(gdq);
+	xfs_dqunlock(udq);
+}
+
+
+/*
+ * Given a locked inode, attach dquot(s) to it, taking U/G/P-QUOTAON
+ * into account.
+ * If XFS_QMOPT_DQALLOC, the dquot(s) will be allocated if needed.
+ * Inode may get unlocked and relocked in here, and the caller must deal with
+ * the consequences.
+ */
+int
+xfs_qm_dqattach_locked(
+	xfs_inode_t	*ip,
+	uint		flags)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	uint		nquotas = 0;
+	int		error = 0;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) ||
+	    !XFS_IS_QUOTA_ON(mp) ||
+	    !XFS_NOT_DQATTACHED(mp, ip) ||
+	    ip->i_ino == mp->m_sb.sb_uquotino ||
+	    ip->i_ino == mp->m_sb.sb_gquotino)
+		return 0;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+	if (XFS_IS_UQUOTA_ON(mp)) {
+		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
+						flags & XFS_QMOPT_DQALLOC,
+						NULL, &ip->i_udquot);
+		if (error)
+			goto done;
+		nquotas++;
+	}
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	if (XFS_IS_OQUOTA_ON(mp)) {
+		error = XFS_IS_GQUOTA_ON(mp) ?
+			xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
+						flags & XFS_QMOPT_DQALLOC,
+						ip->i_udquot, &ip->i_gdquot) :
+			xfs_qm_dqattach_one(ip, xfs_get_projid(ip), XFS_DQ_PROJ,
+						flags & XFS_QMOPT_DQALLOC,
+						ip->i_udquot, &ip->i_gdquot);
+		/*
+		 * Don't worry about the udquot that we may have
+		 * attached above. It'll get detached, if not already.
+		 */
+		if (error)
+			goto done;
+		nquotas++;
+	}
+
+	/*
+	 * Attach this group quota to the user quota as a hint.
+	 * This WON'T, in general, result in a thrash.
+	 */
+	if (nquotas == 2) {
+		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+		ASSERT(ip->i_udquot);
+		ASSERT(ip->i_gdquot);
+
+		/*
+		 * We may or may not have the i_udquot locked at this point,
+		 * but this check is OK since we don't depend on the i_gdquot to
+		 * be accurate 100% all the time. It is just a hint, and this
+		 * will succeed in general.
+		 */
+		if (ip->i_udquot->q_gdquot == ip->i_gdquot)
+			goto done;
+		/*
+		 * Attach i_gdquot to the gdquot hint inside the i_udquot.
+		 */
+		xfs_qm_dqattach_grouphint(ip->i_udquot, ip->i_gdquot);
+	}
+
+ done:
+#ifdef DEBUG
+	if (!error) {
+		if (XFS_IS_UQUOTA_ON(mp))
+			ASSERT(ip->i_udquot);
+		if (XFS_IS_OQUOTA_ON(mp))
+			ASSERT(ip->i_gdquot);
+	}
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+#endif
+	return error;
+}
+
+int
+xfs_qm_dqattach(
+	struct xfs_inode	*ip,
+	uint			flags)
+{
+	int			error;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	error = xfs_qm_dqattach_locked(ip, flags);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	return error;
+}
+
+/*
+ * Release dquots (and their references) if any.
+ * The inode should be locked EXCL except when this's called by
+ * xfs_ireclaim.
+ */
+void
+xfs_qm_dqdetach(
+	xfs_inode_t	*ip)
+{
+	if (!(ip->i_udquot || ip->i_gdquot))
+		return;
+
+	trace_xfs_dquot_dqdetach(ip);
+
+	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_uquotino);
+	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_gquotino);
+	if (ip->i_udquot) {
+		xfs_qm_dqrele(ip->i_udquot);
+		ip->i_udquot = NULL;
+	}
+	if (ip->i_gdquot) {
+		xfs_qm_dqrele(ip->i_gdquot);
+		ip->i_gdquot = NULL;
+	}
+}
+
+int
+xfs_qm_sync(
+	struct xfs_mount	*mp,
+	int			flags)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	int			recl, restarts;
+	struct xfs_dquot	*dqp;
+	int			error;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return 0;
+
+	restarts = 0;
+
+  again:
+	mutex_lock(&q->qi_dqlist_lock);
+	/*
+	 * dqpurge_all() also takes the mplist lock and iterate thru all dquots
+	 * in quotaoff. However, if the QUOTA_ACTIVE bits are not cleared
+	 * when we have the mplist lock, we know that dquots will be consistent
+	 * as long as we have it locked.
+	 */
+	if (!XFS_IS_QUOTA_ON(mp)) {
+		mutex_unlock(&q->qi_dqlist_lock);
+		return 0;
+	}
+	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
+	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
+		/*
+		 * If this is vfs_sync calling, then skip the dquots that
+		 * don't 'seem' to be dirty. ie. don't acquire dqlock.
+		 * This is very similar to what xfs_sync does with inodes.
+		 */
+		if (flags & SYNC_TRYLOCK) {
+			if (!XFS_DQ_IS_DIRTY(dqp))
+				continue;
+			if (!xfs_qm_dqlock_nowait(dqp))
+				continue;
+		} else {
+			xfs_dqlock(dqp);
+		}
+
+		/*
+		 * Now, find out for sure if this dquot is dirty or not.
+		 */
+		if (! XFS_DQ_IS_DIRTY(dqp)) {
+			xfs_dqunlock(dqp);
+			continue;
+		}
+
+		/* XXX a sentinel would be better */
+		recl = q->qi_dqreclaims;
+		if (!xfs_dqflock_nowait(dqp)) {
+			if (flags & SYNC_TRYLOCK) {
+				xfs_dqunlock(dqp);
+				continue;
+			}
+			/*
+			 * If we can't grab the flush lock then if the caller
+			 * really wanted us to give this our best shot, so
+			 * see if we can give a push to the buffer before we wait
+			 * on the flush lock. At this point, we know that
+			 * even though the dquot is being flushed,
+			 * it has (new) dirty data.
+			 */
+			xfs_qm_dqflock_pushbuf_wait(dqp);
+		}
+		/*
+		 * Let go of the mplist lock. We don't want to hold it
+		 * across a disk write
+		 */
+		mutex_unlock(&q->qi_dqlist_lock);
+		error = xfs_qm_dqflush(dqp, flags);
+		xfs_dqunlock(dqp);
+		if (error && XFS_FORCED_SHUTDOWN(mp))
+			return 0;	/* Need to prevent umount failure */
+		else if (error)
+			return error;
+
+		mutex_lock(&q->qi_dqlist_lock);
+		if (recl != q->qi_dqreclaims) {
+			if (++restarts >= XFS_QM_SYNC_MAX_RESTARTS)
+				break;
+
+			mutex_unlock(&q->qi_dqlist_lock);
+			goto again;
+		}
+	}
+
+	mutex_unlock(&q->qi_dqlist_lock);
+	return 0;
+}
+
+/*
+ * The hash chains and the mplist use the same xfs_dqhash structure as
+ * their list head, but we can take the mplist qh_lock and one of the
+ * hash qh_locks at the same time without any problem as they aren't
+ * related.
+ */
+static struct lock_class_key xfs_quota_mplist_class;
+
+/*
+ * This initializes all the quota information that's kept in the
+ * mount structure
+ */
+STATIC int
+xfs_qm_init_quotainfo(
+	xfs_mount_t	*mp)
+{
+	xfs_quotainfo_t *qinf;
+	int		error;
+	xfs_dquot_t	*dqp;
+
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	/*
+	 * Tell XQM that we exist as soon as possible.
+	 */
+	if ((error = xfs_qm_hold_quotafs_ref(mp))) {
+		return error;
+	}
+
+	qinf = mp->m_quotainfo = kmem_zalloc(sizeof(xfs_quotainfo_t), KM_SLEEP);
+
+	/*
+	 * See if quotainodes are setup, and if not, allocate them,
+	 * and change the superblock accordingly.
+	 */
+	if ((error = xfs_qm_init_quotainos(mp))) {
+		kmem_free(qinf);
+		mp->m_quotainfo = NULL;
+		return error;
+	}
+
+	INIT_LIST_HEAD(&qinf->qi_dqlist);
+	mutex_init(&qinf->qi_dqlist_lock);
+	lockdep_set_class(&qinf->qi_dqlist_lock, &xfs_quota_mplist_class);
+
+	qinf->qi_dqreclaims = 0;
+
+	/* mutex used to serialize quotaoffs */
+	mutex_init(&qinf->qi_quotaofflock);
+
+	/* Precalc some constants */
+	qinf->qi_dqchunklen = XFS_FSB_TO_BB(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
+	ASSERT(qinf->qi_dqchunklen);
+	qinf->qi_dqperchunk = BBTOB(qinf->qi_dqchunklen);
+	do_div(qinf->qi_dqperchunk, sizeof(xfs_dqblk_t));
+
+	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);
+
+	/*
+	 * We try to get the limits from the superuser's limits fields.
+	 * This is quite hacky, but it is standard quota practice.
+	 * We look at the USR dquot with id == 0 first, but if user quotas
+	 * are not enabled we goto the GRP dquot with id == 0.
+	 * We don't really care to keep separate default limits for user
+	 * and group quotas, at least not at this point.
+	 */
+	error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)0,
+			     XFS_IS_UQUOTA_RUNNING(mp) ? XFS_DQ_USER : 
+			     (XFS_IS_GQUOTA_RUNNING(mp) ? XFS_DQ_GROUP :
+				XFS_DQ_PROJ),
+			     XFS_QMOPT_DQSUSER|XFS_QMOPT_DOWARN,
+			     &dqp);
+	if (! error) {
+		xfs_disk_dquot_t	*ddqp = &dqp->q_core;
+
+		/*
+		 * The warnings and timers set the grace period given to
+		 * a user or group before he or she can not perform any
+		 * more writing. If it is zero, a default is used.
+		 */
+		qinf->qi_btimelimit = ddqp->d_btimer ?
+			be32_to_cpu(ddqp->d_btimer) : XFS_QM_BTIMELIMIT;
+		qinf->qi_itimelimit = ddqp->d_itimer ?
+			be32_to_cpu(ddqp->d_itimer) : XFS_QM_ITIMELIMIT;
+		qinf->qi_rtbtimelimit = ddqp->d_rtbtimer ?
+			be32_to_cpu(ddqp->d_rtbtimer) : XFS_QM_RTBTIMELIMIT;
+		qinf->qi_bwarnlimit = ddqp->d_bwarns ?
+			be16_to_cpu(ddqp->d_bwarns) : XFS_QM_BWARNLIMIT;
+		qinf->qi_iwarnlimit = ddqp->d_iwarns ?
+			be16_to_cpu(ddqp->d_iwarns) : XFS_QM_IWARNLIMIT;
+		qinf->qi_rtbwarnlimit = ddqp->d_rtbwarns ?
+			be16_to_cpu(ddqp->d_rtbwarns) : XFS_QM_RTBWARNLIMIT;
+		qinf->qi_bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
+		qinf->qi_bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
+		qinf->qi_ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
+		qinf->qi_isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
+		qinf->qi_rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
+		qinf->qi_rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
+ 
+		/*
+		 * We sent the XFS_QMOPT_DQSUSER flag to dqget because
+		 * we don't want this dquot cached. We haven't done a
+		 * quotacheck yet, and quotacheck doesn't like incore dquots.
+		 */
+		xfs_qm_dqdestroy(dqp);
+	} else {
+		qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
+		qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
+		qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
+		qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
+		qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
+		qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
+	}
+
+	return 0;
+}
+
+
+/*
+ * Gets called when unmounting a filesystem or when all quotas get
+ * turned off.
+ * This purges the quota inodes, destroys locks and frees itself.
+ */
+void
+xfs_qm_destroy_quotainfo(
+	xfs_mount_t	*mp)
+{
+	xfs_quotainfo_t *qi;
+
+	qi = mp->m_quotainfo;
+	ASSERT(qi != NULL);
+	ASSERT(xfs_Gqm != NULL);
+
+	/*
+	 * Release the reference that XQM kept, so that we know
+	 * when the XQM structure should be freed. We cannot assume
+	 * that xfs_Gqm is non-null after this point.
+	 */
+	xfs_qm_rele_quotafs_ref(mp);
+
+	ASSERT(list_empty(&qi->qi_dqlist));
+	mutex_destroy(&qi->qi_dqlist_lock);
+
+	if (qi->qi_uquotaip) {
+		IRELE(qi->qi_uquotaip);
+		qi->qi_uquotaip = NULL; /* paranoia */
+	}
+	if (qi->qi_gquotaip) {
+		IRELE(qi->qi_gquotaip);
+		qi->qi_gquotaip = NULL;
+	}
+	mutex_destroy(&qi->qi_quotaofflock);
+	kmem_free(qi);
+	mp->m_quotainfo = NULL;
+}
+
+
+
+/* ------------------- PRIVATE STATIC FUNCTIONS ----------------------- */
+
+/* ARGSUSED */
+STATIC void
+xfs_qm_list_init(
+	xfs_dqlist_t	*list,
+	char		*str,
+	int		n)
+{
+	mutex_init(&list->qh_lock);
+	INIT_LIST_HEAD(&list->qh_list);
+	list->qh_version = 0;
+	list->qh_nelems = 0;
+}
+
+STATIC void
+xfs_qm_list_destroy(
+	xfs_dqlist_t	*list)
+{
+	mutex_destroy(&(list->qh_lock));
+}
+
+/*
+ * Create an inode and return with a reference already taken, but unlocked
+ * This is how we create quota inodes
+ */
+STATIC int
+xfs_qm_qino_alloc(
+	xfs_mount_t	*mp,
+	xfs_inode_t	**ip,
+	__int64_t	sbfields,
+	uint		flags)
+{
+	xfs_trans_t	*tp;
+	int		error;
+	int		committed;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QINOCREATE);
+	if ((error = xfs_trans_reserve(tp,
+				      XFS_QM_QINOCREATE_SPACE_RES(mp),
+				      XFS_CREATE_LOG_RES(mp), 0,
+				      XFS_TRANS_PERM_LOG_RES,
+				      XFS_CREATE_LOG_COUNT))) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+
+	error = xfs_dir_ialloc(&tp, NULL, S_IFREG, 1, 0, 0, 1, ip, &committed);
+	if (error) {
+		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES |
+				 XFS_TRANS_ABORT);
+		return error;
+	}
+
+	/*
+	 * Make the changes in the superblock, and log those too.
+	 * sbfields arg may contain fields other than *QUOTINO;
+	 * VERSIONNUM for example.
+	 */
+	spin_lock(&mp->m_sb_lock);
+	if (flags & XFS_QMOPT_SBVERSION) {
+		ASSERT(!xfs_sb_version_hasquota(&mp->m_sb));
+		ASSERT((sbfields & (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
+				   XFS_SB_GQUOTINO | XFS_SB_QFLAGS)) ==
+		       (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
+			XFS_SB_GQUOTINO | XFS_SB_QFLAGS));
+
+		xfs_sb_version_addquota(&mp->m_sb);
+		mp->m_sb.sb_uquotino = NULLFSINO;
+		mp->m_sb.sb_gquotino = NULLFSINO;
+
+		/* qflags will get updated _after_ quotacheck */
+		mp->m_sb.sb_qflags = 0;
+	}
+	if (flags & XFS_QMOPT_UQUOTA)
+		mp->m_sb.sb_uquotino = (*ip)->i_ino;
+	else
+		mp->m_sb.sb_gquotino = (*ip)->i_ino;
+	spin_unlock(&mp->m_sb_lock);
+	xfs_mod_sb(tp, sbfields);
+
+	if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES))) {
+		xfs_alert(mp, "%s failed (error %d)!", __func__, error);
+		return error;
+	}
+	return 0;
+}
+
+
+STATIC void
+xfs_qm_reset_dqcounts(
+	xfs_mount_t	*mp,
+	xfs_buf_t	*bp,
+	xfs_dqid_t	id,
+	uint		type)
+{
+	xfs_disk_dquot_t	*ddq;
+	int			j;
+
+	trace_xfs_reset_dqcounts(bp, _RET_IP_);
+
+	/*
+	 * Reset all counters and timers. They'll be
+	 * started afresh by xfs_qm_quotacheck.
+	 */
+#ifdef DEBUG
+	j = XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
+	do_div(j, sizeof(xfs_dqblk_t));
+	ASSERT(mp->m_quotainfo->qi_dqperchunk == j);
+#endif
+	ddq = bp->b_addr;
+	for (j = 0; j < mp->m_quotainfo->qi_dqperchunk; j++) {
+		/*
+		 * Do a sanity check, and if needed, repair the dqblk. Don't
+		 * output any warnings because it's perfectly possible to
+		 * find uninitialised dquot blks. See comment in xfs_qm_dqcheck.
+		 */
+		(void) xfs_qm_dqcheck(mp, ddq, id+j, type, XFS_QMOPT_DQREPAIR,
+				      "xfs_quotacheck");
+		ddq->d_bcount = 0;
+		ddq->d_icount = 0;
+		ddq->d_rtbcount = 0;
+		ddq->d_btimer = 0;
+		ddq->d_itimer = 0;
+		ddq->d_rtbtimer = 0;
+		ddq->d_bwarns = 0;
+		ddq->d_iwarns = 0;
+		ddq->d_rtbwarns = 0;
+		ddq = (xfs_disk_dquot_t *) ((xfs_dqblk_t *)ddq + 1);
+	}
+}
+
+STATIC int
+xfs_qm_dqiter_bufs(
+	xfs_mount_t	*mp,
+	xfs_dqid_t	firstid,
+	xfs_fsblock_t	bno,
+	xfs_filblks_t	blkcnt,
+	uint		flags)
+{
+	xfs_buf_t	*bp;
+	int		error;
+	int		type;
+
+	ASSERT(blkcnt > 0);
+	type = flags & XFS_QMOPT_UQUOTA ? XFS_DQ_USER :
+		(flags & XFS_QMOPT_PQUOTA ? XFS_DQ_PROJ : XFS_DQ_GROUP);
+	error = 0;
+
+	/*
+	 * Blkcnt arg can be a very big number, and might even be
+	 * larger than the log itself. So, we have to break it up into
+	 * manageable-sized transactions.
+	 * Note that we don't start a permanent transaction here; we might
+	 * not be able to get a log reservation for the whole thing up front,
+	 * and we don't really care to either, because we just discard
+	 * everything if we were to crash in the middle of this loop.
+	 */
+	while (blkcnt--) {
+		error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
+			      XFS_FSB_TO_DADDR(mp, bno),
+			      mp->m_quotainfo->qi_dqchunklen, 0, &bp);
+		if (error)
+			break;
+
+		xfs_qm_reset_dqcounts(mp, bp, firstid, type);
+		xfs_buf_delwri_queue(bp);
+		xfs_buf_relse(bp);
+		/*
+		 * goto the next block.
+		 */
+		bno++;
+		firstid += mp->m_quotainfo->qi_dqperchunk;
+	}
+	return error;
+}
+
+/*
+ * Iterate over all allocated USR/GRP/PRJ dquots in the system, calling a
+ * caller supplied function for every chunk of dquots that we find.
+ */
+STATIC int
+xfs_qm_dqiterate(
+	xfs_mount_t	*mp,
+	xfs_inode_t	*qip,
+	uint		flags)
+{
+	xfs_bmbt_irec_t		*map;
+	int			i, nmaps;	/* number of map entries */
+	int			error;		/* return value */
+	xfs_fileoff_t		lblkno;
+	xfs_filblks_t		maxlblkcnt;
+	xfs_dqid_t		firstid;
+	xfs_fsblock_t		rablkno;
+	xfs_filblks_t		rablkcnt;
+
+	error = 0;
+	/*
+	 * This looks racy, but we can't keep an inode lock across a
+	 * trans_reserve. But, this gets called during quotacheck, and that
+	 * happens only at mount time which is single threaded.
+	 */
+	if (qip->i_d.di_nblocks == 0)
+		return 0;
+
+	map = kmem_alloc(XFS_DQITER_MAP_SIZE * sizeof(*map), KM_SLEEP);
+
+	lblkno = 0;
+	maxlblkcnt = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp));
+	do {
+		nmaps = XFS_DQITER_MAP_SIZE;
+		/*
+		 * We aren't changing the inode itself. Just changing
+		 * some of its data. No new blocks are added here, and
+		 * the inode is never added to the transaction.
+		 */
+		xfs_ilock(qip, XFS_ILOCK_SHARED);
+		error = xfs_bmapi_read(qip, lblkno, maxlblkcnt - lblkno,
+				       map, &nmaps, 0);
+		xfs_iunlock(qip, XFS_ILOCK_SHARED);
+		if (error)
+			break;
+
+		ASSERT(nmaps <= XFS_DQITER_MAP_SIZE);
+		for (i = 0; i < nmaps; i++) {
+			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
+			ASSERT(map[i].br_blockcount);
+
+
+			lblkno += map[i].br_blockcount;
+
+			if (map[i].br_startblock == HOLESTARTBLOCK)
+				continue;
+
+			firstid = (xfs_dqid_t) map[i].br_startoff *
+				mp->m_quotainfo->qi_dqperchunk;
+			/*
+			 * Do a read-ahead on the next extent.
+			 */
+			if ((i+1 < nmaps) &&
+			    (map[i+1].br_startblock != HOLESTARTBLOCK)) {
+				rablkcnt =  map[i+1].br_blockcount;
+				rablkno = map[i+1].br_startblock;
+				while (rablkcnt--) {
+					xfs_buf_readahead(mp->m_ddev_targp,
+					       XFS_FSB_TO_DADDR(mp, rablkno),
+					       mp->m_quotainfo->qi_dqchunklen);
+					rablkno++;
+				}
+			}
+			/*
+			 * Iterate thru all the blks in the extent and
+			 * reset the counters of all the dquots inside them.
+			 */
+			if ((error = xfs_qm_dqiter_bufs(mp,
+						       firstid,
+						       map[i].br_startblock,
+						       map[i].br_blockcount,
+						       flags))) {
+				break;
+			}
+		}
+
+		if (error)
+			break;
+	} while (nmaps > 0);
+
+	kmem_free(map);
+
+	return error;
+}
+
+/*
+ * Called by dqusage_adjust in doing a quotacheck.
+ *
+ * Given the inode, and a dquot id this updates both the incore dqout as well
+ * as the buffer copy. This is so that once the quotacheck is done, we can
+ * just log all the buffers, as opposed to logging numerous updates to
+ * individual dquots.
+ */
+STATIC int
+xfs_qm_quotacheck_dqadjust(
+	struct xfs_inode	*ip,
+	xfs_dqid_t		id,
+	uint			type,
+	xfs_qcnt_t		nblks,
+	xfs_qcnt_t		rtblks)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_dquot	*dqp;
+	int			error;
+
+	error = xfs_qm_dqget(mp, ip, id, type,
+			     XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN, &dqp);
+	if (error) {
+		/*
+		 * Shouldn't be able to turn off quotas here.
+		 */
+		ASSERT(error != ESRCH);
+		ASSERT(error != ENOENT);
+		return error;
+	}
+
+	trace_xfs_dqadjust(dqp);
+
+	/*
+	 * Adjust the inode count and the block count to reflect this inode's
+	 * resource usage.
+	 */
+	be64_add_cpu(&dqp->q_core.d_icount, 1);
+	dqp->q_res_icount++;
+	if (nblks) {
+		be64_add_cpu(&dqp->q_core.d_bcount, nblks);
+		dqp->q_res_bcount += nblks;
+	}
+	if (rtblks) {
+		be64_add_cpu(&dqp->q_core.d_rtbcount, rtblks);
+		dqp->q_res_rtbcount += rtblks;
+	}
+
+	/*
+	 * Set default limits, adjust timers (since we changed usages)
+	 *
+	 * There are no timers for the default values set in the root dquot.
+	 */
+	if (dqp->q_core.d_id) {
+		xfs_qm_adjust_dqlimits(mp, &dqp->q_core);
+		xfs_qm_adjust_dqtimers(mp, &dqp->q_core);
+	}
+
+	dqp->dq_flags |= XFS_DQ_DIRTY;
+	xfs_qm_dqput(dqp);
+	return 0;
+}
+
+STATIC int
+xfs_qm_get_rtblks(
+	xfs_inode_t	*ip,
+	xfs_qcnt_t	*O_rtblks)
+{
+	xfs_filblks_t	rtblks;			/* total rt blks */
+	xfs_extnum_t	idx;			/* extent record index */
+	xfs_ifork_t	*ifp;			/* inode fork pointer */
+	xfs_extnum_t	nextents;		/* number of extent entries */
+	int		error;
+
+	ASSERT(XFS_IS_REALTIME_INODE(ip));
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		if ((error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK)))
+			return error;
+	}
+	rtblks = 0;
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	for (idx = 0; idx < nextents; idx++)
+		rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx));
+	*O_rtblks = (xfs_qcnt_t)rtblks;
+	return 0;
+}
+
+/*
+ * callback routine supplied to bulkstat(). Given an inumber, find its
+ * dquots and update them to account for resources taken by that inode.
+ */
+/* ARGSUSED */
+STATIC int
+xfs_qm_dqusage_adjust(
+	xfs_mount_t	*mp,		/* mount point for filesystem */
+	xfs_ino_t	ino,		/* inode number to get data for */
+	void		__user *buffer,	/* not used */
+	int		ubsize,		/* not used */
+	int		*ubused,	/* not used */
+	int		*res)		/* result code value */
+{
+	xfs_inode_t	*ip;
+	xfs_qcnt_t	nblks, rtblks = 0;
+	int		error;
+
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	/*
+	 * rootino must have its resources accounted for, not so with the quota
+	 * inodes.
+	 */
+	if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) {
+		*res = BULKSTAT_RV_NOTHING;
+		return XFS_ERROR(EINVAL);
+	}
+
+	/*
+	 * We don't _need_ to take the ilock EXCL. However, the xfs_qm_dqget
+	 * interface expects the inode to be exclusively locked because that's
+	 * the case in all other instances. It's OK that we do this because
+	 * quotacheck is done only at mount time.
+	 */
+	error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_EXCL, &ip);
+	if (error) {
+		*res = BULKSTAT_RV_NOTHING;
+		return error;
+	}
+
+	ASSERT(ip->i_delayed_blks == 0);
+
+	if (XFS_IS_REALTIME_INODE(ip)) {
+		/*
+		 * Walk thru the extent list and count the realtime blocks.
+		 */
+		error = xfs_qm_get_rtblks(ip, &rtblks);
+		if (error)
+			goto error0;
+	}
+
+	nblks = (xfs_qcnt_t)ip->i_d.di_nblocks - rtblks;
+
+	/*
+	 * Add the (disk blocks and inode) resources occupied by this
+	 * inode to its dquots. We do this adjustment in the incore dquot,
+	 * and also copy the changes to its buffer.
+	 * We don't care about putting these changes in a transaction
+	 * envelope because if we crash in the middle of a 'quotacheck'
+	 * we have to start from the beginning anyway.
+	 * Once we're done, we'll log all the dquot bufs.
+	 *
+	 * The *QUOTA_ON checks below may look pretty racy, but quotachecks
+	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
+	 */
+	if (XFS_IS_UQUOTA_ON(mp)) {
+		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_uid,
+						   XFS_DQ_USER, nblks, rtblks);
+		if (error)
+			goto error0;
+	}
+
+	if (XFS_IS_GQUOTA_ON(mp)) {
+		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_gid,
+						   XFS_DQ_GROUP, nblks, rtblks);
+		if (error)
+			goto error0;
+	}
+
+	if (XFS_IS_PQUOTA_ON(mp)) {
+		error = xfs_qm_quotacheck_dqadjust(ip, xfs_get_projid(ip),
+						   XFS_DQ_PROJ, nblks, rtblks);
+		if (error)
+			goto error0;
+	}
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	IRELE(ip);
+	*res = BULKSTAT_RV_DIDONE;
+	return 0;
+
+error0:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	IRELE(ip);
+	*res = BULKSTAT_RV_GIVEUP;
+	return error;
+}
+
+/*
+ * Walk thru all the filesystem inodes and construct a consistent view
+ * of the disk quota world. If the quotacheck fails, disable quotas.
+ */
+int
+xfs_qm_quotacheck(
+	xfs_mount_t	*mp)
+{
+	int		done, count, error;
+	xfs_ino_t	lastino;
+	size_t		structsz;
+	xfs_inode_t	*uip, *gip;
+	uint		flags;
+
+	count = INT_MAX;
+	structsz = 1;
+	lastino = 0;
+	flags = 0;
+
+	ASSERT(mp->m_quotainfo->qi_uquotaip || mp->m_quotainfo->qi_gquotaip);
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	/*
+	 * There should be no cached dquots. The (simplistic) quotacheck
+	 * algorithm doesn't like that.
+	 */
+	ASSERT(list_empty(&mp->m_quotainfo->qi_dqlist));
+
+	xfs_notice(mp, "Quotacheck needed: Please wait.");
+
+	/*
+	 * First we go thru all the dquots on disk, USR and GRP/PRJ, and reset
+	 * their counters to zero. We need a clean slate.
+	 * We don't log our changes till later.
+	 */
+	uip = mp->m_quotainfo->qi_uquotaip;
+	if (uip) {
+		error = xfs_qm_dqiterate(mp, uip, XFS_QMOPT_UQUOTA);
+		if (error)
+			goto error_return;
+		flags |= XFS_UQUOTA_CHKD;
+	}
+
+	gip = mp->m_quotainfo->qi_gquotaip;
+	if (gip) {
+		error = xfs_qm_dqiterate(mp, gip, XFS_IS_GQUOTA_ON(mp) ?
+					XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA);
+		if (error)
+			goto error_return;
+		flags |= XFS_OQUOTA_CHKD;
+	}
+
+	do {
+		/*
+		 * Iterate thru all the inodes in the file system,
+		 * adjusting the corresponding dquot counters in core.
+		 */
+		error = xfs_bulkstat(mp, &lastino, &count,
+				     xfs_qm_dqusage_adjust,
+				     structsz, NULL, &done);
+		if (error)
+			break;
+
+	} while (!done);
+
+	/*
+	 * We've made all the changes that we need to make incore.
+	 * Flush them down to disk buffers if everything was updated
+	 * successfully.
+	 */
+	if (!error)
+		error = xfs_qm_dqflush_all(mp, 0);
+
+	/*
+	 * We can get this error if we couldn't do a dquot allocation inside
+	 * xfs_qm_dqusage_adjust (via bulkstat). We don't care about the
+	 * dirty dquots that might be cached, we just want to get rid of them
+	 * and turn quotaoff. The dquots won't be attached to any of the inodes
+	 * at this point (because we intentionally didn't in dqget_noattach).
+	 */
+	if (error) {
+		xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL);
+		goto error_return;
+	}
+
+	/*
+	 * We didn't log anything, because if we crashed, we'll have to
+	 * start the quotacheck from scratch anyway. However, we must make
+	 * sure that our dquot changes are secure before we put the
+	 * quotacheck'd stamp on the superblock. So, here we do a synchronous
+	 * flush.
+	 */
+	xfs_flush_buftarg(mp->m_ddev_targp, 1);
+
+	/*
+	 * If one type of quotas is off, then it will lose its
+	 * quotachecked status, since we won't be doing accounting for
+	 * that type anymore.
+	 */
+	mp->m_qflags &= ~(XFS_OQUOTA_CHKD | XFS_UQUOTA_CHKD);
+	mp->m_qflags |= flags;
+
+ error_return:
+	if (error) {
+		xfs_warn(mp,
+	"Quotacheck: Unsuccessful (Error %d): Disabling quotas.",
+			error);
+		/*
+		 * We must turn off quotas.
+		 */
+		ASSERT(mp->m_quotainfo != NULL);
+		ASSERT(xfs_Gqm != NULL);
+		xfs_qm_destroy_quotainfo(mp);
+		if (xfs_mount_reset_sbqflags(mp)) {
+			xfs_warn(mp,
+				"Quotacheck: Failed to reset quota flags.");
+		}
+	} else
+		xfs_notice(mp, "Quotacheck: Done.");
+	return (error);
+}
+
+/*
+ * This is called after the superblock has been read in and we're ready to
+ * iget the quota inodes.
+ */
+STATIC int
+xfs_qm_init_quotainos(
+	xfs_mount_t	*mp)
+{
+	xfs_inode_t	*uip, *gip;
+	int		error;
+	__int64_t	sbflags;
+	uint		flags;
+
+	ASSERT(mp->m_quotainfo);
+	uip = gip = NULL;
+	sbflags = 0;
+	flags = 0;
+
+	/*
+	 * Get the uquota and gquota inodes
+	 */
+	if (xfs_sb_version_hasquota(&mp->m_sb)) {
+		if (XFS_IS_UQUOTA_ON(mp) &&
+		    mp->m_sb.sb_uquotino != NULLFSINO) {
+			ASSERT(mp->m_sb.sb_uquotino > 0);
+			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
+					     0, 0, &uip)))
+				return XFS_ERROR(error);
+		}
+		if (XFS_IS_OQUOTA_ON(mp) &&
+		    mp->m_sb.sb_gquotino != NULLFSINO) {
+			ASSERT(mp->m_sb.sb_gquotino > 0);
+			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
+					     0, 0, &gip))) {
+				if (uip)
+					IRELE(uip);
+				return XFS_ERROR(error);
+			}
+		}
+	} else {
+		flags |= XFS_QMOPT_SBVERSION;
+		sbflags |= (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
+			    XFS_SB_GQUOTINO | XFS_SB_QFLAGS);
+	}
+
+	/*
+	 * Create the two inodes, if they don't exist already. The changes
+	 * made above will get added to a transaction and logged in one of
+	 * the qino_alloc calls below.  If the device is readonly,
+	 * temporarily switch to read-write to do this.
+	 */
+	if (XFS_IS_UQUOTA_ON(mp) && uip == NULL) {
+		if ((error = xfs_qm_qino_alloc(mp, &uip,
+					      sbflags | XFS_SB_UQUOTINO,
+					      flags | XFS_QMOPT_UQUOTA)))
+			return XFS_ERROR(error);
+
+		flags &= ~XFS_QMOPT_SBVERSION;
+	}
+	if (XFS_IS_OQUOTA_ON(mp) && gip == NULL) {
+		flags |= (XFS_IS_GQUOTA_ON(mp) ?
+				XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA);
+		error = xfs_qm_qino_alloc(mp, &gip,
+					  sbflags | XFS_SB_GQUOTINO, flags);
+		if (error) {
+			if (uip)
+				IRELE(uip);
+
+			return XFS_ERROR(error);
+		}
+	}
+
+	mp->m_quotainfo->qi_uquotaip = uip;
+	mp->m_quotainfo->qi_gquotaip = gip;
+
+	return 0;
+}
+
+
+
+/*
+ * Just pop the least recently used dquot off the freelist and
+ * recycle it. The returned dquot is locked.
+ */
+STATIC xfs_dquot_t *
+xfs_qm_dqreclaim_one(void)
+{
+	xfs_dquot_t	*dqpout;
+	xfs_dquot_t	*dqp;
+	int		restarts;
+	int		startagain;
+
+	restarts = 0;
+	dqpout = NULL;
+
+	/* lockorder: hashchainlock, freelistlock, mplistlock, dqlock, dqflock */
+again:
+	startagain = 0;
+	mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);
+
+	list_for_each_entry(dqp, &xfs_Gqm->qm_dqfrlist, q_freelist) {
+		struct xfs_mount *mp = dqp->q_mount;
+		xfs_dqlock(dqp);
+
+		/*
+		 * We are racing with dqlookup here. Naturally we don't
+		 * want to reclaim a dquot that lookup wants. We release the
+		 * freelist lock and start over, so that lookup will grab
+		 * both the dquot and the freelistlock.
+		 */
+		if (dqp->dq_flags & XFS_DQ_WANT) {
+			ASSERT(! (dqp->dq_flags & XFS_DQ_INACTIVE));
+
+			trace_xfs_dqreclaim_want(dqp);
+			XQM_STATS_INC(xqmstats.xs_qm_dqwants);
+			restarts++;
+			startagain = 1;
+			goto dqunlock;
+		}
+
+		/*
+		 * If the dquot is inactive, we are assured that it is
+		 * not on the mplist or the hashlist, and that makes our
+		 * life easier.
+		 */
+		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
+			ASSERT(mp == NULL);
+			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
+			ASSERT(list_empty(&dqp->q_hashlist));
+			ASSERT(list_empty(&dqp->q_mplist));
+			list_del_init(&dqp->q_freelist);
+			xfs_Gqm->qm_dqfrlist_cnt--;
+			dqpout = dqp;
+			XQM_STATS_INC(xqmstats.xs_qm_dqinact_reclaims);
+			goto dqunlock;
+		}
+
+		ASSERT(dqp->q_hash);
+		ASSERT(!list_empty(&dqp->q_mplist));
+
+		/*
+		 * Try to grab the flush lock. If this dquot is in the process
+		 * of getting flushed to disk, we don't want to reclaim it.
+		 */
+		if (!xfs_dqflock_nowait(dqp))
+			goto dqunlock;
+
+		/*
+		 * We have the flush lock so we know that this is not in the
+		 * process of being flushed. So, if this is dirty, flush it
+		 * DELWRI so that we don't get a freelist infested with
+		 * dirty dquots.
+		 */
+		if (XFS_DQ_IS_DIRTY(dqp)) {
+			int	error;
+
+			trace_xfs_dqreclaim_dirty(dqp);
+
+			/*
+			 * We flush it delayed write, so don't bother
+			 * releasing the freelist lock.
+			 */
+			error = xfs_qm_dqflush(dqp, 0);
+			if (error) {
+				xfs_warn(mp, "%s: dquot %p flush failed",
+					__func__, dqp);
+			}
+			goto dqunlock;
+		}
+
+		/*
+		 * We're trying to get the hashlock out of order. This races
+		 * with dqlookup; so, we giveup and goto the next dquot if
+		 * we couldn't get the hashlock. This way, we won't starve
+		 * a dqlookup process that holds the hashlock that is
+		 * waiting for the freelist lock.
+		 */
+		if (!mutex_trylock(&dqp->q_hash->qh_lock)) {
+			restarts++;
+			goto dqfunlock;
+		}
+
+		/*
+		 * This races with dquot allocation code as well as dqflush_all
+		 * and reclaim code. So, if we failed to grab the mplist lock,
+		 * giveup everything and start over.
+		 */
+		if (!mutex_trylock(&mp->m_quotainfo->qi_dqlist_lock)) {
+			restarts++;
+			startagain = 1;
+			goto qhunlock;
+		}
+
+		ASSERT(dqp->q_nrefs == 0);
+		list_del_init(&dqp->q_mplist);
+		mp->m_quotainfo->qi_dquots--;
+		mp->m_quotainfo->qi_dqreclaims++;
+		list_del_init(&dqp->q_hashlist);
+		dqp->q_hash->qh_version++;
+		list_del_init(&dqp->q_freelist);
+		xfs_Gqm->qm_dqfrlist_cnt--;
+		dqpout = dqp;
+		mutex_unlock(&mp->m_quotainfo->qi_dqlist_lock);
+qhunlock:
+		mutex_unlock(&dqp->q_hash->qh_lock);
+dqfunlock:
+		xfs_dqfunlock(dqp);
+dqunlock:
+		xfs_dqunlock(dqp);
+		if (dqpout)
+			break;
+		if (restarts >= XFS_QM_RECLAIM_MAX_RESTARTS)
+			break;
+		if (startagain) {
+			mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+			goto again;
+		}
+	}
+	mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
+	return dqpout;
+}
+
+/*
+ * Traverse the freelist of dquots and attempt to reclaim a maximum of
+ * 'howmany' dquots. This operation races with dqlookup(), and attempts to
+ * favor the lookup function ...
+ */
+STATIC int
+xfs_qm_shake_freelist(
+	int	howmany)
+{
+	int		nreclaimed = 0;
+	xfs_dquot_t	*dqp;
+
+	if (howmany <= 0)
+		return 0;
+
+	while (nreclaimed < howmany) {
+		dqp = xfs_qm_dqreclaim_one();
+		if (!dqp)
+			return nreclaimed;
+		xfs_qm_dqdestroy(dqp);
+		nreclaimed++;
+	}
+	return nreclaimed;
+}
+
+/*
+ * The kmem_shake interface is invoked when memory is running low.
+ */
+/* ARGSUSED */
+STATIC int
+xfs_qm_shake(
+	struct shrinker	*shrink,
+	struct shrink_control *sc)
+{
+	int	ndqused, nfree, n;
+	gfp_t gfp_mask = sc->gfp_mask;
+
+	if (!kmem_shake_allow(gfp_mask))
+		return 0;
+	if (!xfs_Gqm)
+		return 0;
+
+	nfree = xfs_Gqm->qm_dqfrlist_cnt; /* free dquots */
+	/* incore dquots in all f/s's */
+	ndqused = atomic_read(&xfs_Gqm->qm_totaldquots) - nfree;
+
+	ASSERT(ndqused >= 0);
+
+	if (nfree <= ndqused && nfree < ndquot)
+		return 0;
+
+	ndqused *= xfs_Gqm->qm_dqfree_ratio;	/* target # of free dquots */
+	n = nfree - ndqused - ndquot;		/* # over target */
+
+	return xfs_qm_shake_freelist(MAX(nfree, n));
+}
+
+
+/*------------------------------------------------------------------*/
+
+/*
+ * Return a new incore dquot. Depending on the number of
+ * dquots in the system, we either allocate a new one on the kernel heap,
+ * or reclaim a free one.
+ * Return value is B_TRUE if we allocated a new dquot, B_FALSE if we managed
+ * to reclaim an existing one from the freelist.
+ */
+boolean_t
+xfs_qm_dqalloc_incore(
+	xfs_dquot_t **O_dqpp)
+{
+	xfs_dquot_t	*dqp;
+
+	/*
+	 * Check against high water mark to see if we want to pop
+	 * a nincompoop dquot off the freelist.
+	 */
+	if (atomic_read(&xfs_Gqm->qm_totaldquots) >= ndquot) {
+		/*
+		 * Try to recycle a dquot from the freelist.
+		 */
+		if ((dqp = xfs_qm_dqreclaim_one())) {
+			XQM_STATS_INC(xqmstats.xs_qm_dqreclaims);
+			/*
+			 * Just zero the core here. The rest will get
+			 * reinitialized by caller. XXX we shouldn't even
+			 * do this zero ...
+			 */
+			memset(&dqp->q_core, 0, sizeof(dqp->q_core));
+			*O_dqpp = dqp;
+			return B_FALSE;
+		}
+		XQM_STATS_INC(xqmstats.xs_qm_dqreclaim_misses);
+	}
+
+	/*
+	 * Allocate a brand new dquot on the kernel heap and return it
+	 * to the caller to initialize.
+	 */
+	ASSERT(xfs_Gqm->qm_dqzone != NULL);
+	*O_dqpp = kmem_zone_zalloc(xfs_Gqm->qm_dqzone, KM_SLEEP);
+	atomic_inc(&xfs_Gqm->qm_totaldquots);
+
+	return B_TRUE;
+}
+
+
+/*
+ * Start a transaction and write the incore superblock changes to
+ * disk. flags parameter indicates which fields have changed.
+ */
+int
+xfs_qm_write_sb_changes(
+	xfs_mount_t	*mp,
+	__int64_t	flags)
+{
+	xfs_trans_t	*tp;
+	int		error;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SBCHANGE);
+	if ((error = xfs_trans_reserve(tp, 0,
+				      mp->m_sb.sb_sectsize + 128, 0,
+				      0,
+				      XFS_DEFAULT_LOG_COUNT))) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+
+	xfs_mod_sb(tp, flags);
+	error = xfs_trans_commit(tp, 0);
+
+	return error;
+}
+
+
+/* --------------- utility functions for vnodeops ---------------- */
+
+
+/*
+ * Given an inode, a uid, gid and prid make sure that we have
+ * allocated relevant dquot(s) on disk, and that we won't exceed inode
+ * quotas by creating this file.
+ * This also attaches dquot(s) to the given inode after locking it,
+ * and returns the dquots corresponding to the uid and/or gid.
+ *
+ * in	: inode (unlocked)
+ * out	: udquot, gdquot with references taken and unlocked
+ */
+int
+xfs_qm_vop_dqalloc(
+	struct xfs_inode	*ip,
+	uid_t			uid,
+	gid_t			gid,
+	prid_t			prid,
+	uint			flags,
+	struct xfs_dquot	**O_udqpp,
+	struct xfs_dquot	**O_gdqpp)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_dquot	*uq, *gq;
+	int			error;
+	uint			lockflags;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return 0;
+
+	lockflags = XFS_ILOCK_EXCL;
+	xfs_ilock(ip, lockflags);
+
+	if ((flags & XFS_QMOPT_INHERIT) && XFS_INHERIT_GID(ip))
+		gid = ip->i_d.di_gid;
+
+	/*
+	 * Attach the dquot(s) to this inode, doing a dquot allocation
+	 * if necessary. The dquot(s) will not be locked.
+	 */
+	if (XFS_NOT_DQATTACHED(mp, ip)) {
+		error = xfs_qm_dqattach_locked(ip, XFS_QMOPT_DQALLOC);
+		if (error) {
+			xfs_iunlock(ip, lockflags);
+			return error;
+		}
+	}
+
+	uq = gq = NULL;
+	if ((flags & XFS_QMOPT_UQUOTA) && XFS_IS_UQUOTA_ON(mp)) {
+		if (ip->i_d.di_uid != uid) {
+			/*
+			 * What we need is the dquot that has this uid, and
+			 * if we send the inode to dqget, the uid of the inode
+			 * takes priority over what's sent in the uid argument.
+			 * We must unlock inode here before calling dqget if
+			 * we're not sending the inode, because otherwise
+			 * we'll deadlock by doing trans_reserve while
+			 * holding ilock.
+			 */
+			xfs_iunlock(ip, lockflags);
+			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t) uid,
+						 XFS_DQ_USER,
+						 XFS_QMOPT_DQALLOC |
+						 XFS_QMOPT_DOWARN,
+						 &uq))) {
+				ASSERT(error != ENOENT);
+				return error;
+			}
+			/*
+			 * Get the ilock in the right order.
+			 */
+			xfs_dqunlock(uq);
+			lockflags = XFS_ILOCK_SHARED;
+			xfs_ilock(ip, lockflags);
+		} else {
+			/*
+			 * Take an extra reference, because we'll return
+			 * this to caller
+			 */
+			ASSERT(ip->i_udquot);
+			uq = ip->i_udquot;
+			xfs_dqlock(uq);
+			XFS_DQHOLD(uq);
+			xfs_dqunlock(uq);
+		}
+	}
+	if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) {
+		if (ip->i_d.di_gid != gid) {
+			xfs_iunlock(ip, lockflags);
+			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)gid,
+						 XFS_DQ_GROUP,
+						 XFS_QMOPT_DQALLOC |
+						 XFS_QMOPT_DOWARN,
+						 &gq))) {
+				if (uq)
+					xfs_qm_dqrele(uq);
+				ASSERT(error != ENOENT);
+				return error;
+			}
+			xfs_dqunlock(gq);
+			lockflags = XFS_ILOCK_SHARED;
+			xfs_ilock(ip, lockflags);
+		} else {
+			ASSERT(ip->i_gdquot);
+			gq = ip->i_gdquot;
+			xfs_dqlock(gq);
+			XFS_DQHOLD(gq);
+			xfs_dqunlock(gq);
+		}
+	} else if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) {
+		if (xfs_get_projid(ip) != prid) {
+			xfs_iunlock(ip, lockflags);
+			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid,
+						 XFS_DQ_PROJ,
+						 XFS_QMOPT_DQALLOC |
+						 XFS_QMOPT_DOWARN,
+						 &gq))) {
+				if (uq)
+					xfs_qm_dqrele(uq);
+				ASSERT(error != ENOENT);
+				return (error);
+			}
+			xfs_dqunlock(gq);
+			lockflags = XFS_ILOCK_SHARED;
+			xfs_ilock(ip, lockflags);
+		} else {
+			ASSERT(ip->i_gdquot);
+			gq = ip->i_gdquot;
+			xfs_dqlock(gq);
+			XFS_DQHOLD(gq);
+			xfs_dqunlock(gq);
+		}
+	}
+	if (uq)
+		trace_xfs_dquot_dqalloc(ip);
+
+	xfs_iunlock(ip, lockflags);
+	if (O_udqpp)
+		*O_udqpp = uq;
+	else if (uq)
+		xfs_qm_dqrele(uq);
+	if (O_gdqpp)
+		*O_gdqpp = gq;
+	else if (gq)
+		xfs_qm_dqrele(gq);
+	return 0;
+}
+
+/*
+ * Actually transfer ownership, and do dquot modifications.
+ * These were already reserved.
+ */
+xfs_dquot_t *
+xfs_qm_vop_chown(
+	xfs_trans_t	*tp,
+	xfs_inode_t	*ip,
+	xfs_dquot_t	**IO_olddq,
+	xfs_dquot_t	*newdq)
+{
+	xfs_dquot_t	*prevdq;
+	uint		bfield = XFS_IS_REALTIME_INODE(ip) ?
+				 XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT;
+
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(XFS_IS_QUOTA_RUNNING(ip->i_mount));
+
+	/* old dquot */
+	prevdq = *IO_olddq;
+	ASSERT(prevdq);
+	ASSERT(prevdq != newdq);
+
+	xfs_trans_mod_dquot(tp, prevdq, bfield, -(ip->i_d.di_nblocks));
+	xfs_trans_mod_dquot(tp, prevdq, XFS_TRANS_DQ_ICOUNT, -1);
+
+	/* the sparkling new dquot */
+	xfs_trans_mod_dquot(tp, newdq, bfield, ip->i_d.di_nblocks);
+	xfs_trans_mod_dquot(tp, newdq, XFS_TRANS_DQ_ICOUNT, 1);
+
+	/*
+	 * Take an extra reference, because the inode
+	 * is going to keep this dquot pointer even
+	 * after the trans_commit.
+	 */
+	xfs_dqlock(newdq);
+	XFS_DQHOLD(newdq);
+	xfs_dqunlock(newdq);
+	*IO_olddq = newdq;
+
+	return prevdq;
+}
+
+/*
+ * Quota reservations for setattr(AT_UID|AT_GID|AT_PROJID).
+ */
+int
+xfs_qm_vop_chown_reserve(
+	xfs_trans_t	*tp,
+	xfs_inode_t	*ip,
+	xfs_dquot_t	*udqp,
+	xfs_dquot_t	*gdqp,
+	uint		flags)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	uint		delblks, blkflags, prjflags = 0;
+	xfs_dquot_t	*unresudq, *unresgdq, *delblksudq, *delblksgdq;
+	int		error;
+
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	delblks = ip->i_delayed_blks;
+	delblksudq = delblksgdq = unresudq = unresgdq = NULL;
+	blkflags = XFS_IS_REALTIME_INODE(ip) ?
+			XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS;
+
+	if (XFS_IS_UQUOTA_ON(mp) && udqp &&
+	    ip->i_d.di_uid != (uid_t)be32_to_cpu(udqp->q_core.d_id)) {
+		delblksudq = udqp;
+		/*
+		 * If there are delayed allocation blocks, then we have to
+		 * unreserve those from the old dquot, and add them to the
+		 * new dquot.
+		 */
+		if (delblks) {
+			ASSERT(ip->i_udquot);
+			unresudq = ip->i_udquot;
+		}
+	}
+	if (XFS_IS_OQUOTA_ON(ip->i_mount) && gdqp) {
+		if (XFS_IS_PQUOTA_ON(ip->i_mount) &&
+		     xfs_get_projid(ip) != be32_to_cpu(gdqp->q_core.d_id))
+			prjflags = XFS_QMOPT_ENOSPC;
+
+		if (prjflags ||
+		    (XFS_IS_GQUOTA_ON(ip->i_mount) &&
+		     ip->i_d.di_gid != be32_to_cpu(gdqp->q_core.d_id))) {
+			delblksgdq = gdqp;
+			if (delblks) {
+				ASSERT(ip->i_gdquot);
+				unresgdq = ip->i_gdquot;
+			}
+		}
+	}
+
+	if ((error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount,
+				delblksudq, delblksgdq, ip->i_d.di_nblocks, 1,
+				flags | blkflags | prjflags)))
+		return (error);
+
+	/*
+	 * Do the delayed blks reservations/unreservations now. Since, these
+	 * are done without the help of a transaction, if a reservation fails
+	 * its previous reservations won't be automatically undone by trans
+	 * code. So, we have to do it manually here.
+	 */
+	if (delblks) {
+		/*
+		 * Do the reservations first. Unreservation can't fail.
+		 */
+		ASSERT(delblksudq || delblksgdq);
+		ASSERT(unresudq || unresgdq);
+		if ((error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
+				delblksudq, delblksgdq, (xfs_qcnt_t)delblks, 0,
+				flags | blkflags | prjflags)))
+			return (error);
+		xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
+				unresudq, unresgdq, -((xfs_qcnt_t)delblks), 0,
+				blkflags);
+	}
+
+	return (0);
+}
+
+int
+xfs_qm_vop_rename_dqattach(
+	struct xfs_inode	**i_tab)
+{
+	struct xfs_mount	*mp = i_tab[0]->i_mount;
+	int			i;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return 0;
+
+	for (i = 0; (i < 4 && i_tab[i]); i++) {
+		struct xfs_inode	*ip = i_tab[i];
+		int			error;
+
+		/*
+		 * Watch out for duplicate entries in the table.
+		 */
+		if (i == 0 || ip != i_tab[i-1]) {
+			if (XFS_NOT_DQATTACHED(mp, ip)) {
+				error = xfs_qm_dqattach(ip, 0);
+				if (error)
+					return error;
+			}
+		}
+	}
+	return 0;
+}
+
+void
+xfs_qm_vop_create_dqattach(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	struct xfs_dquot	*udqp,
+	struct xfs_dquot	*gdqp)
+{
+	struct xfs_mount	*mp = tp->t_mountp;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return;
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
+
+	if (udqp) {
+		xfs_dqlock(udqp);
+		XFS_DQHOLD(udqp);
+		xfs_dqunlock(udqp);
+		ASSERT(ip->i_udquot == NULL);
+		ip->i_udquot = udqp;
+		ASSERT(XFS_IS_UQUOTA_ON(mp));
+		ASSERT(ip->i_d.di_uid == be32_to_cpu(udqp->q_core.d_id));
+		xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1);
+	}
+	if (gdqp) {
+		xfs_dqlock(gdqp);
+		XFS_DQHOLD(gdqp);
+		xfs_dqunlock(gdqp);
+		ASSERT(ip->i_gdquot == NULL);
+		ip->i_gdquot = gdqp;
+		ASSERT(XFS_IS_OQUOTA_ON(mp));
+		ASSERT((XFS_IS_GQUOTA_ON(mp) ?
+			ip->i_d.di_gid : xfs_get_projid(ip)) ==
+				be32_to_cpu(gdqp->q_core.d_id));
+		xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1);
+	}
+}
+
diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h
new file mode 100644
index 0000000..43b9abe
--- /dev/null
+++ b/fs/xfs/xfs_qm.h
@@ -0,0 +1,166 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_QM_H__
+#define __XFS_QM_H__
+
+#include "xfs_dquot_item.h"
+#include "xfs_dquot.h"
+#include "xfs_quota_priv.h"
+#include "xfs_qm_stats.h"
+
+struct xfs_qm;
+struct xfs_inode;
+
+extern uint		ndquot;
+extern struct mutex	xfs_Gqm_lock;
+extern struct xfs_qm	*xfs_Gqm;
+extern kmem_zone_t	*qm_dqzone;
+extern kmem_zone_t	*qm_dqtrxzone;
+
+/*
+ * Used in xfs_qm_sync called by xfs_sync to count the max times that it can
+ * iterate over the mountpt's dquot list in one call.
+ */
+#define XFS_QM_SYNC_MAX_RESTARTS	7
+
+/*
+ * Ditto, for xfs_qm_dqreclaim_one.
+ */
+#define XFS_QM_RECLAIM_MAX_RESTARTS	4
+
+/*
+ * Ideal ratio of free to in use dquots. Quota manager makes an attempt
+ * to keep this balance.
+ */
+#define XFS_QM_DQFREE_RATIO		2
+
+/*
+ * Dquot hashtable constants/threshold values.
+ */
+#define XFS_QM_HASHSIZE_LOW		(PAGE_SIZE / sizeof(xfs_dqhash_t))
+#define XFS_QM_HASHSIZE_HIGH		((PAGE_SIZE * 4) / sizeof(xfs_dqhash_t))
+
+/*
+ * This defines the unit of allocation of dquots.
+ * Currently, it is just one file system block, and a 4K blk contains 30
+ * (136 * 30 = 4080) dquots. It's probably not worth trying to make
+ * this more dynamic.
+ * XXXsup However, if this number is changed, we have to make sure that we don't
+ * implicitly assume that we do allocations in chunks of a single filesystem
+ * block in the dquot/xqm code.
+ */
+#define XFS_DQUOT_CLUSTER_SIZE_FSB	(xfs_filblks_t)1
+
+typedef xfs_dqhash_t	xfs_dqlist_t;
+
+/*
+ * Quota Manager (global) structure. Lives only in core.
+ */
+typedef struct xfs_qm {
+	xfs_dqlist_t	*qm_usr_dqhtable;/* udquot hash table */
+	xfs_dqlist_t	*qm_grp_dqhtable;/* gdquot hash table */
+	uint		 qm_dqhashmask;	 /* # buckets in dq hashtab - 1 */
+	struct list_head qm_dqfrlist;	 /* freelist of dquots */
+	struct mutex	 qm_dqfrlist_lock;
+	int		 qm_dqfrlist_cnt;
+	atomic_t	 qm_totaldquots; /* total incore dquots */
+	uint		 qm_nrefs;	 /* file systems with quota on */
+	int		 qm_dqfree_ratio;/* ratio of free to inuse dquots */
+	kmem_zone_t	*qm_dqzone;	 /* dquot mem-alloc zone */
+	kmem_zone_t	*qm_dqtrxzone;	 /* t_dqinfo of transactions */
+} xfs_qm_t;
+
+/*
+ * Various quota information for individual filesystems.
+ * The mount structure keeps a pointer to this.
+ */
+typedef struct xfs_quotainfo {
+	xfs_inode_t	*qi_uquotaip;	 /* user quota inode */
+	xfs_inode_t	*qi_gquotaip;	 /* group quota inode */
+	struct list_head qi_dqlist;	 /* all dquots in filesys */
+	struct mutex	 qi_dqlist_lock;
+	int		 qi_dquots;
+	int		 qi_dqreclaims;	 /* a change here indicates
+					    a removal in the dqlist */
+	time_t		 qi_btimelimit;	 /* limit for blks timer */
+	time_t		 qi_itimelimit;	 /* limit for inodes timer */
+	time_t		 qi_rtbtimelimit;/* limit for rt blks timer */
+	xfs_qwarncnt_t	 qi_bwarnlimit;	 /* limit for blks warnings */
+	xfs_qwarncnt_t	 qi_iwarnlimit;	 /* limit for inodes warnings */
+	xfs_qwarncnt_t	 qi_rtbwarnlimit;/* limit for rt blks warnings */
+	struct mutex	 qi_quotaofflock;/* to serialize quotaoff */
+	xfs_filblks_t	 qi_dqchunklen;	 /* # BBs in a chunk of dqs */
+	uint		 qi_dqperchunk;	 /* # ondisk dqs in above chunk */
+	xfs_qcnt_t	 qi_bhardlimit;	 /* default data blk hard limit */
+	xfs_qcnt_t	 qi_bsoftlimit;	 /* default data blk soft limit */
+	xfs_qcnt_t	 qi_ihardlimit;	 /* default inode count hard limit */
+	xfs_qcnt_t	 qi_isoftlimit;	 /* default inode count soft limit */
+	xfs_qcnt_t	 qi_rtbhardlimit;/* default realtime blk hard limit */
+	xfs_qcnt_t	 qi_rtbsoftlimit;/* default realtime blk soft limit */
+} xfs_quotainfo_t;
+
+
+extern void	xfs_trans_mod_dquot(xfs_trans_t *, xfs_dquot_t *, uint, long);
+extern int	xfs_trans_reserve_quota_bydquots(xfs_trans_t *, xfs_mount_t *,
+			xfs_dquot_t *, xfs_dquot_t *, long, long, uint);
+extern void	xfs_trans_dqjoin(xfs_trans_t *, xfs_dquot_t *);
+extern void	xfs_trans_log_dquot(xfs_trans_t *, xfs_dquot_t *);
+
+/*
+ * We keep the usr and grp dquots separately so that locking will be easier
+ * to do at commit time. All transactions that we know of at this point
+ * affect no more than two dquots of one type. Hence, the TRANS_MAXDQS value.
+ */
+#define XFS_QM_TRANS_MAXDQS		2
+typedef struct xfs_dquot_acct {
+	xfs_dqtrx_t	dqa_usrdquots[XFS_QM_TRANS_MAXDQS];
+	xfs_dqtrx_t	dqa_grpdquots[XFS_QM_TRANS_MAXDQS];
+} xfs_dquot_acct_t;
+
+/*
+ * Users are allowed to have a usage exceeding their softlimit for
+ * a period this long.
+ */
+#define XFS_QM_BTIMELIMIT	(7 * 24*60*60)          /* 1 week */
+#define XFS_QM_RTBTIMELIMIT	(7 * 24*60*60)          /* 1 week */
+#define XFS_QM_ITIMELIMIT	(7 * 24*60*60)          /* 1 week */
+
+#define XFS_QM_BWARNLIMIT	5
+#define XFS_QM_IWARNLIMIT	5
+#define XFS_QM_RTBWARNLIMIT	5
+
+extern void		xfs_qm_destroy_quotainfo(xfs_mount_t *);
+extern int		xfs_qm_quotacheck(xfs_mount_t *);
+extern int		xfs_qm_write_sb_changes(xfs_mount_t *, __int64_t);
+
+/* dquot stuff */
+extern boolean_t	xfs_qm_dqalloc_incore(xfs_dquot_t **);
+extern int		xfs_qm_dqpurge_all(xfs_mount_t *, uint);
+extern void		xfs_qm_dqrele_all_inodes(xfs_mount_t *, uint);
+
+/* quota ops */
+extern int		xfs_qm_scall_trunc_qfiles(xfs_mount_t *, uint);
+extern int		xfs_qm_scall_getquota(xfs_mount_t *, xfs_dqid_t, uint,
+					fs_disk_quota_t *);
+extern int		xfs_qm_scall_setqlim(xfs_mount_t *, xfs_dqid_t, uint,
+					fs_disk_quota_t *);
+extern int		xfs_qm_scall_getqstat(xfs_mount_t *, fs_quota_stat_t *);
+extern int		xfs_qm_scall_quotaon(xfs_mount_t *, uint);
+extern int		xfs_qm_scall_quotaoff(xfs_mount_t *, uint);
+
+#endif /* __XFS_QM_H__ */
diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c
new file mode 100644
index 0000000..a0a829a
--- /dev/null
+++ b/fs/xfs/xfs_qm_bhv.c
@@ -0,0 +1,176 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_itable.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_qm.h"
+
+
+STATIC void
+xfs_fill_statvfs_from_dquot(
+	struct kstatfs		*statp,
+	xfs_disk_dquot_t	*dp)
+{
+	__uint64_t		limit;
+
+	limit = dp->d_blk_softlimit ?
+		be64_to_cpu(dp->d_blk_softlimit) :
+		be64_to_cpu(dp->d_blk_hardlimit);
+	if (limit && statp->f_blocks > limit) {
+		statp->f_blocks = limit;
+		statp->f_bfree = statp->f_bavail =
+			(statp->f_blocks > be64_to_cpu(dp->d_bcount)) ?
+			 (statp->f_blocks - be64_to_cpu(dp->d_bcount)) : 0;
+	}
+
+	limit = dp->d_ino_softlimit ?
+		be64_to_cpu(dp->d_ino_softlimit) :
+		be64_to_cpu(dp->d_ino_hardlimit);
+	if (limit && statp->f_files > limit) {
+		statp->f_files = limit;
+		statp->f_ffree =
+			(statp->f_files > be64_to_cpu(dp->d_icount)) ?
+			 (statp->f_ffree - be64_to_cpu(dp->d_icount)) : 0;
+	}
+}
+
+
+/*
+ * Directory tree accounting is implemented using project quotas, where
+ * the project identifier is inherited from parent directories.
+ * A statvfs (df, etc.) of a directory that is using project quota should
+ * return a statvfs of the project, not the entire filesystem.
+ * This makes such trees appear as if they are filesystems in themselves.
+ */
+void
+xfs_qm_statvfs(
+	xfs_inode_t		*ip,
+	struct kstatfs		*statp)
+{
+	xfs_mount_t		*mp = ip->i_mount;
+	xfs_dquot_t		*dqp;
+
+	if (!xfs_qm_dqget(mp, NULL, xfs_get_projid(ip), XFS_DQ_PROJ, 0, &dqp)) {
+		xfs_fill_statvfs_from_dquot(statp, &dqp->q_core);
+		xfs_qm_dqput(dqp);
+	}
+}
+
+int
+xfs_qm_newmount(
+	xfs_mount_t	*mp,
+	uint		*needquotamount,
+	uint		*quotaflags)
+{
+	uint		quotaondisk;
+	uint		uquotaondisk = 0, gquotaondisk = 0, pquotaondisk = 0;
+
+	quotaondisk = xfs_sb_version_hasquota(&mp->m_sb) &&
+				(mp->m_sb.sb_qflags & XFS_ALL_QUOTA_ACCT);
+
+	if (quotaondisk) {
+		uquotaondisk = mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT;
+		pquotaondisk = mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT;
+		gquotaondisk = mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT;
+	}
+
+	/*
+	 * If the device itself is read-only, we can't allow
+	 * the user to change the state of quota on the mount -
+	 * this would generate a transaction on the ro device,
+	 * which would lead to an I/O error and shutdown
+	 */
+
+	if (((uquotaondisk && !XFS_IS_UQUOTA_ON(mp)) ||
+	    (!uquotaondisk &&  XFS_IS_UQUOTA_ON(mp)) ||
+	     (pquotaondisk && !XFS_IS_PQUOTA_ON(mp)) ||
+	    (!pquotaondisk &&  XFS_IS_PQUOTA_ON(mp)) ||
+	     (gquotaondisk && !XFS_IS_GQUOTA_ON(mp)) ||
+	    (!gquotaondisk &&  XFS_IS_OQUOTA_ON(mp)))  &&
+	    xfs_dev_is_read_only(mp, "changing quota state")) {
+		xfs_warn(mp, "please mount with%s%s%s%s.",
+			(!quotaondisk ? "out quota" : ""),
+			(uquotaondisk ? " usrquota" : ""),
+			(pquotaondisk ? " prjquota" : ""),
+			(gquotaondisk ? " grpquota" : ""));
+		return XFS_ERROR(EPERM);
+	}
+
+	if (XFS_IS_QUOTA_ON(mp) || quotaondisk) {
+		/*
+		 * Call mount_quotas at this point only if we won't have to do
+		 * a quotacheck.
+		 */
+		if (quotaondisk && !XFS_QM_NEED_QUOTACHECK(mp)) {
+			/*
+			 * If an error occurred, qm_mount_quotas code
+			 * has already disabled quotas. So, just finish
+			 * mounting, and get on with the boring life
+			 * without disk quotas.
+			 */
+			xfs_qm_mount_quotas(mp);
+		} else {
+			/*
+			 * Clear the quota flags, but remember them. This
+			 * is so that the quota code doesn't get invoked
+			 * before we're ready. This can happen when an
+			 * inode goes inactive and wants to free blocks,
+			 * or via xfs_log_mount_finish.
+			 */
+			*needquotamount = B_TRUE;
+			*quotaflags = mp->m_qflags;
+			mp->m_qflags = 0;
+		}
+	}
+
+	return 0;
+}
+
+void __init
+xfs_qm_init(void)
+{
+	printk(KERN_INFO "SGI XFS Quota Management subsystem\n");
+	mutex_init(&xfs_Gqm_lock);
+	xfs_qm_init_procfs();
+}
+
+void __exit
+xfs_qm_exit(void)
+{
+	xfs_qm_cleanup_procfs();
+	if (qm_dqzone)
+		kmem_zone_destroy(qm_dqzone);
+	if (qm_dqtrxzone)
+		kmem_zone_destroy(qm_dqtrxzone);
+}
diff --git a/fs/xfs/xfs_qm_stats.c b/fs/xfs/xfs_qm_stats.c
new file mode 100644
index 0000000..8671a0b
--- /dev/null
+++ b/fs/xfs/xfs_qm_stats.c
@@ -0,0 +1,105 @@
+/*
+ * Copyright (c) 2000-2003 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_itable.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_qm.h"
+
+struct xqmstats xqmstats;
+
+static int xqm_proc_show(struct seq_file *m, void *v)
+{
+	/* maximum; incore; ratio free to inuse; freelist */
+	seq_printf(m, "%d\t%d\t%d\t%u\n",
+			ndquot,
+			xfs_Gqm? atomic_read(&xfs_Gqm->qm_totaldquots) : 0,
+			xfs_Gqm? xfs_Gqm->qm_dqfree_ratio : 0,
+			xfs_Gqm? xfs_Gqm->qm_dqfrlist_cnt : 0);
+	return 0;
+}
+
+static int xqm_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xqm_proc_show, NULL);
+}
+
+static const struct file_operations xqm_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqm_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static int xqmstat_proc_show(struct seq_file *m, void *v)
+{
+	/* quota performance statistics */
+	seq_printf(m, "qm %u %u %u %u %u %u %u %u\n",
+			xqmstats.xs_qm_dqreclaims,
+			xqmstats.xs_qm_dqreclaim_misses,
+			xqmstats.xs_qm_dquot_dups,
+			xqmstats.xs_qm_dqcachemisses,
+			xqmstats.xs_qm_dqcachehits,
+			xqmstats.xs_qm_dqwants,
+			xqmstats.xs_qm_dqshake_reclaims,
+			xqmstats.xs_qm_dqinact_reclaims);
+	return 0;
+}
+
+static int xqmstat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xqmstat_proc_show, NULL);
+}
+
+static const struct file_operations xqmstat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqmstat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+void
+xfs_qm_init_procfs(void)
+{
+	proc_create("fs/xfs/xqmstat", 0, NULL, &xqmstat_proc_fops);
+	proc_create("fs/xfs/xqm", 0, NULL, &xqm_proc_fops);
+}
+
+void
+xfs_qm_cleanup_procfs(void)
+{
+	remove_proc_entry("fs/xfs/xqm", NULL);
+	remove_proc_entry("fs/xfs/xqmstat", NULL);
+}
diff --git a/fs/xfs/xfs_qm_stats.h b/fs/xfs/xfs_qm_stats.h
new file mode 100644
index 0000000..5b964fc
--- /dev/null
+++ b/fs/xfs/xfs_qm_stats.h
@@ -0,0 +1,53 @@
+/*
+ * Copyright (c) 2002 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_QM_STATS_H__
+#define __XFS_QM_STATS_H__
+
+#if defined(CONFIG_PROC_FS) && !defined(XFS_STATS_OFF)
+
+/*
+ * XQM global statistics
+ */
+struct xqmstats {
+	__uint32_t		xs_qm_dqreclaims;
+	__uint32_t		xs_qm_dqreclaim_misses;
+	__uint32_t		xs_qm_dquot_dups;
+	__uint32_t		xs_qm_dqcachemisses;
+	__uint32_t		xs_qm_dqcachehits;
+	__uint32_t		xs_qm_dqwants;
+	__uint32_t		xs_qm_dqshake_reclaims;
+	__uint32_t		xs_qm_dqinact_reclaims;
+};
+
+extern struct xqmstats xqmstats;
+
+# define XQM_STATS_INC(count)	( (count)++ )
+
+extern void xfs_qm_init_procfs(void);
+extern void xfs_qm_cleanup_procfs(void);
+
+#else
+
+# define XQM_STATS_INC(count)	do { } while (0)
+
+static inline void xfs_qm_init_procfs(void) { };
+static inline void xfs_qm_cleanup_procfs(void) { };
+
+#endif
+
+#endif	/* __XFS_QM_STATS_H__ */
diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
new file mode 100644
index 0000000..5cc3dde
--- /dev/null
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -0,0 +1,906 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/capability.h>
+
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_itable.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_utils.h"
+#include "xfs_qm.h"
+#include "xfs_trace.h"
+
+STATIC int	xfs_qm_log_quotaoff(xfs_mount_t *, xfs_qoff_logitem_t **, uint);
+STATIC int	xfs_qm_log_quotaoff_end(xfs_mount_t *, xfs_qoff_logitem_t *,
+					uint);
+STATIC uint	xfs_qm_export_flags(uint);
+STATIC uint	xfs_qm_export_qtype_flags(uint);
+STATIC void	xfs_qm_export_dquot(xfs_mount_t *, xfs_disk_dquot_t *,
+					fs_disk_quota_t *);
+
+
+/*
+ * Turn off quota accounting and/or enforcement for all udquots and/or
+ * gdquots. Called only at unmount time.
+ *
+ * This assumes that there are no dquots of this file system cached
+ * incore, and modifies the ondisk dquot directly. Therefore, for example,
+ * it is an error to call this twice, without purging the cache.
+ */
+int
+xfs_qm_scall_quotaoff(
+	xfs_mount_t		*mp,
+	uint			flags)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	uint			dqtype;
+	int			error;
+	uint			inactivate_flags;
+	xfs_qoff_logitem_t	*qoffstart;
+	int			nculprits;
+
+	/*
+	 * No file system can have quotas enabled on disk but not in core.
+	 * Note that quota utilities (like quotaoff) _expect_
+	 * errno == EEXIST here.
+	 */
+	if ((mp->m_qflags & flags) == 0)
+		return XFS_ERROR(EEXIST);
+	error = 0;
+
+	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
+
+	/*
+	 * We don't want to deal with two quotaoffs messing up each other,
+	 * so we're going to serialize it. quotaoff isn't exactly a performance
+	 * critical thing.
+	 * If quotaoff, then we must be dealing with the root filesystem.
+	 */
+	ASSERT(q);
+	mutex_lock(&q->qi_quotaofflock);
+
+	/*
+	 * If we're just turning off quota enforcement, change mp and go.
+	 */
+	if ((flags & XFS_ALL_QUOTA_ACCT) == 0) {
+		mp->m_qflags &= ~(flags);
+
+		spin_lock(&mp->m_sb_lock);
+		mp->m_sb.sb_qflags = mp->m_qflags;
+		spin_unlock(&mp->m_sb_lock);
+		mutex_unlock(&q->qi_quotaofflock);
+
+		/* XXX what to do if error ? Revert back to old vals incore ? */
+		error = xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS);
+		return (error);
+	}
+
+	dqtype = 0;
+	inactivate_flags = 0;
+	/*
+	 * If accounting is off, we must turn enforcement off, clear the
+	 * quota 'CHKD' certificate to make it known that we have to
+	 * do a quotacheck the next time this quota is turned on.
+	 */
+	if (flags & XFS_UQUOTA_ACCT) {
+		dqtype |= XFS_QMOPT_UQUOTA;
+		flags |= (XFS_UQUOTA_CHKD | XFS_UQUOTA_ENFD);
+		inactivate_flags |= XFS_UQUOTA_ACTIVE;
+	}
+	if (flags & XFS_GQUOTA_ACCT) {
+		dqtype |= XFS_QMOPT_GQUOTA;
+		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
+		inactivate_flags |= XFS_GQUOTA_ACTIVE;
+	} else if (flags & XFS_PQUOTA_ACCT) {
+		dqtype |= XFS_QMOPT_PQUOTA;
+		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
+		inactivate_flags |= XFS_PQUOTA_ACTIVE;
+	}
+
+	/*
+	 * Nothing to do?  Don't complain. This happens when we're just
+	 * turning off quota enforcement.
+	 */
+	if ((mp->m_qflags & flags) == 0)
+		goto out_unlock;
+
+	/*
+	 * Write the LI_QUOTAOFF log record, and do SB changes atomically,
+	 * and synchronously. If we fail to write, we should abort the
+	 * operation as it cannot be recovered safely if we crash.
+	 */
+	error = xfs_qm_log_quotaoff(mp, &qoffstart, flags);
+	if (error)
+		goto out_unlock;
+
+	/*
+	 * Next we clear the XFS_MOUNT_*DQ_ACTIVE bit(s) in the mount struct
+	 * to take care of the race between dqget and quotaoff. We don't take
+	 * any special locks to reset these bits. All processes need to check
+	 * these bits *after* taking inode lock(s) to see if the particular
+	 * quota type is in the process of being turned off. If *ACTIVE, it is
+	 * guaranteed that all dquot structures and all quotainode ptrs will all
+	 * stay valid as long as that inode is kept locked.
+	 *
+	 * There is no turning back after this.
+	 */
+	mp->m_qflags &= ~inactivate_flags;
+
+	/*
+	 * Give back all the dquot reference(s) held by inodes.
+	 * Here we go thru every single incore inode in this file system, and
+	 * do a dqrele on the i_udquot/i_gdquot that it may have.
+	 * Essentially, as long as somebody has an inode locked, this guarantees
+	 * that quotas will not be turned off. This is handy because in a
+	 * transaction once we lock the inode(s) and check for quotaon, we can
+	 * depend on the quota inodes (and other things) being valid as long as
+	 * we keep the lock(s).
+	 */
+	xfs_qm_dqrele_all_inodes(mp, flags);
+
+	/*
+	 * Next we make the changes in the quota flag in the mount struct.
+	 * This isn't protected by a particular lock directly, because we
+	 * don't want to take a mrlock every time we depend on quotas being on.
+	 */
+	mp->m_qflags &= ~(flags);
+
+	/*
+	 * Go through all the dquots of this file system and purge them,
+	 * according to what was turned off. We may not be able to get rid
+	 * of all dquots, because dquots can have temporary references that
+	 * are not attached to inodes. eg. xfs_setattr, xfs_create.
+	 * So, if we couldn't purge all the dquots from the filesystem,
+	 * we can't get rid of the incore data structures.
+	 */
+	while ((nculprits = xfs_qm_dqpurge_all(mp, dqtype)))
+		delay(10 * nculprits);
+
+	/*
+	 * Transactions that had started before ACTIVE state bit was cleared
+	 * could have logged many dquots, so they'd have higher LSNs than
+	 * the first QUOTAOFF log record does. If we happen to crash when
+	 * the tail of the log has gone past the QUOTAOFF record, but
+	 * before the last dquot modification, those dquots __will__
+	 * recover, and that's not good.
+	 *
+	 * So, we have QUOTAOFF start and end logitems; the start
+	 * logitem won't get overwritten until the end logitem appears...
+	 */
+	error = xfs_qm_log_quotaoff_end(mp, qoffstart, flags);
+	if (error) {
+		/* We're screwed now. Shutdown is the only option. */
+		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		goto out_unlock;
+	}
+
+	/*
+	 * If quotas is completely disabled, close shop.
+	 */
+	if (((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET1) ||
+	    ((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET2)) {
+		mutex_unlock(&q->qi_quotaofflock);
+		xfs_qm_destroy_quotainfo(mp);
+		return (0);
+	}
+
+	/*
+	 * Release our quotainode references if we don't need them anymore.
+	 */
+	if ((dqtype & XFS_QMOPT_UQUOTA) && q->qi_uquotaip) {
+		IRELE(q->qi_uquotaip);
+		q->qi_uquotaip = NULL;
+	}
+	if ((dqtype & (XFS_QMOPT_GQUOTA|XFS_QMOPT_PQUOTA)) && q->qi_gquotaip) {
+		IRELE(q->qi_gquotaip);
+		q->qi_gquotaip = NULL;
+	}
+
+out_unlock:
+	mutex_unlock(&q->qi_quotaofflock);
+	return error;
+}
+
+STATIC int
+xfs_qm_scall_trunc_qfile(
+	struct xfs_mount	*mp,
+	xfs_ino_t		ino)
+{
+	struct xfs_inode	*ip;
+	struct xfs_trans	*tp;
+	int			error;
+
+	if (ino == NULLFSINO)
+		return 0;
+
+	error = xfs_iget(mp, NULL, ino, 0, 0, &ip);
+	if (error)
+		return error;
+
+	xfs_ilock(ip, XFS_IOLOCK_EXCL);
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_TRUNCATE_FILE);
+	error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0,
+				  XFS_TRANS_PERM_LOG_RES,
+				  XFS_ITRUNCATE_LOG_COUNT);
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		xfs_iunlock(ip, XFS_IOLOCK_EXCL);
+		goto out_put;
+	}
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, 0);
+
+	error = xfs_itruncate_data(&tp, ip, 0);
+	if (error) {
+		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES |
+				     XFS_TRANS_ABORT);
+		goto out_unlock;
+	}
+
+	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+
+out_unlock:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+out_put:
+	IRELE(ip);
+	return error;
+}
+
+int
+xfs_qm_scall_trunc_qfiles(
+	xfs_mount_t	*mp,
+	uint		flags)
+{
+	int		error = 0, error2 = 0;
+
+	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) {
+		xfs_debug(mp, "%s: flags=%x m_qflags=%x\n",
+			__func__, flags, mp->m_qflags);
+		return XFS_ERROR(EINVAL);
+	}
+
+	if (flags & XFS_DQ_USER)
+		error = xfs_qm_scall_trunc_qfile(mp, mp->m_sb.sb_uquotino);
+	if (flags & (XFS_DQ_GROUP|XFS_DQ_PROJ))
+		error2 = xfs_qm_scall_trunc_qfile(mp, mp->m_sb.sb_gquotino);
+
+	return error ? error : error2;
+}
+
+/*
+ * Switch on (a given) quota enforcement for a filesystem.  This takes
+ * effect immediately.
+ * (Switching on quota accounting must be done at mount time.)
+ */
+int
+xfs_qm_scall_quotaon(
+	xfs_mount_t	*mp,
+	uint		flags)
+{
+	int		error;
+	uint		qf;
+	__int64_t	sbflags;
+
+	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
+	/*
+	 * Switching on quota accounting must be done at mount time.
+	 */
+	flags &= ~(XFS_ALL_QUOTA_ACCT);
+
+	sbflags = 0;
+
+	if (flags == 0) {
+		xfs_debug(mp, "%s: zero flags, m_qflags=%x\n",
+			__func__, mp->m_qflags);
+		return XFS_ERROR(EINVAL);
+	}
+
+	/* No fs can turn on quotas with a delayed effect */
+	ASSERT((flags & XFS_ALL_QUOTA_ACCT) == 0);
+
+	/*
+	 * Can't enforce without accounting. We check the superblock
+	 * qflags here instead of m_qflags because rootfs can have
+	 * quota acct on ondisk without m_qflags' knowing.
+	 */
+	if (((flags & XFS_UQUOTA_ACCT) == 0 &&
+	    (mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) == 0 &&
+	    (flags & XFS_UQUOTA_ENFD))
+	    ||
+	    ((flags & XFS_PQUOTA_ACCT) == 0 &&
+	    (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) == 0 &&
+	    (flags & XFS_GQUOTA_ACCT) == 0 &&
+	    (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) == 0 &&
+	    (flags & XFS_OQUOTA_ENFD))) {
+		xfs_debug(mp,
+			"%s: Can't enforce without acct, flags=%x sbflags=%x\n",
+			__func__, flags, mp->m_sb.sb_qflags);
+		return XFS_ERROR(EINVAL);
+	}
+	/*
+	 * If everything's up to-date incore, then don't waste time.
+	 */
+	if ((mp->m_qflags & flags) == flags)
+		return XFS_ERROR(EEXIST);
+
+	/*
+	 * Change sb_qflags on disk but not incore mp->qflags
+	 * if this is the root filesystem.
+	 */
+	spin_lock(&mp->m_sb_lock);
+	qf = mp->m_sb.sb_qflags;
+	mp->m_sb.sb_qflags = qf | flags;
+	spin_unlock(&mp->m_sb_lock);
+
+	/*
+	 * There's nothing to change if it's the same.
+	 */
+	if ((qf & flags) == flags && sbflags == 0)
+		return XFS_ERROR(EEXIST);
+	sbflags |= XFS_SB_QFLAGS;
+
+	if ((error = xfs_qm_write_sb_changes(mp, sbflags)))
+		return (error);
+	/*
+	 * If we aren't trying to switch on quota enforcement, we are done.
+	 */
+	if  (((mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) !=
+	     (mp->m_qflags & XFS_UQUOTA_ACCT)) ||
+	     ((mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) !=
+	     (mp->m_qflags & XFS_PQUOTA_ACCT)) ||
+	     ((mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) !=
+	     (mp->m_qflags & XFS_GQUOTA_ACCT)) ||
+	    (flags & XFS_ALL_QUOTA_ENFD) == 0)
+		return (0);
+
+	if (! XFS_IS_QUOTA_RUNNING(mp))
+		return XFS_ERROR(ESRCH);
+
+	/*
+	 * Switch on quota enforcement in core.
+	 */
+	mutex_lock(&mp->m_quotainfo->qi_quotaofflock);
+	mp->m_qflags |= (flags & XFS_ALL_QUOTA_ENFD);
+	mutex_unlock(&mp->m_quotainfo->qi_quotaofflock);
+
+	return (0);
+}
+
+
+/*
+ * Return quota status information, such as uquota-off, enforcements, etc.
+ */
+int
+xfs_qm_scall_getqstat(
+	struct xfs_mount	*mp,
+	struct fs_quota_stat	*out)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	struct xfs_inode	*uip, *gip;
+	boolean_t		tempuqip, tempgqip;
+
+	uip = gip = NULL;
+	tempuqip = tempgqip = B_FALSE;
+	memset(out, 0, sizeof(fs_quota_stat_t));
+
+	out->qs_version = FS_QSTAT_VERSION;
+	if (!xfs_sb_version_hasquota(&mp->m_sb)) {
+		out->qs_uquota.qfs_ino = NULLFSINO;
+		out->qs_gquota.qfs_ino = NULLFSINO;
+		return (0);
+	}
+	out->qs_flags = (__uint16_t) xfs_qm_export_flags(mp->m_qflags &
+							(XFS_ALL_QUOTA_ACCT|
+							 XFS_ALL_QUOTA_ENFD));
+	out->qs_pad = 0;
+	out->qs_uquota.qfs_ino = mp->m_sb.sb_uquotino;
+	out->qs_gquota.qfs_ino = mp->m_sb.sb_gquotino;
+
+	if (q) {
+		uip = q->qi_uquotaip;
+		gip = q->qi_gquotaip;
+	}
+	if (!uip && mp->m_sb.sb_uquotino != NULLFSINO) {
+		if (xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
+					0, 0, &uip) == 0)
+			tempuqip = B_TRUE;
+	}
+	if (!gip && mp->m_sb.sb_gquotino != NULLFSINO) {
+		if (xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
+					0, 0, &gip) == 0)
+			tempgqip = B_TRUE;
+	}
+	if (uip) {
+		out->qs_uquota.qfs_nblks = uip->i_d.di_nblocks;
+		out->qs_uquota.qfs_nextents = uip->i_d.di_nextents;
+		if (tempuqip)
+			IRELE(uip);
+	}
+	if (gip) {
+		out->qs_gquota.qfs_nblks = gip->i_d.di_nblocks;
+		out->qs_gquota.qfs_nextents = gip->i_d.di_nextents;
+		if (tempgqip)
+			IRELE(gip);
+	}
+	if (q) {
+		out->qs_incoredqs = q->qi_dquots;
+		out->qs_btimelimit = q->qi_btimelimit;
+		out->qs_itimelimit = q->qi_itimelimit;
+		out->qs_rtbtimelimit = q->qi_rtbtimelimit;
+		out->qs_bwarnlimit = q->qi_bwarnlimit;
+		out->qs_iwarnlimit = q->qi_iwarnlimit;
+	}
+	return 0;
+}
+
+#define XFS_DQ_MASK \
+	(FS_DQ_LIMIT_MASK | FS_DQ_TIMER_MASK | FS_DQ_WARNS_MASK)
+
+/*
+ * Adjust quota limits, and start/stop timers accordingly.
+ */
+int
+xfs_qm_scall_setqlim(
+	xfs_mount_t		*mp,
+	xfs_dqid_t		id,
+	uint			type,
+	fs_disk_quota_t		*newlim)
+{
+	struct xfs_quotainfo	*q = mp->m_quotainfo;
+	xfs_disk_dquot_t	*ddq;
+	xfs_dquot_t		*dqp;
+	xfs_trans_t		*tp;
+	int			error;
+	xfs_qcnt_t		hard, soft;
+
+	if (newlim->d_fieldmask & ~XFS_DQ_MASK)
+		return EINVAL;
+	if ((newlim->d_fieldmask & XFS_DQ_MASK) == 0)
+		return 0;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SETQLIM);
+	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_disk_dquot_t) + 128,
+				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
+		xfs_trans_cancel(tp, 0);
+		return (error);
+	}
+
+	/*
+	 * We don't want to race with a quotaoff so take the quotaoff lock.
+	 * (We don't hold an inode lock, so there's nothing else to stop
+	 * a quotaoff from happening). (XXXThis doesn't currently happen
+	 * because we take the vfslock before calling xfs_qm_sysent).
+	 */
+	mutex_lock(&q->qi_quotaofflock);
+
+	/*
+	 * Get the dquot (locked), and join it to the transaction.
+	 * Allocate the dquot if this doesn't exist.
+	 */
+	if ((error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp))) {
+		xfs_trans_cancel(tp, XFS_TRANS_ABORT);
+		ASSERT(error != ENOENT);
+		goto out_unlock;
+	}
+	xfs_trans_dqjoin(tp, dqp);
+	ddq = &dqp->q_core;
+
+	/*
+	 * Make sure that hardlimits are >= soft limits before changing.
+	 */
+	hard = (newlim->d_fieldmask & FS_DQ_BHARD) ?
+		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_hardlimit) :
+			be64_to_cpu(ddq->d_blk_hardlimit);
+	soft = (newlim->d_fieldmask & FS_DQ_BSOFT) ?
+		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_softlimit) :
+			be64_to_cpu(ddq->d_blk_softlimit);
+	if (hard == 0 || hard >= soft) {
+		ddq->d_blk_hardlimit = cpu_to_be64(hard);
+		ddq->d_blk_softlimit = cpu_to_be64(soft);
+		if (id == 0) {
+			q->qi_bhardlimit = hard;
+			q->qi_bsoftlimit = soft;
+		}
+	} else {
+		xfs_debug(mp, "blkhard %Ld < blksoft %Ld\n", hard, soft);
+	}
+	hard = (newlim->d_fieldmask & FS_DQ_RTBHARD) ?
+		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_hardlimit) :
+			be64_to_cpu(ddq->d_rtb_hardlimit);
+	soft = (newlim->d_fieldmask & FS_DQ_RTBSOFT) ?
+		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_softlimit) :
+			be64_to_cpu(ddq->d_rtb_softlimit);
+	if (hard == 0 || hard >= soft) {
+		ddq->d_rtb_hardlimit = cpu_to_be64(hard);
+		ddq->d_rtb_softlimit = cpu_to_be64(soft);
+		if (id == 0) {
+			q->qi_rtbhardlimit = hard;
+			q->qi_rtbsoftlimit = soft;
+		}
+	} else {
+		xfs_debug(mp, "rtbhard %Ld < rtbsoft %Ld\n", hard, soft);
+	}
+
+	hard = (newlim->d_fieldmask & FS_DQ_IHARD) ?
+		(xfs_qcnt_t) newlim->d_ino_hardlimit :
+			be64_to_cpu(ddq->d_ino_hardlimit);
+	soft = (newlim->d_fieldmask & FS_DQ_ISOFT) ?
+		(xfs_qcnt_t) newlim->d_ino_softlimit :
+			be64_to_cpu(ddq->d_ino_softlimit);
+	if (hard == 0 || hard >= soft) {
+		ddq->d_ino_hardlimit = cpu_to_be64(hard);
+		ddq->d_ino_softlimit = cpu_to_be64(soft);
+		if (id == 0) {
+			q->qi_ihardlimit = hard;
+			q->qi_isoftlimit = soft;
+		}
+	} else {
+		xfs_debug(mp, "ihard %Ld < isoft %Ld\n", hard, soft);
+	}
+
+	/*
+	 * Update warnings counter(s) if requested
+	 */
+	if (newlim->d_fieldmask & FS_DQ_BWARNS)
+		ddq->d_bwarns = cpu_to_be16(newlim->d_bwarns);
+	if (newlim->d_fieldmask & FS_DQ_IWARNS)
+		ddq->d_iwarns = cpu_to_be16(newlim->d_iwarns);
+	if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
+		ddq->d_rtbwarns = cpu_to_be16(newlim->d_rtbwarns);
+
+	if (id == 0) {
+		/*
+		 * Timelimits for the super user set the relative time
+		 * the other users can be over quota for this file system.
+		 * If it is zero a default is used.  Ditto for the default
+		 * soft and hard limit values (already done, above), and
+		 * for warnings.
+		 */
+		if (newlim->d_fieldmask & FS_DQ_BTIMER) {
+			q->qi_btimelimit = newlim->d_btimer;
+			ddq->d_btimer = cpu_to_be32(newlim->d_btimer);
+		}
+		if (newlim->d_fieldmask & FS_DQ_ITIMER) {
+			q->qi_itimelimit = newlim->d_itimer;
+			ddq->d_itimer = cpu_to_be32(newlim->d_itimer);
+		}
+		if (newlim->d_fieldmask & FS_DQ_RTBTIMER) {
+			q->qi_rtbtimelimit = newlim->d_rtbtimer;
+			ddq->d_rtbtimer = cpu_to_be32(newlim->d_rtbtimer);
+		}
+		if (newlim->d_fieldmask & FS_DQ_BWARNS)
+			q->qi_bwarnlimit = newlim->d_bwarns;
+		if (newlim->d_fieldmask & FS_DQ_IWARNS)
+			q->qi_iwarnlimit = newlim->d_iwarns;
+		if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
+			q->qi_rtbwarnlimit = newlim->d_rtbwarns;
+	} else {
+		/*
+		 * If the user is now over quota, start the timelimit.
+		 * The user will not be 'warned'.
+		 * Note that we keep the timers ticking, whether enforcement
+		 * is on or off. We don't really want to bother with iterating
+		 * over all ondisk dquots and turning the timers on/off.
+		 */
+		xfs_qm_adjust_dqtimers(mp, ddq);
+	}
+	dqp->dq_flags |= XFS_DQ_DIRTY;
+	xfs_trans_log_dquot(tp, dqp);
+
+	error = xfs_trans_commit(tp, 0);
+	xfs_qm_dqrele(dqp);
+
+ out_unlock:
+	mutex_unlock(&q->qi_quotaofflock);
+	return error;
+}
+
+int
+xfs_qm_scall_getquota(
+	xfs_mount_t	*mp,
+	xfs_dqid_t	id,
+	uint		type,
+	fs_disk_quota_t *out)
+{
+	xfs_dquot_t	*dqp;
+	int		error;
+
+	/*
+	 * Try to get the dquot. We don't want it allocated on disk, so
+	 * we aren't passing the XFS_QMOPT_DOALLOC flag. If it doesn't
+	 * exist, we'll get ENOENT back.
+	 */
+	if ((error = xfs_qm_dqget(mp, NULL, id, type, 0, &dqp))) {
+		return (error);
+	}
+
+	/*
+	 * If everything's NULL, this dquot doesn't quite exist as far as
+	 * our utility programs are concerned.
+	 */
+	if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
+		xfs_qm_dqput(dqp);
+		return XFS_ERROR(ENOENT);
+	}
+	/*
+	 * Convert the disk dquot to the exportable format
+	 */
+	xfs_qm_export_dquot(mp, &dqp->q_core, out);
+	xfs_qm_dqput(dqp);
+	return (error ? XFS_ERROR(EFAULT) : 0);
+}
+
+
+STATIC int
+xfs_qm_log_quotaoff_end(
+	xfs_mount_t		*mp,
+	xfs_qoff_logitem_t	*startqoff,
+	uint			flags)
+{
+	xfs_trans_t		*tp;
+	int			error;
+	xfs_qoff_logitem_t	*qoffi;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF_END);
+
+	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_qoff_logitem_t) * 2,
+				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
+		xfs_trans_cancel(tp, 0);
+		return (error);
+	}
+
+	qoffi = xfs_trans_get_qoff_item(tp, startqoff,
+					flags & XFS_ALL_QUOTA_ACCT);
+	xfs_trans_log_quotaoff_item(tp, qoffi);
+
+	/*
+	 * We have to make sure that the transaction is secure on disk before we
+	 * return and actually stop quota accounting. So, make it synchronous.
+	 * We don't care about quotoff's performance.
+	 */
+	xfs_trans_set_sync(tp);
+	error = xfs_trans_commit(tp, 0);
+	return (error);
+}
+
+
+STATIC int
+xfs_qm_log_quotaoff(
+	xfs_mount_t	       *mp,
+	xfs_qoff_logitem_t     **qoffstartp,
+	uint		       flags)
+{
+	xfs_trans_t	       *tp;
+	int			error;
+	xfs_qoff_logitem_t     *qoffi=NULL;
+	uint			oldsbqflag=0;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
+	if ((error = xfs_trans_reserve(tp, 0,
+				      sizeof(xfs_qoff_logitem_t) * 2 +
+				      mp->m_sb.sb_sectsize + 128,
+				      0,
+				      0,
+				      XFS_DEFAULT_LOG_COUNT))) {
+		goto error0;
+	}
+
+	qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
+	xfs_trans_log_quotaoff_item(tp, qoffi);
+
+	spin_lock(&mp->m_sb_lock);
+	oldsbqflag = mp->m_sb.sb_qflags;
+	mp->m_sb.sb_qflags = (mp->m_qflags & ~(flags)) & XFS_MOUNT_QUOTA_ALL;
+	spin_unlock(&mp->m_sb_lock);
+
+	xfs_mod_sb(tp, XFS_SB_QFLAGS);
+
+	/*
+	 * We have to make sure that the transaction is secure on disk before we
+	 * return and actually stop quota accounting. So, make it synchronous.
+	 * We don't care about quotoff's performance.
+	 */
+	xfs_trans_set_sync(tp);
+	error = xfs_trans_commit(tp, 0);
+
+error0:
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		/*
+		 * No one else is modifying sb_qflags, so this is OK.
+		 * We still hold the quotaofflock.
+		 */
+		spin_lock(&mp->m_sb_lock);
+		mp->m_sb.sb_qflags = oldsbqflag;
+		spin_unlock(&mp->m_sb_lock);
+	}
+	*qoffstartp = qoffi;
+	return (error);
+}
+
+
+/*
+ * Translate an internal style on-disk-dquot to the exportable format.
+ * The main differences are that the counters/limits are all in Basic
+ * Blocks (BBs) instead of the internal FSBs, and all on-disk data has
+ * to be converted to the native endianness.
+ */
+STATIC void
+xfs_qm_export_dquot(
+	xfs_mount_t		*mp,
+	xfs_disk_dquot_t	*src,
+	struct fs_disk_quota	*dst)
+{
+	memset(dst, 0, sizeof(*dst));
+	dst->d_version = FS_DQUOT_VERSION;  /* different from src->d_version */
+	dst->d_flags = xfs_qm_export_qtype_flags(src->d_flags);
+	dst->d_id = be32_to_cpu(src->d_id);
+	dst->d_blk_hardlimit =
+		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_hardlimit));
+	dst->d_blk_softlimit =
+		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_softlimit));
+	dst->d_ino_hardlimit = be64_to_cpu(src->d_ino_hardlimit);
+	dst->d_ino_softlimit = be64_to_cpu(src->d_ino_softlimit);
+	dst->d_bcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_bcount));
+	dst->d_icount = be64_to_cpu(src->d_icount);
+	dst->d_btimer = be32_to_cpu(src->d_btimer);
+	dst->d_itimer = be32_to_cpu(src->d_itimer);
+	dst->d_iwarns = be16_to_cpu(src->d_iwarns);
+	dst->d_bwarns = be16_to_cpu(src->d_bwarns);
+	dst->d_rtb_hardlimit =
+		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_hardlimit));
+	dst->d_rtb_softlimit =
+		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_softlimit));
+	dst->d_rtbcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtbcount));
+	dst->d_rtbtimer = be32_to_cpu(src->d_rtbtimer);
+	dst->d_rtbwarns = be16_to_cpu(src->d_rtbwarns);
+
+	/*
+	 * Internally, we don't reset all the timers when quota enforcement
+	 * gets turned off. No need to confuse the user level code,
+	 * so return zeroes in that case.
+	 */
+	if ((!XFS_IS_UQUOTA_ENFORCED(mp) && src->d_flags == XFS_DQ_USER) ||
+	    (!XFS_IS_OQUOTA_ENFORCED(mp) &&
+			(src->d_flags & (XFS_DQ_PROJ | XFS_DQ_GROUP)))) {
+		dst->d_btimer = 0;
+		dst->d_itimer = 0;
+		dst->d_rtbtimer = 0;
+	}
+
+#ifdef DEBUG
+	if (((XFS_IS_UQUOTA_ENFORCED(mp) && dst->d_flags == FS_USER_QUOTA) ||
+	     (XFS_IS_OQUOTA_ENFORCED(mp) &&
+			(dst->d_flags & (FS_PROJ_QUOTA | FS_GROUP_QUOTA)))) &&
+	    dst->d_id != 0) {
+		if (((int) dst->d_bcount >= (int) dst->d_blk_softlimit) &&
+		    (dst->d_blk_softlimit > 0)) {
+			ASSERT(dst->d_btimer != 0);
+		}
+		if (((int) dst->d_icount >= (int) dst->d_ino_softlimit) &&
+		    (dst->d_ino_softlimit > 0)) {
+			ASSERT(dst->d_itimer != 0);
+		}
+	}
+#endif
+}
+
+STATIC uint
+xfs_qm_export_qtype_flags(
+	uint flags)
+{
+	/*
+	 * Can't be more than one, or none.
+	 */
+	ASSERT((flags & (FS_PROJ_QUOTA | FS_USER_QUOTA)) !=
+		(FS_PROJ_QUOTA | FS_USER_QUOTA));
+	ASSERT((flags & (FS_PROJ_QUOTA | FS_GROUP_QUOTA)) !=
+		(FS_PROJ_QUOTA | FS_GROUP_QUOTA));
+	ASSERT((flags & (FS_USER_QUOTA | FS_GROUP_QUOTA)) !=
+		(FS_USER_QUOTA | FS_GROUP_QUOTA));
+	ASSERT((flags & (FS_PROJ_QUOTA|FS_USER_QUOTA|FS_GROUP_QUOTA)) != 0);
+
+	return (flags & XFS_DQ_USER) ?
+		FS_USER_QUOTA : (flags & XFS_DQ_PROJ) ?
+			FS_PROJ_QUOTA : FS_GROUP_QUOTA;
+}
+
+STATIC uint
+xfs_qm_export_flags(
+	uint flags)
+{
+	uint uflags;
+
+	uflags = 0;
+	if (flags & XFS_UQUOTA_ACCT)
+		uflags |= FS_QUOTA_UDQ_ACCT;
+	if (flags & XFS_PQUOTA_ACCT)
+		uflags |= FS_QUOTA_PDQ_ACCT;
+	if (flags & XFS_GQUOTA_ACCT)
+		uflags |= FS_QUOTA_GDQ_ACCT;
+	if (flags & XFS_UQUOTA_ENFD)
+		uflags |= FS_QUOTA_UDQ_ENFD;
+	if (flags & (XFS_OQUOTA_ENFD)) {
+		uflags |= (flags & XFS_GQUOTA_ACCT) ?
+			FS_QUOTA_GDQ_ENFD : FS_QUOTA_PDQ_ENFD;
+	}
+	return (uflags);
+}
+
+
+STATIC int
+xfs_dqrele_inode(
+	struct xfs_inode	*ip,
+	struct xfs_perag	*pag,
+	int			flags)
+{
+	/* skip quota inodes */
+	if (ip == ip->i_mount->m_quotainfo->qi_uquotaip ||
+	    ip == ip->i_mount->m_quotainfo->qi_gquotaip) {
+		ASSERT(ip->i_udquot == NULL);
+		ASSERT(ip->i_gdquot == NULL);
+		return 0;
+	}
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	if ((flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
+		xfs_qm_dqrele(ip->i_udquot);
+		ip->i_udquot = NULL;
+	}
+	if (flags & (XFS_PQUOTA_ACCT|XFS_GQUOTA_ACCT) && ip->i_gdquot) {
+		xfs_qm_dqrele(ip->i_gdquot);
+		ip->i_gdquot = NULL;
+	}
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return 0;
+}
+
+
+/*
+ * Go thru all the inodes in the file system, releasing their dquots.
+ *
+ * Note that the mount structure gets modified to indicate that quotas are off
+ * AFTER this, in the case of quotaoff.
+ */
+void
+xfs_qm_dqrele_all_inodes(
+	struct xfs_mount *mp,
+	uint		 flags)
+{
+	ASSERT(mp->m_quotainfo);
+	xfs_inode_ag_iterator(mp, xfs_dqrele_inode, flags);
+}
diff --git a/fs/xfs/xfs_quota_priv.h b/fs/xfs/xfs_quota_priv.h
new file mode 100644
index 0000000..94a3d92
--- /dev/null
+++ b/fs/xfs/xfs_quota_priv.h
@@ -0,0 +1,53 @@
+/*
+ * Copyright (c) 2000-2003 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_QUOTA_PRIV_H__
+#define __XFS_QUOTA_PRIV_H__
+
+/*
+ * Number of bmaps that we ask from bmapi when doing a quotacheck.
+ * We make this restriction to keep the memory usage to a minimum.
+ */
+#define XFS_DQITER_MAP_SIZE	10
+
+/*
+ * Hash into a bucket in the dquot hash table, based on <mp, id>.
+ */
+#define XFS_DQ_HASHVAL(mp, id) (((__psunsigned_t)(mp) + \
+				 (__psunsigned_t)(id)) & \
+				(xfs_Gqm->qm_dqhashmask - 1))
+#define XFS_DQ_HASH(mp, id, type)   (type == XFS_DQ_USER ? \
+				     (xfs_Gqm->qm_usr_dqhtable + \
+				      XFS_DQ_HASHVAL(mp, id)) : \
+				     (xfs_Gqm->qm_grp_dqhtable + \
+				      XFS_DQ_HASHVAL(mp, id)))
+#define XFS_IS_DQUOT_UNINITIALIZED(dqp) ( \
+	!dqp->q_core.d_blk_hardlimit && \
+	!dqp->q_core.d_blk_softlimit && \
+	!dqp->q_core.d_rtb_hardlimit && \
+	!dqp->q_core.d_rtb_softlimit && \
+	!dqp->q_core.d_ino_hardlimit && \
+	!dqp->q_core.d_ino_softlimit && \
+	!dqp->q_core.d_bcount && \
+	!dqp->q_core.d_rtbcount && \
+	!dqp->q_core.d_icount)
+
+#define DQFLAGTO_TYPESTR(d)	(((d)->dq_flags & XFS_DQ_USER) ? "USR" : \
+				 (((d)->dq_flags & XFS_DQ_GROUP) ? "GRP" : \
+				 (((d)->dq_flags & XFS_DQ_PROJ) ? "PRJ":"???")))
+
+#endif	/* __XFS_QUOTA_PRIV_H__ */
diff --git a/fs/xfs/xfs_quotaops.c b/fs/xfs/xfs_quotaops.c
new file mode 100644
index 0000000..7e76f53
--- /dev/null
+++ b/fs/xfs/xfs_quotaops.c
@@ -0,0 +1,139 @@
+/*
+ * Copyright (c) 2008, Christoph Hellwig
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_sb.h"
+#include "xfs_inum.h"
+#include "xfs_log.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_quota.h"
+#include "xfs_trans.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_qm.h"
+#include <linux/quota.h>
+
+
+STATIC int
+xfs_quota_type(int type)
+{
+	switch (type) {
+	case USRQUOTA:
+		return XFS_DQ_USER;
+	case GRPQUOTA:
+		return XFS_DQ_GROUP;
+	default:
+		return XFS_DQ_PROJ;
+	}
+}
+
+STATIC int
+xfs_fs_get_xstate(
+	struct super_block	*sb,
+	struct fs_quota_stat	*fqs)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	if (!XFS_IS_QUOTA_RUNNING(mp))
+		return -ENOSYS;
+	return -xfs_qm_scall_getqstat(mp, fqs);
+}
+
+STATIC int
+xfs_fs_set_xstate(
+	struct super_block	*sb,
+	unsigned int		uflags,
+	int			op)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+	unsigned int		flags = 0;
+
+	if (sb->s_flags & MS_RDONLY)
+		return -EROFS;
+	if (op != Q_XQUOTARM && !XFS_IS_QUOTA_RUNNING(mp))
+		return -ENOSYS;
+
+	if (uflags & FS_QUOTA_UDQ_ACCT)
+		flags |= XFS_UQUOTA_ACCT;
+	if (uflags & FS_QUOTA_PDQ_ACCT)
+		flags |= XFS_PQUOTA_ACCT;
+	if (uflags & FS_QUOTA_GDQ_ACCT)
+		flags |= XFS_GQUOTA_ACCT;
+	if (uflags & FS_QUOTA_UDQ_ENFD)
+		flags |= XFS_UQUOTA_ENFD;
+	if (uflags & (FS_QUOTA_PDQ_ENFD|FS_QUOTA_GDQ_ENFD))
+		flags |= XFS_OQUOTA_ENFD;
+
+	switch (op) {
+	case Q_XQUOTAON:
+		return -xfs_qm_scall_quotaon(mp, flags);
+	case Q_XQUOTAOFF:
+		if (!XFS_IS_QUOTA_ON(mp))
+			return -EINVAL;
+		return -xfs_qm_scall_quotaoff(mp, flags);
+	case Q_XQUOTARM:
+		if (XFS_IS_QUOTA_ON(mp))
+			return -EINVAL;
+		return -xfs_qm_scall_trunc_qfiles(mp, flags);
+	}
+
+	return -EINVAL;
+}
+
+STATIC int
+xfs_fs_get_dqblk(
+	struct super_block	*sb,
+	int			type,
+	qid_t			id,
+	struct fs_disk_quota	*fdq)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	if (!XFS_IS_QUOTA_RUNNING(mp))
+		return -ENOSYS;
+	if (!XFS_IS_QUOTA_ON(mp))
+		return -ESRCH;
+
+	return -xfs_qm_scall_getquota(mp, id, xfs_quota_type(type), fdq);
+}
+
+STATIC int
+xfs_fs_set_dqblk(
+	struct super_block	*sb,
+	int			type,
+	qid_t			id,
+	struct fs_disk_quota	*fdq)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	if (sb->s_flags & MS_RDONLY)
+		return -EROFS;
+	if (!XFS_IS_QUOTA_RUNNING(mp))
+		return -ENOSYS;
+	if (!XFS_IS_QUOTA_ON(mp))
+		return -ESRCH;
+
+	return -xfs_qm_scall_setqlim(mp, id, xfs_quota_type(type), fdq);
+}
+
+const struct quotactl_ops xfs_quotactl_operations = {
+	.get_xstate		= xfs_fs_get_xstate,
+	.set_xstate		= xfs_fs_set_xstate,
+	.get_dqblk		= xfs_fs_get_dqblk,
+	.set_dqblk		= xfs_fs_set_dqblk,
+};
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
new file mode 100644
index 0000000..76fdc58
--- /dev/null
+++ b/fs/xfs/xfs_stats.c
@@ -0,0 +1,122 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include <linux/proc_fs.h>
+
+DEFINE_PER_CPU(struct xfsstats, xfsstats);
+
+static int xfs_stat_proc_show(struct seq_file *m, void *v)
+{
+	int		c, i, j, val;
+	__uint64_t	xs_xstrat_bytes = 0;
+	__uint64_t	xs_write_bytes = 0;
+	__uint64_t	xs_read_bytes = 0;
+
+	static const struct xstats_entry {
+		char	*desc;
+		int	endpoint;
+	} xstats[] = {
+		{ "extent_alloc",	XFSSTAT_END_EXTENT_ALLOC	},
+		{ "abt",		XFSSTAT_END_ALLOC_BTREE		},
+		{ "blk_map",		XFSSTAT_END_BLOCK_MAPPING	},
+		{ "bmbt",		XFSSTAT_END_BLOCK_MAP_BTREE	},
+		{ "dir",		XFSSTAT_END_DIRECTORY_OPS	},
+		{ "trans",		XFSSTAT_END_TRANSACTIONS	},
+		{ "ig",			XFSSTAT_END_INODE_OPS		},
+		{ "log",		XFSSTAT_END_LOG_OPS		},
+		{ "push_ail",		XFSSTAT_END_TAIL_PUSHING	},
+		{ "xstrat",		XFSSTAT_END_WRITE_CONVERT	},
+		{ "rw",			XFSSTAT_END_READ_WRITE_OPS	},
+		{ "attr",		XFSSTAT_END_ATTRIBUTE_OPS	},
+		{ "icluster",		XFSSTAT_END_INODE_CLUSTER	},
+		{ "vnodes",		XFSSTAT_END_VNODE_OPS		},
+		{ "buf",		XFSSTAT_END_BUF			},
+		{ "abtb2",		XFSSTAT_END_ABTB_V2		},
+		{ "abtc2",		XFSSTAT_END_ABTC_V2		},
+		{ "bmbt2",		XFSSTAT_END_BMBT_V2		},
+		{ "ibt2",		XFSSTAT_END_IBT_V2		},
+	};
+
+	/* Loop over all stats groups */
+	for (i=j = 0; i < ARRAY_SIZE(xstats); i++) {
+		seq_printf(m, "%s", xstats[i].desc);
+		/* inner loop does each group */
+		while (j < xstats[i].endpoint) {
+			val = 0;
+			/* sum over all cpus */
+			for_each_possible_cpu(c)
+				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
+			seq_printf(m, " %u", val);
+			j++;
+		}
+		seq_putc(m, '\n');
+	}
+	/* extra precision counters */
+	for_each_possible_cpu(i) {
+		xs_xstrat_bytes += per_cpu(xfsstats, i).xs_xstrat_bytes;
+		xs_write_bytes += per_cpu(xfsstats, i).xs_write_bytes;
+		xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
+	}
+
+	seq_printf(m, "xpc %Lu %Lu %Lu\n",
+			xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
+	seq_printf(m, "debug %u\n",
+#if defined(DEBUG)
+		1);
+#else
+		0);
+#endif
+	return 0;
+}
+
+static int xfs_stat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xfs_stat_proc_show, NULL);
+}
+
+static const struct file_operations xfs_stat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xfs_stat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+int
+xfs_init_procfs(void)
+{
+	if (!proc_mkdir("fs/xfs", NULL))
+		goto out;
+
+	if (!proc_create("fs/xfs/stat", 0, NULL,
+			 &xfs_stat_proc_fops))
+		goto out_remove_entry;
+	return 0;
+
+ out_remove_entry:
+	remove_proc_entry("fs/xfs", NULL);
+ out:
+	return -ENOMEM;
+}
+
+void
+xfs_cleanup_procfs(void)
+{
+	remove_proc_entry("fs/xfs/stat", NULL);
+	remove_proc_entry("fs/xfs", NULL);
+}
diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h
new file mode 100644
index 0000000..736854b
--- /dev/null
+++ b/fs/xfs/xfs_stats.h
@@ -0,0 +1,223 @@
+/*
+ * Copyright (c) 2000,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_STATS_H__
+#define __XFS_STATS_H__
+
+
+#if defined(CONFIG_PROC_FS) && !defined(XFS_STATS_OFF)
+
+#include <linux/percpu.h>
+
+/*
+ * XFS global statistics
+ */
+struct xfsstats {
+# define XFSSTAT_END_EXTENT_ALLOC	4
+	__uint32_t		xs_allocx;
+	__uint32_t		xs_allocb;
+	__uint32_t		xs_freex;
+	__uint32_t		xs_freeb;
+# define XFSSTAT_END_ALLOC_BTREE	(XFSSTAT_END_EXTENT_ALLOC+4)
+	__uint32_t		xs_abt_lookup;
+	__uint32_t		xs_abt_compare;
+	__uint32_t		xs_abt_insrec;
+	__uint32_t		xs_abt_delrec;
+# define XFSSTAT_END_BLOCK_MAPPING	(XFSSTAT_END_ALLOC_BTREE+7)
+	__uint32_t		xs_blk_mapr;
+	__uint32_t		xs_blk_mapw;
+	__uint32_t		xs_blk_unmap;
+	__uint32_t		xs_add_exlist;
+	__uint32_t		xs_del_exlist;
+	__uint32_t		xs_look_exlist;
+	__uint32_t		xs_cmp_exlist;
+# define XFSSTAT_END_BLOCK_MAP_BTREE	(XFSSTAT_END_BLOCK_MAPPING+4)
+	__uint32_t		xs_bmbt_lookup;
+	__uint32_t		xs_bmbt_compare;
+	__uint32_t		xs_bmbt_insrec;
+	__uint32_t		xs_bmbt_delrec;
+# define XFSSTAT_END_DIRECTORY_OPS	(XFSSTAT_END_BLOCK_MAP_BTREE+4)
+	__uint32_t		xs_dir_lookup;
+	__uint32_t		xs_dir_create;
+	__uint32_t		xs_dir_remove;
+	__uint32_t		xs_dir_getdents;
+# define XFSSTAT_END_TRANSACTIONS	(XFSSTAT_END_DIRECTORY_OPS+3)
+	__uint32_t		xs_trans_sync;
+	__uint32_t		xs_trans_async;
+	__uint32_t		xs_trans_empty;
+# define XFSSTAT_END_INODE_OPS		(XFSSTAT_END_TRANSACTIONS+7)
+	__uint32_t		xs_ig_attempts;
+	__uint32_t		xs_ig_found;
+	__uint32_t		xs_ig_frecycle;
+	__uint32_t		xs_ig_missed;
+	__uint32_t		xs_ig_dup;
+	__uint32_t		xs_ig_reclaims;
+	__uint32_t		xs_ig_attrchg;
+# define XFSSTAT_END_LOG_OPS		(XFSSTAT_END_INODE_OPS+5)
+	__uint32_t		xs_log_writes;
+	__uint32_t		xs_log_blocks;
+	__uint32_t		xs_log_noiclogs;
+	__uint32_t		xs_log_force;
+	__uint32_t		xs_log_force_sleep;
+# define XFSSTAT_END_TAIL_PUSHING	(XFSSTAT_END_LOG_OPS+10)
+	__uint32_t		xs_try_logspace;
+	__uint32_t		xs_sleep_logspace;
+	__uint32_t		xs_push_ail;
+	__uint32_t		xs_push_ail_success;
+	__uint32_t		xs_push_ail_pushbuf;
+	__uint32_t		xs_push_ail_pinned;
+	__uint32_t		xs_push_ail_locked;
+	__uint32_t		xs_push_ail_flushing;
+	__uint32_t		xs_push_ail_restarts;
+	__uint32_t		xs_push_ail_flush;
+# define XFSSTAT_END_WRITE_CONVERT	(XFSSTAT_END_TAIL_PUSHING+2)
+	__uint32_t		xs_xstrat_quick;
+	__uint32_t		xs_xstrat_split;
+# define XFSSTAT_END_READ_WRITE_OPS	(XFSSTAT_END_WRITE_CONVERT+2)
+	__uint32_t		xs_write_calls;
+	__uint32_t		xs_read_calls;
+# define XFSSTAT_END_ATTRIBUTE_OPS	(XFSSTAT_END_READ_WRITE_OPS+4)
+	__uint32_t		xs_attr_get;
+	__uint32_t		xs_attr_set;
+	__uint32_t		xs_attr_remove;
+	__uint32_t		xs_attr_list;
+# define XFSSTAT_END_INODE_CLUSTER	(XFSSTAT_END_ATTRIBUTE_OPS+3)
+	__uint32_t		xs_iflush_count;
+	__uint32_t		xs_icluster_flushcnt;
+	__uint32_t		xs_icluster_flushinode;
+# define XFSSTAT_END_VNODE_OPS		(XFSSTAT_END_INODE_CLUSTER+8)
+	__uint32_t		vn_active;	/* # vnodes not on free lists */
+	__uint32_t		vn_alloc;	/* # times vn_alloc called */
+	__uint32_t		vn_get;		/* # times vn_get called */
+	__uint32_t		vn_hold;	/* # times vn_hold called */
+	__uint32_t		vn_rele;	/* # times vn_rele called */
+	__uint32_t		vn_reclaim;	/* # times vn_reclaim called */
+	__uint32_t		vn_remove;	/* # times vn_remove called */
+	__uint32_t		vn_free;	/* # times vn_free called */
+#define XFSSTAT_END_BUF			(XFSSTAT_END_VNODE_OPS+9)
+	__uint32_t		xb_get;
+	__uint32_t		xb_create;
+	__uint32_t		xb_get_locked;
+	__uint32_t		xb_get_locked_waited;
+	__uint32_t		xb_busy_locked;
+	__uint32_t		xb_miss_locked;
+	__uint32_t		xb_page_retries;
+	__uint32_t		xb_page_found;
+	__uint32_t		xb_get_read;
+/* Version 2 btree counters */
+#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF+15)
+	__uint32_t		xs_abtb_2_lookup;
+	__uint32_t		xs_abtb_2_compare;
+	__uint32_t		xs_abtb_2_insrec;
+	__uint32_t		xs_abtb_2_delrec;
+	__uint32_t		xs_abtb_2_newroot;
+	__uint32_t		xs_abtb_2_killroot;
+	__uint32_t		xs_abtb_2_increment;
+	__uint32_t		xs_abtb_2_decrement;
+	__uint32_t		xs_abtb_2_lshift;
+	__uint32_t		xs_abtb_2_rshift;
+	__uint32_t		xs_abtb_2_split;
+	__uint32_t		xs_abtb_2_join;
+	__uint32_t		xs_abtb_2_alloc;
+	__uint32_t		xs_abtb_2_free;
+	__uint32_t		xs_abtb_2_moves;
+#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2+15)
+	__uint32_t		xs_abtc_2_lookup;
+	__uint32_t		xs_abtc_2_compare;
+	__uint32_t		xs_abtc_2_insrec;
+	__uint32_t		xs_abtc_2_delrec;
+	__uint32_t		xs_abtc_2_newroot;
+	__uint32_t		xs_abtc_2_killroot;
+	__uint32_t		xs_abtc_2_increment;
+	__uint32_t		xs_abtc_2_decrement;
+	__uint32_t		xs_abtc_2_lshift;
+	__uint32_t		xs_abtc_2_rshift;
+	__uint32_t		xs_abtc_2_split;
+	__uint32_t		xs_abtc_2_join;
+	__uint32_t		xs_abtc_2_alloc;
+	__uint32_t		xs_abtc_2_free;
+	__uint32_t		xs_abtc_2_moves;
+#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2+15)
+	__uint32_t		xs_bmbt_2_lookup;
+	__uint32_t		xs_bmbt_2_compare;
+	__uint32_t		xs_bmbt_2_insrec;
+	__uint32_t		xs_bmbt_2_delrec;
+	__uint32_t		xs_bmbt_2_newroot;
+	__uint32_t		xs_bmbt_2_killroot;
+	__uint32_t		xs_bmbt_2_increment;
+	__uint32_t		xs_bmbt_2_decrement;
+	__uint32_t		xs_bmbt_2_lshift;
+	__uint32_t		xs_bmbt_2_rshift;
+	__uint32_t		xs_bmbt_2_split;
+	__uint32_t		xs_bmbt_2_join;
+	__uint32_t		xs_bmbt_2_alloc;
+	__uint32_t		xs_bmbt_2_free;
+	__uint32_t		xs_bmbt_2_moves;
+#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2+15)
+	__uint32_t		xs_ibt_2_lookup;
+	__uint32_t		xs_ibt_2_compare;
+	__uint32_t		xs_ibt_2_insrec;
+	__uint32_t		xs_ibt_2_delrec;
+	__uint32_t		xs_ibt_2_newroot;
+	__uint32_t		xs_ibt_2_killroot;
+	__uint32_t		xs_ibt_2_increment;
+	__uint32_t		xs_ibt_2_decrement;
+	__uint32_t		xs_ibt_2_lshift;
+	__uint32_t		xs_ibt_2_rshift;
+	__uint32_t		xs_ibt_2_split;
+	__uint32_t		xs_ibt_2_join;
+	__uint32_t		xs_ibt_2_alloc;
+	__uint32_t		xs_ibt_2_free;
+	__uint32_t		xs_ibt_2_moves;
+/* Extra precision counters */
+	__uint64_t		xs_xstrat_bytes;
+	__uint64_t		xs_write_bytes;
+	__uint64_t		xs_read_bytes;
+};
+
+DECLARE_PER_CPU(struct xfsstats, xfsstats);
+
+/*
+ * We don't disable preempt, not too worried about poking the
+ * wrong CPU's stat for now (also aggregated before reporting).
+ */
+#define XFS_STATS_INC(v)	(per_cpu(xfsstats, current_cpu()).v++)
+#define XFS_STATS_DEC(v)	(per_cpu(xfsstats, current_cpu()).v--)
+#define XFS_STATS_ADD(v, inc)	(per_cpu(xfsstats, current_cpu()).v += (inc))
+
+extern int xfs_init_procfs(void);
+extern void xfs_cleanup_procfs(void);
+
+
+#else	/* !CONFIG_PROC_FS */
+
+# define XFS_STATS_INC(count)
+# define XFS_STATS_DEC(count)
+# define XFS_STATS_ADD(count, inc)
+
+static inline int xfs_init_procfs(void)
+{
+	return 0;
+}
+
+static inline void xfs_cleanup_procfs(void)
+{
+}
+
+#endif	/* !CONFIG_PROC_FS */
+
+#endif /* __XFS_STATS_H__ */
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
new file mode 100644
index 0000000..90ccd1c
--- /dev/null
+++ b/fs/xfs/xfs_super.c
@@ -0,0 +1,1723 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "xfs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_itable.h"
+#include "xfs_fsops.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_utils.h"
+#include "xfs_vnodeops.h"
+#include "xfs_log_priv.h"
+#include "xfs_trans_priv.h"
+#include "xfs_filestream.h"
+#include "xfs_da_btree.h"
+#include "xfs_extfree_item.h"
+#include "xfs_mru_cache.h"
+#include "xfs_inode_item.h"
+#include "xfs_sync.h"
+#include "xfs_trace.h"
+
+#include <linux/namei.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/mount.h>
+#include <linux/mempool.h>
+#include <linux/writeback.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
+#include <linux/parser.h>
+
+static const struct super_operations xfs_super_operations;
+static kmem_zone_t *xfs_ioend_zone;
+mempool_t *xfs_ioend_pool;
+
+#define MNTOPT_LOGBUFS	"logbufs"	/* number of XFS log buffers */
+#define MNTOPT_LOGBSIZE	"logbsize"	/* size of XFS log buffers */
+#define MNTOPT_LOGDEV	"logdev"	/* log device */
+#define MNTOPT_RTDEV	"rtdev"		/* realtime I/O device */
+#define MNTOPT_BIOSIZE	"biosize"	/* log2 of preferred buffered io size */
+#define MNTOPT_WSYNC	"wsync"		/* safe-mode nfs compatible mount */
+#define MNTOPT_NOALIGN	"noalign"	/* turn off stripe alignment */
+#define MNTOPT_SWALLOC	"swalloc"	/* turn on stripe width allocation */
+#define MNTOPT_SUNIT	"sunit"		/* data volume stripe unit */
+#define MNTOPT_SWIDTH	"swidth"	/* data volume stripe width */
+#define MNTOPT_NOUUID	"nouuid"	/* ignore filesystem UUID */
+#define MNTOPT_MTPT	"mtpt"		/* filesystem mount point */
+#define MNTOPT_GRPID	"grpid"		/* group-ID from parent directory */
+#define MNTOPT_NOGRPID	"nogrpid"	/* group-ID from current process */
+#define MNTOPT_BSDGROUPS    "bsdgroups"    /* group-ID from parent directory */
+#define MNTOPT_SYSVGROUPS   "sysvgroups"   /* group-ID from current process */
+#define MNTOPT_ALLOCSIZE    "allocsize"    /* preferred allocation size */
+#define MNTOPT_NORECOVERY   "norecovery"   /* don't run XFS recovery */
+#define MNTOPT_BARRIER	"barrier"	/* use writer barriers for log write and
+					 * unwritten extent conversion */
+#define MNTOPT_NOBARRIER "nobarrier"	/* .. disable */
+#define MNTOPT_64BITINODE   "inode64"	/* inodes can be allocated anywhere */
+#define MNTOPT_IKEEP	"ikeep"		/* do not free empty inode clusters */
+#define MNTOPT_NOIKEEP	"noikeep"	/* free empty inode clusters */
+#define MNTOPT_LARGEIO	   "largeio"	/* report large I/O sizes in stat() */
+#define MNTOPT_NOLARGEIO   "nolargeio"	/* do not report large I/O sizes
+					 * in stat(). */
+#define MNTOPT_ATTR2	"attr2"		/* do use attr2 attribute format */
+#define MNTOPT_NOATTR2	"noattr2"	/* do not use attr2 attribute format */
+#define MNTOPT_FILESTREAM  "filestreams" /* use filestreams allocator */
+#define MNTOPT_QUOTA	"quota"		/* disk quotas (user) */
+#define MNTOPT_NOQUOTA	"noquota"	/* no quotas */
+#define MNTOPT_USRQUOTA	"usrquota"	/* user quota enabled */
+#define MNTOPT_GRPQUOTA	"grpquota"	/* group quota enabled */
+#define MNTOPT_PRJQUOTA	"prjquota"	/* project quota enabled */
+#define MNTOPT_UQUOTA	"uquota"	/* user quota (IRIX variant) */
+#define MNTOPT_GQUOTA	"gquota"	/* group quota (IRIX variant) */
+#define MNTOPT_PQUOTA	"pquota"	/* project quota (IRIX variant) */
+#define MNTOPT_UQUOTANOENF "uqnoenforce"/* user quota limit enforcement */
+#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
+#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
+#define MNTOPT_QUOTANOENF  "qnoenforce"	/* same as uqnoenforce */
+#define MNTOPT_DELAYLOG    "delaylog"	/* Delayed logging enabled */
+#define MNTOPT_NODELAYLOG  "nodelaylog"	/* Delayed logging disabled */
+#define MNTOPT_DISCARD	   "discard"	/* Discard unused blocks */
+#define MNTOPT_NODISCARD   "nodiscard"	/* Do not discard unused blocks */
+
+/*
+ * Table driven mount option parser.
+ *
+ * Currently only used for remount, but it will be used for mount
+ * in the future, too.
+ */
+enum {
+	Opt_barrier, Opt_nobarrier, Opt_err
+};
+
+static const match_table_t tokens = {
+	{Opt_barrier, "barrier"},
+	{Opt_nobarrier, "nobarrier"},
+	{Opt_err, NULL}
+};
+
+
+STATIC unsigned long
+suffix_strtoul(char *s, char **endp, unsigned int base)
+{
+	int	last, shift_left_factor = 0;
+	char	*value = s;
+
+	last = strlen(value) - 1;
+	if (value[last] == 'K' || value[last] == 'k') {
+		shift_left_factor = 10;
+		value[last] = '\0';
+	}
+	if (value[last] == 'M' || value[last] == 'm') {
+		shift_left_factor = 20;
+		value[last] = '\0';
+	}
+	if (value[last] == 'G' || value[last] == 'g') {
+		shift_left_factor = 30;
+		value[last] = '\0';
+	}
+
+	return simple_strtoul((const char *)s, endp, base) << shift_left_factor;
+}
+
+/*
+ * This function fills in xfs_mount_t fields based on mount args.
+ * Note: the superblock has _not_ yet been read in.
+ *
+ * Note that this function leaks the various device name allocations on
+ * failure.  The caller takes care of them.
+ */
+STATIC int
+xfs_parseargs(
+	struct xfs_mount	*mp,
+	char			*options)
+{
+	struct super_block	*sb = mp->m_super;
+	char			*this_char, *value, *eov;
+	int			dsunit = 0;
+	int			dswidth = 0;
+	int			iosize = 0;
+	__uint8_t		iosizelog = 0;
+
+	/*
+	 * set up the mount name first so all the errors will refer to the
+	 * correct device.
+	 */
+	mp->m_fsname = kstrndup(sb->s_id, MAXNAMELEN, GFP_KERNEL);
+	if (!mp->m_fsname)
+		return ENOMEM;
+	mp->m_fsname_len = strlen(mp->m_fsname) + 1;
+
+	/*
+	 * Copy binary VFS mount flags we are interested in.
+	 */
+	if (sb->s_flags & MS_RDONLY)
+		mp->m_flags |= XFS_MOUNT_RDONLY;
+	if (sb->s_flags & MS_DIRSYNC)
+		mp->m_flags |= XFS_MOUNT_DIRSYNC;
+	if (sb->s_flags & MS_SYNCHRONOUS)
+		mp->m_flags |= XFS_MOUNT_WSYNC;
+
+	/*
+	 * Set some default flags that could be cleared by the mount option
+	 * parsing.
+	 */
+	mp->m_flags |= XFS_MOUNT_BARRIER;
+	mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
+	mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
+	mp->m_flags |= XFS_MOUNT_DELAYLOG;
+
+	/*
+	 * These can be overridden by the mount option parsing.
+	 */
+	mp->m_logbufs = -1;
+	mp->m_logbsize = -1;
+
+	if (!options)
+		goto done;
+
+	while ((this_char = strsep(&options, ",")) != NULL) {
+		if (!*this_char)
+			continue;
+		if ((value = strchr(this_char, '=')) != NULL)
+			*value++ = 0;
+
+		if (!strcmp(this_char, MNTOPT_LOGBUFS)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			mp->m_logbufs = simple_strtoul(value, &eov, 10);
+		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			mp->m_logbsize = suffix_strtoul(value, &eov, 10);
+		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			mp->m_logname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
+			if (!mp->m_logname)
+				return ENOMEM;
+		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
+			xfs_warn(mp, "%s option not allowed on this system",
+				this_char);
+			return EINVAL;
+		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			mp->m_rtname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
+			if (!mp->m_rtname)
+				return ENOMEM;
+		} else if (!strcmp(this_char, MNTOPT_BIOSIZE)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			iosize = simple_strtoul(value, &eov, 10);
+			iosizelog = ffs(iosize) - 1;
+		} else if (!strcmp(this_char, MNTOPT_ALLOCSIZE)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			iosize = suffix_strtoul(value, &eov, 10);
+			iosizelog = ffs(iosize) - 1;
+		} else if (!strcmp(this_char, MNTOPT_GRPID) ||
+			   !strcmp(this_char, MNTOPT_BSDGROUPS)) {
+			mp->m_flags |= XFS_MOUNT_GRPID;
+		} else if (!strcmp(this_char, MNTOPT_NOGRPID) ||
+			   !strcmp(this_char, MNTOPT_SYSVGROUPS)) {
+			mp->m_flags &= ~XFS_MOUNT_GRPID;
+		} else if (!strcmp(this_char, MNTOPT_WSYNC)) {
+			mp->m_flags |= XFS_MOUNT_WSYNC;
+		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
+			mp->m_flags |= XFS_MOUNT_NORECOVERY;
+		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
+			mp->m_flags |= XFS_MOUNT_NOALIGN;
+		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
+			mp->m_flags |= XFS_MOUNT_SWALLOC;
+		} else if (!strcmp(this_char, MNTOPT_SUNIT)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			dsunit = simple_strtoul(value, &eov, 10);
+		} else if (!strcmp(this_char, MNTOPT_SWIDTH)) {
+			if (!value || !*value) {
+				xfs_warn(mp, "%s option requires an argument",
+					this_char);
+				return EINVAL;
+			}
+			dswidth = simple_strtoul(value, &eov, 10);
+		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
+			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
+#if !XFS_BIG_INUMS
+			xfs_warn(mp, "%s option not allowed on this system",
+				this_char);
+			return EINVAL;
+#endif
+		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
+			mp->m_flags |= XFS_MOUNT_NOUUID;
+		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
+			mp->m_flags |= XFS_MOUNT_BARRIER;
+		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
+			mp->m_flags &= ~XFS_MOUNT_BARRIER;
+		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
+			mp->m_flags |= XFS_MOUNT_IKEEP;
+		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
+			mp->m_flags &= ~XFS_MOUNT_IKEEP;
+		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
+			mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
+		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
+			mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
+		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
+			mp->m_flags |= XFS_MOUNT_ATTR2;
+		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
+			mp->m_flags &= ~XFS_MOUNT_ATTR2;
+			mp->m_flags |= XFS_MOUNT_NOATTR2;
+		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
+			mp->m_flags |= XFS_MOUNT_FILESTREAMS;
+		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
+			mp->m_qflags &= ~(XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
+					  XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
+					  XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
+					  XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD);
+		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
+			   !strcmp(this_char, MNTOPT_UQUOTA) ||
+			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
+			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
+					 XFS_UQUOTA_ENFD);
+		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
+			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
+			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_UQUOTA_ENFD;
+		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
+			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
+			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
+					 XFS_OQUOTA_ENFD);
+		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
+			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
+		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
+			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
+			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
+					 XFS_OQUOTA_ENFD);
+		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
+			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
+		} else if (!strcmp(this_char, MNTOPT_DELAYLOG)) {
+			mp->m_flags |= XFS_MOUNT_DELAYLOG;
+		} else if (!strcmp(this_char, MNTOPT_NODELAYLOG)) {
+			mp->m_flags &= ~XFS_MOUNT_DELAYLOG;
+			xfs_warn(mp,
+	"nodelaylog is deprecated and will be removed in Linux 3.3");
+		} else if (!strcmp(this_char, MNTOPT_DISCARD)) {
+			mp->m_flags |= XFS_MOUNT_DISCARD;
+		} else if (!strcmp(this_char, MNTOPT_NODISCARD)) {
+			mp->m_flags &= ~XFS_MOUNT_DISCARD;
+		} else if (!strcmp(this_char, "ihashsize")) {
+			xfs_warn(mp,
+	"ihashsize no longer used, option is deprecated.");
+		} else if (!strcmp(this_char, "osyncisdsync")) {
+			xfs_warn(mp,
+	"osyncisdsync has no effect, option is deprecated.");
+		} else if (!strcmp(this_char, "osyncisosync")) {
+			xfs_warn(mp,
+	"osyncisosync has no effect, option is deprecated.");
+		} else if (!strcmp(this_char, "irixsgid")) {
+			xfs_warn(mp,
+	"irixsgid is now a sysctl(2) variable, option is deprecated.");
+		} else {
+			xfs_warn(mp, "unknown mount option [%s].", this_char);
+			return EINVAL;
+		}
+	}
+
+	/*
+	 * no recovery flag requires a read-only mount
+	 */
+	if ((mp->m_flags & XFS_MOUNT_NORECOVERY) &&
+	    !(mp->m_flags & XFS_MOUNT_RDONLY)) {
+		xfs_warn(mp, "no-recovery mounts must be read-only.");
+		return EINVAL;
+	}
+
+	if ((mp->m_flags & XFS_MOUNT_NOALIGN) && (dsunit || dswidth)) {
+		xfs_warn(mp,
+	"sunit and swidth options incompatible with the noalign option");
+		return EINVAL;
+	}
+
+	if ((mp->m_flags & XFS_MOUNT_DISCARD) &&
+	    !(mp->m_flags & XFS_MOUNT_DELAYLOG)) {
+		xfs_warn(mp,
+	"the discard option is incompatible with the nodelaylog option");
+		return EINVAL;
+	}
+
+#ifndef CONFIG_XFS_QUOTA
+	if (XFS_IS_QUOTA_RUNNING(mp)) {
+		xfs_warn(mp, "quota support not available in this kernel.");
+		return EINVAL;
+	}
+#endif
+
+	if ((mp->m_qflags & (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE)) &&
+	    (mp->m_qflags & (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE))) {
+		xfs_warn(mp, "cannot mount with both project and group quota");
+		return EINVAL;
+	}
+
+	if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
+		xfs_warn(mp, "sunit and swidth must be specified together");
+		return EINVAL;
+	}
+
+	if (dsunit && (dswidth % dsunit != 0)) {
+		xfs_warn(mp,
+	"stripe width (%d) must be a multiple of the stripe unit (%d)",
+			dswidth, dsunit);
+		return EINVAL;
+	}
+
+done:
+	if (!(mp->m_flags & XFS_MOUNT_NOALIGN)) {
+		/*
+		 * At this point the superblock has not been read
+		 * in, therefore we do not know the block size.
+		 * Before the mount call ends we will convert
+		 * these to FSBs.
+		 */
+		if (dsunit) {
+			mp->m_dalign = dsunit;
+			mp->m_flags |= XFS_MOUNT_RETERR;
+		}
+
+		if (dswidth)
+			mp->m_swidth = dswidth;
+	}
+
+	if (mp->m_logbufs != -1 &&
+	    mp->m_logbufs != 0 &&
+	    (mp->m_logbufs < XLOG_MIN_ICLOGS ||
+	     mp->m_logbufs > XLOG_MAX_ICLOGS)) {
+		xfs_warn(mp, "invalid logbufs value: %d [not %d-%d]",
+			mp->m_logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
+		return XFS_ERROR(EINVAL);
+	}
+	if (mp->m_logbsize != -1 &&
+	    mp->m_logbsize !=  0 &&
+	    (mp->m_logbsize < XLOG_MIN_RECORD_BSIZE ||
+	     mp->m_logbsize > XLOG_MAX_RECORD_BSIZE ||
+	     !is_power_of_2(mp->m_logbsize))) {
+		xfs_warn(mp,
+			"invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
+			mp->m_logbsize);
+		return XFS_ERROR(EINVAL);
+	}
+
+	if (iosizelog) {
+		if (iosizelog > XFS_MAX_IO_LOG ||
+		    iosizelog < XFS_MIN_IO_LOG) {
+			xfs_warn(mp, "invalid log iosize: %d [not %d-%d]",
+				iosizelog, XFS_MIN_IO_LOG,
+				XFS_MAX_IO_LOG);
+			return XFS_ERROR(EINVAL);
+		}
+
+		mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
+		mp->m_readio_log = iosizelog;
+		mp->m_writeio_log = iosizelog;
+	}
+
+	return 0;
+}
+
+struct proc_xfs_info {
+	int	flag;
+	char	*str;
+};
+
+STATIC int
+xfs_showargs(
+	struct xfs_mount	*mp,
+	struct seq_file		*m)
+{
+	static struct proc_xfs_info xfs_info_set[] = {
+		/* the few simple ones we can get from the mount struct */
+		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
+		{ XFS_MOUNT_WSYNC,		"," MNTOPT_WSYNC },
+		{ XFS_MOUNT_NOALIGN,		"," MNTOPT_NOALIGN },
+		{ XFS_MOUNT_SWALLOC,		"," MNTOPT_SWALLOC },
+		{ XFS_MOUNT_NOUUID,		"," MNTOPT_NOUUID },
+		{ XFS_MOUNT_NORECOVERY,		"," MNTOPT_NORECOVERY },
+		{ XFS_MOUNT_ATTR2,		"," MNTOPT_ATTR2 },
+		{ XFS_MOUNT_FILESTREAMS,	"," MNTOPT_FILESTREAM },
+		{ XFS_MOUNT_GRPID,		"," MNTOPT_GRPID },
+		{ XFS_MOUNT_DELAYLOG,		"," MNTOPT_DELAYLOG },
+		{ XFS_MOUNT_DISCARD,		"," MNTOPT_DISCARD },
+		{ 0, NULL }
+	};
+	static struct proc_xfs_info xfs_info_unset[] = {
+		/* the few simple ones we can get from the mount struct */
+		{ XFS_MOUNT_COMPAT_IOSIZE,	"," MNTOPT_LARGEIO },
+		{ XFS_MOUNT_BARRIER,		"," MNTOPT_NOBARRIER },
+		{ XFS_MOUNT_SMALL_INUMS,	"," MNTOPT_64BITINODE },
+		{ 0, NULL }
+	};
+	struct proc_xfs_info	*xfs_infop;
+
+	for (xfs_infop = xfs_info_set; xfs_infop->flag; xfs_infop++) {
+		if (mp->m_flags & xfs_infop->flag)
+			seq_puts(m, xfs_infop->str);
+	}
+	for (xfs_infop = xfs_info_unset; xfs_infop->flag; xfs_infop++) {
+		if (!(mp->m_flags & xfs_infop->flag))
+			seq_puts(m, xfs_infop->str);
+	}
+
+	if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE)
+		seq_printf(m, "," MNTOPT_ALLOCSIZE "=%dk",
+				(int)(1 << mp->m_writeio_log) >> 10);
+
+	if (mp->m_logbufs > 0)
+		seq_printf(m, "," MNTOPT_LOGBUFS "=%d", mp->m_logbufs);
+	if (mp->m_logbsize > 0)
+		seq_printf(m, "," MNTOPT_LOGBSIZE "=%dk", mp->m_logbsize >> 10);
+
+	if (mp->m_logname)
+		seq_show_option(m, MNTOPT_LOGDEV, mp->m_logname);
+	if (mp->m_rtname)
+		seq_show_option(m, MNTOPT_RTDEV, mp->m_rtname);
+
+	if (mp->m_dalign > 0)
+		seq_printf(m, "," MNTOPT_SUNIT "=%d",
+				(int)XFS_FSB_TO_BB(mp, mp->m_dalign));
+	if (mp->m_swidth > 0)
+		seq_printf(m, "," MNTOPT_SWIDTH "=%d",
+				(int)XFS_FSB_TO_BB(mp, mp->m_swidth));
+
+	if (mp->m_qflags & (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD))
+		seq_puts(m, "," MNTOPT_USRQUOTA);
+	else if (mp->m_qflags & XFS_UQUOTA_ACCT)
+		seq_puts(m, "," MNTOPT_UQUOTANOENF);
+
+	/* Either project or group quotas can be active, not both */
+
+	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_PRJQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_PQUOTANOENF);
+	} else if (mp->m_qflags & XFS_GQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_GRPQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	}
+
+	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
+		seq_puts(m, "," MNTOPT_NOQUOTA);
+
+	return 0;
+}
+__uint64_t
+xfs_max_file_offset(
+	unsigned int		blockshift)
+{
+	unsigned int		pagefactor = 1;
+	unsigned int		bitshift = BITS_PER_LONG - 1;
+
+	/* Figure out maximum filesize, on Linux this can depend on
+	 * the filesystem blocksize (on 32 bit platforms).
+	 * __block_write_begin does this in an [unsigned] long...
+	 *      page->index << (PAGE_CACHE_SHIFT - bbits)
+	 * So, for page sized blocks (4K on 32 bit platforms),
+	 * this wraps at around 8Tb (hence MAX_LFS_FILESIZE which is
+	 *      (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
+	 * but for smaller blocksizes it is less (bbits = log2 bsize).
+	 * Note1: get_block_t takes a long (implicit cast from above)
+	 * Note2: The Large Block Device (LBD and HAVE_SECTOR_T) patch
+	 * can optionally convert the [unsigned] long from above into
+	 * an [unsigned] long long.
+	 */
+
+#if BITS_PER_LONG == 32
+# if defined(CONFIG_LBDAF)
+	ASSERT(sizeof(sector_t) == 8);
+	pagefactor = PAGE_CACHE_SIZE;
+	bitshift = BITS_PER_LONG;
+# else
+	pagefactor = PAGE_CACHE_SIZE >> (PAGE_CACHE_SHIFT - blockshift);
+# endif
+#endif
+
+	return (((__uint64_t)pagefactor) << bitshift) - 1;
+}
+
+STATIC int
+xfs_blkdev_get(
+	xfs_mount_t		*mp,
+	const char		*name,
+	struct block_device	**bdevp)
+{
+	int			error = 0;
+
+	*bdevp = blkdev_get_by_path(name, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
+				    mp);
+	if (IS_ERR(*bdevp)) {
+		error = PTR_ERR(*bdevp);
+		xfs_warn(mp, "Invalid device [%s], error=%d\n", name, error);
+	}
+
+	return -error;
+}
+
+STATIC void
+xfs_blkdev_put(
+	struct block_device	*bdev)
+{
+	if (bdev)
+		blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
+}
+
+void
+xfs_blkdev_issue_flush(
+	xfs_buftarg_t		*buftarg)
+{
+	blkdev_issue_flush(buftarg->bt_bdev, GFP_KERNEL, NULL);
+}
+
+STATIC void
+xfs_close_devices(
+	struct xfs_mount	*mp)
+{
+	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
+		struct block_device *logdev = mp->m_logdev_targp->bt_bdev;
+		xfs_free_buftarg(mp, mp->m_logdev_targp);
+		xfs_blkdev_put(logdev);
+	}
+	if (mp->m_rtdev_targp) {
+		struct block_device *rtdev = mp->m_rtdev_targp->bt_bdev;
+		xfs_free_buftarg(mp, mp->m_rtdev_targp);
+		xfs_blkdev_put(rtdev);
+	}
+	xfs_free_buftarg(mp, mp->m_ddev_targp);
+}
+
+/*
+ * The file system configurations are:
+ *	(1) device (partition) with data and internal log
+ *	(2) logical volume with data and log subvolumes.
+ *	(3) logical volume with data, log, and realtime subvolumes.
+ *
+ * We only have to handle opening the log and realtime volumes here if
+ * they are present.  The data subvolume has already been opened by
+ * get_sb_bdev() and is stored in sb->s_bdev.
+ */
+STATIC int
+xfs_open_devices(
+	struct xfs_mount	*mp)
+{
+	struct block_device	*ddev = mp->m_super->s_bdev;
+	struct block_device	*logdev = NULL, *rtdev = NULL;
+	int			error;
+
+	/*
+	 * Open real time and log devices - order is important.
+	 */
+	if (mp->m_logname) {
+		error = xfs_blkdev_get(mp, mp->m_logname, &logdev);
+		if (error)
+			goto out;
+	}
+
+	if (mp->m_rtname) {
+		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev);
+		if (error)
+			goto out_close_logdev;
+
+		if (rtdev == ddev || rtdev == logdev) {
+			xfs_warn(mp,
+	"Cannot mount filesystem with identical rtdev and ddev/logdev.");
+			error = EINVAL;
+			goto out_close_rtdev;
+		}
+	}
+
+	/*
+	 * Setup xfs_mount buffer target pointers
+	 */
+	error = ENOMEM;
+	mp->m_ddev_targp = xfs_alloc_buftarg(mp, ddev, 0, mp->m_fsname);
+	if (!mp->m_ddev_targp)
+		goto out_close_rtdev;
+
+	if (rtdev) {
+		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev, 1,
+							mp->m_fsname);
+		if (!mp->m_rtdev_targp)
+			goto out_free_ddev_targ;
+	}
+
+	if (logdev && logdev != ddev) {
+		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev, 1,
+							mp->m_fsname);
+		if (!mp->m_logdev_targp)
+			goto out_free_rtdev_targ;
+	} else {
+		mp->m_logdev_targp = mp->m_ddev_targp;
+	}
+
+	return 0;
+
+ out_free_rtdev_targ:
+	if (mp->m_rtdev_targp)
+		xfs_free_buftarg(mp, mp->m_rtdev_targp);
+ out_free_ddev_targ:
+	xfs_free_buftarg(mp, mp->m_ddev_targp);
+ out_close_rtdev:
+	if (rtdev)
+		xfs_blkdev_put(rtdev);
+ out_close_logdev:
+	if (logdev && logdev != ddev)
+		xfs_blkdev_put(logdev);
+ out:
+	return error;
+}
+
+/*
+ * Setup xfs_mount buffer target pointers based on superblock
+ */
+STATIC int
+xfs_setup_devices(
+	struct xfs_mount	*mp)
+{
+	int			error;
+
+	error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_blocksize,
+				    mp->m_sb.sb_sectsize);
+	if (error)
+		return error;
+
+	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
+		unsigned int	log_sector_size = BBSIZE;
+
+		if (xfs_sb_version_hassector(&mp->m_sb))
+			log_sector_size = mp->m_sb.sb_logsectsize;
+		error = xfs_setsize_buftarg(mp->m_logdev_targp,
+					    mp->m_sb.sb_blocksize,
+					    log_sector_size);
+		if (error)
+			return error;
+	}
+	if (mp->m_rtdev_targp) {
+		error = xfs_setsize_buftarg(mp->m_rtdev_targp,
+					    mp->m_sb.sb_blocksize,
+					    mp->m_sb.sb_sectsize);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Catch misguided souls that try to use this interface on XFS */
+STATIC struct inode *
+xfs_fs_alloc_inode(
+	struct super_block	*sb)
+{
+	BUG();
+	return NULL;
+}
+
+/*
+ * Now that the generic code is guaranteed not to be accessing
+ * the linux inode, we can reclaim the inode.
+ */
+STATIC void
+xfs_fs_destroy_inode(
+	struct inode		*inode)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+
+	trace_xfs_destroy_inode(ip);
+
+	XFS_STATS_INC(vn_reclaim);
+
+	/* bad inode, get out here ASAP */
+	if (is_bad_inode(inode))
+		goto out_reclaim;
+
+	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
+
+	/*
+	 * We should never get here with one of the reclaim flags already set.
+	 */
+	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
+	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIM));
+
+	/*
+	 * We always use background reclaim here because even if the
+	 * inode is clean, it still may be under IO and hence we have
+	 * to take the flush lock. The background reclaim path handles
+	 * this more efficiently than we can here, so simply let background
+	 * reclaim tear down all inodes.
+	 */
+out_reclaim:
+	xfs_inode_set_reclaim_tag(ip);
+}
+
+/*
+ * Slab object creation initialisation for the XFS inode.
+ * This covers only the idempotent fields in the XFS inode;
+ * all other fields need to be initialised on allocation
+ * from the slab. This avoids the need to repeatedly initialise
+ * fields in the xfs inode that left in the initialise state
+ * when freeing the inode.
+ */
+STATIC void
+xfs_fs_inode_init_once(
+	void			*inode)
+{
+	struct xfs_inode	*ip = inode;
+
+	memset(ip, 0, sizeof(struct xfs_inode));
+
+	/* vfs inode */
+	inode_init_once(VFS_I(ip));
+
+	/* xfs inode */
+	atomic_set(&ip->i_pincount, 0);
+	spin_lock_init(&ip->i_flags_lock);
+	init_waitqueue_head(&ip->i_ipin_wait);
+	/*
+	 * Because we want to use a counting completion, complete
+	 * the flush completion once to allow a single access to
+	 * the flush completion without blocking.
+	 */
+	init_completion(&ip->i_flush);
+	complete(&ip->i_flush);
+
+	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
+		     "xfsino", ip->i_ino);
+}
+
+/*
+ * Dirty the XFS inode when mark_inode_dirty_sync() is called so that
+ * we catch unlogged VFS level updates to the inode.
+ *
+ * We need the barrier() to maintain correct ordering between unlogged
+ * updates and the transaction commit code that clears the i_update_core
+ * field. This requires all updates to be completed before marking the
+ * inode dirty.
+ */
+STATIC void
+xfs_fs_dirty_inode(
+	struct inode	*inode,
+	int		flags)
+{
+	barrier();
+	XFS_I(inode)->i_update_core = 1;
+}
+
+STATIC int
+xfs_fs_write_inode(
+	struct inode		*inode,
+	struct writeback_control *wbc)
+{
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	int			error = EAGAIN;
+
+	trace_xfs_write_inode(ip);
+
+	if (XFS_FORCED_SHUTDOWN(mp))
+		return -XFS_ERROR(EIO);
+
+	if (wbc->sync_mode == WB_SYNC_ALL || wbc->for_kupdate) {
+		/*
+		 * Make sure the inode has made it it into the log.  Instead
+		 * of forcing it all the way to stable storage using a
+		 * synchronous transaction we let the log force inside the
+		 * ->sync_fs call do that for thus, which reduces the number
+		 * of synchronous log forces dramatically.
+		 */
+		error = xfs_log_dirty_inode(ip, NULL, 0);
+		if (error)
+			goto out;
+		return 0;
+	} else {
+		if (!ip->i_update_core)
+			return 0;
+
+		/*
+		 * We make this non-blocking if the inode is contended, return
+		 * EAGAIN to indicate to the caller that they did not succeed.
+		 * This prevents the flush path from blocking on inodes inside
+		 * another operation right now, they get caught later by
+		 * xfs_sync.
+		 */
+		if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED))
+			goto out;
+
+		if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip))
+			goto out_unlock;
+
+		/*
+		 * Now we have the flush lock and the inode is not pinned, we
+		 * can check if the inode is really clean as we know that
+		 * there are no pending transaction completions, it is not
+		 * waiting on the delayed write queue and there is no IO in
+		 * progress.
+		 */
+		if (xfs_inode_clean(ip)) {
+			xfs_ifunlock(ip);
+			error = 0;
+			goto out_unlock;
+		}
+		error = xfs_iflush(ip, SYNC_TRYLOCK);
+	}
+
+ out_unlock:
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+ out:
+	/*
+	 * if we failed to write out the inode then mark
+	 * it dirty again so we'll try again later.
+	 */
+	if (error)
+		xfs_mark_inode_dirty_sync(ip);
+	return -error;
+}
+
+STATIC void
+xfs_fs_evict_inode(
+	struct inode		*inode)
+{
+	xfs_inode_t		*ip = XFS_I(inode);
+
+	trace_xfs_evict_inode(ip);
+
+	truncate_inode_pages(&inode->i_data, 0);
+	end_writeback(inode);
+	XFS_STATS_INC(vn_rele);
+	XFS_STATS_INC(vn_remove);
+	XFS_STATS_DEC(vn_active);
+
+	/*
+	 * The iolock is used by the file system to coordinate reads,
+	 * writes, and block truncates.  Up to this point the lock
+	 * protected concurrent accesses by users of the inode.  But
+	 * from here forward we're doing some final processing of the
+	 * inode because we're done with it, and although we reuse the
+	 * iolock for protection it is really a distinct lock class
+	 * (in the lockdep sense) from before.  To keep lockdep happy
+	 * (and basically indicate what we are doing), we explicitly
+	 * re-init the iolock here.
+	 */
+	ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));
+	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
+	lockdep_set_class_and_name(&ip->i_iolock.mr_lock,
+			&xfs_iolock_reclaimable, "xfs_iolock_reclaimable");
+
+	xfs_inactive(ip);
+}
+
+STATIC void
+xfs_free_fsname(
+	struct xfs_mount	*mp)
+{
+	kfree(mp->m_fsname);
+	kfree(mp->m_rtname);
+	kfree(mp->m_logname);
+}
+
+STATIC void
+xfs_fs_put_super(
+	struct super_block	*sb)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	xfs_syncd_stop(mp);
+
+	/*
+	 * Blow away any referenced inode in the filestreams cache.
+	 * This can and will cause log traffic as inodes go inactive
+	 * here.
+	 */
+	xfs_filestream_unmount(mp);
+
+	xfs_flush_buftarg(mp->m_ddev_targp, 1);
+
+	xfs_unmountfs(mp);
+	xfs_freesb(mp);
+	xfs_icsb_destroy_counters(mp);
+	xfs_close_devices(mp);
+	xfs_free_fsname(mp);
+	kfree(mp);
+}
+
+STATIC int
+xfs_fs_sync_fs(
+	struct super_block	*sb,
+	int			wait)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+	int			error;
+
+	/*
+	 * Not much we can do for the first async pass.  Writing out the
+	 * superblock would be counter-productive as we are going to redirty
+	 * when writing out other data and metadata (and writing out a single
+	 * block is quite fast anyway).
+	 *
+	 * Try to asynchronously kick off quota syncing at least.
+	 */
+	if (!wait) {
+		xfs_qm_sync(mp, SYNC_TRYLOCK);
+		return 0;
+	}
+
+	error = xfs_quiesce_data(mp);
+	if (error)
+		return -error;
+
+	if (laptop_mode) {
+		/*
+		 * The disk must be active because we're syncing.
+		 * We schedule xfssyncd now (now that the disk is
+		 * active) instead of later (when it might not be).
+		 */
+		flush_delayed_work_sync(&mp->m_sync_work);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_fs_statfs(
+	struct dentry		*dentry,
+	struct kstatfs		*statp)
+{
+	struct xfs_mount	*mp = XFS_M(dentry->d_sb);
+	xfs_sb_t		*sbp = &mp->m_sb;
+	struct xfs_inode	*ip = XFS_I(dentry->d_inode);
+	__uint64_t		fakeinos, id;
+	xfs_extlen_t		lsize;
+	__int64_t		ffree;
+
+	statp->f_type = XFS_SB_MAGIC;
+	statp->f_namelen = MAXNAMELEN - 1;
+
+	id = huge_encode_dev(mp->m_ddev_targp->bt_dev);
+	statp->f_fsid.val[0] = (u32)id;
+	statp->f_fsid.val[1] = (u32)(id >> 32);
+
+	xfs_icsb_sync_counters(mp, XFS_ICSB_LAZY_COUNT);
+
+	spin_lock(&mp->m_sb_lock);
+	statp->f_bsize = sbp->sb_blocksize;
+	lsize = sbp->sb_logstart ? sbp->sb_logblocks : 0;
+	statp->f_blocks = sbp->sb_dblocks - lsize;
+	statp->f_bfree = statp->f_bavail =
+				sbp->sb_fdblocks - XFS_ALLOC_SET_ASIDE(mp);
+	fakeinos = statp->f_bfree << sbp->sb_inopblog;
+	statp->f_files =
+	    MIN(sbp->sb_icount + fakeinos, (__uint64_t)XFS_MAXINUMBER);
+	if (mp->m_maxicount)
+		statp->f_files = min_t(typeof(statp->f_files),
+					statp->f_files,
+					mp->m_maxicount);
+
+	/* make sure statp->f_ffree does not underflow */
+	ffree = statp->f_files - (sbp->sb_icount - sbp->sb_ifree);
+	statp->f_ffree = max_t(__int64_t, ffree, 0);
+
+	spin_unlock(&mp->m_sb_lock);
+
+	if ((ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) ||
+	    ((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))) ==
+			      (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
+		xfs_qm_statvfs(ip, statp);
+	return 0;
+}
+
+STATIC void
+xfs_save_resvblks(struct xfs_mount *mp)
+{
+	__uint64_t resblks = 0;
+
+	mp->m_resblks_save = mp->m_resblks;
+	xfs_reserve_blocks(mp, &resblks, NULL);
+}
+
+STATIC void
+xfs_restore_resvblks(struct xfs_mount *mp)
+{
+	__uint64_t resblks;
+
+	if (mp->m_resblks_save) {
+		resblks = mp->m_resblks_save;
+		mp->m_resblks_save = 0;
+	} else
+		resblks = xfs_default_resblks(mp);
+
+	xfs_reserve_blocks(mp, &resblks, NULL);
+}
+
+STATIC int
+xfs_fs_remount(
+	struct super_block	*sb,
+	int			*flags,
+	char			*options)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+	substring_t		args[MAX_OPT_ARGS];
+	char			*p;
+	int			error;
+
+	while ((p = strsep(&options, ",")) != NULL) {
+		int token;
+
+		if (!*p)
+			continue;
+
+		token = match_token(p, tokens, args);
+		switch (token) {
+		case Opt_barrier:
+			mp->m_flags |= XFS_MOUNT_BARRIER;
+			break;
+		case Opt_nobarrier:
+			mp->m_flags &= ~XFS_MOUNT_BARRIER;
+			break;
+		default:
+			/*
+			 * Logically we would return an error here to prevent
+			 * users from believing they might have changed
+			 * mount options using remount which can't be changed.
+			 *
+			 * But unfortunately mount(8) adds all options from
+			 * mtab and fstab to the mount arguments in some cases
+			 * so we can't blindly reject options, but have to
+			 * check for each specified option if it actually
+			 * differs from the currently set option and only
+			 * reject it if that's the case.
+			 *
+			 * Until that is implemented we return success for
+			 * every remount request, and silently ignore all
+			 * options that we can't actually change.
+			 */
+#if 0
+			xfs_info(mp,
+		"mount option \"%s\" not supported for remount\n", p);
+			return -EINVAL;
+#else
+			break;
+#endif
+		}
+	}
+
+	/* ro -> rw */
+	if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) {
+		mp->m_flags &= ~XFS_MOUNT_RDONLY;
+
+		/*
+		 * If this is the first remount to writeable state we
+		 * might have some superblock changes to update.
+		 */
+		if (mp->m_update_flags) {
+			error = xfs_mount_log_sb(mp, mp->m_update_flags);
+			if (error) {
+				xfs_warn(mp, "failed to write sb changes");
+				return error;
+			}
+			mp->m_update_flags = 0;
+		}
+
+		/*
+		 * Fill out the reserve pool if it is empty. Use the stashed
+		 * value if it is non-zero, otherwise go with the default.
+		 */
+		xfs_restore_resvblks(mp);
+	}
+
+	/* rw -> ro */
+	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) {
+		/*
+		 * After we have synced the data but before we sync the
+		 * metadata, we need to free up the reserve block pool so that
+		 * the used block count in the superblock on disk is correct at
+		 * the end of the remount. Stash the current reserve pool size
+		 * so that if we get remounted rw, we can return it to the same
+		 * size.
+		 */
+
+		xfs_quiesce_data(mp);
+		xfs_save_resvblks(mp);
+		xfs_quiesce_attr(mp);
+		mp->m_flags |= XFS_MOUNT_RDONLY;
+	}
+
+	return 0;
+}
+
+/*
+ * Second stage of a freeze. The data is already frozen so we only
+ * need to take care of the metadata. Once that's done write a dummy
+ * record to dirty the log in case of a crash while frozen.
+ */
+STATIC int
+xfs_fs_freeze(
+	struct super_block	*sb)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	xfs_save_resvblks(mp);
+	xfs_quiesce_attr(mp);
+	return -xfs_fs_log_dummy(mp);
+}
+
+STATIC int
+xfs_fs_unfreeze(
+	struct super_block	*sb)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+
+	xfs_restore_resvblks(mp);
+	return 0;
+}
+
+STATIC int
+xfs_fs_show_options(
+	struct seq_file		*m,
+	struct vfsmount		*mnt)
+{
+	return -xfs_showargs(XFS_M(mnt->mnt_sb), m);
+}
+
+/*
+ * This function fills in xfs_mount_t fields based on mount args.
+ * Note: the superblock _has_ now been read in.
+ */
+STATIC int
+xfs_finish_flags(
+	struct xfs_mount	*mp)
+{
+	int			ronly = (mp->m_flags & XFS_MOUNT_RDONLY);
+
+	/* Fail a mount where the logbuf is smaller than the log stripe */
+	if (xfs_sb_version_haslogv2(&mp->m_sb)) {
+		if (mp->m_logbsize <= 0 &&
+		    mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE) {
+			mp->m_logbsize = mp->m_sb.sb_logsunit;
+		} else if (mp->m_logbsize > 0 &&
+			   mp->m_logbsize < mp->m_sb.sb_logsunit) {
+			xfs_warn(mp,
+		"logbuf size must be greater than or equal to log stripe size");
+			return XFS_ERROR(EINVAL);
+		}
+	} else {
+		/* Fail a mount if the logbuf is larger than 32K */
+		if (mp->m_logbsize > XLOG_BIG_RECORD_BSIZE) {
+			xfs_warn(mp,
+		"logbuf size for version 1 logs must be 16K or 32K");
+			return XFS_ERROR(EINVAL);
+		}
+	}
+
+	/*
+	 * mkfs'ed attr2 will turn on attr2 mount unless explicitly
+	 * told by noattr2 to turn it off
+	 */
+	if (xfs_sb_version_hasattr2(&mp->m_sb) &&
+	    !(mp->m_flags & XFS_MOUNT_NOATTR2))
+		mp->m_flags |= XFS_MOUNT_ATTR2;
+
+	/*
+	 * prohibit r/w mounts of read-only filesystems
+	 */
+	if ((mp->m_sb.sb_flags & XFS_SBF_READONLY) && !ronly) {
+		xfs_warn(mp,
+			"cannot mount a read-only filesystem as read-write");
+		return XFS_ERROR(EROFS);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_fs_fill_super(
+	struct super_block	*sb,
+	void			*data,
+	int			silent)
+{
+	struct inode		*root;
+	struct xfs_mount	*mp = NULL;
+	int			flags = 0, error = ENOMEM;
+
+	mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL);
+	if (!mp)
+		goto out;
+
+	spin_lock_init(&mp->m_sb_lock);
+	mutex_init(&mp->m_growlock);
+	atomic_set(&mp->m_active_trans, 0);
+
+	mp->m_super = sb;
+	sb->s_fs_info = mp;
+
+	error = xfs_parseargs(mp, (char *)data);
+	if (error)
+		goto out_free_fsname;
+
+	sb_min_blocksize(sb, BBSIZE);
+	sb->s_xattr = xfs_xattr_handlers;
+	sb->s_export_op = &xfs_export_operations;
+#ifdef CONFIG_XFS_QUOTA
+	sb->s_qcop = &xfs_quotactl_operations;
+#endif
+	sb->s_op = &xfs_super_operations;
+
+	if (silent)
+		flags |= XFS_MFSI_QUIET;
+
+	error = xfs_open_devices(mp);
+	if (error)
+		goto out_free_fsname;
+
+	error = xfs_icsb_init_counters(mp);
+	if (error)
+		goto out_close_devices;
+
+	error = xfs_readsb(mp, flags);
+	if (error)
+		goto out_destroy_counters;
+
+	error = xfs_finish_flags(mp);
+	if (error)
+		goto out_free_sb;
+
+	error = xfs_setup_devices(mp);
+	if (error)
+		goto out_free_sb;
+
+	error = xfs_filestream_mount(mp);
+	if (error)
+		goto out_free_sb;
+
+	/*
+	 * we must configure the block size in the superblock before we run the
+	 * full mount process as the mount process can lookup and cache inodes.
+	 * For the same reason we must also initialise the syncd and register
+	 * the inode cache shrinker so that inodes can be reclaimed during
+	 * operations like a quotacheck that iterate all inodes in the
+	 * filesystem.
+	 */
+	sb->s_magic = XFS_SB_MAGIC;
+	sb->s_blocksize = mp->m_sb.sb_blocksize;
+	sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1;
+	sb->s_maxbytes = xfs_max_file_offset(sb->s_blocksize_bits);
+	sb->s_time_gran = 1;
+	set_posix_acl_flag(sb);
+
+	error = xfs_mountfs(mp);
+	if (error)
+		goto out_filestream_unmount;
+
+	error = xfs_syncd_init(mp);
+	if (error)
+		goto out_unmount;
+
+	root = igrab(VFS_I(mp->m_rootip));
+	if (!root) {
+		error = ENOENT;
+		goto out_syncd_stop;
+	}
+	if (is_bad_inode(root)) {
+		error = EINVAL;
+		goto out_syncd_stop;
+	}
+	sb->s_root = d_alloc_root(root);
+	if (!sb->s_root) {
+		error = ENOMEM;
+		goto out_iput;
+	}
+
+	return 0;
+
+ out_filestream_unmount:
+	xfs_filestream_unmount(mp);
+ out_free_sb:
+	xfs_freesb(mp);
+ out_destroy_counters:
+	xfs_icsb_destroy_counters(mp);
+ out_close_devices:
+	xfs_close_devices(mp);
+ out_free_fsname:
+	xfs_free_fsname(mp);
+	kfree(mp);
+ out:
+	return -error;
+
+ out_iput:
+	iput(root);
+ out_syncd_stop:
+	xfs_syncd_stop(mp);
+ out_unmount:
+	/*
+	 * Blow away any referenced inode in the filestreams cache.
+	 * This can and will cause log traffic as inodes go inactive
+	 * here.
+	 */
+	xfs_filestream_unmount(mp);
+
+	xfs_flush_buftarg(mp->m_ddev_targp, 1);
+
+	xfs_unmountfs(mp);
+	goto out_free_sb;
+}
+
+STATIC struct dentry *
+xfs_fs_mount(
+	struct file_system_type	*fs_type,
+	int			flags,
+	const char		*dev_name,
+	void			*data)
+{
+	return mount_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super);
+}
+
+static int
+xfs_fs_nr_cached_objects(
+	struct super_block	*sb)
+{
+	return xfs_reclaim_inodes_count(XFS_M(sb));
+}
+
+static void
+xfs_fs_free_cached_objects(
+	struct super_block	*sb,
+	int			nr_to_scan)
+{
+	xfs_reclaim_inodes_nr(XFS_M(sb), nr_to_scan);
+}
+
+static const struct super_operations xfs_super_operations = {
+	.alloc_inode		= xfs_fs_alloc_inode,
+	.destroy_inode		= xfs_fs_destroy_inode,
+	.dirty_inode		= xfs_fs_dirty_inode,
+	.write_inode		= xfs_fs_write_inode,
+	.evict_inode		= xfs_fs_evict_inode,
+	.put_super		= xfs_fs_put_super,
+	.sync_fs		= xfs_fs_sync_fs,
+	.freeze_fs		= xfs_fs_freeze,
+	.unfreeze_fs		= xfs_fs_unfreeze,
+	.statfs			= xfs_fs_statfs,
+	.remount_fs		= xfs_fs_remount,
+	.show_options		= xfs_fs_show_options,
+	.nr_cached_objects	= xfs_fs_nr_cached_objects,
+	.free_cached_objects	= xfs_fs_free_cached_objects,
+};
+
+static struct file_system_type xfs_fs_type = {
+	.owner			= THIS_MODULE,
+	.name			= "xfs",
+	.mount			= xfs_fs_mount,
+	.kill_sb		= kill_block_super,
+	.fs_flags		= FS_REQUIRES_DEV,
+};
+
+STATIC int __init
+xfs_init_zones(void)
+{
+
+	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
+	if (!xfs_ioend_zone)
+		goto out;
+
+	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
+						  xfs_ioend_zone);
+	if (!xfs_ioend_pool)
+		goto out_destroy_ioend_zone;
+
+	xfs_log_ticket_zone = kmem_zone_init(sizeof(xlog_ticket_t),
+						"xfs_log_ticket");
+	if (!xfs_log_ticket_zone)
+		goto out_destroy_ioend_pool;
+
+	xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t),
+						"xfs_bmap_free_item");
+	if (!xfs_bmap_free_item_zone)
+		goto out_destroy_log_ticket_zone;
+
+	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
+						"xfs_btree_cur");
+	if (!xfs_btree_cur_zone)
+		goto out_destroy_bmap_free_item_zone;
+
+	xfs_da_state_zone = kmem_zone_init(sizeof(xfs_da_state_t),
+						"xfs_da_state");
+	if (!xfs_da_state_zone)
+		goto out_destroy_btree_cur_zone;
+
+	xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
+	if (!xfs_dabuf_zone)
+		goto out_destroy_da_state_zone;
+
+	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
+	if (!xfs_ifork_zone)
+		goto out_destroy_dabuf_zone;
+
+	xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
+	if (!xfs_trans_zone)
+		goto out_destroy_ifork_zone;
+
+	xfs_log_item_desc_zone =
+		kmem_zone_init(sizeof(struct xfs_log_item_desc),
+			       "xfs_log_item_desc");
+	if (!xfs_log_item_desc_zone)
+		goto out_destroy_trans_zone;
+
+	/*
+	 * The size of the zone allocated buf log item is the maximum
+	 * size possible under XFS.  This wastes a little bit of memory,
+	 * but it is much faster.
+	 */
+	xfs_buf_item_zone = kmem_zone_init((sizeof(xfs_buf_log_item_t) +
+				(((XFS_MAX_BLOCKSIZE / XFS_BLF_CHUNK) /
+				  NBWORD) * sizeof(int))), "xfs_buf_item");
+	if (!xfs_buf_item_zone)
+		goto out_destroy_log_item_desc_zone;
+
+	xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
+			((XFS_EFD_MAX_FAST_EXTENTS - 1) *
+				 sizeof(xfs_extent_t))), "xfs_efd_item");
+	if (!xfs_efd_zone)
+		goto out_destroy_buf_item_zone;
+
+	xfs_efi_zone = kmem_zone_init((sizeof(xfs_efi_log_item_t) +
+			((XFS_EFI_MAX_FAST_EXTENTS - 1) *
+				sizeof(xfs_extent_t))), "xfs_efi_item");
+	if (!xfs_efi_zone)
+		goto out_destroy_efd_zone;
+
+	xfs_inode_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
+			KM_ZONE_HWALIGN | KM_ZONE_RECLAIM | KM_ZONE_SPREAD,
+			xfs_fs_inode_init_once);
+	if (!xfs_inode_zone)
+		goto out_destroy_efi_zone;
+
+	xfs_ili_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
+					KM_ZONE_SPREAD, NULL);
+	if (!xfs_ili_zone)
+		goto out_destroy_inode_zone;
+
+	return 0;
+
+ out_destroy_inode_zone:
+	kmem_zone_destroy(xfs_inode_zone);
+ out_destroy_efi_zone:
+	kmem_zone_destroy(xfs_efi_zone);
+ out_destroy_efd_zone:
+	kmem_zone_destroy(xfs_efd_zone);
+ out_destroy_buf_item_zone:
+	kmem_zone_destroy(xfs_buf_item_zone);
+ out_destroy_log_item_desc_zone:
+	kmem_zone_destroy(xfs_log_item_desc_zone);
+ out_destroy_trans_zone:
+	kmem_zone_destroy(xfs_trans_zone);
+ out_destroy_ifork_zone:
+	kmem_zone_destroy(xfs_ifork_zone);
+ out_destroy_dabuf_zone:
+	kmem_zone_destroy(xfs_dabuf_zone);
+ out_destroy_da_state_zone:
+	kmem_zone_destroy(xfs_da_state_zone);
+ out_destroy_btree_cur_zone:
+	kmem_zone_destroy(xfs_btree_cur_zone);
+ out_destroy_bmap_free_item_zone:
+	kmem_zone_destroy(xfs_bmap_free_item_zone);
+ out_destroy_log_ticket_zone:
+	kmem_zone_destroy(xfs_log_ticket_zone);
+ out_destroy_ioend_pool:
+	mempool_destroy(xfs_ioend_pool);
+ out_destroy_ioend_zone:
+	kmem_zone_destroy(xfs_ioend_zone);
+ out:
+	return -ENOMEM;
+}
+
+STATIC void
+xfs_destroy_zones(void)
+{
+	kmem_zone_destroy(xfs_ili_zone);
+	kmem_zone_destroy(xfs_inode_zone);
+	kmem_zone_destroy(xfs_efi_zone);
+	kmem_zone_destroy(xfs_efd_zone);
+	kmem_zone_destroy(xfs_buf_item_zone);
+	kmem_zone_destroy(xfs_log_item_desc_zone);
+	kmem_zone_destroy(xfs_trans_zone);
+	kmem_zone_destroy(xfs_ifork_zone);
+	kmem_zone_destroy(xfs_dabuf_zone);
+	kmem_zone_destroy(xfs_da_state_zone);
+	kmem_zone_destroy(xfs_btree_cur_zone);
+	kmem_zone_destroy(xfs_bmap_free_item_zone);
+	kmem_zone_destroy(xfs_log_ticket_zone);
+	mempool_destroy(xfs_ioend_pool);
+	kmem_zone_destroy(xfs_ioend_zone);
+
+}
+
+STATIC int __init
+xfs_init_workqueues(void)
+{
+	/*
+	 * max_active is set to 8 to give enough concurency to allow
+	 * multiple work operations on each CPU to run. This allows multiple
+	 * filesystems to be running sync work concurrently, and scales with
+	 * the number of CPUs in the system.
+	 */
+	xfs_syncd_wq = alloc_workqueue("xfssyncd", WQ_CPU_INTENSIVE, 8);
+	if (!xfs_syncd_wq)
+		return -ENOMEM;
+	return 0;
+}
+
+STATIC void
+xfs_destroy_workqueues(void)
+{
+	destroy_workqueue(xfs_syncd_wq);
+}
+
+STATIC int __init
+init_xfs_fs(void)
+{
+	int			error;
+
+	printk(KERN_INFO XFS_VERSION_STRING " with "
+			 XFS_BUILD_OPTIONS " enabled\n");
+
+	xfs_dir_startup();
+
+	error = xfs_init_zones();
+	if (error)
+		goto out;
+
+	error = xfs_init_workqueues();
+	if (error)
+		goto out_destroy_zones;
+
+	error = xfs_mru_cache_init();
+	if (error)
+		goto out_destroy_wq;
+
+	error = xfs_filestream_init();
+	if (error)
+		goto out_mru_cache_uninit;
+
+	error = xfs_buf_init();
+	if (error)
+		goto out_filestream_uninit;
+
+	error = xfs_init_procfs();
+	if (error)
+		goto out_buf_terminate;
+
+	error = xfs_sysctl_register();
+	if (error)
+		goto out_cleanup_procfs;
+
+	vfs_initquota();
+
+	error = register_filesystem(&xfs_fs_type);
+	if (error)
+		goto out_sysctl_unregister;
+	return 0;
+
+ out_sysctl_unregister:
+	xfs_sysctl_unregister();
+ out_cleanup_procfs:
+	xfs_cleanup_procfs();
+ out_buf_terminate:
+	xfs_buf_terminate();
+ out_filestream_uninit:
+	xfs_filestream_uninit();
+ out_mru_cache_uninit:
+	xfs_mru_cache_uninit();
+ out_destroy_wq:
+	xfs_destroy_workqueues();
+ out_destroy_zones:
+	xfs_destroy_zones();
+ out:
+	return error;
+}
+
+STATIC void __exit
+exit_xfs_fs(void)
+{
+	vfs_exitquota();
+	unregister_filesystem(&xfs_fs_type);
+	xfs_sysctl_unregister();
+	xfs_cleanup_procfs();
+	xfs_buf_terminate();
+	xfs_filestream_uninit();
+	xfs_mru_cache_uninit();
+	xfs_destroy_workqueues();
+	xfs_destroy_zones();
+}
+
+module_init(init_xfs_fs);
+module_exit(exit_xfs_fs);
+
+MODULE_AUTHOR("Silicon Graphics, Inc.");
+MODULE_DESCRIPTION(XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled");
+MODULE_LICENSE("GPL");
diff --git a/fs/xfs/xfs_super.h b/fs/xfs/xfs_super.h
new file mode 100644
index 0000000..50a3266
--- /dev/null
+++ b/fs/xfs/xfs_super.h
@@ -0,0 +1,87 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SUPER_H__
+#define __XFS_SUPER_H__
+
+#include <linux/exportfs.h>
+
+#ifdef CONFIG_XFS_QUOTA
+extern void xfs_qm_init(void);
+extern void xfs_qm_exit(void);
+# define vfs_initquota()	xfs_qm_init()
+# define vfs_exitquota()	xfs_qm_exit()
+#else
+# define vfs_initquota()	do { } while (0)
+# define vfs_exitquota()	do { } while (0)
+#endif
+
+#ifdef CONFIG_XFS_POSIX_ACL
+# define XFS_ACL_STRING		"ACLs, "
+# define set_posix_acl_flag(sb)	((sb)->s_flags |= MS_POSIXACL)
+#else
+# define XFS_ACL_STRING
+# define set_posix_acl_flag(sb)	do { } while (0)
+#endif
+
+#define XFS_SECURITY_STRING	"security attributes, "
+
+#ifdef CONFIG_XFS_RT
+# define XFS_REALTIME_STRING	"realtime, "
+#else
+# define XFS_REALTIME_STRING
+#endif
+
+#if XFS_BIG_BLKNOS
+# if XFS_BIG_INUMS
+#  define XFS_BIGFS_STRING	"large block/inode numbers, "
+# else
+#  define XFS_BIGFS_STRING	"large block numbers, "
+# endif
+#else
+# define XFS_BIGFS_STRING
+#endif
+
+#ifdef DEBUG
+# define XFS_DBG_STRING		"debug"
+#else
+# define XFS_DBG_STRING		"no debug"
+#endif
+
+#define XFS_VERSION_STRING	"SGI XFS"
+#define XFS_BUILD_OPTIONS	XFS_ACL_STRING \
+				XFS_SECURITY_STRING \
+				XFS_REALTIME_STRING \
+				XFS_BIGFS_STRING \
+				XFS_DBG_STRING /* DBG must be last */
+
+struct xfs_inode;
+struct xfs_mount;
+struct xfs_buftarg;
+struct block_device;
+
+extern __uint64_t xfs_max_file_offset(unsigned int);
+
+extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
+
+extern const struct export_operations xfs_export_operations;
+extern const struct xattr_handler *xfs_xattr_handlers[];
+extern const struct quotactl_ops xfs_quotactl_operations;
+
+#define XFS_M(sb)		((struct xfs_mount *)((sb)->s_fs_info))
+
+#endif	/* __XFS_SUPER_H__ */
diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c
new file mode 100644
index 0000000..f0994aedc
--- /dev/null
+++ b/fs/xfs/xfs_sync.c
@@ -0,0 +1,1110 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_types.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_trans_priv.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_dinode.h"
+#include "xfs_error.h"
+#include "xfs_filestream.h"
+#include "xfs_vnodeops.h"
+#include "xfs_inode_item.h"
+#include "xfs_quota.h"
+#include "xfs_trace.h"
+#include "xfs_fsops.h"
+
+#include <linux/kthread.h>
+#include <linux/freezer.h>
+
+struct workqueue_struct	*xfs_syncd_wq;	/* sync workqueue */
+
+/*
+ * The inode lookup is done in batches to keep the amount of lock traffic and
+ * radix tree lookups to a minimum. The batch size is a trade off between
+ * lookup reduction and stack usage. This is in the reclaim path, so we can't
+ * be too greedy.
+ */
+#define XFS_LOOKUP_BATCH	32
+
+STATIC int
+xfs_inode_ag_walk_grab(
+	struct xfs_inode	*ip)
+{
+	struct inode		*inode = VFS_I(ip);
+
+	ASSERT(rcu_read_lock_held());
+
+	/*
+	 * check for stale RCU freed inode
+	 *
+	 * If the inode has been reallocated, it doesn't matter if it's not in
+	 * the AG we are walking - we are walking for writeback, so if it
+	 * passes all the "valid inode" checks and is dirty, then we'll write
+	 * it back anyway.  If it has been reallocated and still being
+	 * initialised, the XFS_INEW check below will catch it.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	if (!ip->i_ino)
+		goto out_unlock_noent;
+
+	/* avoid new or reclaimable inodes. Leave for reclaim code to flush */
+	if (__xfs_iflags_test(ip, XFS_INEW | XFS_IRECLAIMABLE | XFS_IRECLAIM))
+		goto out_unlock_noent;
+	spin_unlock(&ip->i_flags_lock);
+
+	/* nothing to sync during shutdown */
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return EFSCORRUPTED;
+
+	/* If we can't grab the inode, it must on it's way to reclaim. */
+	if (!igrab(inode))
+		return ENOENT;
+
+	if (is_bad_inode(inode)) {
+		IRELE(ip);
+		return ENOENT;
+	}
+
+	/* inode is valid */
+	return 0;
+
+out_unlock_noent:
+	spin_unlock(&ip->i_flags_lock);
+	return ENOENT;
+}
+
+STATIC int
+xfs_inode_ag_walk(
+	struct xfs_mount	*mp,
+	struct xfs_perag	*pag,
+	int			(*execute)(struct xfs_inode *ip,
+					   struct xfs_perag *pag, int flags),
+	int			flags)
+{
+	uint32_t		first_index;
+	int			last_error = 0;
+	int			skipped;
+	int			done;
+	int			nr_found;
+
+restart:
+	done = 0;
+	skipped = 0;
+	first_index = 0;
+	nr_found = 0;
+	do {
+		struct xfs_inode *batch[XFS_LOOKUP_BATCH];
+		int		error = 0;
+		int		i;
+
+		rcu_read_lock();
+		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
+					(void **)batch, first_index,
+					XFS_LOOKUP_BATCH);
+		if (!nr_found) {
+			rcu_read_unlock();
+			break;
+		}
+
+		/*
+		 * Grab the inodes before we drop the lock. if we found
+		 * nothing, nr == 0 and the loop will be skipped.
+		 */
+		for (i = 0; i < nr_found; i++) {
+			struct xfs_inode *ip = batch[i];
+
+			if (done || xfs_inode_ag_walk_grab(ip))
+				batch[i] = NULL;
+
+			/*
+			 * Update the index for the next lookup. Catch
+			 * overflows into the next AG range which can occur if
+			 * we have inodes in the last block of the AG and we
+			 * are currently pointing to the last inode.
+			 *
+			 * Because we may see inodes that are from the wrong AG
+			 * due to RCU freeing and reallocation, only update the
+			 * index if it lies in this AG. It was a race that lead
+			 * us to see this inode, so another lookup from the
+			 * same index will not find it again.
+			 */
+			if (XFS_INO_TO_AGNO(mp, ip->i_ino) != pag->pag_agno)
+				continue;
+			first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+			if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
+				done = 1;
+		}
+
+		/* unlock now we've grabbed the inodes. */
+		rcu_read_unlock();
+
+		for (i = 0; i < nr_found; i++) {
+			if (!batch[i])
+				continue;
+			error = execute(batch[i], pag, flags);
+			IRELE(batch[i]);
+			if (error == EAGAIN) {
+				skipped++;
+				continue;
+			}
+			if (error && last_error != EFSCORRUPTED)
+				last_error = error;
+		}
+
+		/* bail out if the filesystem is corrupted.  */
+		if (error == EFSCORRUPTED)
+			break;
+
+		cond_resched();
+
+	} while (nr_found && !done);
+
+	if (skipped) {
+		delay(1);
+		goto restart;
+	}
+	return last_error;
+}
+
+int
+xfs_inode_ag_iterator(
+	struct xfs_mount	*mp,
+	int			(*execute)(struct xfs_inode *ip,
+					   struct xfs_perag *pag, int flags),
+	int			flags)
+{
+	struct xfs_perag	*pag;
+	int			error = 0;
+	int			last_error = 0;
+	xfs_agnumber_t		ag;
+
+	ag = 0;
+	while ((pag = xfs_perag_get(mp, ag))) {
+		ag = pag->pag_agno + 1;
+		error = xfs_inode_ag_walk(mp, pag, execute, flags);
+		xfs_perag_put(pag);
+		if (error) {
+			last_error = error;
+			if (error == EFSCORRUPTED)
+				break;
+		}
+	}
+	return XFS_ERROR(last_error);
+}
+
+STATIC int
+xfs_sync_inode_data(
+	struct xfs_inode	*ip,
+	struct xfs_perag	*pag,
+	int			flags)
+{
+	struct inode		*inode = VFS_I(ip);
+	struct address_space *mapping = inode->i_mapping;
+	int			error = 0;
+
+	if (!mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
+		return 0;
+
+	if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED)) {
+		if (flags & SYNC_TRYLOCK)
+			return 0;
+		xfs_ilock(ip, XFS_IOLOCK_SHARED);
+	}
+
+	error = xfs_flush_pages(ip, 0, -1, (flags & SYNC_WAIT) ?
+				0 : XBF_ASYNC, FI_NONE);
+	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
+	return error;
+}
+
+STATIC int
+xfs_sync_inode_attr(
+	struct xfs_inode	*ip,
+	struct xfs_perag	*pag,
+	int			flags)
+{
+	int			error = 0;
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	if (xfs_inode_clean(ip))
+		goto out_unlock;
+	if (!xfs_iflock_nowait(ip)) {
+		if (!(flags & SYNC_WAIT))
+			goto out_unlock;
+		xfs_iflock(ip);
+	}
+
+	if (xfs_inode_clean(ip)) {
+		xfs_ifunlock(ip);
+		goto out_unlock;
+	}
+
+	error = xfs_iflush(ip, flags);
+
+	/*
+	 * We don't want to try again on non-blocking flushes that can't run
+	 * again immediately. If an inode really must be written, then that's
+	 * what the SYNC_WAIT flag is for.
+	 */
+	if (error == EAGAIN) {
+		ASSERT(!(flags & SYNC_WAIT));
+		error = 0;
+	}
+
+ out_unlock:
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	return error;
+}
+
+/*
+ * Write out pagecache data for the whole filesystem.
+ */
+STATIC int
+xfs_sync_data(
+	struct xfs_mount	*mp,
+	int			flags)
+{
+	int			error;
+
+	ASSERT((flags & ~(SYNC_TRYLOCK|SYNC_WAIT)) == 0);
+
+	error = xfs_inode_ag_iterator(mp, xfs_sync_inode_data, flags);
+	if (error)
+		return XFS_ERROR(error);
+
+	xfs_log_force(mp, (flags & SYNC_WAIT) ? XFS_LOG_SYNC : 0);
+	return 0;
+}
+
+/*
+ * Write out inode metadata (attributes) for the whole filesystem.
+ */
+STATIC int
+xfs_sync_attr(
+	struct xfs_mount	*mp,
+	int			flags)
+{
+	ASSERT((flags & ~SYNC_WAIT) == 0);
+
+	return xfs_inode_ag_iterator(mp, xfs_sync_inode_attr, flags);
+}
+
+STATIC int
+xfs_sync_fsdata(
+	struct xfs_mount	*mp)
+{
+	struct xfs_buf		*bp;
+	int			error;
+
+	/*
+	 * If the buffer is pinned then push on the log so we won't get stuck
+	 * waiting in the write for someone, maybe ourselves, to flush the log.
+	 *
+	 * Even though we just pushed the log above, we did not have the
+	 * superblock buffer locked at that point so it can become pinned in
+	 * between there and here.
+	 */
+	bp = xfs_getsb(mp, 0);
+	if (xfs_buf_ispinned(bp))
+		xfs_log_force(mp, 0);
+	error = xfs_bwrite(bp);
+	xfs_buf_relse(bp);
+	return error;
+}
+
+int
+xfs_log_dirty_inode(
+	struct xfs_inode	*ip,
+	struct xfs_perag	*pag,
+	int			flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_trans	*tp;
+	int			error;
+
+	if (!ip->i_update_core)
+		return 0;
+
+	tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS);
+	error = xfs_trans_reserve(tp, 0, XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0);
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	return xfs_trans_commit(tp, 0);
+}
+
+/*
+ * When remounting a filesystem read-only or freezing the filesystem, we have
+ * two phases to execute. This first phase is syncing the data before we
+ * quiesce the filesystem, and the second is flushing all the inodes out after
+ * we've waited for all the transactions created by the first phase to
+ * complete. The second phase ensures that the inodes are written to their
+ * location on disk rather than just existing in transactions in the log. This
+ * means after a quiesce there is no log replay required to write the inodes to
+ * disk (this is the main difference between a sync and a quiesce).
+ */
+/*
+ * First stage of freeze - no writers will make progress now we are here,
+ * so we flush delwri and delalloc buffers here, then wait for all I/O to
+ * complete.  Data is frozen at that point. Metadata is not frozen,
+ * transactions can still occur here so don't bother flushing the buftarg
+ * because it'll just get dirty again.
+ */
+int
+xfs_quiesce_data(
+	struct xfs_mount	*mp)
+{
+	int			error, error2 = 0;
+
+	/*
+	 * Log all pending size and timestamp updates.  The vfs writeback
+	 * code is supposed to do this, but due to its overagressive
+	 * livelock detection it will skip inodes where appending writes
+	 * were written out in the first non-blocking sync phase if their
+	 * completion took long enough that it happened after taking the
+	 * timestamp for the cut-off in the blocking phase.
+	 */
+	xfs_inode_ag_iterator(mp, xfs_log_dirty_inode, 0);
+
+	xfs_qm_sync(mp, SYNC_TRYLOCK);
+	xfs_qm_sync(mp, SYNC_WAIT);
+
+	/* force out the newly dirtied log buffers */
+	xfs_log_force(mp, XFS_LOG_SYNC);
+
+	/* write superblock and hoover up shutdown errors */
+	error = xfs_sync_fsdata(mp);
+
+	/* make sure all delwri buffers are written out */
+	xfs_flush_buftarg(mp->m_ddev_targp, 1);
+
+	/* mark the log as covered if needed */
+	if (xfs_log_need_covered(mp))
+		error2 = xfs_fs_log_dummy(mp);
+
+	/* flush data-only devices */
+	if (mp->m_rtdev_targp)
+		xfs_flush_buftarg(mp->m_rtdev_targp, 1);
+
+	return error ? error : error2;
+}
+
+STATIC void
+xfs_quiesce_fs(
+	struct xfs_mount	*mp)
+{
+	int	count = 0, pincount;
+
+	xfs_reclaim_inodes(mp, 0);
+	xfs_flush_buftarg(mp->m_ddev_targp, 0);
+
+	/*
+	 * This loop must run at least twice.  The first instance of the loop
+	 * will flush most meta data but that will generate more meta data
+	 * (typically directory updates).  Which then must be flushed and
+	 * logged before we can write the unmount record. We also so sync
+	 * reclaim of inodes to catch any that the above delwri flush skipped.
+	 */
+	do {
+		xfs_reclaim_inodes(mp, SYNC_WAIT);
+		xfs_sync_attr(mp, SYNC_WAIT);
+		pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
+		if (!pincount) {
+			delay(50);
+			count++;
+		}
+	} while (count < 2);
+}
+
+/*
+ * Second stage of a quiesce. The data is already synced, now we have to take
+ * care of the metadata. New transactions are already blocked, so we need to
+ * wait for any remaining transactions to drain out before proceeding.
+ */
+void
+xfs_quiesce_attr(
+	struct xfs_mount	*mp)
+{
+	int	error = 0;
+
+	/* wait for all modifications to complete */
+	while (atomic_read(&mp->m_active_trans) > 0)
+		delay(100);
+
+	/* flush inodes and push all remaining buffers out to disk */
+	xfs_quiesce_fs(mp);
+
+	/*
+	 * Just warn here till VFS can correctly support
+	 * read-only remount without racing.
+	 */
+	WARN_ON(atomic_read(&mp->m_active_trans) != 0);
+
+	/* Push the superblock and write an unmount record */
+	error = xfs_log_sbcount(mp);
+	if (error)
+		xfs_warn(mp, "xfs_attr_quiesce: failed to log sb changes. "
+				"Frozen image may not be consistent.");
+	xfs_log_unmount_write(mp);
+	xfs_unmountfs_writesb(mp);
+}
+
+static void
+xfs_syncd_queue_sync(
+	struct xfs_mount        *mp)
+{
+	queue_delayed_work(xfs_syncd_wq, &mp->m_sync_work,
+				msecs_to_jiffies(xfs_syncd_centisecs * 10));
+}
+
+/*
+ * Every sync period we need to unpin all items, reclaim inodes and sync
+ * disk quotas.  We might need to cover the log to indicate that the
+ * filesystem is idle and not frozen.
+ */
+STATIC void
+xfs_sync_worker(
+	struct work_struct *work)
+{
+	struct xfs_mount *mp = container_of(to_delayed_work(work),
+					struct xfs_mount, m_sync_work);
+	int		error;
+
+	if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {
+		/* dgc: errors ignored here */
+		if (mp->m_super->s_frozen == SB_UNFROZEN &&
+		    xfs_log_need_covered(mp))
+			error = xfs_fs_log_dummy(mp);
+		else
+			xfs_log_force(mp, 0);
+		error = xfs_qm_sync(mp, SYNC_TRYLOCK);
+
+		/* start pushing all the metadata that is currently dirty */
+		xfs_ail_push_all(mp->m_ail);
+	}
+
+	/* queue us up again */
+	xfs_syncd_queue_sync(mp);
+}
+
+/*
+ * Queue a new inode reclaim pass if there are reclaimable inodes and there
+ * isn't a reclaim pass already in progress. By default it runs every 5s based
+ * on the xfs syncd work default of 30s. Perhaps this should have it's own
+ * tunable, but that can be done if this method proves to be ineffective or too
+ * aggressive.
+ */
+static void
+xfs_syncd_queue_reclaim(
+	struct xfs_mount        *mp)
+{
+
+	/*
+	 * We can have inodes enter reclaim after we've shut down the syncd
+	 * workqueue during unmount, so don't allow reclaim work to be queued
+	 * during unmount.
+	 */
+	if (!(mp->m_super->s_flags & MS_ACTIVE))
+		return;
+
+	rcu_read_lock();
+	if (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_RECLAIM_TAG)) {
+		queue_delayed_work(xfs_syncd_wq, &mp->m_reclaim_work,
+			msecs_to_jiffies(xfs_syncd_centisecs / 6 * 10));
+	}
+	rcu_read_unlock();
+}
+
+/*
+ * This is a fast pass over the inode cache to try to get reclaim moving on as
+ * many inodes as possible in a short period of time. It kicks itself every few
+ * seconds, as well as being kicked by the inode cache shrinker when memory
+ * goes low. It scans as quickly as possible avoiding locked inodes or those
+ * already being flushed, and once done schedules a future pass.
+ */
+STATIC void
+xfs_reclaim_worker(
+	struct work_struct *work)
+{
+	struct xfs_mount *mp = container_of(to_delayed_work(work),
+					struct xfs_mount, m_reclaim_work);
+
+	xfs_reclaim_inodes(mp, SYNC_TRYLOCK);
+	xfs_syncd_queue_reclaim(mp);
+}
+
+/*
+ * Flush delayed allocate data, attempting to free up reserved space
+ * from existing allocations.  At this point a new allocation attempt
+ * has failed with ENOSPC and we are in the process of scratching our
+ * heads, looking about for more room.
+ *
+ * Queue a new data flush if there isn't one already in progress and
+ * wait for completion of the flush. This means that we only ever have one
+ * inode flush in progress no matter how many ENOSPC events are occurring and
+ * so will prevent the system from bogging down due to every concurrent
+ * ENOSPC event scanning all the active inodes in the system for writeback.
+ */
+void
+xfs_flush_inodes(
+	struct xfs_inode	*ip)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+
+	queue_work(xfs_syncd_wq, &mp->m_flush_work);
+	flush_work_sync(&mp->m_flush_work);
+}
+
+STATIC void
+xfs_flush_worker(
+	struct work_struct *work)
+{
+	struct xfs_mount *mp = container_of(work,
+					struct xfs_mount, m_flush_work);
+
+	xfs_sync_data(mp, SYNC_TRYLOCK);
+	xfs_sync_data(mp, SYNC_TRYLOCK | SYNC_WAIT);
+}
+
+int
+xfs_syncd_init(
+	struct xfs_mount	*mp)
+{
+	INIT_WORK(&mp->m_flush_work, xfs_flush_worker);
+	INIT_DELAYED_WORK(&mp->m_sync_work, xfs_sync_worker);
+	INIT_DELAYED_WORK(&mp->m_reclaim_work, xfs_reclaim_worker);
+
+	xfs_syncd_queue_sync(mp);
+	xfs_syncd_queue_reclaim(mp);
+
+	return 0;
+}
+
+void
+xfs_syncd_stop(
+	struct xfs_mount	*mp)
+{
+	cancel_delayed_work_sync(&mp->m_sync_work);
+	cancel_delayed_work_sync(&mp->m_reclaim_work);
+	cancel_work_sync(&mp->m_flush_work);
+}
+
+void
+__xfs_inode_set_reclaim_tag(
+	struct xfs_perag	*pag,
+	struct xfs_inode	*ip)
+{
+	radix_tree_tag_set(&pag->pag_ici_root,
+			   XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino),
+			   XFS_ICI_RECLAIM_TAG);
+
+	if (!pag->pag_ici_reclaimable) {
+		/* propagate the reclaim tag up into the perag radix tree */
+		spin_lock(&ip->i_mount->m_perag_lock);
+		radix_tree_tag_set(&ip->i_mount->m_perag_tree,
+				XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino),
+				XFS_ICI_RECLAIM_TAG);
+		spin_unlock(&ip->i_mount->m_perag_lock);
+
+		/* schedule periodic background inode reclaim */
+		xfs_syncd_queue_reclaim(ip->i_mount);
+
+		trace_xfs_perag_set_reclaim(ip->i_mount, pag->pag_agno,
+							-1, _RET_IP_);
+	}
+	pag->pag_ici_reclaimable++;
+}
+
+/*
+ * We set the inode flag atomically with the radix tree tag.
+ * Once we get tag lookups on the radix tree, this inode flag
+ * can go away.
+ */
+void
+xfs_inode_set_reclaim_tag(
+	xfs_inode_t	*ip)
+{
+	struct xfs_mount *mp = ip->i_mount;
+	struct xfs_perag *pag;
+
+	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
+	spin_lock(&pag->pag_ici_lock);
+	spin_lock(&ip->i_flags_lock);
+	__xfs_inode_set_reclaim_tag(pag, ip);
+	__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
+	spin_unlock(&ip->i_flags_lock);
+	spin_unlock(&pag->pag_ici_lock);
+	xfs_perag_put(pag);
+}
+
+STATIC void
+__xfs_inode_clear_reclaim(
+	xfs_perag_t	*pag,
+	xfs_inode_t	*ip)
+{
+	pag->pag_ici_reclaimable--;
+	if (!pag->pag_ici_reclaimable) {
+		/* clear the reclaim tag from the perag radix tree */
+		spin_lock(&ip->i_mount->m_perag_lock);
+		radix_tree_tag_clear(&ip->i_mount->m_perag_tree,
+				XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino),
+				XFS_ICI_RECLAIM_TAG);
+		spin_unlock(&ip->i_mount->m_perag_lock);
+		trace_xfs_perag_clear_reclaim(ip->i_mount, pag->pag_agno,
+							-1, _RET_IP_);
+	}
+}
+
+void
+__xfs_inode_clear_reclaim_tag(
+	xfs_mount_t	*mp,
+	xfs_perag_t	*pag,
+	xfs_inode_t	*ip)
+{
+	radix_tree_tag_clear(&pag->pag_ici_root,
+			XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+	__xfs_inode_clear_reclaim(pag, ip);
+}
+
+/*
+ * Grab the inode for reclaim exclusively.
+ * Return 0 if we grabbed it, non-zero otherwise.
+ */
+STATIC int
+xfs_reclaim_inode_grab(
+	struct xfs_inode	*ip,
+	int			flags)
+{
+	ASSERT(rcu_read_lock_held());
+
+	/* quick check for stale RCU freed inode */
+	if (!ip->i_ino)
+		return 1;
+
+	/*
+	 * do some unlocked checks first to avoid unnecessary lock traffic.
+	 * The first is a flush lock check, the second is a already in reclaim
+	 * check. Only do these checks if we are not going to block on locks.
+	 */
+	if ((flags & SYNC_TRYLOCK) &&
+	    (!ip->i_flush.done || __xfs_iflags_test(ip, XFS_IRECLAIM))) {
+		return 1;
+	}
+
+	/*
+	 * The radix tree lock here protects a thread in xfs_iget from racing
+	 * with us starting reclaim on the inode.  Once we have the
+	 * XFS_IRECLAIM flag set it will not touch us.
+	 *
+	 * Due to RCU lookup, we may find inodes that have been freed and only
+	 * have XFS_IRECLAIM set.  Indeed, we may see reallocated inodes that
+	 * aren't candidates for reclaim at all, so we must check the
+	 * XFS_IRECLAIMABLE is set first before proceeding to reclaim.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	if (!__xfs_iflags_test(ip, XFS_IRECLAIMABLE) ||
+	    __xfs_iflags_test(ip, XFS_IRECLAIM)) {
+		/* not a reclaim candidate. */
+		spin_unlock(&ip->i_flags_lock);
+		return 1;
+	}
+	__xfs_iflags_set(ip, XFS_IRECLAIM);
+	spin_unlock(&ip->i_flags_lock);
+	return 0;
+}
+
+/*
+ * Inodes in different states need to be treated differently, and the return
+ * value of xfs_iflush is not sufficient to get this right. The following table
+ * lists the inode states and the reclaim actions necessary for non-blocking
+ * reclaim:
+ *
+ *
+ *	inode state	     iflush ret		required action
+ *      ---------------      ----------         ---------------
+ *	bad			-		reclaim
+ *	shutdown		EIO		unpin and reclaim
+ *	clean, unpinned		0		reclaim
+ *	stale, unpinned		0		reclaim
+ *	clean, pinned(*)	0		requeue
+ *	stale, pinned		EAGAIN		requeue
+ *	dirty, delwri ok	0		requeue
+ *	dirty, delwri blocked	EAGAIN		requeue
+ *	dirty, sync flush	0		reclaim
+ *
+ * (*) dgc: I don't think the clean, pinned state is possible but it gets
+ * handled anyway given the order of checks implemented.
+ *
+ * As can be seen from the table, the return value of xfs_iflush() is not
+ * sufficient to correctly decide the reclaim action here. The checks in
+ * xfs_iflush() might look like duplicates, but they are not.
+ *
+ * Also, because we get the flush lock first, we know that any inode that has
+ * been flushed delwri has had the flush completed by the time we check that
+ * the inode is clean. The clean inode check needs to be done before flushing
+ * the inode delwri otherwise we would loop forever requeuing clean inodes as
+ * we cannot tell apart a successful delwri flush and a clean inode from the
+ * return value of xfs_iflush().
+ *
+ * Note that because the inode is flushed delayed write by background
+ * writeback, the flush lock may already be held here and waiting on it can
+ * result in very long latencies. Hence for sync reclaims, where we wait on the
+ * flush lock, the caller should push out delayed write inodes first before
+ * trying to reclaim them to minimise the amount of time spent waiting. For
+ * background relaim, we just requeue the inode for the next pass.
+ *
+ * Hence the order of actions after gaining the locks should be:
+ *	bad		=> reclaim
+ *	shutdown	=> unpin and reclaim
+ *	pinned, delwri	=> requeue
+ *	pinned, sync	=> unpin
+ *	stale		=> reclaim
+ *	clean		=> reclaim
+ *	dirty, delwri	=> flush and requeue
+ *	dirty, sync	=> flush, wait and reclaim
+ */
+STATIC int
+xfs_reclaim_inode(
+	struct xfs_inode	*ip,
+	struct xfs_perag	*pag,
+	int			sync_mode)
+{
+	int	error;
+
+restart:
+	error = 0;
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	if (!xfs_iflock_nowait(ip)) {
+		if (!(sync_mode & SYNC_WAIT))
+			goto out;
+
+		/*
+		 * If we only have a single dirty inode in a cluster there is
+		 * a fair chance that the AIL push may have pushed it into
+		 * the buffer, but xfsbufd won't touch it until 30 seconds
+		 * from now, and thus we will lock up here.
+		 *
+		 * Promote the inode buffer to the front of the delwri list
+		 * and wake up xfsbufd now.
+		 */
+		xfs_promote_inode(ip);
+		xfs_iflock(ip);
+	}
+
+	if (is_bad_inode(VFS_I(ip)))
+		goto reclaim;
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+		xfs_iunpin_wait(ip);
+		goto reclaim;
+	}
+	if (xfs_ipincount(ip)) {
+		if (!(sync_mode & SYNC_WAIT)) {
+			xfs_ifunlock(ip);
+			goto out;
+		}
+		xfs_iunpin_wait(ip);
+	}
+	if (xfs_iflags_test(ip, XFS_ISTALE))
+		goto reclaim;
+	if (xfs_inode_clean(ip))
+		goto reclaim;
+
+	/*
+	 * Now we have an inode that needs flushing.
+	 *
+	 * We do a nonblocking flush here even if we are doing a SYNC_WAIT
+	 * reclaim as we can deadlock with inode cluster removal.
+	 * xfs_ifree_cluster() can lock the inode buffer before it locks the
+	 * ip->i_lock, and we are doing the exact opposite here. As a result,
+	 * doing a blocking xfs_itobp() to get the cluster buffer will result
+	 * in an ABBA deadlock with xfs_ifree_cluster().
+	 *
+	 * As xfs_ifree_cluser() must gather all inodes that are active in the
+	 * cache to mark them stale, if we hit this case we don't actually want
+	 * to do IO here - we want the inode marked stale so we can simply
+	 * reclaim it. Hence if we get an EAGAIN error on a SYNC_WAIT flush,
+	 * just unlock the inode, back off and try again. Hopefully the next
+	 * pass through will see the stale flag set on the inode.
+	 */
+	error = xfs_iflush(ip, SYNC_TRYLOCK | sync_mode);
+	if (sync_mode & SYNC_WAIT) {
+		if (error == EAGAIN) {
+			xfs_iunlock(ip, XFS_ILOCK_EXCL);
+			/* backoff longer than in xfs_ifree_cluster */
+			delay(2);
+			goto restart;
+		}
+		xfs_iflock(ip);
+		goto reclaim;
+	}
+
+	/*
+	 * When we have to flush an inode but don't have SYNC_WAIT set, we
+	 * flush the inode out using a delwri buffer and wait for the next
+	 * call into reclaim to find it in a clean state instead of waiting for
+	 * it now. We also don't return errors here - if the error is transient
+	 * then the next reclaim pass will flush the inode, and if the error
+	 * is permanent then the next sync reclaim will reclaim the inode and
+	 * pass on the error.
+	 */
+	if (error && error != EAGAIN && !XFS_FORCED_SHUTDOWN(ip->i_mount)) {
+		xfs_warn(ip->i_mount,
+			"inode 0x%llx background reclaim flush failed with %d",
+			(long long)ip->i_ino, error);
+	}
+out:
+	xfs_iflags_clear(ip, XFS_IRECLAIM);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	/*
+	 * We could return EAGAIN here to make reclaim rescan the inode tree in
+	 * a short while. However, this just burns CPU time scanning the tree
+	 * waiting for IO to complete and xfssyncd never goes back to the idle
+	 * state. Instead, return 0 to let the next scheduled background reclaim
+	 * attempt to reclaim the inode again.
+	 */
+	return 0;
+
+reclaim:
+	xfs_ifunlock(ip);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	XFS_STATS_INC(xs_ig_reclaims);
+	/*
+	 * Remove the inode from the per-AG radix tree.
+	 *
+	 * Because radix_tree_delete won't complain even if the item was never
+	 * added to the tree assert that it's been there before to catch
+	 * problems with the inode life time early on.
+	 */
+	spin_lock(&pag->pag_ici_lock);
+	if (!radix_tree_delete(&pag->pag_ici_root,
+				XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino)))
+		ASSERT(0);
+	__xfs_inode_clear_reclaim(pag, ip);
+	spin_unlock(&pag->pag_ici_lock);
+
+	/*
+	 * Here we do an (almost) spurious inode lock in order to coordinate
+	 * with inode cache radix tree lookups.  This is because the lookup
+	 * can reference the inodes in the cache without taking references.
+	 *
+	 * We make that OK here by ensuring that we wait until the inode is
+	 * unlocked after the lookup before we go ahead and free it.  We get
+	 * both the ilock and the iolock because the code may need to drop the
+	 * ilock one but will still hold the iolock.
+	 */
+	xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+	xfs_qm_dqdetach(ip);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+
+	xfs_inode_free(ip);
+	return error;
+
+}
+
+/*
+ * Walk the AGs and reclaim the inodes in them. Even if the filesystem is
+ * corrupted, we still want to try to reclaim all the inodes. If we don't,
+ * then a shut down during filesystem unmount reclaim walk leak all the
+ * unreclaimed inodes.
+ */
+int
+xfs_reclaim_inodes_ag(
+	struct xfs_mount	*mp,
+	int			flags,
+	int			*nr_to_scan)
+{
+	struct xfs_perag	*pag;
+	int			error = 0;
+	int			last_error = 0;
+	xfs_agnumber_t		ag;
+	int			trylock = flags & SYNC_TRYLOCK;
+	int			skipped;
+
+restart:
+	ag = 0;
+	skipped = 0;
+	while ((pag = xfs_perag_get_tag(mp, ag, XFS_ICI_RECLAIM_TAG))) {
+		unsigned long	first_index = 0;
+		int		done = 0;
+		int		nr_found = 0;
+
+		ag = pag->pag_agno + 1;
+
+		if (trylock) {
+			if (!mutex_trylock(&pag->pag_ici_reclaim_lock)) {
+				skipped++;
+				xfs_perag_put(pag);
+				continue;
+			}
+			first_index = pag->pag_ici_reclaim_cursor;
+		} else
+			mutex_lock(&pag->pag_ici_reclaim_lock);
+
+		do {
+			struct xfs_inode *batch[XFS_LOOKUP_BATCH];
+			int	i;
+
+			rcu_read_lock();
+			nr_found = radix_tree_gang_lookup_tag(
+					&pag->pag_ici_root,
+					(void **)batch, first_index,
+					XFS_LOOKUP_BATCH,
+					XFS_ICI_RECLAIM_TAG);
+			if (!nr_found) {
+				done = 1;
+				rcu_read_unlock();
+				break;
+			}
+
+			/*
+			 * Grab the inodes before we drop the lock. if we found
+			 * nothing, nr == 0 and the loop will be skipped.
+			 */
+			for (i = 0; i < nr_found; i++) {
+				struct xfs_inode *ip = batch[i];
+
+				if (done || xfs_reclaim_inode_grab(ip, flags))
+					batch[i] = NULL;
+
+				/*
+				 * Update the index for the next lookup. Catch
+				 * overflows into the next AG range which can
+				 * occur if we have inodes in the last block of
+				 * the AG and we are currently pointing to the
+				 * last inode.
+				 *
+				 * Because we may see inodes that are from the
+				 * wrong AG due to RCU freeing and
+				 * reallocation, only update the index if it
+				 * lies in this AG. It was a race that lead us
+				 * to see this inode, so another lookup from
+				 * the same index will not find it again.
+				 */
+				if (XFS_INO_TO_AGNO(mp, ip->i_ino) !=
+								pag->pag_agno)
+					continue;
+				first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+				if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
+					done = 1;
+			}
+
+			/* unlock now we've grabbed the inodes. */
+			rcu_read_unlock();
+
+			for (i = 0; i < nr_found; i++) {
+				if (!batch[i])
+					continue;
+				error = xfs_reclaim_inode(batch[i], pag, flags);
+				if (error && last_error != EFSCORRUPTED)
+					last_error = error;
+			}
+
+			*nr_to_scan -= XFS_LOOKUP_BATCH;
+
+			cond_resched();
+
+		} while (nr_found && !done && *nr_to_scan > 0);
+
+		if (trylock && !done)
+			pag->pag_ici_reclaim_cursor = first_index;
+		else
+			pag->pag_ici_reclaim_cursor = 0;
+		mutex_unlock(&pag->pag_ici_reclaim_lock);
+		xfs_perag_put(pag);
+	}
+
+	/*
+	 * if we skipped any AG, and we still have scan count remaining, do
+	 * another pass this time using blocking reclaim semantics (i.e
+	 * waiting on the reclaim locks and ignoring the reclaim cursors). This
+	 * ensure that when we get more reclaimers than AGs we block rather
+	 * than spin trying to execute reclaim.
+	 */
+	if (skipped && (flags & SYNC_WAIT) && *nr_to_scan > 0) {
+		trylock = 0;
+		goto restart;
+	}
+	return XFS_ERROR(last_error);
+}
+
+int
+xfs_reclaim_inodes(
+	xfs_mount_t	*mp,
+	int		mode)
+{
+	int		nr_to_scan = INT_MAX;
+
+	return xfs_reclaim_inodes_ag(mp, mode, &nr_to_scan);
+}
+
+/*
+ * Scan a certain number of inodes for reclaim.
+ *
+ * When called we make sure that there is a background (fast) inode reclaim in
+ * progress, while we will throttle the speed of reclaim via doing synchronous
+ * reclaim of inodes. That means if we come across dirty inodes, we wait for
+ * them to be cleaned, which we hope will not be very long due to the
+ * background walker having already kicked the IO off on those dirty inodes.
+ */
+void
+xfs_reclaim_inodes_nr(
+	struct xfs_mount	*mp,
+	int			nr_to_scan)
+{
+	/* kick background reclaimer and push the AIL */
+	xfs_syncd_queue_reclaim(mp);
+	xfs_ail_push_all(mp->m_ail);
+
+	xfs_reclaim_inodes_ag(mp, SYNC_TRYLOCK | SYNC_WAIT, &nr_to_scan);
+}
+
+/*
+ * Return the number of reclaimable inodes in the filesystem for
+ * the shrinker to determine how much to reclaim.
+ */
+int
+xfs_reclaim_inodes_count(
+	struct xfs_mount	*mp)
+{
+	struct xfs_perag	*pag;
+	xfs_agnumber_t		ag = 0;
+	int			reclaimable = 0;
+
+	while ((pag = xfs_perag_get_tag(mp, ag, XFS_ICI_RECLAIM_TAG))) {
+		ag = pag->pag_agno + 1;
+		reclaimable += pag->pag_ici_reclaimable;
+		xfs_perag_put(pag);
+	}
+	return reclaimable;
+}
+
diff --git a/fs/xfs/xfs_sync.h b/fs/xfs/xfs_sync.h
new file mode 100644
index 0000000..fa96547
--- /dev/null
+++ b/fs/xfs/xfs_sync.h
@@ -0,0 +1,53 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef XFS_SYNC_H
+#define XFS_SYNC_H 1
+
+struct xfs_mount;
+struct xfs_perag;
+
+#define SYNC_WAIT		0x0001	/* wait for i/o to complete */
+#define SYNC_TRYLOCK		0x0002  /* only try to lock inodes */
+
+extern struct workqueue_struct	*xfs_syncd_wq;	/* sync workqueue */
+
+int xfs_syncd_init(struct xfs_mount *mp);
+void xfs_syncd_stop(struct xfs_mount *mp);
+
+int xfs_quiesce_data(struct xfs_mount *mp);
+void xfs_quiesce_attr(struct xfs_mount *mp);
+
+void xfs_flush_inodes(struct xfs_inode *ip);
+
+int xfs_log_dirty_inode(struct xfs_inode *ip, struct xfs_perag *pag, int flags);
+
+int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
+int xfs_reclaim_inodes_count(struct xfs_mount *mp);
+void xfs_reclaim_inodes_nr(struct xfs_mount *mp, int nr_to_scan);
+
+void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
+void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
+void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
+				struct xfs_inode *ip);
+
+int xfs_sync_inode_grab(struct xfs_inode *ip);
+int xfs_inode_ag_iterator(struct xfs_mount *mp,
+	int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag, int flags),
+	int flags);
+
+#endif
diff --git a/fs/xfs/xfs_sysctl.c b/fs/xfs/xfs_sysctl.c
new file mode 100644
index 0000000..ee2d2ad
--- /dev/null
+++ b/fs/xfs/xfs_sysctl.c
@@ -0,0 +1,252 @@
+/*
+ * Copyright (c) 2001-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include <linux/sysctl.h>
+#include <linux/proc_fs.h>
+#include "xfs_error.h"
+
+static struct ctl_table_header *xfs_table_header;
+
+#ifdef CONFIG_PROC_FS
+STATIC int
+xfs_stats_clear_proc_handler(
+	ctl_table	*ctl,
+	int		write,
+	void		__user *buffer,
+	size_t		*lenp,
+	loff_t		*ppos)
+{
+	int		c, ret, *valp = ctl->data;
+	__uint32_t	vn_active;
+
+	ret = proc_dointvec_minmax(ctl, write, buffer, lenp, ppos);
+
+	if (!ret && write && *valp) {
+		xfs_notice(NULL, "Clearing xfsstats");
+		for_each_possible_cpu(c) {
+			preempt_disable();
+			/* save vn_active, it's a universal truth! */
+			vn_active = per_cpu(xfsstats, c).vn_active;
+			memset(&per_cpu(xfsstats, c), 0,
+			       sizeof(struct xfsstats));
+			per_cpu(xfsstats, c).vn_active = vn_active;
+			preempt_enable();
+		}
+		xfs_stats_clear = 0;
+	}
+
+	return ret;
+}
+
+STATIC int
+xfs_panic_mask_proc_handler(
+	ctl_table	*ctl,
+	int		write,
+	void		__user *buffer,
+	size_t		*lenp,
+	loff_t		*ppos)
+{
+	int		ret, *valp = ctl->data;
+
+	ret = proc_dointvec_minmax(ctl, write, buffer, lenp, ppos);
+	if (!ret && write) {
+		xfs_panic_mask = *valp;
+#ifdef DEBUG
+		xfs_panic_mask |= (XFS_PTAG_SHUTDOWN_CORRUPT | XFS_PTAG_LOGRES);
+#endif
+	}
+	return ret;
+}
+#endif /* CONFIG_PROC_FS */
+
+static ctl_table xfs_table[] = {
+	{
+		.procname	= "irix_sgid_inherit",
+		.data		= &xfs_params.sgid_inherit.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.sgid_inherit.min,
+		.extra2		= &xfs_params.sgid_inherit.max
+	},
+	{
+		.procname	= "irix_symlink_mode",
+		.data		= &xfs_params.symlink_mode.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.symlink_mode.min,
+		.extra2		= &xfs_params.symlink_mode.max
+	},
+	{
+		.procname	= "panic_mask",
+		.data		= &xfs_params.panic_mask.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= xfs_panic_mask_proc_handler,
+		.extra1		= &xfs_params.panic_mask.min,
+		.extra2		= &xfs_params.panic_mask.max
+	},
+
+	{
+		.procname	= "error_level",
+		.data		= &xfs_params.error_level.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.error_level.min,
+		.extra2		= &xfs_params.error_level.max
+	},
+	{
+		.procname	= "xfssyncd_centisecs",
+		.data		= &xfs_params.syncd_timer.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.syncd_timer.min,
+		.extra2		= &xfs_params.syncd_timer.max
+	},
+	{
+		.procname	= "inherit_sync",
+		.data		= &xfs_params.inherit_sync.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.inherit_sync.min,
+		.extra2		= &xfs_params.inherit_sync.max
+	},
+	{
+		.procname	= "inherit_nodump",
+		.data		= &xfs_params.inherit_nodump.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.inherit_nodump.min,
+		.extra2		= &xfs_params.inherit_nodump.max
+	},
+	{
+		.procname	= "inherit_noatime",
+		.data		= &xfs_params.inherit_noatim.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.inherit_noatim.min,
+		.extra2		= &xfs_params.inherit_noatim.max
+	},
+	{
+		.procname	= "xfsbufd_centisecs",
+		.data		= &xfs_params.xfs_buf_timer.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.xfs_buf_timer.min,
+		.extra2		= &xfs_params.xfs_buf_timer.max
+	},
+	{
+		.procname	= "age_buffer_centisecs",
+		.data		= &xfs_params.xfs_buf_age.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.xfs_buf_age.min,
+		.extra2		= &xfs_params.xfs_buf_age.max
+	},
+	{
+		.procname	= "inherit_nosymlinks",
+		.data		= &xfs_params.inherit_nosym.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.inherit_nosym.min,
+		.extra2		= &xfs_params.inherit_nosym.max
+	},
+	{
+		.procname	= "rotorstep",
+		.data		= &xfs_params.rotorstep.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.rotorstep.min,
+		.extra2		= &xfs_params.rotorstep.max
+	},
+	{
+		.procname	= "inherit_nodefrag",
+		.data		= &xfs_params.inherit_nodfrg.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.inherit_nodfrg.min,
+		.extra2		= &xfs_params.inherit_nodfrg.max
+	},
+	{
+		.procname	= "filestream_centisecs",
+		.data		= &xfs_params.fstrm_timer.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &xfs_params.fstrm_timer.min,
+		.extra2		= &xfs_params.fstrm_timer.max,
+	},
+	/* please keep this the last entry */
+#ifdef CONFIG_PROC_FS
+	{
+		.procname	= "stats_clear",
+		.data		= &xfs_params.stats_clear.val,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= xfs_stats_clear_proc_handler,
+		.extra1		= &xfs_params.stats_clear.min,
+		.extra2		= &xfs_params.stats_clear.max
+	},
+#endif /* CONFIG_PROC_FS */
+
+	{}
+};
+
+static ctl_table xfs_dir_table[] = {
+	{
+		.procname	= "xfs",
+		.mode		= 0555,
+		.child		= xfs_table
+	},
+	{}
+};
+
+static ctl_table xfs_root_table[] = {
+	{
+		.procname	= "fs",
+		.mode		= 0555,
+		.child		= xfs_dir_table
+	},
+	{}
+};
+
+int
+xfs_sysctl_register(void)
+{
+	xfs_table_header = register_sysctl_table(xfs_root_table);
+	if (!xfs_table_header)
+		return -ENOMEM;
+	return 0;
+}
+
+void
+xfs_sysctl_unregister(void)
+{
+	unregister_sysctl_table(xfs_table_header);
+}
diff --git a/fs/xfs/xfs_sysctl.h b/fs/xfs/xfs_sysctl.h
new file mode 100644
index 0000000..b9937d4
--- /dev/null
+++ b/fs/xfs/xfs_sysctl.h
@@ -0,0 +1,102 @@
+/*
+ * Copyright (c) 2001-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_SYSCTL_H__
+#define __XFS_SYSCTL_H__
+
+#include <linux/sysctl.h>
+
+/*
+ * Tunable xfs parameters
+ */
+
+typedef struct xfs_sysctl_val {
+	int min;
+	int val;
+	int max;
+} xfs_sysctl_val_t;
+
+typedef struct xfs_param {
+	xfs_sysctl_val_t sgid_inherit;	/* Inherit S_ISGID if process' GID is
+					 * not a member of parent dir GID. */
+	xfs_sysctl_val_t symlink_mode;	/* Link creat mode affected by umask */
+	xfs_sysctl_val_t panic_mask;	/* bitmask to cause panic on errors. */
+	xfs_sysctl_val_t error_level;	/* Degree of reporting for problems  */
+	xfs_sysctl_val_t syncd_timer;	/* Interval between xfssyncd wakeups */
+	xfs_sysctl_val_t stats_clear;	/* Reset all XFS statistics to zero. */
+	xfs_sysctl_val_t inherit_sync;	/* Inherit the "sync" inode flag. */
+	xfs_sysctl_val_t inherit_nodump;/* Inherit the "nodump" inode flag. */
+	xfs_sysctl_val_t inherit_noatim;/* Inherit the "noatime" inode flag. */
+	xfs_sysctl_val_t xfs_buf_timer;	/* Interval between xfsbufd wakeups. */
+	xfs_sysctl_val_t xfs_buf_age;	/* Metadata buffer age before flush. */
+	xfs_sysctl_val_t inherit_nosym;	/* Inherit the "nosymlinks" flag. */
+	xfs_sysctl_val_t rotorstep;	/* inode32 AG rotoring control knob */
+	xfs_sysctl_val_t inherit_nodfrg;/* Inherit the "nodefrag" inode flag. */
+	xfs_sysctl_val_t fstrm_timer;	/* Filestream dir-AG assoc'n timeout. */
+} xfs_param_t;
+
+/*
+ * xfs_error_level:
+ *
+ * How much error reporting will be done when internal problems are
+ * encountered.  These problems normally return an EFSCORRUPTED to their
+ * caller, with no other information reported.
+ *
+ * 0	No error reports
+ * 1	Report EFSCORRUPTED errors that will cause a filesystem shutdown
+ * 5	Report all EFSCORRUPTED errors (all of the above errors, plus any
+ *	additional errors that are known to not cause shutdowns)
+ *
+ * xfs_panic_mask bit 0x8 turns the error reports into panics
+ */
+
+enum {
+	/* XFS_REFCACHE_SIZE = 1 */
+	/* XFS_REFCACHE_PURGE = 2 */
+	/* XFS_RESTRICT_CHOWN = 3 */
+	XFS_SGID_INHERIT = 4,
+	XFS_SYMLINK_MODE = 5,
+	XFS_PANIC_MASK = 6,
+	XFS_ERRLEVEL = 7,
+	XFS_SYNCD_TIMER = 8,
+	/* XFS_PROBE_DMAPI = 9 */
+	/* XFS_PROBE_IOOPS = 10 */
+	/* XFS_PROBE_QUOTA = 11 */
+	XFS_STATS_CLEAR = 12,
+	XFS_INHERIT_SYNC = 13,
+	XFS_INHERIT_NODUMP = 14,
+	XFS_INHERIT_NOATIME = 15,
+	XFS_BUF_TIMER = 16,
+	XFS_BUF_AGE = 17,
+	/* XFS_IO_BYPASS = 18 */
+	XFS_INHERIT_NOSYM = 19,
+	XFS_ROTORSTEP = 20,
+	XFS_INHERIT_NODFRG = 21,
+	XFS_FILESTREAM_TIMER = 22,
+};
+
+extern xfs_param_t	xfs_params;
+
+#ifdef CONFIG_SYSCTL
+extern int xfs_sysctl_register(void);
+extern void xfs_sysctl_unregister(void);
+#else
+# define xfs_sysctl_register()		(0)
+# define xfs_sysctl_unregister()	do { } while (0)
+#endif /* CONFIG_SYSCTL */
+
+#endif /* __XFS_SYSCTL_H__ */
diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c
new file mode 100644
index 0000000..9010ce8
--- /dev/null
+++ b/fs/xfs/xfs_trace.c
@@ -0,0 +1,56 @@
+/*
+ * Copyright (c) 2009, Christoph Hellwig
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_types.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_da_btree.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_dinode.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_mount.h"
+#include "xfs_ialloc.h"
+#include "xfs_itable.h"
+#include "xfs_alloc.h"
+#include "xfs_bmap.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_log_priv.h"
+#include "xfs_buf_item.h"
+#include "xfs_quota.h"
+#include "xfs_iomap.h"
+#include "xfs_aops.h"
+#include "xfs_dquot_item.h"
+#include "xfs_dquot.h"
+#include "xfs_log_recover.h"
+#include "xfs_inode_item.h"
+
+/*
+ * We include this last to have the helpers above available for the trace
+ * event implementations.
+ */
+#define CREATE_TRACE_POINTS
+#include "xfs_trace.h"
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
new file mode 100644
index 0000000..4940357
--- /dev/null
+++ b/fs/xfs/xfs_trace.h
@@ -0,0 +1,1779 @@
+/*
+ * Copyright (c) 2009, Christoph Hellwig
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM xfs
+
+#if !defined(_TRACE_XFS_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_XFS_H
+
+#include <linux/tracepoint.h>
+
+struct xfs_agf;
+struct xfs_alloc_arg;
+struct xfs_attr_list_context;
+struct xfs_buf_log_item;
+struct xfs_da_args;
+struct xfs_da_node_entry;
+struct xfs_dquot;
+struct xfs_log_item;
+struct xlog_ticket;
+struct log;
+struct xlog_recover;
+struct xlog_recover_item;
+struct xfs_buf_log_format;
+struct xfs_inode_log_format;
+
+DECLARE_EVENT_CLASS(xfs_attr_list_class,
+	TP_PROTO(struct xfs_attr_list_context *ctx),
+	TP_ARGS(ctx),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(u32, hashval)
+		__field(u32, blkno)
+		__field(u32, offset)
+		__field(void *, alist)
+		__field(int, bufsize)
+		__field(int, count)
+		__field(int, firstu)
+		__field(int, dupcnt)
+		__field(int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ctx->dp)->i_sb->s_dev;
+		__entry->ino = ctx->dp->i_ino;
+		__entry->hashval = ctx->cursor->hashval;
+		__entry->blkno = ctx->cursor->blkno;
+		__entry->offset = ctx->cursor->offset;
+		__entry->alist = ctx->alist;
+		__entry->bufsize = ctx->bufsize;
+		__entry->count = ctx->count;
+		__entry->firstu = ctx->firstu;
+		__entry->flags = ctx->flags;
+	),
+	TP_printk("dev %d:%d ino 0x%llx cursor h/b/o 0x%x/0x%x/%u dupcnt %u "
+		  "alist 0x%p size %u count %u firstu %u flags %d %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		   __entry->ino,
+		   __entry->hashval,
+		   __entry->blkno,
+		   __entry->offset,
+		   __entry->dupcnt,
+		   __entry->alist,
+		   __entry->bufsize,
+		   __entry->count,
+		   __entry->firstu,
+		   __entry->flags,
+		   __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS)
+	)
+)
+
+#define DEFINE_ATTR_LIST_EVENT(name) \
+DEFINE_EVENT(xfs_attr_list_class, name, \
+	TP_PROTO(struct xfs_attr_list_context *ctx), \
+	TP_ARGS(ctx))
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_sf);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_sf_all);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_leaf);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_leaf_end);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_full);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_add);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_wrong_blk);
+DEFINE_ATTR_LIST_EVENT(xfs_attr_list_notfound);
+
+DECLARE_EVENT_CLASS(xfs_perag_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int refcount,
+		 unsigned long caller_ip),
+	TP_ARGS(mp, agno, refcount, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(int, refcount)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->refcount = refcount;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d agno %u refcount %d caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->refcount,
+		  (char *)__entry->caller_ip)
+);
+
+#define DEFINE_PERAG_REF_EVENT(name)	\
+DEFINE_EVENT(xfs_perag_class, name,	\
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int refcount,	\
+		 unsigned long caller_ip),					\
+	TP_ARGS(mp, agno, refcount, caller_ip))
+DEFINE_PERAG_REF_EVENT(xfs_perag_get);
+DEFINE_PERAG_REF_EVENT(xfs_perag_get_tag);
+DEFINE_PERAG_REF_EVENT(xfs_perag_put);
+DEFINE_PERAG_REF_EVENT(xfs_perag_set_reclaim);
+DEFINE_PERAG_REF_EVENT(xfs_perag_clear_reclaim);
+
+TRACE_EVENT(xfs_attr_list_node_descend,
+	TP_PROTO(struct xfs_attr_list_context *ctx,
+		 struct xfs_da_node_entry *btree),
+	TP_ARGS(ctx, btree),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(u32, hashval)
+		__field(u32, blkno)
+		__field(u32, offset)
+		__field(void *, alist)
+		__field(int, bufsize)
+		__field(int, count)
+		__field(int, firstu)
+		__field(int, dupcnt)
+		__field(int, flags)
+		__field(u32, bt_hashval)
+		__field(u32, bt_before)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ctx->dp)->i_sb->s_dev;
+		__entry->ino = ctx->dp->i_ino;
+		__entry->hashval = ctx->cursor->hashval;
+		__entry->blkno = ctx->cursor->blkno;
+		__entry->offset = ctx->cursor->offset;
+		__entry->alist = ctx->alist;
+		__entry->bufsize = ctx->bufsize;
+		__entry->count = ctx->count;
+		__entry->firstu = ctx->firstu;
+		__entry->flags = ctx->flags;
+		__entry->bt_hashval = be32_to_cpu(btree->hashval);
+		__entry->bt_before = be32_to_cpu(btree->before);
+	),
+	TP_printk("dev %d:%d ino 0x%llx cursor h/b/o 0x%x/0x%x/%u dupcnt %u "
+		  "alist 0x%p size %u count %u firstu %u flags %d %s "
+		  "node hashval %u, node before %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		   __entry->ino,
+		   __entry->hashval,
+		   __entry->blkno,
+		   __entry->offset,
+		   __entry->dupcnt,
+		   __entry->alist,
+		   __entry->bufsize,
+		   __entry->count,
+		   __entry->firstu,
+		   __entry->flags,
+		   __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS),
+		   __entry->bt_hashval,
+		   __entry->bt_before)
+);
+
+TRACE_EVENT(xfs_iext_insert,
+	TP_PROTO(struct xfs_inode *ip, xfs_extnum_t idx,
+		 struct xfs_bmbt_irec *r, int state, unsigned long caller_ip),
+	TP_ARGS(ip, idx, r, state, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_extnum_t, idx)
+		__field(xfs_fileoff_t, startoff)
+		__field(xfs_fsblock_t, startblock)
+		__field(xfs_filblks_t, blockcount)
+		__field(xfs_exntst_t, state)
+		__field(int, bmap_state)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->idx = idx;
+		__entry->startoff = r->br_startoff;
+		__entry->startblock = r->br_startblock;
+		__entry->blockcount = r->br_blockcount;
+		__entry->state = r->br_state;
+		__entry->bmap_state = state;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx state %s idx %ld "
+		  "offset %lld block %lld count %lld flag %d caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_flags(__entry->bmap_state, "|", XFS_BMAP_EXT_FLAGS),
+		  (long)__entry->idx,
+		  __entry->startoff,
+		  (__int64_t)__entry->startblock,
+		  __entry->blockcount,
+		  __entry->state,
+		  (char *)__entry->caller_ip)
+);
+
+DECLARE_EVENT_CLASS(xfs_bmap_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_extnum_t idx, int state,
+		 unsigned long caller_ip),
+	TP_ARGS(ip, idx, state, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_extnum_t, idx)
+		__field(xfs_fileoff_t, startoff)
+		__field(xfs_fsblock_t, startblock)
+		__field(xfs_filblks_t, blockcount)
+		__field(xfs_exntst_t, state)
+		__field(int, bmap_state)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		struct xfs_ifork	*ifp = (state & BMAP_ATTRFORK) ?
+						ip->i_afp : &ip->i_df;
+		struct xfs_bmbt_irec	r;
+
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &r);
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->idx = idx;
+		__entry->startoff = r.br_startoff;
+		__entry->startblock = r.br_startblock;
+		__entry->blockcount = r.br_blockcount;
+		__entry->state = r.br_state;
+		__entry->bmap_state = state;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx state %s idx %ld "
+		  "offset %lld block %lld count %lld flag %d caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_flags(__entry->bmap_state, "|", XFS_BMAP_EXT_FLAGS),
+		  (long)__entry->idx,
+		  __entry->startoff,
+		  (__int64_t)__entry->startblock,
+		  __entry->blockcount,
+		  __entry->state,
+		  (char *)__entry->caller_ip)
+)
+
+#define DEFINE_BMAP_EVENT(name) \
+DEFINE_EVENT(xfs_bmap_class, name, \
+	TP_PROTO(struct xfs_inode *ip, xfs_extnum_t idx, int state, \
+		 unsigned long caller_ip), \
+	TP_ARGS(ip, idx, state, caller_ip))
+DEFINE_BMAP_EVENT(xfs_iext_remove);
+DEFINE_BMAP_EVENT(xfs_bmap_pre_update);
+DEFINE_BMAP_EVENT(xfs_bmap_post_update);
+DEFINE_BMAP_EVENT(xfs_extlist);
+
+DECLARE_EVENT_CLASS(xfs_buf_class,
+	TP_PROTO(struct xfs_buf *bp, unsigned long caller_ip),
+	TP_ARGS(bp, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_daddr_t, bno)
+		__field(size_t, buffer_length)
+		__field(int, hold)
+		__field(int, pincount)
+		__field(unsigned, lockval)
+		__field(unsigned, flags)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = bp->b_target->bt_dev;
+		__entry->bno = bp->b_bn;
+		__entry->buffer_length = bp->b_buffer_length;
+		__entry->hold = atomic_read(&bp->b_hold);
+		__entry->pincount = atomic_read(&bp->b_pin_count);
+		__entry->lockval = bp->b_sema.count;
+		__entry->flags = bp->b_flags;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d bno 0x%llx len 0x%zx hold %d pincount %d "
+		  "lock %d flags %s caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->bno,
+		  __entry->buffer_length,
+		  __entry->hold,
+		  __entry->pincount,
+		  __entry->lockval,
+		  __print_flags(__entry->flags, "|", XFS_BUF_FLAGS),
+		  (void *)__entry->caller_ip)
+)
+
+#define DEFINE_BUF_EVENT(name) \
+DEFINE_EVENT(xfs_buf_class, name, \
+	TP_PROTO(struct xfs_buf *bp, unsigned long caller_ip), \
+	TP_ARGS(bp, caller_ip))
+DEFINE_BUF_EVENT(xfs_buf_init);
+DEFINE_BUF_EVENT(xfs_buf_free);
+DEFINE_BUF_EVENT(xfs_buf_hold);
+DEFINE_BUF_EVENT(xfs_buf_rele);
+DEFINE_BUF_EVENT(xfs_buf_iodone);
+DEFINE_BUF_EVENT(xfs_buf_iorequest);
+DEFINE_BUF_EVENT(xfs_buf_bawrite);
+DEFINE_BUF_EVENT(xfs_buf_lock);
+DEFINE_BUF_EVENT(xfs_buf_lock_done);
+DEFINE_BUF_EVENT(xfs_buf_trylock);
+DEFINE_BUF_EVENT(xfs_buf_unlock);
+DEFINE_BUF_EVENT(xfs_buf_iowait);
+DEFINE_BUF_EVENT(xfs_buf_iowait_done);
+DEFINE_BUF_EVENT(xfs_buf_delwri_queue);
+DEFINE_BUF_EVENT(xfs_buf_delwri_dequeue);
+DEFINE_BUF_EVENT(xfs_buf_delwri_split);
+DEFINE_BUF_EVENT(xfs_buf_get_uncached);
+DEFINE_BUF_EVENT(xfs_bdstrat_shut);
+DEFINE_BUF_EVENT(xfs_buf_item_relse);
+DEFINE_BUF_EVENT(xfs_buf_item_iodone);
+DEFINE_BUF_EVENT(xfs_buf_item_iodone_async);
+DEFINE_BUF_EVENT(xfs_buf_error_relse);
+DEFINE_BUF_EVENT(xfs_trans_read_buf_io);
+DEFINE_BUF_EVENT(xfs_trans_read_buf_shut);
+
+/* not really buffer traces, but the buf provides useful information */
+DEFINE_BUF_EVENT(xfs_btree_corrupt);
+DEFINE_BUF_EVENT(xfs_da_btree_corrupt);
+DEFINE_BUF_EVENT(xfs_reset_dqcounts);
+DEFINE_BUF_EVENT(xfs_inode_item_push);
+
+/* pass flags explicitly */
+DECLARE_EVENT_CLASS(xfs_buf_flags_class,
+	TP_PROTO(struct xfs_buf *bp, unsigned flags, unsigned long caller_ip),
+	TP_ARGS(bp, flags, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_daddr_t, bno)
+		__field(size_t, buffer_length)
+		__field(int, hold)
+		__field(int, pincount)
+		__field(unsigned, lockval)
+		__field(unsigned, flags)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = bp->b_target->bt_dev;
+		__entry->bno = bp->b_bn;
+		__entry->buffer_length = bp->b_buffer_length;
+		__entry->flags = flags;
+		__entry->hold = atomic_read(&bp->b_hold);
+		__entry->pincount = atomic_read(&bp->b_pin_count);
+		__entry->lockval = bp->b_sema.count;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d bno 0x%llx len 0x%zx hold %d pincount %d "
+		  "lock %d flags %s caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->bno,
+		  __entry->buffer_length,
+		  __entry->hold,
+		  __entry->pincount,
+		  __entry->lockval,
+		  __print_flags(__entry->flags, "|", XFS_BUF_FLAGS),
+		  (void *)__entry->caller_ip)
+)
+
+#define DEFINE_BUF_FLAGS_EVENT(name) \
+DEFINE_EVENT(xfs_buf_flags_class, name, \
+	TP_PROTO(struct xfs_buf *bp, unsigned flags, unsigned long caller_ip), \
+	TP_ARGS(bp, flags, caller_ip))
+DEFINE_BUF_FLAGS_EVENT(xfs_buf_find);
+DEFINE_BUF_FLAGS_EVENT(xfs_buf_get);
+DEFINE_BUF_FLAGS_EVENT(xfs_buf_read);
+
+TRACE_EVENT(xfs_buf_ioerror,
+	TP_PROTO(struct xfs_buf *bp, int error, unsigned long caller_ip),
+	TP_ARGS(bp, error, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_daddr_t, bno)
+		__field(size_t, buffer_length)
+		__field(unsigned, flags)
+		__field(int, hold)
+		__field(int, pincount)
+		__field(unsigned, lockval)
+		__field(int, error)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = bp->b_target->bt_dev;
+		__entry->bno = bp->b_bn;
+		__entry->buffer_length = bp->b_buffer_length;
+		__entry->hold = atomic_read(&bp->b_hold);
+		__entry->pincount = atomic_read(&bp->b_pin_count);
+		__entry->lockval = bp->b_sema.count;
+		__entry->error = error;
+		__entry->flags = bp->b_flags;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d bno 0x%llx len 0x%zx hold %d pincount %d "
+		  "lock %d error %d flags %s caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->bno,
+		  __entry->buffer_length,
+		  __entry->hold,
+		  __entry->pincount,
+		  __entry->lockval,
+		  __entry->error,
+		  __print_flags(__entry->flags, "|", XFS_BUF_FLAGS),
+		  (void *)__entry->caller_ip)
+);
+
+DECLARE_EVENT_CLASS(xfs_buf_item_class,
+	TP_PROTO(struct xfs_buf_log_item *bip),
+	TP_ARGS(bip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_daddr_t, buf_bno)
+		__field(size_t, buf_len)
+		__field(int, buf_hold)
+		__field(int, buf_pincount)
+		__field(int, buf_lockval)
+		__field(unsigned, buf_flags)
+		__field(unsigned, bli_recur)
+		__field(int, bli_refcount)
+		__field(unsigned, bli_flags)
+		__field(void *, li_desc)
+		__field(unsigned, li_flags)
+	),
+	TP_fast_assign(
+		__entry->dev = bip->bli_buf->b_target->bt_dev;
+		__entry->bli_flags = bip->bli_flags;
+		__entry->bli_recur = bip->bli_recur;
+		__entry->bli_refcount = atomic_read(&bip->bli_refcount);
+		__entry->buf_bno = bip->bli_buf->b_bn;
+		__entry->buf_len = bip->bli_buf->b_buffer_length;
+		__entry->buf_flags = bip->bli_buf->b_flags;
+		__entry->buf_hold = atomic_read(&bip->bli_buf->b_hold);
+		__entry->buf_pincount = atomic_read(&bip->bli_buf->b_pin_count);
+		__entry->buf_lockval = bip->bli_buf->b_sema.count;
+		__entry->li_desc = bip->bli_item.li_desc;
+		__entry->li_flags = bip->bli_item.li_flags;
+	),
+	TP_printk("dev %d:%d bno 0x%llx len 0x%zx hold %d pincount %d "
+		  "lock %d flags %s recur %d refcount %d bliflags %s "
+		  "lidesc 0x%p liflags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  (unsigned long long)__entry->buf_bno,
+		  __entry->buf_len,
+		  __entry->buf_hold,
+		  __entry->buf_pincount,
+		  __entry->buf_lockval,
+		  __print_flags(__entry->buf_flags, "|", XFS_BUF_FLAGS),
+		  __entry->bli_recur,
+		  __entry->bli_refcount,
+		  __print_flags(__entry->bli_flags, "|", XFS_BLI_FLAGS),
+		  __entry->li_desc,
+		  __print_flags(__entry->li_flags, "|", XFS_LI_FLAGS))
+)
+
+#define DEFINE_BUF_ITEM_EVENT(name) \
+DEFINE_EVENT(xfs_buf_item_class, name, \
+	TP_PROTO(struct xfs_buf_log_item *bip), \
+	TP_ARGS(bip))
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_size_stale);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_format_stale);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_pin);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unpin_stale);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_trylock);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unlock);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_unlock_stale);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_committed);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_push);
+DEFINE_BUF_ITEM_EVENT(xfs_buf_item_pushbuf);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_get_buf);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_get_buf_recur);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_getsb);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_getsb_recur);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf_recur);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_log_buf);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_brelse);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release);
+DEFINE_BUF_ITEM_EVENT(xfs_trans_binval);
+
+DECLARE_EVENT_CLASS(xfs_lock_class,
+	TP_PROTO(struct xfs_inode *ip, unsigned lock_flags,
+		 unsigned long caller_ip),
+	TP_ARGS(ip,  lock_flags, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, lock_flags)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->lock_flags = lock_flags;
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx flags %s caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_flags(__entry->lock_flags, "|", XFS_LOCK_FLAGS),
+		  (void *)__entry->caller_ip)
+)
+
+#define DEFINE_LOCK_EVENT(name) \
+DEFINE_EVENT(xfs_lock_class, name, \
+	TP_PROTO(struct xfs_inode *ip, unsigned lock_flags, \
+		 unsigned long caller_ip), \
+	TP_ARGS(ip,  lock_flags, caller_ip))
+DEFINE_LOCK_EVENT(xfs_ilock);
+DEFINE_LOCK_EVENT(xfs_ilock_nowait);
+DEFINE_LOCK_EVENT(xfs_ilock_demote);
+DEFINE_LOCK_EVENT(xfs_iunlock);
+
+DECLARE_EVENT_CLASS(xfs_inode_class,
+	TP_PROTO(struct xfs_inode *ip),
+	TP_ARGS(ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+	),
+	TP_printk("dev %d:%d ino 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino)
+)
+
+#define DEFINE_INODE_EVENT(name) \
+DEFINE_EVENT(xfs_inode_class, name, \
+	TP_PROTO(struct xfs_inode *ip), \
+	TP_ARGS(ip))
+DEFINE_INODE_EVENT(xfs_iget_skip);
+DEFINE_INODE_EVENT(xfs_iget_reclaim);
+DEFINE_INODE_EVENT(xfs_iget_reclaim_fail);
+DEFINE_INODE_EVENT(xfs_iget_hit);
+DEFINE_INODE_EVENT(xfs_iget_miss);
+
+DEFINE_INODE_EVENT(xfs_getattr);
+DEFINE_INODE_EVENT(xfs_setattr);
+DEFINE_INODE_EVENT(xfs_readlink);
+DEFINE_INODE_EVENT(xfs_alloc_file_space);
+DEFINE_INODE_EVENT(xfs_free_file_space);
+DEFINE_INODE_EVENT(xfs_readdir);
+#ifdef CONFIG_XFS_POSIX_ACL
+DEFINE_INODE_EVENT(xfs_get_acl);
+#endif
+DEFINE_INODE_EVENT(xfs_vm_bmap);
+DEFINE_INODE_EVENT(xfs_file_ioctl);
+DEFINE_INODE_EVENT(xfs_file_compat_ioctl);
+DEFINE_INODE_EVENT(xfs_ioctl_setattr);
+DEFINE_INODE_EVENT(xfs_dir_fsync);
+DEFINE_INODE_EVENT(xfs_file_fsync);
+DEFINE_INODE_EVENT(xfs_destroy_inode);
+DEFINE_INODE_EVENT(xfs_write_inode);
+DEFINE_INODE_EVENT(xfs_evict_inode);
+
+DEFINE_INODE_EVENT(xfs_dquot_dqalloc);
+DEFINE_INODE_EVENT(xfs_dquot_dqdetach);
+
+DECLARE_EVENT_CLASS(xfs_iref_class,
+	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip),
+	TP_ARGS(ip, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, count)
+		__field(int, pincount)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->count = atomic_read(&VFS_I(ip)->i_count);
+		__entry->pincount = atomic_read(&ip->i_pincount);
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx count %d pincount %d caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->count,
+		  __entry->pincount,
+		  (char *)__entry->caller_ip)
+)
+
+#define DEFINE_IREF_EVENT(name) \
+DEFINE_EVENT(xfs_iref_class, name, \
+	TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip), \
+	TP_ARGS(ip, caller_ip))
+DEFINE_IREF_EVENT(xfs_ihold);
+DEFINE_IREF_EVENT(xfs_irele);
+DEFINE_IREF_EVENT(xfs_inode_pin);
+DEFINE_IREF_EVENT(xfs_inode_unpin);
+DEFINE_IREF_EVENT(xfs_inode_unpin_nowait);
+
+DECLARE_EVENT_CLASS(xfs_namespace_class,
+	TP_PROTO(struct xfs_inode *dp, struct xfs_name *name),
+	TP_ARGS(dp, name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, dp_ino)
+		__dynamic_array(char, name, name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(dp)->i_sb->s_dev;
+		__entry->dp_ino = dp->i_ino;
+		memcpy(__get_str(name), name->name, name->len);
+	),
+	TP_printk("dev %d:%d dp ino 0x%llx name %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dp_ino,
+		  __get_str(name))
+)
+
+#define DEFINE_NAMESPACE_EVENT(name) \
+DEFINE_EVENT(xfs_namespace_class, name, \
+	TP_PROTO(struct xfs_inode *dp, struct xfs_name *name), \
+	TP_ARGS(dp, name))
+DEFINE_NAMESPACE_EVENT(xfs_remove);
+DEFINE_NAMESPACE_EVENT(xfs_link);
+DEFINE_NAMESPACE_EVENT(xfs_lookup);
+DEFINE_NAMESPACE_EVENT(xfs_create);
+DEFINE_NAMESPACE_EVENT(xfs_symlink);
+
+TRACE_EVENT(xfs_rename,
+	TP_PROTO(struct xfs_inode *src_dp, struct xfs_inode *target_dp,
+		 struct xfs_name *src_name, struct xfs_name *target_name),
+	TP_ARGS(src_dp, target_dp, src_name, target_name),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, src_dp_ino)
+		__field(xfs_ino_t, target_dp_ino)
+		__dynamic_array(char, src_name, src_name->len)
+		__dynamic_array(char, target_name, target_name->len)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(src_dp)->i_sb->s_dev;
+		__entry->src_dp_ino = src_dp->i_ino;
+		__entry->target_dp_ino = target_dp->i_ino;
+		memcpy(__get_str(src_name), src_name->name, src_name->len);
+		memcpy(__get_str(target_name), target_name->name, target_name->len);
+	),
+	TP_printk("dev %d:%d src dp ino 0x%llx target dp ino 0x%llx"
+		  " src name %s target name %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->src_dp_ino,
+		  __entry->target_dp_ino,
+		  __get_str(src_name),
+		  __get_str(target_name))
+)
+
+DECLARE_EVENT_CLASS(xfs_dquot_class,
+	TP_PROTO(struct xfs_dquot *dqp),
+	TP_ARGS(dqp),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(u32, id)
+		__field(unsigned, flags)
+		__field(unsigned, nrefs)
+		__field(unsigned long long, res_bcount)
+		__field(unsigned long long, bcount)
+		__field(unsigned long long, icount)
+		__field(unsigned long long, blk_hardlimit)
+		__field(unsigned long long, blk_softlimit)
+		__field(unsigned long long, ino_hardlimit)
+		__field(unsigned long long, ino_softlimit)
+	), \
+	TP_fast_assign(
+		__entry->dev = dqp->q_mount->m_super->s_dev;
+		__entry->id = be32_to_cpu(dqp->q_core.d_id);
+		__entry->flags = dqp->dq_flags;
+		__entry->nrefs = dqp->q_nrefs;
+		__entry->res_bcount = dqp->q_res_bcount;
+		__entry->bcount = be64_to_cpu(dqp->q_core.d_bcount);
+		__entry->icount = be64_to_cpu(dqp->q_core.d_icount);
+		__entry->blk_hardlimit =
+			be64_to_cpu(dqp->q_core.d_blk_hardlimit);
+		__entry->blk_softlimit =
+			be64_to_cpu(dqp->q_core.d_blk_softlimit);
+		__entry->ino_hardlimit =
+			be64_to_cpu(dqp->q_core.d_ino_hardlimit);
+		__entry->ino_softlimit =
+			be64_to_cpu(dqp->q_core.d_ino_softlimit);
+	),
+	TP_printk("dev %d:%d id 0x%x flags %s nrefs %u res_bc 0x%llx "
+		  "bcnt 0x%llx bhardlimit 0x%llx bsoftlimit 0x%llx "
+		  "icnt 0x%llx ihardlimit 0x%llx isoftlimit 0x%llx]",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->id,
+		  __print_flags(__entry->flags, "|", XFS_DQ_FLAGS),
+		  __entry->nrefs,
+		  __entry->res_bcount,
+		  __entry->bcount,
+		  __entry->blk_hardlimit,
+		  __entry->blk_softlimit,
+		  __entry->icount,
+		  __entry->ino_hardlimit,
+		  __entry->ino_softlimit)
+)
+
+#define DEFINE_DQUOT_EVENT(name) \
+DEFINE_EVENT(xfs_dquot_class, name, \
+	TP_PROTO(struct xfs_dquot *dqp), \
+	TP_ARGS(dqp))
+DEFINE_DQUOT_EVENT(xfs_dqadjust);
+DEFINE_DQUOT_EVENT(xfs_dqreclaim_want);
+DEFINE_DQUOT_EVENT(xfs_dqreclaim_dirty);
+DEFINE_DQUOT_EVENT(xfs_dqreclaim_unlink);
+DEFINE_DQUOT_EVENT(xfs_dqattach_found);
+DEFINE_DQUOT_EVENT(xfs_dqattach_get);
+DEFINE_DQUOT_EVENT(xfs_dqinit);
+DEFINE_DQUOT_EVENT(xfs_dqreuse);
+DEFINE_DQUOT_EVENT(xfs_dqalloc);
+DEFINE_DQUOT_EVENT(xfs_dqtobp_read);
+DEFINE_DQUOT_EVENT(xfs_dqread);
+DEFINE_DQUOT_EVENT(xfs_dqread_fail);
+DEFINE_DQUOT_EVENT(xfs_dqlookup_found);
+DEFINE_DQUOT_EVENT(xfs_dqlookup_want);
+DEFINE_DQUOT_EVENT(xfs_dqlookup_freelist);
+DEFINE_DQUOT_EVENT(xfs_dqlookup_done);
+DEFINE_DQUOT_EVENT(xfs_dqget_hit);
+DEFINE_DQUOT_EVENT(xfs_dqget_miss);
+DEFINE_DQUOT_EVENT(xfs_dqput);
+DEFINE_DQUOT_EVENT(xfs_dqput_wait);
+DEFINE_DQUOT_EVENT(xfs_dqput_free);
+DEFINE_DQUOT_EVENT(xfs_dqrele);
+DEFINE_DQUOT_EVENT(xfs_dqflush);
+DEFINE_DQUOT_EVENT(xfs_dqflush_force);
+DEFINE_DQUOT_EVENT(xfs_dqflush_done);
+
+DECLARE_EVENT_CLASS(xfs_loggrant_class,
+	TP_PROTO(struct log *log, struct xlog_ticket *tic),
+	TP_ARGS(log, tic),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned, trans_type)
+		__field(char, ocnt)
+		__field(char, cnt)
+		__field(int, curr_res)
+		__field(int, unit_res)
+		__field(unsigned int, flags)
+		__field(int, reserveq)
+		__field(int, writeq)
+		__field(int, grant_reserve_cycle)
+		__field(int, grant_reserve_bytes)
+		__field(int, grant_write_cycle)
+		__field(int, grant_write_bytes)
+		__field(int, curr_cycle)
+		__field(int, curr_block)
+		__field(xfs_lsn_t, tail_lsn)
+	),
+	TP_fast_assign(
+		__entry->dev = log->l_mp->m_super->s_dev;
+		__entry->trans_type = tic->t_trans_type;
+		__entry->ocnt = tic->t_ocnt;
+		__entry->cnt = tic->t_cnt;
+		__entry->curr_res = tic->t_curr_res;
+		__entry->unit_res = tic->t_unit_res;
+		__entry->flags = tic->t_flags;
+		__entry->reserveq = list_empty(&log->l_reserveq);
+		__entry->writeq = list_empty(&log->l_writeq);
+		xlog_crack_grant_head(&log->l_grant_reserve_head,
+				&__entry->grant_reserve_cycle,
+				&__entry->grant_reserve_bytes);
+		xlog_crack_grant_head(&log->l_grant_write_head,
+				&__entry->grant_write_cycle,
+				&__entry->grant_write_bytes);
+		__entry->curr_cycle = log->l_curr_cycle;
+		__entry->curr_block = log->l_curr_block;
+		__entry->tail_lsn = atomic64_read(&log->l_tail_lsn);
+	),
+	TP_printk("dev %d:%d type %s t_ocnt %u t_cnt %u t_curr_res %u "
+		  "t_unit_res %u t_flags %s reserveq %s "
+		  "writeq %s grant_reserve_cycle %d "
+		  "grant_reserve_bytes %d grant_write_cycle %d "
+		  "grant_write_bytes %d curr_cycle %d curr_block %d "
+		  "tail_cycle %d tail_block %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->trans_type, XFS_TRANS_TYPES),
+		  __entry->ocnt,
+		  __entry->cnt,
+		  __entry->curr_res,
+		  __entry->unit_res,
+		  __print_flags(__entry->flags, "|", XLOG_TIC_FLAGS),
+		  __entry->reserveq ? "empty" : "active",
+		  __entry->writeq ? "empty" : "active",
+		  __entry->grant_reserve_cycle,
+		  __entry->grant_reserve_bytes,
+		  __entry->grant_write_cycle,
+		  __entry->grant_write_bytes,
+		  __entry->curr_cycle,
+		  __entry->curr_block,
+		  CYCLE_LSN(__entry->tail_lsn),
+		  BLOCK_LSN(__entry->tail_lsn)
+	)
+)
+
+#define DEFINE_LOGGRANT_EVENT(name) \
+DEFINE_EVENT(xfs_loggrant_class, name, \
+	TP_PROTO(struct log *log, struct xlog_ticket *tic), \
+	TP_ARGS(log, tic))
+DEFINE_LOGGRANT_EVENT(xfs_log_done_nonperm);
+DEFINE_LOGGRANT_EVENT(xfs_log_done_perm);
+DEFINE_LOGGRANT_EVENT(xfs_log_reserve);
+DEFINE_LOGGRANT_EVENT(xfs_log_umount_write);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_enter);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_error);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_sleep);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake);
+DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake_up);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_enter);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_error);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_sleep);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake_up);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_enter);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_sub);
+DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_enter);
+DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_sub);
+
+DECLARE_EVENT_CLASS(xfs_log_item_class,
+	TP_PROTO(struct xfs_log_item *lip),
+	TP_ARGS(lip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(void *, lip)
+		__field(uint, type)
+		__field(uint, flags)
+		__field(xfs_lsn_t, lsn)
+	),
+	TP_fast_assign(
+		__entry->dev = lip->li_mountp->m_super->s_dev;
+		__entry->lip = lip;
+		__entry->type = lip->li_type;
+		__entry->flags = lip->li_flags;
+		__entry->lsn = lip->li_lsn;
+	),
+	TP_printk("dev %d:%d lip 0x%p lsn %d/%d type %s flags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->lip,
+		  CYCLE_LSN(__entry->lsn), BLOCK_LSN(__entry->lsn),
+		  __print_symbolic(__entry->type, XFS_LI_TYPE_DESC),
+		  __print_flags(__entry->flags, "|", XFS_LI_FLAGS))
+)
+
+#define DEFINE_LOG_ITEM_EVENT(name) \
+DEFINE_EVENT(xfs_log_item_class, name, \
+	TP_PROTO(struct xfs_log_item *lip), \
+	TP_ARGS(lip))
+DEFINE_LOG_ITEM_EVENT(xfs_ail_push);
+DEFINE_LOG_ITEM_EVENT(xfs_ail_pushbuf);
+DEFINE_LOG_ITEM_EVENT(xfs_ail_pushbuf_pinned);
+DEFINE_LOG_ITEM_EVENT(xfs_ail_pinned);
+DEFINE_LOG_ITEM_EVENT(xfs_ail_locked);
+
+
+DECLARE_EVENT_CLASS(xfs_file_class,
+	TP_PROTO(struct xfs_inode *ip, size_t count, loff_t offset, int flags),
+	TP_ARGS(ip, count, offset, flags),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_fsize_t, size)
+		__field(xfs_fsize_t, new_size)
+		__field(loff_t, offset)
+		__field(size_t, count)
+		__field(int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->size = ip->i_d.di_size;
+		__entry->new_size = ip->i_new_size;
+		__entry->offset = offset;
+		__entry->count = count;
+		__entry->flags = flags;
+	),
+	TP_printk("dev %d:%d ino 0x%llx size 0x%llx new_size 0x%llx "
+		  "offset 0x%llx count 0x%zx ioflags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->new_size,
+		  __entry->offset,
+		  __entry->count,
+		  __print_flags(__entry->flags, "|", XFS_IO_FLAGS))
+)
+
+#define DEFINE_RW_EVENT(name)		\
+DEFINE_EVENT(xfs_file_class, name,	\
+	TP_PROTO(struct xfs_inode *ip, size_t count, loff_t offset, int flags),	\
+	TP_ARGS(ip, count, offset, flags))
+DEFINE_RW_EVENT(xfs_file_read);
+DEFINE_RW_EVENT(xfs_file_buffered_write);
+DEFINE_RW_EVENT(xfs_file_direct_write);
+DEFINE_RW_EVENT(xfs_file_splice_read);
+DEFINE_RW_EVENT(xfs_file_splice_write);
+
+DECLARE_EVENT_CLASS(xfs_page_class,
+	TP_PROTO(struct inode *inode, struct page *page, unsigned long off),
+	TP_ARGS(inode, page, off),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(pgoff_t, pgoff)
+		__field(loff_t, size)
+		__field(unsigned long, offset)
+		__field(int, delalloc)
+		__field(int, unwritten)
+	),
+	TP_fast_assign(
+		int delalloc = -1, unwritten = -1;
+
+		if (page_has_buffers(page))
+			xfs_count_page_state(page, &delalloc, &unwritten);
+		__entry->dev = inode->i_sb->s_dev;
+		__entry->ino = XFS_I(inode)->i_ino;
+		__entry->pgoff = page_offset(page);
+		__entry->size = i_size_read(inode);
+		__entry->offset = off;
+		__entry->delalloc = delalloc;
+		__entry->unwritten = unwritten;
+	),
+	TP_printk("dev %d:%d ino 0x%llx pgoff 0x%lx size 0x%llx offset %lx "
+		  "delalloc %d unwritten %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->pgoff,
+		  __entry->size,
+		  __entry->offset,
+		  __entry->delalloc,
+		  __entry->unwritten)
+)
+
+#define DEFINE_PAGE_EVENT(name)		\
+DEFINE_EVENT(xfs_page_class, name,	\
+	TP_PROTO(struct inode *inode, struct page *page, unsigned long off),	\
+	TP_ARGS(inode, page, off))
+DEFINE_PAGE_EVENT(xfs_writepage);
+DEFINE_PAGE_EVENT(xfs_releasepage);
+DEFINE_PAGE_EVENT(xfs_invalidatepage);
+
+DECLARE_EVENT_CLASS(xfs_imap_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_off_t offset, ssize_t count,
+		 int type, struct xfs_bmbt_irec *irec),
+	TP_ARGS(ip, offset, count, type, irec),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(loff_t, size)
+		__field(loff_t, new_size)
+		__field(loff_t, offset)
+		__field(size_t, count)
+		__field(int, type)
+		__field(xfs_fileoff_t, startoff)
+		__field(xfs_fsblock_t, startblock)
+		__field(xfs_filblks_t, blockcount)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->size = ip->i_d.di_size;
+		__entry->new_size = ip->i_new_size;
+		__entry->offset = offset;
+		__entry->count = count;
+		__entry->type = type;
+		__entry->startoff = irec ? irec->br_startoff : 0;
+		__entry->startblock = irec ? irec->br_startblock : 0;
+		__entry->blockcount = irec ? irec->br_blockcount : 0;
+	),
+	TP_printk("dev %d:%d ino 0x%llx size 0x%llx new_size 0x%llx "
+		  "offset 0x%llx count %zd type %s "
+		  "startoff 0x%llx startblock %lld blockcount 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->new_size,
+		  __entry->offset,
+		  __entry->count,
+		  __print_symbolic(__entry->type, XFS_IO_TYPES),
+		  __entry->startoff,
+		  (__int64_t)__entry->startblock,
+		  __entry->blockcount)
+)
+
+#define DEFINE_IOMAP_EVENT(name)	\
+DEFINE_EVENT(xfs_imap_class, name,	\
+	TP_PROTO(struct xfs_inode *ip, xfs_off_t offset, ssize_t count,	\
+		 int type, struct xfs_bmbt_irec *irec),		\
+	TP_ARGS(ip, offset, count, type, irec))
+DEFINE_IOMAP_EVENT(xfs_map_blocks_found);
+DEFINE_IOMAP_EVENT(xfs_map_blocks_alloc);
+DEFINE_IOMAP_EVENT(xfs_get_blocks_found);
+DEFINE_IOMAP_EVENT(xfs_get_blocks_alloc);
+
+DECLARE_EVENT_CLASS(xfs_simple_io_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_off_t offset, ssize_t count),
+	TP_ARGS(ip, offset, count),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(loff_t, isize)
+		__field(loff_t, disize)
+		__field(loff_t, new_size)
+		__field(loff_t, offset)
+		__field(size_t, count)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->isize = ip->i_size;
+		__entry->disize = ip->i_d.di_size;
+		__entry->new_size = ip->i_new_size;
+		__entry->offset = offset;
+		__entry->count = count;
+	),
+	TP_printk("dev %d:%d ino 0x%llx isize 0x%llx disize 0x%llx new_size 0x%llx "
+		  "offset 0x%llx count %zd",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->isize,
+		  __entry->disize,
+		  __entry->new_size,
+		  __entry->offset,
+		  __entry->count)
+);
+
+#define DEFINE_SIMPLE_IO_EVENT(name)	\
+DEFINE_EVENT(xfs_simple_io_class, name,	\
+	TP_PROTO(struct xfs_inode *ip, xfs_off_t offset, ssize_t count),	\
+	TP_ARGS(ip, offset, count))
+DEFINE_SIMPLE_IO_EVENT(xfs_delalloc_enospc);
+DEFINE_SIMPLE_IO_EVENT(xfs_unwritten_convert);
+DEFINE_SIMPLE_IO_EVENT(xfs_get_blocks_notfound);
+DEFINE_SIMPLE_IO_EVENT(xfs_setfilesize);
+
+DECLARE_EVENT_CLASS(xfs_itrunc_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_fsize_t new_size),
+	TP_ARGS(ip, new_size),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_fsize_t, size)
+		__field(xfs_fsize_t, new_size)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->size = ip->i_d.di_size;
+		__entry->new_size = new_size;
+	),
+	TP_printk("dev %d:%d ino 0x%llx size 0x%llx new_size 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->new_size)
+)
+
+#define DEFINE_ITRUNC_EVENT(name) \
+DEFINE_EVENT(xfs_itrunc_class, name, \
+	TP_PROTO(struct xfs_inode *ip, xfs_fsize_t new_size), \
+	TP_ARGS(ip, new_size))
+DEFINE_ITRUNC_EVENT(xfs_itruncate_data_start);
+DEFINE_ITRUNC_EVENT(xfs_itruncate_data_end);
+
+TRACE_EVENT(xfs_pagecache_inval,
+	TP_PROTO(struct xfs_inode *ip, xfs_off_t start, xfs_off_t finish),
+	TP_ARGS(ip, start, finish),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_fsize_t, size)
+		__field(xfs_off_t, start)
+		__field(xfs_off_t, finish)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->size = ip->i_d.di_size;
+		__entry->start = start;
+		__entry->finish = finish;
+	),
+	TP_printk("dev %d:%d ino 0x%llx size 0x%llx start 0x%llx finish 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->start,
+		  __entry->finish)
+);
+
+TRACE_EVENT(xfs_bunmap,
+	TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len,
+		 int flags, unsigned long caller_ip),
+	TP_ARGS(ip, bno, len, flags, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_fsize_t, size)
+		__field(xfs_fileoff_t, bno)
+		__field(xfs_filblks_t, len)
+		__field(unsigned long, caller_ip)
+		__field(int, flags)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->size = ip->i_d.di_size;
+		__entry->bno = bno;
+		__entry->len = len;
+		__entry->caller_ip = caller_ip;
+		__entry->flags = flags;
+	),
+	TP_printk("dev %d:%d ino 0x%llx size 0x%llx bno 0x%llx len 0x%llx"
+		  "flags %s caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->bno,
+		  __entry->len,
+		  __print_flags(__entry->flags, "|", XFS_BMAPI_FLAGS),
+		  (void *)__entry->caller_ip)
+
+);
+
+DECLARE_EVENT_CLASS(xfs_busy_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len),
+	TP_ARGS(mp, agno, agbno, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len)
+);
+#define DEFINE_BUSY_EVENT(name) \
+DEFINE_EVENT(xfs_busy_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 xfs_agblock_t agbno, xfs_extlen_t len), \
+	TP_ARGS(mp, agno, agbno, len))
+DEFINE_BUSY_EVENT(xfs_alloc_busy);
+DEFINE_BUSY_EVENT(xfs_alloc_busy_enomem);
+DEFINE_BUSY_EVENT(xfs_alloc_busy_force);
+DEFINE_BUSY_EVENT(xfs_alloc_busy_reuse);
+DEFINE_BUSY_EVENT(xfs_alloc_busy_clear);
+
+TRACE_EVENT(xfs_alloc_busy_trim,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len,
+		 xfs_agblock_t tbno, xfs_extlen_t tlen),
+	TP_ARGS(mp, agno, agbno, len, tbno, tlen),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+		__field(xfs_agblock_t, tbno)
+		__field(xfs_extlen_t, tlen)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+		__entry->tbno = tbno;
+		__entry->tlen = tlen;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u tbno %u tlen %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->tbno,
+		  __entry->tlen)
+);
+
+TRACE_EVENT(xfs_trans_commit_lsn,
+	TP_PROTO(struct xfs_trans *trans),
+	TP_ARGS(trans),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(struct xfs_trans *, tp)
+		__field(xfs_lsn_t, lsn)
+	),
+	TP_fast_assign(
+		__entry->dev = trans->t_mountp->m_super->s_dev;
+		__entry->tp = trans;
+		__entry->lsn = trans->t_commit_lsn;
+	),
+	TP_printk("dev %d:%d trans 0x%p commit_lsn 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tp,
+		  __entry->lsn)
+);
+
+TRACE_EVENT(xfs_agf,
+	TP_PROTO(struct xfs_mount *mp, struct xfs_agf *agf, int flags,
+		 unsigned long caller_ip),
+	TP_ARGS(mp, agf, flags, caller_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(int, flags)
+		__field(__u32, length)
+		__field(__u32, bno_root)
+		__field(__u32, cnt_root)
+		__field(__u32, bno_level)
+		__field(__u32, cnt_level)
+		__field(__u32, flfirst)
+		__field(__u32, fllast)
+		__field(__u32, flcount)
+		__field(__u32, freeblks)
+		__field(__u32, longest)
+		__field(unsigned long, caller_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = be32_to_cpu(agf->agf_seqno),
+		__entry->flags = flags;
+		__entry->length = be32_to_cpu(agf->agf_length),
+		__entry->bno_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]),
+		__entry->cnt_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]),
+		__entry->bno_level =
+				be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]),
+		__entry->cnt_level =
+				be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]),
+		__entry->flfirst = be32_to_cpu(agf->agf_flfirst),
+		__entry->fllast = be32_to_cpu(agf->agf_fllast),
+		__entry->flcount = be32_to_cpu(agf->agf_flcount),
+		__entry->freeblks = be32_to_cpu(agf->agf_freeblks),
+		__entry->longest = be32_to_cpu(agf->agf_longest);
+		__entry->caller_ip = caller_ip;
+	),
+	TP_printk("dev %d:%d agno %u flags %s length %u roots b %u c %u "
+		  "levels b %u c %u flfirst %u fllast %u flcount %u "
+		  "freeblks %u longest %u caller %pf",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __print_flags(__entry->flags, "|", XFS_AGF_FLAGS),
+		  __entry->length,
+		  __entry->bno_root,
+		  __entry->cnt_root,
+		  __entry->bno_level,
+		  __entry->cnt_level,
+		  __entry->flfirst,
+		  __entry->fllast,
+		  __entry->flcount,
+		  __entry->freeblks,
+		  __entry->longest,
+		  (void *)__entry->caller_ip)
+);
+
+TRACE_EVENT(xfs_free_extent,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno,
+		 xfs_extlen_t len, bool isfl, int haveleft, int haveright),
+	TP_ARGS(mp, agno, agbno, len, isfl, haveleft, haveright),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+		__field(int, isfl)
+		__field(int, haveleft)
+		__field(int, haveright)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+		__entry->isfl = isfl;
+		__entry->haveleft = haveleft;
+		__entry->haveright = haveright;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u isfl %d %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len,
+		  __entry->isfl,
+		  __entry->haveleft ?
+			(__entry->haveright ? "both" : "left") :
+			(__entry->haveright ? "right" : "none"))
+
+);
+
+DECLARE_EVENT_CLASS(xfs_alloc_class,
+	TP_PROTO(struct xfs_alloc_arg *args),
+	TP_ARGS(args),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, minlen)
+		__field(xfs_extlen_t, maxlen)
+		__field(xfs_extlen_t, mod)
+		__field(xfs_extlen_t, prod)
+		__field(xfs_extlen_t, minleft)
+		__field(xfs_extlen_t, total)
+		__field(xfs_extlen_t, alignment)
+		__field(xfs_extlen_t, minalignslop)
+		__field(xfs_extlen_t, len)
+		__field(short, type)
+		__field(short, otype)
+		__field(char, wasdel)
+		__field(char, wasfromfl)
+		__field(char, isfl)
+		__field(char, userdata)
+		__field(xfs_fsblock_t, firstblock)
+	),
+	TP_fast_assign(
+		__entry->dev = args->mp->m_super->s_dev;
+		__entry->agno = args->agno;
+		__entry->agbno = args->agbno;
+		__entry->minlen = args->minlen;
+		__entry->maxlen = args->maxlen;
+		__entry->mod = args->mod;
+		__entry->prod = args->prod;
+		__entry->minleft = args->minleft;
+		__entry->total = args->total;
+		__entry->alignment = args->alignment;
+		__entry->minalignslop = args->minalignslop;
+		__entry->len = args->len;
+		__entry->type = args->type;
+		__entry->otype = args->otype;
+		__entry->wasdel = args->wasdel;
+		__entry->wasfromfl = args->wasfromfl;
+		__entry->isfl = args->isfl;
+		__entry->userdata = args->userdata;
+		__entry->firstblock = args->firstblock;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u minlen %u maxlen %u mod %u "
+		  "prod %u minleft %u total %u alignment %u minalignslop %u "
+		  "len %u type %s otype %s wasdel %d wasfromfl %d isfl %d "
+		  "userdata %d firstblock 0x%llx",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->minlen,
+		  __entry->maxlen,
+		  __entry->mod,
+		  __entry->prod,
+		  __entry->minleft,
+		  __entry->total,
+		  __entry->alignment,
+		  __entry->minalignslop,
+		  __entry->len,
+		  __print_symbolic(__entry->type, XFS_ALLOC_TYPES),
+		  __print_symbolic(__entry->otype, XFS_ALLOC_TYPES),
+		  __entry->wasdel,
+		  __entry->wasfromfl,
+		  __entry->isfl,
+		  __entry->userdata,
+		  (unsigned long long)__entry->firstblock)
+)
+
+#define DEFINE_ALLOC_EVENT(name) \
+DEFINE_EVENT(xfs_alloc_class, name, \
+	TP_PROTO(struct xfs_alloc_arg *args), \
+	TP_ARGS(args))
+DEFINE_ALLOC_EVENT(xfs_alloc_exact_done);
+DEFINE_ALLOC_EVENT(xfs_alloc_exact_notfound);
+DEFINE_ALLOC_EVENT(xfs_alloc_exact_error);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_nominleft);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_first);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_greater);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_lesser);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_error);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_noentry);
+DEFINE_ALLOC_EVENT(xfs_alloc_near_busy);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_neither);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_noentry);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_nominleft);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_done);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_error);
+DEFINE_ALLOC_EVENT(xfs_alloc_size_busy);
+DEFINE_ALLOC_EVENT(xfs_alloc_small_freelist);
+DEFINE_ALLOC_EVENT(xfs_alloc_small_notenough);
+DEFINE_ALLOC_EVENT(xfs_alloc_small_done);
+DEFINE_ALLOC_EVENT(xfs_alloc_small_error);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_badargs);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_nofix);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_noagbp);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_loopfailed);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_allfailed);
+
+DECLARE_EVENT_CLASS(xfs_dir2_class,
+	TP_PROTO(struct xfs_da_args *args),
+	TP_ARGS(args),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__dynamic_array(char, name, args->namelen)
+		__field(int, namelen)
+		__field(xfs_dahash_t, hashval)
+		__field(xfs_ino_t, inumber)
+		__field(int, op_flags)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(args->dp)->i_sb->s_dev;
+		__entry->ino = args->dp->i_ino;
+		if (args->namelen)
+			memcpy(__get_str(name), args->name, args->namelen);
+		__entry->namelen = args->namelen;
+		__entry->hashval = args->hashval;
+		__entry->inumber = args->inumber;
+		__entry->op_flags = args->op_flags;
+	),
+	TP_printk("dev %d:%d ino 0x%llx name %.*s namelen %d hashval 0x%x "
+		  "inumber 0x%llx op_flags %s",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->namelen,
+		  __entry->namelen ? __get_str(name) : NULL,
+		  __entry->namelen,
+		  __entry->hashval,
+		  __entry->inumber,
+		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS))
+)
+
+#define DEFINE_DIR2_EVENT(name) \
+DEFINE_EVENT(xfs_dir2_class, name, \
+	TP_PROTO(struct xfs_da_args *args), \
+	TP_ARGS(args))
+DEFINE_DIR2_EVENT(xfs_dir2_sf_addname);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_create);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_lookup);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_replace);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_removename);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_toino4);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_toino8);
+DEFINE_DIR2_EVENT(xfs_dir2_sf_to_block);
+DEFINE_DIR2_EVENT(xfs_dir2_block_addname);
+DEFINE_DIR2_EVENT(xfs_dir2_block_lookup);
+DEFINE_DIR2_EVENT(xfs_dir2_block_replace);
+DEFINE_DIR2_EVENT(xfs_dir2_block_removename);
+DEFINE_DIR2_EVENT(xfs_dir2_block_to_sf);
+DEFINE_DIR2_EVENT(xfs_dir2_block_to_leaf);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_addname);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_lookup);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_replace);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_removename);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_to_block);
+DEFINE_DIR2_EVENT(xfs_dir2_leaf_to_node);
+DEFINE_DIR2_EVENT(xfs_dir2_node_addname);
+DEFINE_DIR2_EVENT(xfs_dir2_node_lookup);
+DEFINE_DIR2_EVENT(xfs_dir2_node_replace);
+DEFINE_DIR2_EVENT(xfs_dir2_node_removename);
+DEFINE_DIR2_EVENT(xfs_dir2_node_to_leaf);
+
+DECLARE_EVENT_CLASS(xfs_dir2_space_class,
+	TP_PROTO(struct xfs_da_args *args, int idx),
+	TP_ARGS(args, idx),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, op_flags)
+		__field(int, idx)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(args->dp)->i_sb->s_dev;
+		__entry->ino = args->dp->i_ino;
+		__entry->op_flags = args->op_flags;
+		__entry->idx = idx;
+	),
+	TP_printk("dev %d:%d ino 0x%llx op_flags %s index %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS),
+		  __entry->idx)
+)
+
+#define DEFINE_DIR2_SPACE_EVENT(name) \
+DEFINE_EVENT(xfs_dir2_space_class, name, \
+	TP_PROTO(struct xfs_da_args *args, int idx), \
+	TP_ARGS(args, idx))
+DEFINE_DIR2_SPACE_EVENT(xfs_dir2_leafn_add);
+DEFINE_DIR2_SPACE_EVENT(xfs_dir2_leafn_remove);
+DEFINE_DIR2_SPACE_EVENT(xfs_dir2_grow_inode);
+DEFINE_DIR2_SPACE_EVENT(xfs_dir2_shrink_inode);
+
+TRACE_EVENT(xfs_dir2_leafn_moveents,
+	TP_PROTO(struct xfs_da_args *args, int src_idx, int dst_idx, int count),
+	TP_ARGS(args, src_idx, dst_idx, count),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, op_flags)
+		__field(int, src_idx)
+		__field(int, dst_idx)
+		__field(int, count)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(args->dp)->i_sb->s_dev;
+		__entry->ino = args->dp->i_ino;
+		__entry->op_flags = args->op_flags;
+		__entry->src_idx = src_idx;
+		__entry->dst_idx = dst_idx;
+		__entry->count = count;
+	),
+	TP_printk("dev %d:%d ino 0x%llx op_flags %s "
+		  "src_idx %d dst_idx %d count %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS),
+		  __entry->src_idx,
+		  __entry->dst_idx,
+		  __entry->count)
+);
+
+#define XFS_SWAPEXT_INODES \
+	{ 0,	"target" }, \
+	{ 1,	"temp" }
+
+#define XFS_INODE_FORMAT_STR \
+	{ 0,	"invalid" }, \
+	{ 1,	"local" }, \
+	{ 2,	"extent" }, \
+	{ 3,	"btree" }
+
+DECLARE_EVENT_CLASS(xfs_swap_extent_class,
+	TP_PROTO(struct xfs_inode *ip, int which),
+	TP_ARGS(ip, which),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(int, which)
+		__field(xfs_ino_t, ino)
+		__field(int, format)
+		__field(int, nex)
+		__field(int, max_nex)
+		__field(int, broot_size)
+		__field(int, fork_off)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->which = which;
+		__entry->ino = ip->i_ino;
+		__entry->format = ip->i_d.di_format;
+		__entry->nex = ip->i_d.di_nextents;
+		__entry->max_nex = ip->i_df.if_ext_max;
+		__entry->broot_size = ip->i_df.if_broot_bytes;
+		__entry->fork_off = XFS_IFORK_BOFF(ip);
+	),
+	TP_printk("dev %d:%d ino 0x%llx (%s), %s format, num_extents %d, "
+		  "Max in-fork extents %d, broot size %d, fork offset %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __print_symbolic(__entry->which, XFS_SWAPEXT_INODES),
+		  __print_symbolic(__entry->format, XFS_INODE_FORMAT_STR),
+		  __entry->nex,
+		  __entry->max_nex,
+		  __entry->broot_size,
+		  __entry->fork_off)
+)
+
+#define DEFINE_SWAPEXT_EVENT(name) \
+DEFINE_EVENT(xfs_swap_extent_class, name, \
+	TP_PROTO(struct xfs_inode *ip, int which), \
+	TP_ARGS(ip, which))
+
+DEFINE_SWAPEXT_EVENT(xfs_swap_extent_before);
+DEFINE_SWAPEXT_EVENT(xfs_swap_extent_after);
+
+DECLARE_EVENT_CLASS(xfs_log_recover_item_class,
+	TP_PROTO(struct log *log, struct xlog_recover *trans,
+		struct xlog_recover_item *item, int pass),
+	TP_ARGS(log, trans, item, pass),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned long, item)
+		__field(xlog_tid_t, tid)
+		__field(int, type)
+		__field(int, pass)
+		__field(int, count)
+		__field(int, total)
+	),
+	TP_fast_assign(
+		__entry->dev = log->l_mp->m_super->s_dev;
+		__entry->item = (unsigned long)item;
+		__entry->tid = trans->r_log_tid;
+		__entry->type = ITEM_TYPE(item);
+		__entry->pass = pass;
+		__entry->count = item->ri_cnt;
+		__entry->total = item->ri_total;
+	),
+	TP_printk("dev %d:%d trans 0x%x, pass %d, item 0x%p, item type %s "
+		  "item region count/total %d/%d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->tid,
+		  __entry->pass,
+		  (void *)__entry->item,
+		  __print_symbolic(__entry->type, XFS_LI_TYPE_DESC),
+		  __entry->count,
+		  __entry->total)
+)
+
+#define DEFINE_LOG_RECOVER_ITEM(name) \
+DEFINE_EVENT(xfs_log_recover_item_class, name, \
+	TP_PROTO(struct log *log, struct xlog_recover *trans, \
+		struct xlog_recover_item *item, int pass), \
+	TP_ARGS(log, trans, item, pass))
+
+DEFINE_LOG_RECOVER_ITEM(xfs_log_recover_item_add);
+DEFINE_LOG_RECOVER_ITEM(xfs_log_recover_item_add_cont);
+DEFINE_LOG_RECOVER_ITEM(xfs_log_recover_item_reorder_head);
+DEFINE_LOG_RECOVER_ITEM(xfs_log_recover_item_reorder_tail);
+DEFINE_LOG_RECOVER_ITEM(xfs_log_recover_item_recover);
+
+DECLARE_EVENT_CLASS(xfs_log_recover_buf_item_class,
+	TP_PROTO(struct log *log, struct xfs_buf_log_format *buf_f),
+	TP_ARGS(log, buf_f),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(__int64_t, blkno)
+		__field(unsigned short, len)
+		__field(unsigned short, flags)
+		__field(unsigned short, size)
+		__field(unsigned int, map_size)
+	),
+	TP_fast_assign(
+		__entry->dev = log->l_mp->m_super->s_dev;
+		__entry->blkno = buf_f->blf_blkno;
+		__entry->len = buf_f->blf_len;
+		__entry->flags = buf_f->blf_flags;
+		__entry->size = buf_f->blf_size;
+		__entry->map_size = buf_f->blf_map_size;
+	),
+	TP_printk("dev %d:%d blkno 0x%llx, len %u, flags 0x%x, size %d, "
+			"map_size %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->blkno,
+		  __entry->len,
+		  __entry->flags,
+		  __entry->size,
+		  __entry->map_size)
+)
+
+#define DEFINE_LOG_RECOVER_BUF_ITEM(name) \
+DEFINE_EVENT(xfs_log_recover_buf_item_class, name, \
+	TP_PROTO(struct log *log, struct xfs_buf_log_format *buf_f), \
+	TP_ARGS(log, buf_f))
+
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_not_cancel);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel_add);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel_ref_inc);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_recover);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_inode_buf);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_reg_buf);
+DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_dquot_buf);
+
+DECLARE_EVENT_CLASS(xfs_log_recover_ino_item_class,
+	TP_PROTO(struct log *log, struct xfs_inode_log_format *in_f),
+	TP_ARGS(log, in_f),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned short, size)
+		__field(int, fields)
+		__field(unsigned short, asize)
+		__field(unsigned short, dsize)
+		__field(__int64_t, blkno)
+		__field(int, len)
+		__field(int, boffset)
+	),
+	TP_fast_assign(
+		__entry->dev = log->l_mp->m_super->s_dev;
+		__entry->ino = in_f->ilf_ino;
+		__entry->size = in_f->ilf_size;
+		__entry->fields = in_f->ilf_fields;
+		__entry->asize = in_f->ilf_asize;
+		__entry->dsize = in_f->ilf_dsize;
+		__entry->blkno = in_f->ilf_blkno;
+		__entry->len = in_f->ilf_len;
+		__entry->boffset = in_f->ilf_boffset;
+	),
+	TP_printk("dev %d:%d ino 0x%llx, size %u, fields 0x%x, asize %d, "
+			"dsize %d, blkno 0x%llx, len %d, boffset %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->size,
+		  __entry->fields,
+		  __entry->asize,
+		  __entry->dsize,
+		  __entry->blkno,
+		  __entry->len,
+		  __entry->boffset)
+)
+#define DEFINE_LOG_RECOVER_INO_ITEM(name) \
+DEFINE_EVENT(xfs_log_recover_ino_item_class, name, \
+	TP_PROTO(struct log *log, struct xfs_inode_log_format *in_f), \
+	TP_ARGS(log, in_f))
+
+DEFINE_LOG_RECOVER_INO_ITEM(xfs_log_recover_inode_recover);
+DEFINE_LOG_RECOVER_INO_ITEM(xfs_log_recover_inode_cancel);
+DEFINE_LOG_RECOVER_INO_ITEM(xfs_log_recover_inode_skip);
+
+DECLARE_EVENT_CLASS(xfs_discard_class,
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno,
+		 xfs_agblock_t agbno, xfs_extlen_t len),
+	TP_ARGS(mp, agno, agbno, len),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(xfs_extlen_t, len)
+	),
+	TP_fast_assign(
+		__entry->dev = mp->m_super->s_dev;
+		__entry->agno = agno;
+		__entry->agbno = agbno;
+		__entry->len = len;
+	),
+	TP_printk("dev %d:%d agno %u agbno %u len %u\n",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->len)
+)
+
+#define DEFINE_DISCARD_EVENT(name) \
+DEFINE_EVENT(xfs_discard_class, name, \
+	TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \
+		 xfs_agblock_t agbno, xfs_extlen_t len), \
+	TP_ARGS(mp, agno, agbno, len))
+DEFINE_DISCARD_EVENT(xfs_discard_extent);
+DEFINE_DISCARD_EVENT(xfs_discard_toosmall);
+DEFINE_DISCARD_EVENT(xfs_discard_exclude);
+DEFINE_DISCARD_EVENT(xfs_discard_busy);
+
+#endif /* _TRACE_XFS_H */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#define TRACE_INCLUDE_FILE xfs_trace
+#include <trace/define_trace.h>
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
new file mode 100644
index 0000000..4d00ee6
--- /dev/null
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -0,0 +1,890 @@
+/*
+ * Copyright (c) 2000-2002 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_alloc.h"
+#include "xfs_quota.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_itable.h"
+#include "xfs_bmap.h"
+#include "xfs_rtalloc.h"
+#include "xfs_error.h"
+#include "xfs_attr.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_priv.h"
+#include "xfs_qm.h"
+
+STATIC void	xfs_trans_alloc_dqinfo(xfs_trans_t *);
+
+/*
+ * Add the locked dquot to the transaction.
+ * The dquot must be locked, and it cannot be associated with any
+ * transaction.
+ */
+void
+xfs_trans_dqjoin(
+	xfs_trans_t	*tp,
+	xfs_dquot_t	*dqp)
+{
+	ASSERT(dqp->q_transp != tp);
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+	ASSERT(dqp->q_logitem.qli_dquot == dqp);
+
+	/*
+	 * Get a log_item_desc to point at the new item.
+	 */
+	xfs_trans_add_item(tp, &dqp->q_logitem.qli_item);
+
+	/*
+	 * Initialize d_transp so we can later determine if this dquot is
+	 * associated with this transaction.
+	 */
+	dqp->q_transp = tp;
+}
+
+
+/*
+ * This is called to mark the dquot as needing
+ * to be logged when the transaction is committed.  The dquot must
+ * already be associated with the given transaction.
+ * Note that it marks the entire transaction as dirty. In the ordinary
+ * case, this gets called via xfs_trans_commit, after the transaction
+ * is already dirty. However, there's nothing stop this from getting
+ * called directly, as done by xfs_qm_scall_setqlim. Hence, the TRANS_DIRTY
+ * flag.
+ */
+void
+xfs_trans_log_dquot(
+	xfs_trans_t	*tp,
+	xfs_dquot_t	*dqp)
+{
+	ASSERT(dqp->q_transp == tp);
+	ASSERT(XFS_DQ_IS_LOCKED(dqp));
+
+	tp->t_flags |= XFS_TRANS_DIRTY;
+	dqp->q_logitem.qli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+}
+
+/*
+ * Carry forward whatever is left of the quota blk reservation to
+ * the spanky new transaction
+ */
+void
+xfs_trans_dup_dqinfo(
+	xfs_trans_t	*otp,
+	xfs_trans_t	*ntp)
+{
+	xfs_dqtrx_t	*oq, *nq;
+	int		i,j;
+	xfs_dqtrx_t	*oqa, *nqa;
+
+	if (!otp->t_dqinfo)
+		return;
+
+	xfs_trans_alloc_dqinfo(ntp);
+	oqa = otp->t_dqinfo->dqa_usrdquots;
+	nqa = ntp->t_dqinfo->dqa_usrdquots;
+
+	/*
+	 * Because the quota blk reservation is carried forward,
+	 * it is also necessary to carry forward the DQ_DIRTY flag.
+	 */
+	if(otp->t_flags & XFS_TRANS_DQ_DIRTY)
+		ntp->t_flags |= XFS_TRANS_DQ_DIRTY;
+
+	for (j = 0; j < 2; j++) {
+		for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+			if (oqa[i].qt_dquot == NULL)
+				break;
+			oq = &oqa[i];
+			nq = &nqa[i];
+
+			nq->qt_dquot = oq->qt_dquot;
+			nq->qt_bcount_delta = nq->qt_icount_delta = 0;
+			nq->qt_rtbcount_delta = 0;
+
+			/*
+			 * Transfer whatever is left of the reservations.
+			 */
+			nq->qt_blk_res = oq->qt_blk_res - oq->qt_blk_res_used;
+			oq->qt_blk_res = oq->qt_blk_res_used;
+
+			nq->qt_rtblk_res = oq->qt_rtblk_res -
+				oq->qt_rtblk_res_used;
+			oq->qt_rtblk_res = oq->qt_rtblk_res_used;
+
+			nq->qt_ino_res = oq->qt_ino_res - oq->qt_ino_res_used;
+			oq->qt_ino_res = oq->qt_ino_res_used;
+
+		}
+		oqa = otp->t_dqinfo->dqa_grpdquots;
+		nqa = ntp->t_dqinfo->dqa_grpdquots;
+	}
+}
+
+/*
+ * Wrap around mod_dquot to account for both user and group quotas.
+ */
+void
+xfs_trans_mod_dquot_byino(
+	xfs_trans_t	*tp,
+	xfs_inode_t	*ip,
+	uint		field,
+	long		delta)
+{
+	xfs_mount_t	*mp = tp->t_mountp;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) ||
+	    !XFS_IS_QUOTA_ON(mp) ||
+	    ip->i_ino == mp->m_sb.sb_uquotino ||
+	    ip->i_ino == mp->m_sb.sb_gquotino)
+		return;
+
+	if (tp->t_dqinfo == NULL)
+		xfs_trans_alloc_dqinfo(tp);
+
+	if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot)
+		(void) xfs_trans_mod_dquot(tp, ip->i_udquot, field, delta);
+	if (XFS_IS_OQUOTA_ON(mp) && ip->i_gdquot)
+		(void) xfs_trans_mod_dquot(tp, ip->i_gdquot, field, delta);
+}
+
+STATIC xfs_dqtrx_t *
+xfs_trans_get_dqtrx(
+	xfs_trans_t	*tp,
+	xfs_dquot_t	*dqp)
+{
+	int		i;
+	xfs_dqtrx_t	*qa;
+
+	qa = XFS_QM_ISUDQ(dqp) ?
+		tp->t_dqinfo->dqa_usrdquots : tp->t_dqinfo->dqa_grpdquots;
+
+	for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+		if (qa[i].qt_dquot == NULL ||
+		    qa[i].qt_dquot == dqp)
+			return &qa[i];
+	}
+
+	return NULL;
+}
+
+/*
+ * Make the changes in the transaction structure.
+ * The moral equivalent to xfs_trans_mod_sb().
+ * We don't touch any fields in the dquot, so we don't care
+ * if it's locked or not (most of the time it won't be).
+ */
+void
+xfs_trans_mod_dquot(
+	xfs_trans_t	*tp,
+	xfs_dquot_t	*dqp,
+	uint		field,
+	long		delta)
+{
+	xfs_dqtrx_t	*qtrx;
+
+	ASSERT(tp);
+	ASSERT(XFS_IS_QUOTA_RUNNING(tp->t_mountp));
+	qtrx = NULL;
+
+	if (tp->t_dqinfo == NULL)
+		xfs_trans_alloc_dqinfo(tp);
+	/*
+	 * Find either the first free slot or the slot that belongs
+	 * to this dquot.
+	 */
+	qtrx = xfs_trans_get_dqtrx(tp, dqp);
+	ASSERT(qtrx);
+	if (qtrx->qt_dquot == NULL)
+		qtrx->qt_dquot = dqp;
+
+	switch (field) {
+
+		/*
+		 * regular disk blk reservation
+		 */
+	      case XFS_TRANS_DQ_RES_BLKS:
+		qtrx->qt_blk_res += (ulong)delta;
+		break;
+
+		/*
+		 * inode reservation
+		 */
+	      case XFS_TRANS_DQ_RES_INOS:
+		qtrx->qt_ino_res += (ulong)delta;
+		break;
+
+		/*
+		 * disk blocks used.
+		 */
+	      case XFS_TRANS_DQ_BCOUNT:
+		if (qtrx->qt_blk_res && delta > 0) {
+			qtrx->qt_blk_res_used += (ulong)delta;
+			ASSERT(qtrx->qt_blk_res >= qtrx->qt_blk_res_used);
+		}
+		qtrx->qt_bcount_delta += delta;
+		break;
+
+	      case XFS_TRANS_DQ_DELBCOUNT:
+		qtrx->qt_delbcnt_delta += delta;
+		break;
+
+		/*
+		 * Inode Count
+		 */
+	      case XFS_TRANS_DQ_ICOUNT:
+		if (qtrx->qt_ino_res && delta > 0) {
+			qtrx->qt_ino_res_used += (ulong)delta;
+			ASSERT(qtrx->qt_ino_res >= qtrx->qt_ino_res_used);
+		}
+		qtrx->qt_icount_delta += delta;
+		break;
+
+		/*
+		 * rtblk reservation
+		 */
+	      case XFS_TRANS_DQ_RES_RTBLKS:
+		qtrx->qt_rtblk_res += (ulong)delta;
+		break;
+
+		/*
+		 * rtblk count
+		 */
+	      case XFS_TRANS_DQ_RTBCOUNT:
+		if (qtrx->qt_rtblk_res && delta > 0) {
+			qtrx->qt_rtblk_res_used += (ulong)delta;
+			ASSERT(qtrx->qt_rtblk_res >= qtrx->qt_rtblk_res_used);
+		}
+		qtrx->qt_rtbcount_delta += delta;
+		break;
+
+	      case XFS_TRANS_DQ_DELRTBCOUNT:
+		qtrx->qt_delrtb_delta += delta;
+		break;
+
+	      default:
+		ASSERT(0);
+	}
+	tp->t_flags |= XFS_TRANS_DQ_DIRTY;
+}
+
+
+/*
+ * Given an array of dqtrx structures, lock all the dquots associated
+ * and join them to the transaction, provided they have been modified.
+ * We know that the highest number of dquots (of one type - usr OR grp),
+ * involved in a transaction is 2 and that both usr and grp combined - 3.
+ * So, we don't attempt to make this very generic.
+ */
+STATIC void
+xfs_trans_dqlockedjoin(
+	xfs_trans_t	*tp,
+	xfs_dqtrx_t	*q)
+{
+	ASSERT(q[0].qt_dquot != NULL);
+	if (q[1].qt_dquot == NULL) {
+		xfs_dqlock(q[0].qt_dquot);
+		xfs_trans_dqjoin(tp, q[0].qt_dquot);
+	} else {
+		ASSERT(XFS_QM_TRANS_MAXDQS == 2);
+		xfs_dqlock2(q[0].qt_dquot, q[1].qt_dquot);
+		xfs_trans_dqjoin(tp, q[0].qt_dquot);
+		xfs_trans_dqjoin(tp, q[1].qt_dquot);
+	}
+}
+
+
+/*
+ * Called by xfs_trans_commit() and similar in spirit to
+ * xfs_trans_apply_sb_deltas().
+ * Go thru all the dquots belonging to this transaction and modify the
+ * INCORE dquot to reflect the actual usages.
+ * Unreserve just the reservations done by this transaction.
+ * dquot is still left locked at exit.
+ */
+void
+xfs_trans_apply_dquot_deltas(
+	xfs_trans_t		*tp)
+{
+	int			i, j;
+	xfs_dquot_t		*dqp;
+	xfs_dqtrx_t		*qtrx, *qa;
+	xfs_disk_dquot_t	*d;
+	long			totalbdelta;
+	long			totalrtbdelta;
+
+	if (!(tp->t_flags & XFS_TRANS_DQ_DIRTY))
+		return;
+
+	ASSERT(tp->t_dqinfo);
+	qa = tp->t_dqinfo->dqa_usrdquots;
+	for (j = 0; j < 2; j++) {
+		if (qa[0].qt_dquot == NULL) {
+			qa = tp->t_dqinfo->dqa_grpdquots;
+			continue;
+		}
+
+		/*
+		 * Lock all of the dquots and join them to the transaction.
+		 */
+		xfs_trans_dqlockedjoin(tp, qa);
+
+		for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+			qtrx = &qa[i];
+			/*
+			 * The array of dquots is filled
+			 * sequentially, not sparsely.
+			 */
+			if ((dqp = qtrx->qt_dquot) == NULL)
+				break;
+
+			ASSERT(XFS_DQ_IS_LOCKED(dqp));
+			ASSERT(dqp->q_transp == tp);
+
+			/*
+			 * adjust the actual number of blocks used
+			 */
+			d = &dqp->q_core;
+
+			/*
+			 * The issue here is - sometimes we don't make a blkquota
+			 * reservation intentionally to be fair to users
+			 * (when the amount is small). On the other hand,
+			 * delayed allocs do make reservations, but that's
+			 * outside of a transaction, so we have no
+			 * idea how much was really reserved.
+			 * So, here we've accumulated delayed allocation blks and
+			 * non-delay blks. The assumption is that the
+			 * delayed ones are always reserved (outside of a
+			 * transaction), and the others may or may not have
+			 * quota reservations.
+			 */
+			totalbdelta = qtrx->qt_bcount_delta +
+				qtrx->qt_delbcnt_delta;
+			totalrtbdelta = qtrx->qt_rtbcount_delta +
+				qtrx->qt_delrtb_delta;
+#ifdef DEBUG
+			if (totalbdelta < 0)
+				ASSERT(be64_to_cpu(d->d_bcount) >=
+				       -totalbdelta);
+
+			if (totalrtbdelta < 0)
+				ASSERT(be64_to_cpu(d->d_rtbcount) >=
+				       -totalrtbdelta);
+
+			if (qtrx->qt_icount_delta < 0)
+				ASSERT(be64_to_cpu(d->d_icount) >=
+				       -qtrx->qt_icount_delta);
+#endif
+			if (totalbdelta)
+				be64_add_cpu(&d->d_bcount, (xfs_qcnt_t)totalbdelta);
+
+			if (qtrx->qt_icount_delta)
+				be64_add_cpu(&d->d_icount, (xfs_qcnt_t)qtrx->qt_icount_delta);
+
+			if (totalrtbdelta)
+				be64_add_cpu(&d->d_rtbcount, (xfs_qcnt_t)totalrtbdelta);
+
+			/*
+			 * Get any default limits in use.
+			 * Start/reset the timer(s) if needed.
+			 */
+			if (d->d_id) {
+				xfs_qm_adjust_dqlimits(tp->t_mountp, d);
+				xfs_qm_adjust_dqtimers(tp->t_mountp, d);
+			}
+
+			dqp->dq_flags |= XFS_DQ_DIRTY;
+			/*
+			 * add this to the list of items to get logged
+			 */
+			xfs_trans_log_dquot(tp, dqp);
+			/*
+			 * Take off what's left of the original reservation.
+			 * In case of delayed allocations, there's no
+			 * reservation that a transaction structure knows of.
+			 */
+			if (qtrx->qt_blk_res != 0) {
+				if (qtrx->qt_blk_res != qtrx->qt_blk_res_used) {
+					if (qtrx->qt_blk_res >
+					    qtrx->qt_blk_res_used)
+						dqp->q_res_bcount -= (xfs_qcnt_t)
+							(qtrx->qt_blk_res -
+							 qtrx->qt_blk_res_used);
+					else
+						dqp->q_res_bcount -= (xfs_qcnt_t)
+							(qtrx->qt_blk_res_used -
+							 qtrx->qt_blk_res);
+				}
+			} else {
+				/*
+				 * These blks were never reserved, either inside
+				 * a transaction or outside one (in a delayed
+				 * allocation). Also, this isn't always a
+				 * negative number since we sometimes
+				 * deliberately skip quota reservations.
+				 */
+				if (qtrx->qt_bcount_delta) {
+					dqp->q_res_bcount +=
+					      (xfs_qcnt_t)qtrx->qt_bcount_delta;
+				}
+			}
+			/*
+			 * Adjust the RT reservation.
+			 */
+			if (qtrx->qt_rtblk_res != 0) {
+				if (qtrx->qt_rtblk_res != qtrx->qt_rtblk_res_used) {
+					if (qtrx->qt_rtblk_res >
+					    qtrx->qt_rtblk_res_used)
+					       dqp->q_res_rtbcount -= (xfs_qcnt_t)
+						       (qtrx->qt_rtblk_res -
+							qtrx->qt_rtblk_res_used);
+					else
+					       dqp->q_res_rtbcount -= (xfs_qcnt_t)
+						       (qtrx->qt_rtblk_res_used -
+							qtrx->qt_rtblk_res);
+				}
+			} else {
+				if (qtrx->qt_rtbcount_delta)
+					dqp->q_res_rtbcount +=
+					    (xfs_qcnt_t)qtrx->qt_rtbcount_delta;
+			}
+
+			/*
+			 * Adjust the inode reservation.
+			 */
+			if (qtrx->qt_ino_res != 0) {
+				ASSERT(qtrx->qt_ino_res >=
+				       qtrx->qt_ino_res_used);
+				if (qtrx->qt_ino_res > qtrx->qt_ino_res_used)
+					dqp->q_res_icount -= (xfs_qcnt_t)
+						(qtrx->qt_ino_res -
+						 qtrx->qt_ino_res_used);
+			} else {
+				if (qtrx->qt_icount_delta)
+					dqp->q_res_icount +=
+					    (xfs_qcnt_t)qtrx->qt_icount_delta;
+			}
+
+			ASSERT(dqp->q_res_bcount >=
+				be64_to_cpu(dqp->q_core.d_bcount));
+			ASSERT(dqp->q_res_icount >=
+				be64_to_cpu(dqp->q_core.d_icount));
+			ASSERT(dqp->q_res_rtbcount >=
+				be64_to_cpu(dqp->q_core.d_rtbcount));
+		}
+		/*
+		 * Do the group quotas next
+		 */
+		qa = tp->t_dqinfo->dqa_grpdquots;
+	}
+}
+
+/*
+ * Release the reservations, and adjust the dquots accordingly.
+ * This is called only when the transaction is being aborted. If by
+ * any chance we have done dquot modifications incore (ie. deltas) already,
+ * we simply throw those away, since that's the expected behavior
+ * when a transaction is curtailed without a commit.
+ */
+void
+xfs_trans_unreserve_and_mod_dquots(
+	xfs_trans_t		*tp)
+{
+	int			i, j;
+	xfs_dquot_t		*dqp;
+	xfs_dqtrx_t		*qtrx, *qa;
+	boolean_t		locked;
+
+	if (!tp->t_dqinfo || !(tp->t_flags & XFS_TRANS_DQ_DIRTY))
+		return;
+
+	qa = tp->t_dqinfo->dqa_usrdquots;
+
+	for (j = 0; j < 2; j++) {
+		for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+			qtrx = &qa[i];
+			/*
+			 * We assume that the array of dquots is filled
+			 * sequentially, not sparsely.
+			 */
+			if ((dqp = qtrx->qt_dquot) == NULL)
+				break;
+			/*
+			 * Unreserve the original reservation. We don't care
+			 * about the number of blocks used field, or deltas.
+			 * Also we don't bother to zero the fields.
+			 */
+			locked = B_FALSE;
+			if (qtrx->qt_blk_res) {
+				xfs_dqlock(dqp);
+				locked = B_TRUE;
+				dqp->q_res_bcount -=
+					(xfs_qcnt_t)qtrx->qt_blk_res;
+			}
+			if (qtrx->qt_ino_res) {
+				if (!locked) {
+					xfs_dqlock(dqp);
+					locked = B_TRUE;
+				}
+				dqp->q_res_icount -=
+					(xfs_qcnt_t)qtrx->qt_ino_res;
+			}
+
+			if (qtrx->qt_rtblk_res) {
+				if (!locked) {
+					xfs_dqlock(dqp);
+					locked = B_TRUE;
+				}
+				dqp->q_res_rtbcount -=
+					(xfs_qcnt_t)qtrx->qt_rtblk_res;
+			}
+			if (locked)
+				xfs_dqunlock(dqp);
+
+		}
+		qa = tp->t_dqinfo->dqa_grpdquots;
+	}
+}
+
+STATIC void
+xfs_quota_warn(
+	struct xfs_mount	*mp,
+	struct xfs_dquot	*dqp,
+	int			type)
+{
+	/* no warnings for project quotas - we just return ENOSPC later */
+	if (dqp->dq_flags & XFS_DQ_PROJ)
+		return;
+	quota_send_warning((dqp->dq_flags & XFS_DQ_USER) ? USRQUOTA : GRPQUOTA,
+			   be32_to_cpu(dqp->q_core.d_id), mp->m_super->s_dev,
+			   type);
+}
+
+/*
+ * This reserves disk blocks and inodes against a dquot.
+ * Flags indicate if the dquot is to be locked here and also
+ * if the blk reservation is for RT or regular blocks.
+ * Sending in XFS_QMOPT_FORCE_RES flag skips the quota check.
+ */
+STATIC int
+xfs_trans_dqresv(
+	xfs_trans_t	*tp,
+	xfs_mount_t	*mp,
+	xfs_dquot_t	*dqp,
+	long		nblks,
+	long		ninos,
+	uint		flags)
+{
+	xfs_qcnt_t	hardlimit;
+	xfs_qcnt_t	softlimit;
+	time_t		timer;
+	xfs_qwarncnt_t	warns;
+	xfs_qwarncnt_t	warnlimit;
+	xfs_qcnt_t	count;
+	xfs_qcnt_t	*resbcountp;
+	xfs_quotainfo_t	*q = mp->m_quotainfo;
+
+
+	xfs_dqlock(dqp);
+
+	if (flags & XFS_TRANS_DQ_RES_BLKS) {
+		hardlimit = be64_to_cpu(dqp->q_core.d_blk_hardlimit);
+		if (!hardlimit)
+			hardlimit = q->qi_bhardlimit;
+		softlimit = be64_to_cpu(dqp->q_core.d_blk_softlimit);
+		if (!softlimit)
+			softlimit = q->qi_bsoftlimit;
+		timer = be32_to_cpu(dqp->q_core.d_btimer);
+		warns = be16_to_cpu(dqp->q_core.d_bwarns);
+		warnlimit = dqp->q_mount->m_quotainfo->qi_bwarnlimit;
+		resbcountp = &dqp->q_res_bcount;
+	} else {
+		ASSERT(flags & XFS_TRANS_DQ_RES_RTBLKS);
+		hardlimit = be64_to_cpu(dqp->q_core.d_rtb_hardlimit);
+		if (!hardlimit)
+			hardlimit = q->qi_rtbhardlimit;
+		softlimit = be64_to_cpu(dqp->q_core.d_rtb_softlimit);
+		if (!softlimit)
+			softlimit = q->qi_rtbsoftlimit;
+		timer = be32_to_cpu(dqp->q_core.d_rtbtimer);
+		warns = be16_to_cpu(dqp->q_core.d_rtbwarns);
+		warnlimit = dqp->q_mount->m_quotainfo->qi_rtbwarnlimit;
+		resbcountp = &dqp->q_res_rtbcount;
+	}
+
+	if ((flags & XFS_QMOPT_FORCE_RES) == 0 &&
+	    dqp->q_core.d_id &&
+	    ((XFS_IS_UQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISUDQ(dqp)) ||
+	     (XFS_IS_OQUOTA_ENFORCED(dqp->q_mount) &&
+	      (XFS_QM_ISPDQ(dqp) || XFS_QM_ISGDQ(dqp))))) {
+		if (nblks > 0) {
+			/*
+			 * dquot is locked already. See if we'd go over the
+			 * hardlimit or exceed the timelimit if we allocate
+			 * nblks.
+			 */
+			if (hardlimit > 0ULL &&
+			    hardlimit <= nblks + *resbcountp) {
+				xfs_quota_warn(mp, dqp, QUOTA_NL_BHARDWARN);
+				goto error_return;
+			}
+			if (softlimit > 0ULL &&
+			    softlimit <= nblks + *resbcountp) {
+				if ((timer != 0 && get_seconds() > timer) ||
+				    (warns != 0 && warns >= warnlimit)) {
+					xfs_quota_warn(mp, dqp,
+						       QUOTA_NL_BSOFTLONGWARN);
+					goto error_return;
+				}
+
+				xfs_quota_warn(mp, dqp, QUOTA_NL_BSOFTWARN);
+			}
+		}
+		if (ninos > 0) {
+			count = be64_to_cpu(dqp->q_core.d_icount);
+			timer = be32_to_cpu(dqp->q_core.d_itimer);
+			warns = be16_to_cpu(dqp->q_core.d_iwarns);
+			warnlimit = dqp->q_mount->m_quotainfo->qi_iwarnlimit;
+			hardlimit = be64_to_cpu(dqp->q_core.d_ino_hardlimit);
+			if (!hardlimit)
+				hardlimit = q->qi_ihardlimit;
+			softlimit = be64_to_cpu(dqp->q_core.d_ino_softlimit);
+			if (!softlimit)
+				softlimit = q->qi_isoftlimit;
+
+			if (hardlimit > 0ULL && count >= hardlimit) {
+				xfs_quota_warn(mp, dqp, QUOTA_NL_IHARDWARN);
+				goto error_return;
+			}
+			if (softlimit > 0ULL && count >= softlimit) {
+				if  ((timer != 0 && get_seconds() > timer) ||
+				     (warns != 0 && warns >= warnlimit)) {
+					xfs_quota_warn(mp, dqp,
+						       QUOTA_NL_ISOFTLONGWARN);
+					goto error_return;
+				}
+				xfs_quota_warn(mp, dqp, QUOTA_NL_ISOFTWARN);
+			}
+		}
+	}
+
+	/*
+	 * Change the reservation, but not the actual usage.
+	 * Note that q_res_bcount = q_core.d_bcount + resv
+	 */
+	(*resbcountp) += (xfs_qcnt_t)nblks;
+	if (ninos != 0)
+		dqp->q_res_icount += (xfs_qcnt_t)ninos;
+
+	/*
+	 * note the reservation amt in the trans struct too,
+	 * so that the transaction knows how much was reserved by
+	 * it against this particular dquot.
+	 * We don't do this when we are reserving for a delayed allocation,
+	 * because we don't have the luxury of a transaction envelope then.
+	 */
+	if (tp) {
+		ASSERT(tp->t_dqinfo);
+		ASSERT(flags & XFS_QMOPT_RESBLK_MASK);
+		if (nblks != 0)
+			xfs_trans_mod_dquot(tp, dqp,
+					    flags & XFS_QMOPT_RESBLK_MASK,
+					    nblks);
+		if (ninos != 0)
+			xfs_trans_mod_dquot(tp, dqp,
+					    XFS_TRANS_DQ_RES_INOS,
+					    ninos);
+	}
+	ASSERT(dqp->q_res_bcount >= be64_to_cpu(dqp->q_core.d_bcount));
+	ASSERT(dqp->q_res_rtbcount >= be64_to_cpu(dqp->q_core.d_rtbcount));
+	ASSERT(dqp->q_res_icount >= be64_to_cpu(dqp->q_core.d_icount));
+
+	xfs_dqunlock(dqp);
+	return 0;
+
+error_return:
+	xfs_dqunlock(dqp);
+	if (flags & XFS_QMOPT_ENOSPC)
+		return ENOSPC;
+	return EDQUOT;
+}
+
+
+/*
+ * Given dquot(s), make disk block and/or inode reservations against them.
+ * The fact that this does the reservation against both the usr and
+ * grp/prj quotas is important, because this follows a both-or-nothing
+ * approach.
+ *
+ * flags = XFS_QMOPT_FORCE_RES evades limit enforcement. Used by chown.
+ *	   XFS_QMOPT_ENOSPC returns ENOSPC not EDQUOT.  Used by pquota.
+ *	   XFS_TRANS_DQ_RES_BLKS reserves regular disk blocks
+ *	   XFS_TRANS_DQ_RES_RTBLKS reserves realtime disk blocks
+ * dquots are unlocked on return, if they were not locked by caller.
+ */
+int
+xfs_trans_reserve_quota_bydquots(
+	xfs_trans_t	*tp,
+	xfs_mount_t	*mp,
+	xfs_dquot_t	*udqp,
+	xfs_dquot_t	*gdqp,
+	long		nblks,
+	long		ninos,
+	uint		flags)
+{
+	int		resvd = 0, error;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return 0;
+
+	if (tp && tp->t_dqinfo == NULL)
+		xfs_trans_alloc_dqinfo(tp);
+
+	ASSERT(flags & XFS_QMOPT_RESBLK_MASK);
+
+	if (udqp) {
+		error = xfs_trans_dqresv(tp, mp, udqp, nblks, ninos,
+					(flags & ~XFS_QMOPT_ENOSPC));
+		if (error)
+			return error;
+		resvd = 1;
+	}
+
+	if (gdqp) {
+		error = xfs_trans_dqresv(tp, mp, gdqp, nblks, ninos, flags);
+		if (error) {
+			/*
+			 * can't do it, so backout previous reservation
+			 */
+			if (resvd) {
+				flags |= XFS_QMOPT_FORCE_RES;
+				xfs_trans_dqresv(tp, mp, udqp,
+						 -nblks, -ninos, flags);
+			}
+			return error;
+		}
+	}
+
+	/*
+	 * Didn't change anything critical, so, no need to log
+	 */
+	return 0;
+}
+
+
+/*
+ * Lock the dquot and change the reservation if we can.
+ * This doesn't change the actual usage, just the reservation.
+ * The inode sent in is locked.
+ */
+int
+xfs_trans_reserve_quota_nblks(
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	long			nblks,
+	long			ninos,
+	uint			flags)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return 0;
+	if (XFS_IS_PQUOTA_ON(mp))
+		flags |= XFS_QMOPT_ENOSPC;
+
+	ASSERT(ip->i_ino != mp->m_sb.sb_uquotino);
+	ASSERT(ip->i_ino != mp->m_sb.sb_gquotino);
+
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT((flags & ~(XFS_QMOPT_FORCE_RES | XFS_QMOPT_ENOSPC)) ==
+				XFS_TRANS_DQ_RES_RTBLKS ||
+	       (flags & ~(XFS_QMOPT_FORCE_RES | XFS_QMOPT_ENOSPC)) ==
+				XFS_TRANS_DQ_RES_BLKS);
+
+	/*
+	 * Reserve nblks against these dquots, with trans as the mediator.
+	 */
+	return xfs_trans_reserve_quota_bydquots(tp, mp,
+						ip->i_udquot, ip->i_gdquot,
+						nblks, ninos, flags);
+}
+
+/*
+ * This routine is called to allocate a quotaoff log item.
+ */
+xfs_qoff_logitem_t *
+xfs_trans_get_qoff_item(
+	xfs_trans_t		*tp,
+	xfs_qoff_logitem_t	*startqoff,
+	uint			flags)
+{
+	xfs_qoff_logitem_t	*q;
+
+	ASSERT(tp != NULL);
+
+	q = xfs_qm_qoff_logitem_init(tp->t_mountp, startqoff, flags);
+	ASSERT(q != NULL);
+
+	/*
+	 * Get a log_item_desc to point at the new item.
+	 */
+	xfs_trans_add_item(tp, &q->qql_item);
+	return q;
+}
+
+
+/*
+ * This is called to mark the quotaoff logitem as needing
+ * to be logged when the transaction is committed.  The logitem must
+ * already be associated with the given transaction.
+ */
+void
+xfs_trans_log_quotaoff_item(
+	xfs_trans_t		*tp,
+	xfs_qoff_logitem_t	*qlp)
+{
+	tp->t_flags |= XFS_TRANS_DIRTY;
+	qlp->qql_item.li_desc->lid_flags |= XFS_LID_DIRTY;
+}
+
+STATIC void
+xfs_trans_alloc_dqinfo(
+	xfs_trans_t	*tp)
+{
+	tp->t_dqinfo = kmem_zone_zalloc(xfs_Gqm->qm_dqtrxzone, KM_SLEEP);
+}
+
+void
+xfs_trans_free_dqinfo(
+	xfs_trans_t	*tp)
+{
+	if (!tp->t_dqinfo)
+		return;
+	kmem_zone_free(xfs_Gqm->qm_dqtrxzone, tp->t_dqinfo);
+	tp->t_dqinfo = NULL;
+}
diff --git a/fs/xfs/xfs_vnode.h b/fs/xfs/xfs_vnode.h
new file mode 100644
index 0000000..7c220b4
--- /dev/null
+++ b/fs/xfs/xfs_vnode.h
@@ -0,0 +1,64 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_VNODE_H__
+#define __XFS_VNODE_H__
+
+#include "xfs_fs.h"
+
+struct file;
+struct xfs_inode;
+struct xfs_iomap;
+struct attrlist_cursor_kern;
+
+/*
+ * Return values for xfs_inactive.  A return value of
+ * VN_INACTIVE_NOCACHE implies that the file system behavior
+ * has disassociated its state and bhv_desc_t from the vnode.
+ */
+#define	VN_INACTIVE_CACHE	0
+#define	VN_INACTIVE_NOCACHE	1
+
+/*
+ * Flags for read/write calls - same values as IRIX
+ */
+#define IO_ISDIRECT	0x00004		/* bypass page cache */
+#define IO_INVIS	0x00020		/* don't update inode timestamps */
+
+#define XFS_IO_FLAGS \
+	{ IO_ISDIRECT,	"DIRECT" }, \
+	{ IO_INVIS,	"INVIS"}
+
+/*
+ * Flush/Invalidate options for vop_toss/flush/flushinval_pages.
+ */
+#define FI_NONE			0	/* none */
+#define FI_REMAPF		1	/* Do a remapf prior to the operation */
+#define FI_REMAPF_LOCKED	2	/* Do a remapf prior to the operation.
+					   Prevent VM access to the pages until
+					   the operation completes. */
+
+/*
+ * Some useful predicates.
+ */
+#define VN_MAPPED(vp)	mapping_mapped(vp->i_mapping)
+#define VN_CACHED(vp)	(vp->i_mapping->nrpages)
+#define VN_DIRTY(vp)	mapping_tagged(vp->i_mapping, \
+					PAGECACHE_TAG_DIRTY)
+
+
+#endif	/* __XFS_VNODE_H__ */
diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c
new file mode 100644
index 0000000..87d3e03
--- /dev/null
+++ b/fs/xfs/xfs_xattr.c
@@ -0,0 +1,241 @@
+/*
+ * Copyright (C) 2008 Christoph Hellwig.
+ * Portions Copyright (C) 2000-2008 Silicon Graphics, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "xfs.h"
+#include "xfs_da_btree.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_inode.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "xfs_acl.h"
+#include "xfs_vnodeops.h"
+
+#include <linux/posix_acl_xattr.h>
+#include <linux/xattr.h>
+
+
+static int
+xfs_xattr_get(struct dentry *dentry, const char *name,
+		void *value, size_t size, int xflags)
+{
+	struct xfs_inode *ip = XFS_I(dentry->d_inode);
+	int error, asize = size;
+
+	if (strcmp(name, "") == 0)
+		return -EINVAL;
+
+	/* Convert Linux syscall to XFS internal ATTR flags */
+	if (!size) {
+		xflags |= ATTR_KERNOVAL;
+		value = NULL;
+	}
+
+	error = -xfs_attr_get(ip, (unsigned char *)name, value, &asize, xflags);
+	if (error)
+		return error;
+	return asize;
+}
+
+static int
+xfs_xattr_set(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags, int xflags)
+{
+	struct xfs_inode *ip = XFS_I(dentry->d_inode);
+
+	if (strcmp(name, "") == 0)
+		return -EINVAL;
+
+	/* Convert Linux syscall to XFS internal ATTR flags */
+	if (flags & XATTR_CREATE)
+		xflags |= ATTR_CREATE;
+	if (flags & XATTR_REPLACE)
+		xflags |= ATTR_REPLACE;
+
+	if (!value)
+		return -xfs_attr_remove(ip, (unsigned char *)name, xflags);
+	return -xfs_attr_set(ip, (unsigned char *)name,
+				(void *)value, size, xflags);
+}
+
+static const struct xattr_handler xfs_xattr_user_handler = {
+	.prefix	= XATTR_USER_PREFIX,
+	.flags	= 0, /* no flags implies user namespace */
+	.get	= xfs_xattr_get,
+	.set	= xfs_xattr_set,
+};
+
+static const struct xattr_handler xfs_xattr_trusted_handler = {
+	.prefix	= XATTR_TRUSTED_PREFIX,
+	.flags	= ATTR_ROOT,
+	.get	= xfs_xattr_get,
+	.set	= xfs_xattr_set,
+};
+
+static const struct xattr_handler xfs_xattr_security_handler = {
+	.prefix	= XATTR_SECURITY_PREFIX,
+	.flags	= ATTR_SECURE,
+	.get	= xfs_xattr_get,
+	.set	= xfs_xattr_set,
+};
+
+const struct xattr_handler *xfs_xattr_handlers[] = {
+	&xfs_xattr_user_handler,
+	&xfs_xattr_trusted_handler,
+	&xfs_xattr_security_handler,
+#ifdef CONFIG_XFS_POSIX_ACL
+	&xfs_xattr_acl_access_handler,
+	&xfs_xattr_acl_default_handler,
+#endif
+	NULL
+};
+
+static unsigned int xfs_xattr_prefix_len(int flags)
+{
+	if (flags & XFS_ATTR_SECURE)
+		return sizeof("security");
+	else if (flags & XFS_ATTR_ROOT)
+		return sizeof("trusted");
+	else
+		return sizeof("user");
+}
+
+static const char *xfs_xattr_prefix(int flags)
+{
+	if (flags & XFS_ATTR_SECURE)
+		return xfs_xattr_security_handler.prefix;
+	else if (flags & XFS_ATTR_ROOT)
+		return xfs_xattr_trusted_handler.prefix;
+	else
+		return xfs_xattr_user_handler.prefix;
+}
+
+static int
+xfs_xattr_put_listent(
+	struct xfs_attr_list_context *context,
+	int		flags,
+	unsigned char	*name,
+	int		namelen,
+	int		valuelen,
+	unsigned char	*value)
+{
+	unsigned int prefix_len = xfs_xattr_prefix_len(flags);
+	char *offset;
+	int arraytop;
+
+	ASSERT(context->count >= 0);
+
+	/*
+	 * Only show root namespace entries if we are actually allowed to
+	 * see them.
+	 */
+	if ((flags & XFS_ATTR_ROOT) && !capable(CAP_SYS_ADMIN))
+		return 0;
+
+	arraytop = context->count + prefix_len + namelen + 1;
+	if (arraytop > context->firstu) {
+		context->count = -1;	/* insufficient space */
+		return 1;
+	}
+	offset = (char *)context->alist + context->count;
+	strncpy(offset, xfs_xattr_prefix(flags), prefix_len);
+	offset += prefix_len;
+	strncpy(offset, (char *)name, namelen);			/* real name */
+	offset += namelen;
+	*offset = '\0';
+	context->count += prefix_len + namelen + 1;
+	return 0;
+}
+
+static int
+xfs_xattr_put_listent_sizes(
+	struct xfs_attr_list_context *context,
+	int		flags,
+	unsigned char	*name,
+	int		namelen,
+	int		valuelen,
+	unsigned char	*value)
+{
+	context->count += xfs_xattr_prefix_len(flags) + namelen + 1;
+	return 0;
+}
+
+static int
+list_one_attr(const char *name, const size_t len, void *data,
+		size_t size, ssize_t *result)
+{
+	char *p = data + *result;
+
+	*result += len;
+	if (!size)
+		return 0;
+	if (*result > size)
+		return -ERANGE;
+
+	strcpy(p, name);
+	return 0;
+}
+
+ssize_t
+xfs_vn_listxattr(struct dentry *dentry, char *data, size_t size)
+{
+	struct xfs_attr_list_context context;
+	struct attrlist_cursor_kern cursor = { 0 };
+	struct inode		*inode = dentry->d_inode;
+	int			error;
+
+	/*
+	 * First read the regular on-disk attributes.
+	 */
+	memset(&context, 0, sizeof(context));
+	context.dp = XFS_I(inode);
+	context.cursor = &cursor;
+	context.resynch = 1;
+	context.alist = data;
+	context.bufsize = size;
+	context.firstu = context.bufsize;
+
+	if (size)
+		context.put_listent = xfs_xattr_put_listent;
+	else
+		context.put_listent = xfs_xattr_put_listent_sizes;
+
+	xfs_attr_list_int(&context);
+	if (context.count < 0)
+		return -ERANGE;
+
+	/*
+	 * Then add the two synthetic ACL attributes.
+	 */
+	if (posix_acl_access_exists(inode)) {
+		error = list_one_attr(POSIX_ACL_XATTR_ACCESS,
+				strlen(POSIX_ACL_XATTR_ACCESS) + 1,
+				data, size, &context.count);
+		if (error)
+			return error;
+	}
+
+	if (posix_acl_default_exists(inode)) {
+		error = list_one_attr(POSIX_ACL_XATTR_DEFAULT,
+				strlen(POSIX_ACL_XATTR_DEFAULT) + 1,
+				data, size, &context.count);
+		if (error)
+			return error;
+	}
+
+	return context.count;
+}