Merge branch 'linus'

175 files changed, 5948 insertions, 5445 deletions

diff --git a/fs/9p/mux.c b/fs/9p/mux.c
index ea1134eb47c8..8e8356c1c229 100644
--- a/fs/9p/mux.c
+++ b/fs/9p/mux.c
@@ -31,6 +31,7 @@
 #include <linux/poll.h>
 #include <linux/kthread.h>
 #include <linux/idr.h>
+#include <linux/mutex.h>
 
 #include "debug.h"
 #include "v9fs.h"
@@ -110,7 +111,7 @@ static void v9fs_pollwait(struct file *filp, wait_queue_head_t * wait_address,
 static u16 v9fs_mux_get_tag(struct v9fs_mux_data *);
 static void v9fs_mux_put_tag(struct v9fs_mux_data *, u16);
 
-static DECLARE_MUTEX(v9fs_mux_task_lock);
+static DEFINE_MUTEX(v9fs_mux_task_lock);
 static struct workqueue_struct *v9fs_mux_wq;
 
 static int v9fs_mux_num;
@@ -166,7 +167,7 @@ static int v9fs_mux_poll_start(struct v9fs_mux_data *m)
 
 	dprintk(DEBUG_MUX, "mux %p muxnum %d procnum %d\n", m, v9fs_mux_num,
 		v9fs_mux_poll_task_num);
-	up(&v9fs_mux_task_lock);
+	mutex_lock(&v9fs_mux_task_lock);
 
 	n = v9fs_mux_calc_poll_procs(v9fs_mux_num + 1);
 	if (n > v9fs_mux_poll_task_num) {
@@ -225,7 +226,7 @@ static int v9fs_mux_poll_start(struct v9fs_mux_data *m)
 	}
 
 	v9fs_mux_num++;
-	down(&v9fs_mux_task_lock);
+	mutex_unlock(&v9fs_mux_task_lock);
 
 	return 0;
 }
@@ -235,7 +236,7 @@ static void v9fs_mux_poll_stop(struct v9fs_mux_data *m)
 	int i;
 	struct v9fs_mux_poll_task *vpt;
 
-	up(&v9fs_mux_task_lock);
+	mutex_lock(&v9fs_mux_task_lock);
 	vpt = m->poll_task;
 	list_del(&m->mux_list);
 	for(i = 0; i < ARRAY_SIZE(m->poll_waddr); i++) {
@@ -252,7 +253,7 @@ static void v9fs_mux_poll_stop(struct v9fs_mux_data *m)
 		v9fs_mux_poll_task_num--;
 	}
 	v9fs_mux_num--;
-	down(&v9fs_mux_task_lock);
+	mutex_unlock(&v9fs_mux_task_lock);
 }
 
 /**
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 3ad8455f8577..651a9e14d9a9 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -614,6 +614,7 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 
 	sb = dir->i_sb;
 	v9ses = v9fs_inode2v9ses(dir);
+	dentry->d_op = &v9fs_dentry_operations;
 	dirfid = v9fs_fid_lookup(dentry->d_parent);
 
 	if (!dirfid) {
@@ -681,8 +682,6 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 		goto FreeFcall;
 
 	fid->qid = fcall->params.rstat.stat.qid;
-
-	dentry->d_op = &v9fs_dentry_operations;
 	v9fs_stat2inode(&fcall->params.rstat.stat, inode, inode->i_sb);
 
 	d_add(dentry, inode);
diff --git a/fs/Kconfig b/fs/Kconfig
index 9a4158f2a3ac..e207be68d4ca 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -859,18 +859,6 @@ config RAMFS
 	  To compile this as a module, choose M here: the module will be called
 	  ramfs.
 
-config RELAYFS_FS
-	tristate "Relayfs file system support"
-	---help---
-	  Relayfs is a high-speed data relay filesystem designed to provide
-	  an efficient mechanism for tools and facilities to relay large
-	  amounts of data from kernel space to user space.
-
-	  To compile this code as a module, choose M here: the module will be
-	  called relayfs.
-
-	  If unsure, say N.
-
 config CONFIGFS_FS
 	tristate "Userspace-driven configuration filesystem (EXPERIMENTAL)"
 	depends on EXPERIMENTAL
diff --git a/fs/Makefile b/fs/Makefile
index 1db711319c80..080b3867be4d 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -91,7 +91,6 @@ obj-$(CONFIG_AUTOFS4_FS)	+= autofs4/
 obj-$(CONFIG_ADFS_FS)		+= adfs/
 obj-$(CONFIG_FUSE_FS)		+= fuse/
 obj-$(CONFIG_UDF_FS)		+= udf/
-obj-$(CONFIG_RELAYFS_FS)	+= relayfs/
 obj-$(CONFIG_SUN_OPENPROMFS)	+= openpromfs/
 obj-$(CONFIG_JFS_FS)		+= jfs/
 obj-$(CONFIG_XFS_FS)		+= xfs/
diff --git a/fs/adfs/file.c b/fs/adfs/file.c
index afebbfde6968..6af10885f9d6 100644
--- a/fs/adfs/file.c
+++ b/fs/adfs/file.c
@@ -19,11 +19,7 @@
  *
  *  adfs regular file handling primitives           
  */
-#include <linux/errno.h>
 #include <linux/fs.h>
-#include <linux/fcntl.h>
-#include <linux/time.h>
-#include <linux/stat.h>
 #include <linux/buffer_head.h>			/* for file_fsync() */
 #include <linux/adfs_fs.h>
 
diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index 385bed09b0d8..f54c5b21f876 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -13,6 +13,7 @@
 /* Internal header file for autofs */
 
 #include <linux/auto_fs4.h>
+#include <linux/mutex.h>
 #include <linux/list.h>
 
 /* This is the range of ioctl() numbers we claim as ours */
@@ -102,7 +103,7 @@ struct autofs_sb_info {
 	int reghost_enabled;
 	int needs_reghost;
 	struct super_block *sb;
-	struct semaphore wq_sem;
+	struct mutex wq_mutex;
 	spinlock_t fs_lock;
 	struct autofs_wait_queue *queues; /* Wait queue pointer */
 };
diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c
index 2d3082854a29..1ad98d48e550 100644
--- a/fs/autofs4/inode.c
+++ b/fs/autofs4/inode.c
@@ -269,7 +269,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	sbi->sb = s;
 	sbi->version = 0;
 	sbi->sub_version = 0;
-	init_MUTEX(&sbi->wq_sem);
+	mutex_init(&sbi->wq_mutex);
 	spin_lock_init(&sbi->fs_lock);
 	sbi->queues = NULL;
 	s->s_blocksize = 1024;
diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c
index 394ff36ef8f1..be78e9378c03 100644
--- a/fs/autofs4/waitq.c
+++ b/fs/autofs4/waitq.c
@@ -178,7 +178,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		return -ENOENT;
 	}
 
-	if (down_interruptible(&sbi->wq_sem)) {
+	if (mutex_lock_interruptible(&sbi->wq_mutex)) {
 		kfree(name);
 		return -EINTR;
 	}
@@ -194,7 +194,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		/* Can't wait for an expire if there's no mount */
 		if (notify == NFY_NONE && !d_mountpoint(dentry)) {
 			kfree(name);
-			up(&sbi->wq_sem);
+			mutex_unlock(&sbi->wq_mutex);
 			return -ENOENT;
 		}
 
@@ -202,7 +202,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		wq = kmalloc(sizeof(struct autofs_wait_queue),GFP_KERNEL);
 		if ( !wq ) {
 			kfree(name);
-			up(&sbi->wq_sem);
+			mutex_unlock(&sbi->wq_mutex);
 			return -ENOMEM;
 		}
 
@@ -218,10 +218,10 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		wq->status = -EINTR; /* Status return if interrupted */
 		atomic_set(&wq->wait_ctr, 2);
 		atomic_set(&wq->notified, 1);
-		up(&sbi->wq_sem);
+		mutex_unlock(&sbi->wq_mutex);
 	} else {
 		atomic_inc(&wq->wait_ctr);
-		up(&sbi->wq_sem);
+		mutex_unlock(&sbi->wq_mutex);
 		kfree(name);
 		DPRINTK("existing wait id = 0x%08lx, name = %.*s, nfy=%d",
 			(unsigned long) wq->wait_queue_token, wq->len, wq->name, notify);
@@ -282,19 +282,19 @@ int autofs4_wait_release(struct autofs_sb_info *sbi, autofs_wqt_t wait_queue_tok
 {
 	struct autofs_wait_queue *wq, **wql;
 
-	down(&sbi->wq_sem);
+	mutex_lock(&sbi->wq_mutex);
 	for ( wql = &sbi->queues ; (wq = *wql) != 0 ; wql = &wq->next ) {
 		if ( wq->wait_queue_token == wait_queue_token )
 			break;
 	}
 
 	if ( !wq ) {
-		up(&sbi->wq_sem);
+		mutex_unlock(&sbi->wq_mutex);
 		return -EINVAL;
 	}
 
 	*wql = wq->next;	/* Unlink from chain */
-	up(&sbi->wq_sem);
+	mutex_unlock(&sbi->wq_mutex);
 	kfree(wq->name);
 	wq->name = NULL;	/* Do not wait on this queue */
 
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index 2d365cb8eec6..dd6048ce0532 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -561,7 +561,7 @@ befs_utf2nls(struct super_block *sb, const char *in,
  * @sb: Superblock
  * @src: Input string buffer in NLS format
  * @srclen: Length of input string in bytes
- * @dest: The output string in UTF8 format
+ * @dest: The output string in UTF-8 format
  * @destlen: Length of the output buffer
  * 
  * Converts input string @src, which is in the format of the loaded NLS map,
diff --git a/fs/bio.c b/fs/bio.c
index 1f3bb501c262..0a8c59cb68f5 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -25,6 +25,7 @@
 #include <linux/module.h>
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
+#include <linux/blktrace_api.h>
 #include <scsi/sg.h>		/* for struct sg_iovec */
 
 #define BIO_POOL_SIZE 256
@@ -1095,6 +1096,9 @@ struct bio_pair *bio_split(struct bio *bi, mempool_t *pool, int first_sectors)
 	if (!bp)
 		return bp;
 
+	blk_add_trace_pdu_int(bdev_get_queue(bi->bi_bdev), BLK_TA_SPLIT, bi,
+				bi->bi_sector + first_sectors);
+
 	BUG_ON(bi->bi_vcnt != 1);
 	BUG_ON(bi->bi_idx != 0);
 	atomic_set(&bp->cnt, 3);
@@ -1243,11 +1247,11 @@ static int __init init_bio(void)
 		scale = 4;
 
 	/*
-	 * scale number of entries
+	 * Limit number of entries reserved -- mempools are only used when
+	 * the system is completely unable to allocate memory, so we only
+	 * need enough to make progress.
 	 */
-	bvec_pool_entries = megabytes * 2;
-	if (bvec_pool_entries > 256)
-		bvec_pool_entries = 256;
+	bvec_pool_entries = 1 + scale;
 
 	fs_bio_set = bioset_create(BIO_POOL_SIZE, bvec_pool_entries, scale);
 	if (!fs_bio_set)
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 6e50346fb1ee..44d05e6e34db 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -265,8 +265,8 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 	    SLAB_CTOR_CONSTRUCTOR)
 	{
 		memset(bdev, 0, sizeof(*bdev));
-		sema_init(&bdev->bd_sem, 1);
-		sema_init(&bdev->bd_mount_sem, 1);
+		mutex_init(&bdev->bd_mutex);
+		mutex_init(&bdev->bd_mount_mutex);
 		INIT_LIST_HEAD(&bdev->bd_inodes);
 		INIT_LIST_HEAD(&bdev->bd_list);
 		inode_init_once(&ei->vfs_inode);
@@ -574,7 +574,7 @@ static int do_open(struct block_device *bdev, struct file *file)
 	}
 	owner = disk->fops->owner;
 
-	down(&bdev->bd_sem);
+	mutex_lock(&bdev->bd_mutex);
 	if (!bdev->bd_openers) {
 		bdev->bd_disk = disk;
 		bdev->bd_contains = bdev;
@@ -605,21 +605,21 @@ static int do_open(struct block_device *bdev, struct file *file)
 			if (ret)
 				goto out_first;
 			bdev->bd_contains = whole;
-			down(&whole->bd_sem);
+			mutex_lock(&whole->bd_mutex);
 			whole->bd_part_count++;
 			p = disk->part[part - 1];
 			bdev->bd_inode->i_data.backing_dev_info =
 			   whole->bd_inode->i_data.backing_dev_info;
 			if (!(disk->flags & GENHD_FL_UP) || !p || !p->nr_sects) {
 				whole->bd_part_count--;
-				up(&whole->bd_sem);
+				mutex_unlock(&whole->bd_mutex);
 				ret = -ENXIO;
 				goto out_first;
 			}
 			kobject_get(&p->kobj);
 			bdev->bd_part = p;
 			bd_set_size(bdev, (loff_t) p->nr_sects << 9);
-			up(&whole->bd_sem);
+			mutex_unlock(&whole->bd_mutex);
 		}
 	} else {
 		put_disk(disk);
@@ -633,13 +633,13 @@ static int do_open(struct block_device *bdev, struct file *file)
 			if (bdev->bd_invalidated)
 				rescan_partitions(bdev->bd_disk, bdev);
 		} else {
-			down(&bdev->bd_contains->bd_sem);
+			mutex_lock(&bdev->bd_contains->bd_mutex);
 			bdev->bd_contains->bd_part_count++;
-			up(&bdev->bd_contains->bd_sem);
+			mutex_unlock(&bdev->bd_contains->bd_mutex);
 		}
 	}
 	bdev->bd_openers++;
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	unlock_kernel();
 	return 0;
 
@@ -652,7 +652,7 @@ out_first:
 	put_disk(disk);
 	module_put(owner);
 out:
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	unlock_kernel();
 	if (ret)
 		bdput(bdev);
@@ -714,7 +714,7 @@ int blkdev_put(struct block_device *bdev)
 	struct inode *bd_inode = bdev->bd_inode;
 	struct gendisk *disk = bdev->bd_disk;
 
-	down(&bdev->bd_sem);
+	mutex_lock(&bdev->bd_mutex);
 	lock_kernel();
 	if (!--bdev->bd_openers) {
 		sync_blockdev(bdev);
@@ -724,9 +724,9 @@ int blkdev_put(struct block_device *bdev)
 		if (disk->fops->release)
 			ret = disk->fops->release(bd_inode, NULL);
 	} else {
-		down(&bdev->bd_contains->bd_sem);
+		mutex_lock(&bdev->bd_contains->bd_mutex);
 		bdev->bd_contains->bd_part_count--;
-		up(&bdev->bd_contains->bd_sem);
+		mutex_unlock(&bdev->bd_contains->bd_mutex);
 	}
 	if (!bdev->bd_openers) {
 		struct module *owner = disk->fops->owner;
@@ -746,7 +746,7 @@ int blkdev_put(struct block_device *bdev)
 		bdev->bd_contains = NULL;
 	}
 	unlock_kernel();
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	bdput(bdev);
 	return ret;
 }
diff --git a/fs/buffer.c b/fs/buffer.c
index a9b399402007..0d6ca7bac6c8 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -201,7 +201,7 @@ int fsync_bdev(struct block_device *bdev)
  * freeze_bdev  --  lock a filesystem and force it into a consistent state
  * @bdev:	blockdevice to lock
  *
- * This takes the block device bd_mount_sem to make sure no new mounts
+ * This takes the block device bd_mount_mutex to make sure no new mounts
  * happen on bdev until thaw_bdev() is called.
  * If a superblock is found on this device, we take the s_umount semaphore
  * on it to make sure nobody unmounts until the snapshot creation is done.
@@ -210,7 +210,7 @@ struct super_block *freeze_bdev(struct block_device *bdev)
 {
 	struct super_block *sb;
 
-	down(&bdev->bd_mount_sem);
+	mutex_lock(&bdev->bd_mount_mutex);
 	sb = get_super(bdev);
 	if (sb && !(sb->s_flags & MS_RDONLY)) {
 		sb->s_frozen = SB_FREEZE_WRITE;
@@ -264,7 +264,7 @@ void thaw_bdev(struct block_device *bdev, struct super_block *sb)
 		drop_super(sb);
 	}
 
-	up(&bdev->bd_mount_sem);
+	mutex_unlock(&bdev->bd_mount_mutex);
 }
 EXPORT_SYMBOL(thaw_bdev);
 
@@ -3051,68 +3051,6 @@ asmlinkage long sys_bdflush(int func, long data)
 }
 
 /*
- * Migration function for pages with buffers. This function can only be used
- * if the underlying filesystem guarantees that no other references to "page"
- * exist.
- */
-#ifdef CONFIG_MIGRATION
-int buffer_migrate_page(struct page *newpage, struct page *page)
-{
-	struct address_space *mapping = page->mapping;
-	struct buffer_head *bh, *head;
-	int rc;
-
-	if (!mapping)
-		return -EAGAIN;
-
-	if (!page_has_buffers(page))
-		return migrate_page(newpage, page);
-
-	head = page_buffers(page);
-
-	rc = migrate_page_remove_references(newpage, page, 3);
-	if (rc)
-		return rc;
-
-	bh = head;
-	do {
-		get_bh(bh);
-		lock_buffer(bh);
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	ClearPagePrivate(page);
-	set_page_private(newpage, page_private(page));
-	set_page_private(page, 0);
-	put_page(page);
-	get_page(newpage);
-
-	bh = head;
-	do {
-		set_bh_page(bh, newpage, bh_offset(bh));
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	SetPagePrivate(newpage);
-
-	migrate_page_copy(newpage, page);
-
-	bh = head;
-	do {
-		unlock_buffer(bh);
- 		put_bh(bh);
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	return 0;
-}
-EXPORT_SYMBOL(buffer_migrate_page);
-#endif
-
-/*
  * Buffer-head allocation
  */
 static kmem_cache_t *bh_cachep;
diff --git a/fs/char_dev.c b/fs/char_dev.c
index 21195c481637..5c36345c9bf7 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -19,6 +19,7 @@
 #include <linux/kobject.h>
 #include <linux/kobj_map.h>
 #include <linux/cdev.h>
+#include <linux/mutex.h>
 
 #ifdef CONFIG_KMOD
 #include <linux/kmod.h>
@@ -28,7 +29,7 @@ static struct kobj_map *cdev_map;
 
 #define MAX_PROBE_HASH 255	/* random */
 
-static DECLARE_MUTEX(chrdevs_lock);
+static DEFINE_MUTEX(chrdevs_lock);
 
 static struct char_device_struct {
 	struct char_device_struct *next;
@@ -88,13 +89,13 @@ out:
 
 void *acquire_chrdev_list(void)
 {
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 	return get_next_chrdev(NULL);
 }
 
 void release_chrdev_list(void *dev)
 {
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	kfree(dev);
 }
 
@@ -151,7 +152,7 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 
 	memset(cd, 0, sizeof(struct char_device_struct));
 
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 
 	/* temporary */
 	if (major == 0) {
@@ -186,10 +187,10 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 	}
 	cd->next = *cp;
 	*cp = cd;
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	return cd;
 out:
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	kfree(cd);
 	return ERR_PTR(ret);
 }
@@ -200,7 +201,7 @@ __unregister_chrdev_region(unsigned major, unsigned baseminor, int minorct)
 	struct char_device_struct *cd = NULL, **cp;
 	int i = major_to_index(major);
 
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 	for (cp = &chrdevs[i]; *cp; cp = &(*cp)->next)
 		if ((*cp)->major == major &&
 		    (*cp)->baseminor == baseminor &&
@@ -210,7 +211,7 @@ __unregister_chrdev_region(unsigned major, unsigned baseminor, int minorct)
 		cd = *cp;
 		*cp = cd->next;
 	}
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	return cd;
 }
 
diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index d335015473a5..cb68efba35db 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -160,7 +160,7 @@ improperly zeroed buffer in CIFS Unix extensions set times call.
 Version 1.25
 ------------
 Fix internationalization problem in cifs readdir with filenames that map to 
-longer UTF8 strings than the string on the wire was in Unicode.  Add workaround
+longer UTF-8 strings than the string on the wire was in Unicode.  Add workaround
 for readdir to netapp servers. Fix search rewind (seek into readdir to return 
 non-consecutive entries).  Do not do readdir when server negotiates 
 buffer size to small to fit filename. Add support for reading POSIX ACLs from
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index fed55e3c53df..632561dd9c50 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -138,9 +138,9 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -317,9 +317,9 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode,
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if(full_path == NULL)
 		rc = -ENOMEM;
 	else if (pTcon->ses->capabilities & CAP_UNIX) {
diff --git a/fs/cifs/fcntl.c b/fs/cifs/fcntl.c
index a7a47bb36bf3..ec4dfe9bf5ef 100644
--- a/fs/cifs/fcntl.c
+++ b/fs/cifs/fcntl.c
@@ -86,9 +86,9 @@ int cifs_dir_notify(struct file * file, unsigned long arg)
 	cifs_sb = CIFS_SB(file->f_dentry->d_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		rc = -ENOMEM;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 675bd2568297..165d67426381 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -203,9 +203,9 @@ int cifs_open(struct inode *inode, struct file *file)
 		}
 	}
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 59359911f481..ff93a9f81d1c 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -574,9 +574,9 @@ int cifs_unlink(struct inode *inode, struct dentry *direntry)
 
 	/* Unlink can be called from rename so we can not grab the sem here
 	   since we deadlock otherwise */
-/*	down(&direntry->d_sb->s_vfs_rename_sem);*/
+/*	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);*/
 	full_path = build_path_from_dentry(direntry);
-/*	up(&direntry->d_sb->s_vfs_rename_sem);*/
+/*	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);*/
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -718,9 +718,9 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -803,9 +803,9 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -1137,9 +1137,9 @@ int cifs_setattr(struct dentry *direntry, struct iattr *attrs)
 			rc = 0;
 	}
 		
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
diff --git a/fs/cifs/link.c b/fs/cifs/link.c
index 0f99aae33162..8d0da7c87c7b 100644
--- a/fs/cifs/link.c
+++ b/fs/cifs/link.c
@@ -48,10 +48,10 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 /* No need to check for cross device links since server will do that
    BB note DFS case in future though (when we may have to check) */
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	fromName = build_path_from_dentry(old_file);
 	toName = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if((fromName == NULL) || (toName == NULL)) {
 		rc = -ENOMEM;
 		goto cifs_hl_exit;
@@ -103,9 +103,9 @@ cifs_follow_link(struct dentry *direntry, struct nameidata *nd)
 
 	xid = GetXid();
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 
 	if (!full_path)
 		goto out_no_free;
@@ -164,9 +164,9 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		FreeXid(xid);
@@ -232,9 +232,9 @@ cifs_readlink(struct dentry *direntry, char __user *pBuffer, int buflen)
 
 /* BB would it be safe against deadlock to grab this sem 
       even though rename itself grabs the sem and calls lookup? */
-/*       down(&inode->i_sb->s_vfs_rename_sem);*/
+/*       mutex_lock(&inode->i_sb->s_vfs_rename_mutex);*/
 	full_path = build_path_from_dentry(direntry);
-/*       up(&inode->i_sb->s_vfs_rename_sem);*/
+/*       mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);*/
 
 	if(full_path == NULL) {
 		FreeXid(xid);
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 288cc048d37f..edb3b6eb34bc 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -404,9 +404,9 @@ static int initiate_cifs_search(const int xid, struct file *file)
 	if(pTcon == NULL)
 		return -EINVAL;
 
-	down(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		return -ENOMEM;
diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c
index 777e3363c2a4..3938444d87b2 100644
--- a/fs/cifs/xattr.c
+++ b/fs/cifs/xattr.c
@@ -62,9 +62,9 @@ int cifs_removexattr(struct dentry * direntry, const char * ea_name)
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
                                                                                      
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -116,9 +116,9 @@ int cifs_setxattr(struct dentry * direntry, const char * ea_name,
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -223,9 +223,9 @@ ssize_t cifs_getxattr(struct dentry * direntry, const char * ea_name,
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -341,9 +341,9 @@ ssize_t cifs_listxattr(struct dentry * direntry, char * data, size_t buf_size)
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
index c666769a875d..7c031f00fd79 100644
--- a/fs/compat_ioctl.c
+++ b/fs/compat_ioctl.c
@@ -72,6 +72,7 @@
 #include <linux/i2c-dev.h>
 #include <linux/wireless.h>
 #include <linux/atalk.h>
+#include <linux/blktrace_api.h>
 
 #include <net/sock.h>          /* siocdevprivate_ioctl */
 #include <net/bluetooth/bluetooth.h>
diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c
index d575452cd9f7..40c4fc973fad 100644
--- a/fs/debugfs/file.c
+++ b/fs/debugfs/file.c
@@ -251,3 +251,49 @@ struct dentry *debugfs_create_bool(const char *name, mode_t mode,
 }
 EXPORT_SYMBOL_GPL(debugfs_create_bool);
 
+static ssize_t read_file_blob(struct file *file, char __user *user_buf,
+			      size_t count, loff_t *ppos)
+{
+	struct debugfs_blob_wrapper *blob = file->private_data;
+	return simple_read_from_buffer(user_buf, count, ppos, blob->data,
+			blob->size);
+}
+
+static struct file_operations fops_blob = {
+	.read =		read_file_blob,
+	.open =		default_open,
+};
+
+/**
+ * debugfs_create_blob - create a file in the debugfs filesystem that is
+ * used to read and write a binary blob.
+ *
+ * @name: a pointer to a string containing the name of the file to create.
+ * @mode: the permission that the file should have
+ * @parent: a pointer to the parent dentry for this file.  This should be a
+ *          directory dentry if set.  If this paramater is NULL, then the
+ *          file will be created in the root of the debugfs filesystem.
+ * @blob: a pointer to a struct debugfs_blob_wrapper which contains a pointer
+ *        to the blob data and the size of the data.
+ *
+ * This function creates a file in debugfs with the given name that exports
+ * @blob->data as a binary blob. If the @mode variable is so set it can be
+ * read from. Writing is not supported.
+ *
+ * This function will return a pointer to a dentry if it succeeds.  This
+ * pointer must be passed to the debugfs_remove() function when the file is
+ * to be removed (no automatic cleanup happens if your module is unloaded,
+ * you are responsible here.)  If an error occurs, NULL will be returned.
+ *
+ * If debugfs is not enabled in the kernel, the value -ENODEV will be
+ * returned.  It is not wise to check for this value, but rather, check for
+ * NULL or !NULL instead as to eliminate the need for #ifdef in the calling
+ * code.
+ */
+struct dentry *debugfs_create_blob(const char *name, mode_t mode,
+				   struct dentry *parent,
+				   struct debugfs_blob_wrapper *blob)
+{
+	return debugfs_create_file(name, mode, parent, blob, &fops_blob);
+}
+EXPORT_SYMBOL_GPL(debugfs_create_blob);
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index bfb8a230bac9..14c5620b5cab 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -18,6 +18,7 @@
 #include <linux/mount.h>
 #include <linux/tty.h>
 #include <linux/devpts_fs.h>
+#include <linux/parser.h>
 
 #define DEVPTS_SUPER_MAGIC 0x1cd1
 
@@ -32,39 +33,60 @@ static struct {
 	umode_t mode;
 } config = {.mode = 0600};
 
+enum {
+	Opt_uid, Opt_gid, Opt_mode,
+	Opt_err
+};
+
+static match_table_t tokens = {
+	{Opt_uid, "uid=%u"},
+	{Opt_gid, "gid=%u"},
+	{Opt_mode, "mode=%o"},
+	{Opt_err, NULL}
+};
+
 static int devpts_remount(struct super_block *sb, int *flags, char *data)
 {
-	int setuid = 0;
-	int setgid = 0;
-	uid_t uid = 0;
-	gid_t gid = 0;
-	umode_t mode = 0600;
-	char *this_char;
-
-	this_char = NULL;
-	while ((this_char = strsep(&data, ",")) != NULL) {
-		int n;
-		char dummy;
-		if (!*this_char)
+	char *p;
+
+	config.setuid  = 0;
+	config.setgid  = 0;
+	config.uid     = 0;
+	config.gid     = 0;
+	config.mode    = 0600;
+
+	while ((p = strsep(&data, ",")) != NULL) {
+		substring_t args[MAX_OPT_ARGS];
+		int token;
+		int option;
+
+		if (!*p)
 			continue;
-		if (sscanf(this_char, "uid=%i%c", &n, &dummy) == 1) {
-			setuid = 1;
-			uid = n;
-		} else if (sscanf(this_char, "gid=%i%c", &n, &dummy) == 1) {
-			setgid = 1;
-			gid = n;
-		} else if (sscanf(this_char, "mode=%o%c", &n, &dummy) == 1)
-			mode = n & ~S_IFMT;
-		else {
-			printk("devpts: called with bogus options\n");
+
+		token = match_token(p, tokens, args);
+		switch (token) {
+		case Opt_uid:
+			if (match_int(&args[0], &option))
+				return -EINVAL;
+			config.uid = option;
+			config.setuid = 1;
+			break;
+		case Opt_gid:
+			if (match_int(&args[0], &option))
+				return -EINVAL;
+			config.gid = option;
+			config.setgid = 1;
+			break;
+		case Opt_mode:
+			if (match_octal(&args[0], &option))
+				return -EINVAL;
+			config.mode = option & ~S_IFMT;
+			break;
+		default:
+			printk(KERN_ERR "devpts: called with bogus options\n");
 			return -EINVAL;
 		}
 	}
-	config.setuid  = setuid;
-	config.setgid  = setgid;
-	config.uid     = uid;
-	config.gid     = gid;
-	config.mode    = mode;
 
 	return 0;
 }
diff --git a/fs/dquot.c b/fs/dquot.c
index 1966c890b48d..acf07e581f8c 100644
--- a/fs/dquot.c
+++ b/fs/dquot.c
@@ -103,12 +103,12 @@
  * (these locking rules also apply for S_NOQUOTA flag in the inode - note that
  * for altering the flag i_mutex is also needed).  If operation is holding
  * reference to dquot in other way (e.g. quotactl ops) it must be guarded by
- * dqonoff_sem.
+ * dqonoff_mutex.
  * This locking assures that:
  *   a) update/access to dquot pointers in inode is serialized
  *   b) everyone is guarded against invalidate_dquots()
  *
- * Each dquot has its dq_lock semaphore. Locked dquots might not be referenced
+ * Each dquot has its dq_lock mutex. Locked dquots might not be referenced
  * from inodes (dquot_alloc_space() and such don't check the dq_lock).
  * Currently dquot is locked only when it is being read to memory (or space for
  * it is being allocated) on the first dqget() and when it is being released on
@@ -118,9 +118,9 @@
  * spinlock to internal buffers before writing.
  *
  * Lock ordering (including related VFS locks) is the following:
- *   i_mutex > dqonoff_sem > iprune_sem > journal_lock > dqptr_sem >
- *   > dquot->dq_lock > dqio_sem
- * i_mutex on quota files is special (it's below dqio_sem)
+ *   i_mutex > dqonoff_sem > journal_lock > dqptr_sem > dquot->dq_lock >
+ *   dqio_mutex
+ * i_mutex on quota files is special (it's below dqio_mutex)
  */
 
 static DEFINE_SPINLOCK(dq_list_lock);
@@ -281,8 +281,8 @@ static inline void remove_inuse(struct dquot *dquot)
 
 static void wait_on_dquot(struct dquot *dquot)
 {
-	down(&dquot->dq_lock);
-	up(&dquot->dq_lock);
+	mutex_lock(&dquot->dq_lock);
+	mutex_unlock(&dquot->dq_lock);
 }
 
 #define mark_dquot_dirty(dquot) ((dquot)->dq_sb->dq_op->mark_dirty(dquot))
@@ -321,8 +321,8 @@ int dquot_acquire(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dquot->dq_lock);
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dquot->dq_lock);
+	mutex_lock(&dqopt->dqio_mutex);
 	if (!test_bit(DQ_READ_B, &dquot->dq_flags))
 		ret = dqopt->ops[dquot->dq_type]->read_dqblk(dquot);
 	if (ret < 0)
@@ -343,8 +343,8 @@ int dquot_acquire(struct dquot *dquot)
 	}
 	set_bit(DQ_ACTIVE_B, &dquot->dq_flags);
 out_iolock:
-	up(&dqopt->dqio_sem);
-	up(&dquot->dq_lock);
+	mutex_unlock(&dqopt->dqio_mutex);
+	mutex_unlock(&dquot->dq_lock);
 	return ret;
 }
 
@@ -356,7 +356,7 @@ int dquot_commit(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	spin_lock(&dq_list_lock);
 	if (!clear_dquot_dirty(dquot)) {
 		spin_unlock(&dq_list_lock);
@@ -373,7 +373,7 @@ int dquot_commit(struct dquot *dquot)
 			ret = ret2;
 	}
 out_sem:
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	return ret;
 }
 
@@ -385,11 +385,11 @@ int dquot_release(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dquot->dq_lock);
+	mutex_lock(&dquot->dq_lock);
 	/* Check whether we are not racing with some other dqget() */
 	if (atomic_read(&dquot->dq_count) > 1)
 		goto out_dqlock;
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	if (dqopt->ops[dquot->dq_type]->release_dqblk) {
 		ret = dqopt->ops[dquot->dq_type]->release_dqblk(dquot);
 		/* Write the info */
@@ -399,31 +399,57 @@ int dquot_release(struct dquot *dquot)
 			ret = ret2;
 	}
 	clear_bit(DQ_ACTIVE_B, &dquot->dq_flags);
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 out_dqlock:
-	up(&dquot->dq_lock);
+	mutex_unlock(&dquot->dq_lock);
 	return ret;
 }
 
 /* Invalidate all dquots on the list. Note that this function is called after
  * quota is disabled and pointers from inodes removed so there cannot be new
- * quota users. Also because we hold dqonoff_sem there can be no quota users
- * for this sb+type at all. */
+ * quota users. There can still be some users of quotas due to inodes being
+ * just deleted or pruned by prune_icache() (those are not attached to any
+ * list). We have to wait for such users.
+ */
 static void invalidate_dquots(struct super_block *sb, int type)
 {
 	struct dquot *dquot, *tmp;
 
+restart:
 	spin_lock(&dq_list_lock);
 	list_for_each_entry_safe(dquot, tmp, &inuse_list, dq_inuse) {
 		if (dquot->dq_sb != sb)
 			continue;
 		if (dquot->dq_type != type)
 			continue;
-#ifdef __DQUOT_PARANOIA
-		if (atomic_read(&dquot->dq_count))
-			BUG();
-#endif
-		/* Quota now has no users and it has been written on last dqput() */
+		/* Wait for dquot users */
+		if (atomic_read(&dquot->dq_count)) {
+			DEFINE_WAIT(wait);
+
+			atomic_inc(&dquot->dq_count);
+			prepare_to_wait(&dquot->dq_wait_unused, &wait,
+					TASK_UNINTERRUPTIBLE);
+			spin_unlock(&dq_list_lock);
+			/* Once dqput() wakes us up, we know it's time to free
+			 * the dquot.
+			 * IMPORTANT: we rely on the fact that there is always
+			 * at most one process waiting for dquot to free.
+			 * Otherwise dq_count would be > 1 and we would never
+			 * wake up.
+			 */
+			if (atomic_read(&dquot->dq_count) > 1)
+				schedule();
+			finish_wait(&dquot->dq_wait_unused, &wait);
+			dqput(dquot);
+			/* At this moment dquot() need not exist (it could be
+			 * reclaimed by prune_dqcache(). Hence we must
+			 * restart. */
+			goto restart;
+		}
+		/*
+		 * Quota now has no users and it has been written on last
+		 * dqput()
+		 */
 		remove_dquot_hash(dquot);
 		remove_free_dquot(dquot);
 		remove_inuse(dquot);
@@ -439,7 +465,7 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	struct quota_info *dqopt = sb_dqopt(sb);
 	int cnt;
 
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (type != -1 && cnt != type)
 			continue;
@@ -474,7 +500,7 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	spin_lock(&dq_list_lock);
 	dqstats.syncs++;
 	spin_unlock(&dq_list_lock);
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 
 	return 0;
 }
@@ -515,7 +541,7 @@ static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
 /*
  * Put reference to dquot
  * NOTE: If you change this function please check whether dqput_blocks() works right...
- * MUST be called with either dqptr_sem or dqonoff_sem held
+ * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
 static void dqput(struct dquot *dquot)
 {
@@ -540,6 +566,10 @@ we_slept:
 	if (atomic_read(&dquot->dq_count) > 1) {
 		/* We have more than one user... nothing to do */
 		atomic_dec(&dquot->dq_count);
+		/* Releasing dquot during quotaoff phase? */
+		if (!sb_has_quota_enabled(dquot->dq_sb, dquot->dq_type) &&
+		    atomic_read(&dquot->dq_count) == 1)
+			wake_up(&dquot->dq_wait_unused);
 		spin_unlock(&dq_list_lock);
 		return;
 	}
@@ -576,11 +606,12 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 		return NODQUOT;
 
 	memset((caddr_t)dquot, 0, sizeof(struct dquot));
-	sema_init(&dquot->dq_lock, 1);
+	mutex_init(&dquot->dq_lock);
 	INIT_LIST_HEAD(&dquot->dq_free);
 	INIT_LIST_HEAD(&dquot->dq_inuse);
 	INIT_HLIST_NODE(&dquot->dq_hash);
 	INIT_LIST_HEAD(&dquot->dq_dirty);
+	init_waitqueue_head(&dquot->dq_wait_unused);
 	dquot->dq_sb = sb;
 	dquot->dq_type = type;
 	atomic_set(&dquot->dq_count, 1);
@@ -590,7 +621,7 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 
 /*
  * Get reference to dquot
- * MUST be called with either dqptr_sem or dqonoff_sem held
+ * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
 static struct dquot *dqget(struct super_block *sb, unsigned int id, int type)
 {
@@ -656,7 +687,7 @@ static int dqinit_needed(struct inode *inode, int type)
 	return 0;
 }
 
-/* This routine is guarded by dqonoff_sem semaphore */
+/* This routine is guarded by dqonoff_mutex mutex */
 static void add_dquot_ref(struct super_block *sb, int type)
 {
 	struct list_head *p;
@@ -732,13 +763,9 @@ static void drop_dquot_ref(struct super_block *sb, int type)
 {
 	LIST_HEAD(tofree_head);
 
-	/* We need to be guarded against prune_icache to reach all the
-	 * inodes - otherwise some can be on the local list of prune_icache */
-	down(&iprune_sem);
 	down_write(&sb_dqopt(sb)->dqptr_sem);
 	remove_dquot_ref(sb, type, &tofree_head);
 	up_write(&sb_dqopt(sb)->dqptr_sem);
-	up(&iprune_sem);
 	put_dquot_list(&tofree_head);
 }
 
@@ -938,8 +965,8 @@ int dquot_initialize(struct inode *inode, int type)
 	unsigned int id = 0;
 	int cnt, ret = 0;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return 0;
 	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
@@ -1002,8 +1029,8 @@ int dquot_alloc_space(struct inode *inode, qsize_t number, int warn)
 	int cnt, ret = NO_QUOTA;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode)) {
 out_add:
 		inode_add_bytes(inode, number);
@@ -1051,8 +1078,8 @@ int dquot_alloc_inode(const struct inode *inode, unsigned long number)
 	int cnt, ret = NO_QUOTA;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
@@ -1095,8 +1122,8 @@ int dquot_free_space(struct inode *inode, qsize_t number)
 {
 	unsigned int cnt;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode)) {
 out_sub:
 		inode_sub_bytes(inode, number);
@@ -1131,8 +1158,8 @@ int dquot_free_inode(const struct inode *inode, unsigned long number)
 {
 	unsigned int cnt;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
@@ -1171,8 +1198,8 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr)
 	    chgid = (iattr->ia_valid & ATTR_GID) && inode->i_gid != iattr->ia_gid;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	/* Clear the arrays */
@@ -1266,9 +1293,9 @@ int dquot_commit_info(struct super_block *sb, int type)
 	int ret;
 	struct quota_info *dqopt = sb_dqopt(sb);
 
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	ret = dqopt->ops[type]->write_file_info(sb, type);
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	return ret;
 }
 
@@ -1324,7 +1351,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 	struct inode *toputinode[MAXQUOTAS];
 
 	/* We need to serialize quota_off() for device */
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		toputinode[cnt] = NULL;
 		if (type != -1 && cnt != type)
@@ -1353,7 +1380,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 		dqopt->info[cnt].dqi_bgrace = 0;
 		dqopt->ops[cnt] = NULL;
 	}
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 	/* Sync the superblock so that buffers with quota data are written to
 	 * disk (and so userspace sees correct data afterwards). */
 	if (sb->s_op->sync_fs)
@@ -1366,7 +1393,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 	 * changes done by userspace on the next quotaon() */
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
 		if (toputinode[cnt]) {
-			down(&dqopt->dqonoff_sem);
+			mutex_lock(&dqopt->dqonoff_mutex);
 			/* If quota was reenabled in the meantime, we have
 			 * nothing to do */
 			if (!sb_has_quota_enabled(sb, cnt)) {
@@ -1378,7 +1405,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 				mark_inode_dirty(toputinode[cnt]);
 				iput(toputinode[cnt]);
 			}
-			up(&dqopt->dqonoff_sem);
+			mutex_unlock(&dqopt->dqonoff_mutex);
 		}
 	if (sb->s_bdev)
 		invalidate_bdev(sb->s_bdev, 0);
@@ -1419,7 +1446,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 	/* And now flush the block cache so that kernel sees the changes */
 	invalidate_bdev(sb->s_bdev, 0);
 	mutex_lock(&inode->i_mutex);
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	if (sb_has_quota_enabled(sb, type)) {
 		error = -EBUSY;
 		goto out_lock;
@@ -1444,17 +1471,17 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 	dqopt->ops[type] = fmt->qf_ops;
 	dqopt->info[type].dqi_format = fmt;
 	INIT_LIST_HEAD(&dqopt->info[type].dqi_dirty_list);
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	if ((error = dqopt->ops[type]->read_file_info(sb, type)) < 0) {
-		up(&dqopt->dqio_sem);
+		mutex_unlock(&dqopt->dqio_mutex);
 		goto out_file_init;
 	}
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	mutex_unlock(&inode->i_mutex);
 	set_enable_flags(dqopt, type);
 
 	add_dquot_ref(sb, type);
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 
 	return 0;
 
@@ -1462,7 +1489,7 @@ out_file_init:
 	dqopt->files[type] = NULL;
 	iput(inode);
 out_lock:
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 	if (oldflags != -1) {
 		down_write(&dqopt->dqptr_sem);
 		/* Set the flags back (in the case of accidental quotaon()
@@ -1550,14 +1577,14 @@ int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d
 {
 	struct dquot *dquot;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!(dquot = dqget(sb, id, type))) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	do_get_dqblk(dquot, di);
 	dqput(dquot);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1619,14 +1646,14 @@ int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d
 {
 	struct dquot *dquot;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!(dquot = dqget(sb, id, type))) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	do_set_dqblk(dquot, di);
 	dqput(dquot);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1635,9 +1662,9 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
   
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!sb_has_quota_enabled(sb, type)) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	mi = sb_dqopt(sb)->info + type;
@@ -1647,7 +1674,7 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	ii->dqi_flags = mi->dqi_flags & DQF_MASK;
 	ii->dqi_valid = IIF_ALL;
 	spin_unlock(&dq_data_lock);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1656,9 +1683,9 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!sb_has_quota_enabled(sb, type)) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	mi = sb_dqopt(sb)->info + type;
@@ -1673,7 +1700,7 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	mark_info_dirty(sb, type);
 	/* Force write to disk */
 	sb->dq_op->write_info(sb, type);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 4284cd31eba6..1c2b16fda13a 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -34,6 +34,7 @@
 #include <linux/eventpoll.h>
 #include <linux/mount.h>
 #include <linux/bitops.h>
+#include <linux/mutex.h>
 #include <asm/uaccess.h>
 #include <asm/system.h>
 #include <asm/io.h>
@@ -46,7 +47,7 @@
  * LOCKING:
  * There are three level of locking required by epoll :
  *
- * 1) epsem (semaphore)
+ * 1) epmutex (mutex)
  * 2) ep->sem (rw_semaphore)
  * 3) ep->lock (rw_lock)
  *
@@ -67,9 +68,9 @@
  * if a file has been pushed inside an epoll set and it is then
  * close()d without a previous call toepoll_ctl(EPOLL_CTL_DEL).
  * It is possible to drop the "ep->sem" and to use the global
- * semaphore "epsem" (together with "ep->lock") to have it working,
+ * semaphore "epmutex" (together with "ep->lock") to have it working,
  * but having "ep->sem" will make the interface more scalable.
- * Events that require holding "epsem" are very rare, while for
+ * Events that require holding "epmutex" are very rare, while for
  * normal operations the epoll private "ep->sem" will guarantee
  * a greater scalability.
  */
@@ -274,7 +275,7 @@ static struct super_block *eventpollfs_get_sb(struct file_system_type *fs_type,
 /*
  * This semaphore is used to serialize ep_free() and eventpoll_release_file().
  */
-static struct semaphore epsem;
+static struct mutex epmutex;
 
 /* Safe wake up implementation */
 static struct poll_safewake psw;
@@ -451,15 +452,6 @@ static void ep_poll_safewake(struct poll_safewake *psw, wait_queue_head_t *wq)
 }
 
 
-/* Used to initialize the epoll bits inside the "struct file" */
-void eventpoll_init_file(struct file *file)
-{
-
-	INIT_LIST_HEAD(&file->f_ep_links);
-	spin_lock_init(&file->f_ep_lock);
-}
-
-
 /*
  * This is called from eventpoll_release() to unlink files from the eventpoll
  * interface. We need to have this facility to cleanup correctly files that are
@@ -477,10 +469,10 @@ void eventpoll_release_file(struct file *file)
 	 * cleanup path, and this means that noone is using this file anymore.
 	 * The only hit might come from ep_free() but by holding the semaphore
 	 * will correctly serialize the operation. We do need to acquire
-	 * "ep->sem" after "epsem" because ep_remove() requires it when called
+	 * "ep->sem" after "epmutex" because ep_remove() requires it when called
 	 * from anywhere but ep_free().
 	 */
-	down(&epsem);
+	mutex_lock(&epmutex);
 
 	while (!list_empty(lsthead)) {
 		epi = list_entry(lsthead->next, struct epitem, fllink);
@@ -492,7 +484,7 @@ void eventpoll_release_file(struct file *file)
 		up_write(&ep->sem);
 	}
 
-	up(&epsem);
+	mutex_unlock(&epmutex);
 }
 
 
@@ -819,9 +811,9 @@ static void ep_free(struct eventpoll *ep)
 	 * We do not need to hold "ep->sem" here because the epoll file
 	 * is on the way to be removed and no one has references to it
 	 * anymore. The only hit might come from eventpoll_release_file() but
-	 * holding "epsem" is sufficent here.
+	 * holding "epmutex" is sufficent here.
 	 */
-	down(&epsem);
+	mutex_lock(&epmutex);
 
 	/*
 	 * Walks through the whole tree by unregistering poll callbacks.
@@ -843,7 +835,7 @@ static void ep_free(struct eventpoll *ep)
 		ep_remove(ep, epi);
 	}
 
-	up(&epsem);
+	mutex_unlock(&epmutex);
 }
 
 
@@ -1615,7 +1607,7 @@ static int __init eventpoll_init(void)
 {
 	int error;
 
-	init_MUTEX(&epsem);
+	mutex_init(&epmutex);
 
 	/* Initialize the structure used to perform safe poll wait head wake ups */
 	ep_poll_safewake_init(&psw);
diff --git a/fs/ext2/namei.c b/fs/ext2/namei.c
index ad1432a2a62e..4ca824985321 100644
--- a/fs/ext2/namei.c
+++ b/fs/ext2/namei.c
@@ -36,22 +36,6 @@
 #include "acl.h"
 #include "xip.h"
 
-/*
- * Couple of helper functions - make the code slightly cleaner.
- */
-
-static inline void ext2_inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void ext2_dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static inline int ext2_add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = ext2_add_link(dentry, inode);
@@ -59,7 +43,7 @@ static inline int ext2_add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -201,7 +185,7 @@ out:
 	return err;
 
 out_fail:
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput (inode);
 	goto out;
 }
@@ -215,7 +199,7 @@ static int ext2_link (struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	ext2_inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	return ext2_add_nondir(dentry, inode);
@@ -229,7 +213,7 @@ static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	if (dir->i_nlink >= EXT2_LINK_MAX)
 		goto out;
 
-	ext2_inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = ext2_new_inode (dir, S_IFDIR | mode);
 	err = PTR_ERR(inode);
@@ -243,7 +227,7 @@ static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	else
 		inode->i_mapping->a_ops = &ext2_aops;
 
-	ext2_inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = ext2_make_empty(inode, dir);
 	if (err)
@@ -258,11 +242,11 @@ out:
 	return err;
 
 out_fail:
-	ext2_dec_count(inode);
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	ext2_dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -282,7 +266,7 @@ static int ext2_unlink(struct inode * dir, struct dentry *dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	err = 0;
 out:
 	return err;
@@ -297,8 +281,8 @@ static int ext2_rmdir (struct inode * dir, struct dentry *dentry)
 		err = ext2_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			ext2_dec_count(inode);
-			ext2_dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	return err;
@@ -338,41 +322,41 @@ static int ext2_rename (struct inode * old_dir, struct dentry * old_dentry,
 		new_de = ext2_find_entry (new_dir, new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		ext2_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		ext2_set_link(new_dir, new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		ext2_dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= EXT2_LINK_MAX)
 				goto out_dir;
 		}
-		ext2_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = ext2_add_link(new_dentry, old_inode);
 		if (err) {
-			ext2_dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			ext2_inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	/*
 	 * Like most other Unix systems, set the ctime for inodes on a
  	 * rename.
-	 * ext2_dec_count() will mark the inode dirty.
+	 * inode_dec_link_count() will mark the inode dirty.
 	 */
 	old_inode->i_ctime = CURRENT_TIME_SEC;
 
 	ext2_delete_entry (old_de, old_page);
-	ext2_dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		ext2_set_link(old_inode, dir_de, dir_page, new_dir);
-		ext2_dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 832867aef3dc..773459164bb2 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -95,11 +95,10 @@ static int ext3_readdir(struct file * filp,
 			 void * dirent, filldir_t filldir)
 {
 	int error = 0;
-	unsigned long offset, blk;
-	int i, num, stored;
-	struct buffer_head * bh, * tmp, * bha[16];
-	struct ext3_dir_entry_2 * de;
-	struct super_block * sb;
+	unsigned long offset;
+	int i, stored;
+	struct ext3_dir_entry_2 *de;
+	struct super_block *sb;
 	int err;
 	struct inode *inode = filp->f_dentry->d_inode;
 	int ret = 0;
@@ -124,12 +123,29 @@ static int ext3_readdir(struct file * filp,
 	}
 #endif
 	stored = 0;
-	bh = NULL;
 	offset = filp->f_pos & (sb->s_blocksize - 1);
 
 	while (!error && !stored && filp->f_pos < inode->i_size) {
-		blk = (filp->f_pos) >> EXT3_BLOCK_SIZE_BITS(sb);
-		bh = ext3_bread(NULL, inode, blk, 0, &err);
+		unsigned long blk = filp->f_pos >> EXT3_BLOCK_SIZE_BITS(sb);
+		struct buffer_head map_bh;
+		struct buffer_head *bh = NULL;
+
+		map_bh.b_state = 0;
+		err = ext3_get_block_handle(NULL, inode, blk, &map_bh, 0, 0);
+		if (!err) {
+			page_cache_readahead(sb->s_bdev->bd_inode->i_mapping,
+				&filp->f_ra,
+				filp,
+				map_bh.b_blocknr >>
+					(PAGE_CACHE_SHIFT - inode->i_blkbits),
+				1);
+			bh = ext3_bread(NULL, inode, blk, 0, &err);
+		}
+
+		/*
+		 * We ignore I/O errors on directories so users have a chance
+		 * of recovering data when there's a bad sector
+		 */
 		if (!bh) {
 			ext3_error (sb, "ext3_readdir",
 				"directory #%lu contains a hole at offset %lu",
@@ -138,26 +154,6 @@ static int ext3_readdir(struct file * filp,
 			continue;
 		}
 
-		/*
-		 * Do the readahead
-		 */
-		if (!offset) {
-			for (i = 16 >> (EXT3_BLOCK_SIZE_BITS(sb) - 9), num = 0;
-			     i > 0; i--) {
-				tmp = ext3_getblk (NULL, inode, ++blk, 0, &err);
-				if (tmp && !buffer_uptodate(tmp) &&
-						!buffer_locked(tmp))
-					bha[num++] = tmp;
-				else
-					brelse (tmp);
-			}
-			if (num) {
-				ll_rw_block (READA, num, bha);
-				for (i = 0; i < num; i++)
-					brelse (bha[i]);
-			}
-		}
-
 revalidate:
 		/* If the dir block has changed since the last call to
 		 * readdir(2), then we might be pointing to an invalid
diff --git a/fs/ext3/file.c b/fs/ext3/file.c
index 98e78345ead9..59098ea56711 100644
--- a/fs/ext3/file.c
+++ b/fs/ext3/file.c
@@ -37,9 +37,9 @@ static int ext3_release_file (struct inode * inode, struct file * filp)
 	if ((filp->f_mode & FMODE_WRITE) &&
 			(atomic_read(&inode->i_writecount) == 1))
 	{
-		down(&EXT3_I(inode)->truncate_sem);
+		mutex_lock(&EXT3_I(inode)->truncate_mutex);
 		ext3_discard_reservation(inode);
-		up(&EXT3_I(inode)->truncate_sem);
+		mutex_unlock(&EXT3_I(inode)->truncate_mutex);
 	}
 	if (is_dx(inode) && filp->private_data)
 		ext3_htree_free_dir_info(filp->private_data);
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index 0384e539b88f..2c361377e0a5 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -671,7 +671,7 @@ err_out:
  * The BKL may not be held on entry here.  Be sure to take it early.
  */
 
-static int
+int
 ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 		struct buffer_head *bh_result, int create, int extend_disksize)
 {
@@ -702,7 +702,7 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 	if (!create || err == -EIO)
 		goto cleanup;
 
-	down(&ei->truncate_sem);
+	mutex_lock(&ei->truncate_mutex);
 
 	/*
 	 * If the indirect block is missing while we are reading
@@ -723,7 +723,7 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 		}
 		partial = ext3_get_branch(inode, depth, offsets, chain, &err);
 		if (!partial) {
-			up(&ei->truncate_sem);
+			mutex_unlock(&ei->truncate_mutex);
 			if (err)
 				goto cleanup;
 			clear_buffer_new(bh_result);
@@ -759,13 +759,13 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 		err = ext3_splice_branch(handle, inode, iblock, chain,
 					 partial, left);
 	/*
-	 * i_disksize growing is protected by truncate_sem.  Don't forget to
+	 * i_disksize growing is protected by truncate_mutex.  Don't forget to
 	 * protect it if you're about to implement concurrent
 	 * ext3_get_block() -bzzz
 	*/
 	if (!err && extend_disksize && inode->i_size > ei->i_disksize)
 		ei->i_disksize = inode->i_size;
-	up(&ei->truncate_sem);
+	mutex_unlock(&ei->truncate_mutex);
 	if (err)
 		goto cleanup;
 
@@ -1227,7 +1227,7 @@ static int journal_dirty_data_fn(handle_t *handle, struct buffer_head *bh)
  *	ext3_file_write() -> generic_file_write() -> __alloc_pages() -> ...
  *
  * Same applies to ext3_get_block().  We will deadlock on various things like
- * lock_journal and i_truncate_sem.
+ * lock_journal and i_truncate_mutex.
  *
  * Setting PF_MEMALLOC here doesn't work - too many internal memory
  * allocations fail.
@@ -2161,7 +2161,7 @@ void ext3_truncate(struct inode * inode)
 	 * From here we block out all ext3_get_block() callers who want to
 	 * modify the block allocation tree.
 	 */
-	down(&ei->truncate_sem);
+	mutex_lock(&ei->truncate_mutex);
 
 	if (n == 1) {		/* direct blocks */
 		ext3_free_data(handle, inode, NULL, i_data+offsets[0],
@@ -2228,7 +2228,7 @@ do_indirects:
 
 	ext3_discard_reservation(inode);
 
-	up(&ei->truncate_sem);
+	mutex_unlock(&ei->truncate_mutex);
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
 	ext3_mark_inode_dirty(handle, inode);
 
diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c
index 556cd5510078..aaf1da17b6d4 100644
--- a/fs/ext3/ioctl.c
+++ b/fs/ext3/ioctl.c
@@ -182,7 +182,7 @@ flags_err:
 		 * need to allocate reservation structure for this inode
 		 * before set the window size
 		 */
-		down(&ei->truncate_sem);
+		mutex_lock(&ei->truncate_mutex);
 		if (!ei->i_block_alloc_info)
 			ext3_init_block_alloc_info(inode);
 
@@ -190,7 +190,7 @@ flags_err:
 			struct ext3_reserve_window_node *rsv = &ei->i_block_alloc_info->rsv_window_node;
 			rsv->rsv_goal_size = rsv_window_size;
 		}
-		up(&ei->truncate_sem);
+		mutex_unlock(&ei->truncate_mutex);
 		return 0;
 	}
 	case EXT3_IOC_GROUP_EXTEND: {
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 56bf76586019..efe5b20d7a5a 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -472,7 +472,7 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 #ifdef CONFIG_EXT3_FS_XATTR
 		init_rwsem(&ei->xattr_sem);
 #endif
-		init_MUTEX(&ei->truncate_sem);
+		mutex_init(&ei->truncate_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -2382,8 +2382,8 @@ static int ext3_statfs (struct super_block * sb, struct kstatfs * buf)
  * Process 1                         Process 2
  * ext3_create()                     quota_sync()
  *   journal_start()                   write_dquot()
- *   DQUOT_INIT()                        down(dqio_sem)
- *     down(dqio_sem)                    journal_start()
+ *   DQUOT_INIT()                        down(dqio_mutex)
+ *     down(dqio_mutex)                    journal_start()
  *
  */
 
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index db0de5c621c7..4095bc149eb1 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -114,7 +114,7 @@ static inline int fat_get_entry(struct inode *dir, loff_t *pos,
 }
 
 /*
- * Convert Unicode 16 to UTF8, translated Unicode, or ASCII.
+ * Convert Unicode 16 to UTF-8, translated Unicode, or ASCII.
  * If uni_xlate is enabled and we can't get a 1:1 conversion, use a
  * colon as an escape character since it is normally invalid on the vfat
  * filesystem. The following four characters are the hexadecimal digits
diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c
index a1a9e0451217..ab171ea8e869 100644
--- a/fs/fat/fatent.c
+++ b/fs/fat/fatent.c
@@ -267,19 +267,19 @@ static struct fatent_operations fat32_ops = {
 
 static inline void lock_fat(struct msdos_sb_info *sbi)
 {
-	down(&sbi->fat_lock);
+	mutex_lock(&sbi->fat_lock);
 }
 
 static inline void unlock_fat(struct msdos_sb_info *sbi)
 {
-	up(&sbi->fat_lock);
+	mutex_unlock(&sbi->fat_lock);
 }
 
 void fat_ent_access_init(struct super_block *sb)
 {
 	struct msdos_sb_info *sbi = MSDOS_SB(sb);
 
-	init_MUTEX(&sbi->fat_lock);
+	mutex_init(&sbi->fat_lock);
 
 	switch (sbi->fat_bits) {
 	case 32:
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index e7f4aa7fc686..e78d7b4842cc 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -1101,7 +1101,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug,
 			return -EINVAL;
 		}
 	}
-	/* UTF8 doesn't provide FAT semantics */
+	/* UTF-8 doesn't provide FAT semantics */
 	if (!strcmp(opts->iocharset, "utf8")) {
 		printk(KERN_ERR "FAT: utf8 is not a recommended IO charset"
 		       " for FAT filesystems, filesystem will be case sensitive!\n");
diff --git a/fs/fcntl.c b/fs/fcntl.c
index dc4a7007f4e7..03c789560fb8 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -73,8 +73,8 @@ repeat:
 	 * orig_start..fdt->next_fd
 	 */
 	start = orig_start;
-	if (start < fdt->next_fd)
-		start = fdt->next_fd;
+	if (start < files->next_fd)
+		start = files->next_fd;
 
 	newfd = start;
 	if (start < fdt->max_fdset) {
@@ -102,9 +102,8 @@ repeat:
 	 * we reacquire the fdtable pointer and use it while holding
 	 * the lock, no one can free it during that time.
 	 */
-	fdt = files_fdtable(files);
-	if (start <= fdt->next_fd)
-		fdt->next_fd = newfd + 1;
+	if (start <= files->next_fd)
+		files->next_fd = newfd + 1;
 
 	error = newfd;
 	
diff --git a/fs/file.c b/fs/file.c
index cea7cbea11d0..bbc743314730 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -125,7 +125,8 @@ static void free_fdtable_rcu(struct rcu_head *rcu)
 		kmem_cache_free(files_cachep, fdt->free_files);
 		return;
 	}
-	if (fdt->max_fdset <= __FD_SETSIZE && fdt->max_fds <= NR_OPEN_DEFAULT) {
+	if (fdt->max_fdset <= EMBEDDED_FD_SET_SIZE &&
+		fdt->max_fds <= NR_OPEN_DEFAULT) {
 		/*
 		 * The fdtable was embedded
 		 */
@@ -155,8 +156,9 @@ static void free_fdtable_rcu(struct rcu_head *rcu)
 
 void free_fdtable(struct fdtable *fdt)
 {
-	if (fdt->free_files || fdt->max_fdset > __FD_SETSIZE ||
-					fdt->max_fds > NR_OPEN_DEFAULT)
+	if (fdt->free_files ||
+		fdt->max_fdset > EMBEDDED_FD_SET_SIZE ||
+		fdt->max_fds > NR_OPEN_DEFAULT)
 		call_rcu(&fdt->rcu, free_fdtable_rcu);
 }
 
@@ -199,7 +201,6 @@ static void copy_fdtable(struct fdtable *nfdt, struct fdtable *fdt)
 		       (nfdt->max_fds - fdt->max_fds) *
 					sizeof(struct file *));
 	}
-	nfdt->next_fd = fdt->next_fd;
 }
 
 /*
@@ -220,11 +221,9 @@ fd_set * alloc_fdset(int num)
 
 void free_fdset(fd_set *array, int num)
 {
-	int size = num / 8;
-
-	if (num <= __FD_SETSIZE) /* Don't free an embedded fdset */
+	if (num <= EMBEDDED_FD_SET_SIZE) /* Don't free an embedded fdset */
 		return;
-	else if (size <= PAGE_SIZE)
+	else if (num <= 8 * PAGE_SIZE)
 		kfree(array);
 	else
 		vfree(array);
@@ -237,22 +236,17 @@ static struct fdtable *alloc_fdtable(int nr)
   	fd_set *new_openset = NULL, *new_execset = NULL;
 	struct file **new_fds;
 
-	fdt = kmalloc(sizeof(*fdt), GFP_KERNEL);
+	fdt = kzalloc(sizeof(*fdt), GFP_KERNEL);
 	if (!fdt)
   		goto out;
-	memset(fdt, 0, sizeof(*fdt));
 
-	nfds = __FD_SETSIZE;
+	nfds = 8 * L1_CACHE_BYTES;
   	/* Expand to the max in easy steps */
-  	do {
-		if (nfds < (PAGE_SIZE * 8))
-			nfds = PAGE_SIZE * 8;
-		else {
-			nfds = nfds * 2;
-			if (nfds > NR_OPEN)
-				nfds = NR_OPEN;
-		}
-	} while (nfds <= nr);
+  	while (nfds <= nr) {
+		nfds = nfds * 2;
+		if (nfds > NR_OPEN)
+			nfds = NR_OPEN;
+	}
 
   	new_openset = alloc_fdset(nfds);
   	new_execset = alloc_fdset(nfds);
diff --git a/fs/file_table.c b/fs/file_table.c
index 44fabeaa9415..bcea1998b4de 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -88,6 +88,7 @@ int proc_nr_files(ctl_table *table, int write, struct file *filp,
  */
 struct file *get_empty_filp(void)
 {
+	struct task_struct *tsk;
 	static int old_max;
 	struct file * f;
 
@@ -112,13 +113,14 @@ struct file *get_empty_filp(void)
 	if (security_file_alloc(f))
 		goto fail_sec;
 
-	eventpoll_init_file(f);
+	tsk = current;
+	INIT_LIST_HEAD(&f->f_u.fu_list);
 	atomic_set(&f->f_count, 1);
-	f->f_uid = current->fsuid;
-	f->f_gid = current->fsgid;
 	rwlock_init(&f->f_owner.lock);
+	f->f_uid = tsk->fsuid;
+	f->f_gid = tsk->fsgid;
+	eventpoll_init_file(f);
 	/* f->f_version: 0 */
-	INIT_LIST_HEAD(&f->f_u.fu_list);
 	return f;
 
 over:
diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h
index 6628c3b352cb..4c6473ab3b34 100644
--- a/fs/hpfs/hpfs_fn.h
+++ b/fs/hpfs/hpfs_fn.h
@@ -9,6 +9,7 @@
 //#define DBG
 //#define DEBUG_LOCKS
 
+#include <linux/mutex.h>
 #include <linux/pagemap.h>
 #include <linux/buffer_head.h>
 #include <linux/hpfs_fs.h>
@@ -57,8 +58,8 @@ struct hpfs_inode_info {
 	unsigned i_ea_uid : 1;	/* file's uid is stored in ea */
 	unsigned i_ea_gid : 1;	/* file's gid is stored in ea */
 	unsigned i_dirty : 1;
-	struct semaphore i_sem;
-	struct semaphore i_parent;
+	struct mutex i_mutex;
+	struct mutex i_parent_mutex;
 	loff_t **i_rddir_off;
 	struct inode vfs_inode;
 };
diff --git a/fs/hpfs/inode.c b/fs/hpfs/inode.c
index e3d17e9ea6c1..56f2c338c4d9 100644
--- a/fs/hpfs/inode.c
+++ b/fs/hpfs/inode.c
@@ -186,9 +186,9 @@ void hpfs_write_inode(struct inode *i)
 		kfree(hpfs_inode->i_rddir_off);
 		hpfs_inode->i_rddir_off = NULL;
 	}
-	down(&hpfs_inode->i_parent);
+	mutex_lock(&hpfs_inode->i_parent_mutex);
 	if (!i->i_nlink) {
-		up(&hpfs_inode->i_parent);
+		mutex_unlock(&hpfs_inode->i_parent_mutex);
 		return;
 	}
 	parent = iget_locked(i->i_sb, hpfs_inode->i_parent_dir);
@@ -199,14 +199,14 @@ void hpfs_write_inode(struct inode *i)
 			hpfs_read_inode(parent);
 			unlock_new_inode(parent);
 		}
-		down(&hpfs_inode->i_sem);
+		mutex_lock(&hpfs_inode->i_mutex);
 		hpfs_write_inode_nolock(i);
-		up(&hpfs_inode->i_sem);
+		mutex_unlock(&hpfs_inode->i_mutex);
 		iput(parent);
 	} else {
 		mark_inode_dirty(i);
 	}
-	up(&hpfs_inode->i_parent);
+	mutex_unlock(&hpfs_inode->i_parent_mutex);
 }
 
 void hpfs_write_inode_nolock(struct inode *i)
diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c
index 8ff8fc433fc1..a03abb12c610 100644
--- a/fs/hpfs/namei.c
+++ b/fs/hpfs/namei.c
@@ -60,7 +60,7 @@ static int hpfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	if (dee.read_only)
 		result->i_mode &= ~0222;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail3;
@@ -101,11 +101,11 @@ static int hpfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 		hpfs_write_inode_nolock(result);
 	}
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 bail3:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail2:
 	hpfs_brelse4(&qbh0);
@@ -168,7 +168,7 @@ static int hpfs_create(struct inode *dir, struct dentry *dentry, int mode, struc
 	result->i_data.a_ops = &hpfs_aops;
 	hpfs_i(result)->mmu_private = 0;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -193,12 +193,12 @@ static int hpfs_create(struct inode *dir, struct dentry *dentry, int mode, struc
 		hpfs_write_inode_nolock(result);
 	}
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -254,7 +254,7 @@ static int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t
 	result->i_blocks = 1;
 	init_special_inode(result, mode, rdev);
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -271,12 +271,12 @@ static int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t
 
 	hpfs_write_inode_nolock(result);
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	brelse(bh);
 	unlock_kernel();
 	return 0;
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -333,7 +333,7 @@ static int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *sy
 	result->i_op = &page_symlink_inode_operations;
 	result->i_data.a_ops = &hpfs_symlink_aops;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -352,11 +352,11 @@ static int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *sy
 
 	hpfs_write_inode_nolock(result);
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -382,8 +382,8 @@ static int hpfs_unlink(struct inode *dir, struct dentry *dentry)
 	lock_kernel();
 	hpfs_adjust_length((char *)name, &len);
 again:
-	down(&hpfs_i(inode)->i_parent);
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	err = -ENOENT;
 	de = map_dirent(dir, hpfs_i(dir)->i_dno, (char *)name, len, &dno, &qbh);
 	if (!de)
@@ -410,8 +410,8 @@ again:
 		if (rep++)
 			break;
 
-		up(&hpfs_i(dir)->i_sem);
-		up(&hpfs_i(inode)->i_parent);
+		mutex_unlock(&hpfs_i(dir)->i_mutex);
+		mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 		d_drop(dentry);
 		spin_lock(&dentry->d_lock);
 		if (atomic_read(&dentry->d_count) > 1 ||
@@ -442,8 +442,8 @@ again:
 out1:
 	hpfs_brelse4(&qbh);
 out:
-	up(&hpfs_i(dir)->i_sem);
-	up(&hpfs_i(inode)->i_parent);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
+	mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
@@ -463,8 +463,8 @@ static int hpfs_rmdir(struct inode *dir, struct dentry *dentry)
 
 	hpfs_adjust_length((char *)name, &len);
 	lock_kernel();
-	down(&hpfs_i(inode)->i_parent);
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	err = -ENOENT;
 	de = map_dirent(dir, hpfs_i(dir)->i_dno, (char *)name, len, &dno, &qbh);
 	if (!de)
@@ -502,8 +502,8 @@ static int hpfs_rmdir(struct inode *dir, struct dentry *dentry)
 out1:
 	hpfs_brelse4(&qbh);
 out:
-	up(&hpfs_i(dir)->i_sem);
-	up(&hpfs_i(inode)->i_parent);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
+	mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
@@ -565,12 +565,12 @@ static int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 	lock_kernel();
 	/* order doesn't matter, due to VFS exclusion */
-	down(&hpfs_i(i)->i_parent);
+	mutex_lock(&hpfs_i(i)->i_parent_mutex);
 	if (new_inode)
-		down(&hpfs_i(new_inode)->i_parent);
-	down(&hpfs_i(old_dir)->i_sem);
+		mutex_lock(&hpfs_i(new_inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(old_dir)->i_mutex);
 	if (new_dir != old_dir)
-		down(&hpfs_i(new_dir)->i_sem);
+		mutex_lock(&hpfs_i(new_dir)->i_mutex);
 	
 	/* Erm? Moving over the empty non-busy directory is perfectly legal */
 	if (new_inode && S_ISDIR(new_inode->i_mode)) {
@@ -650,11 +650,11 @@ static int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	hpfs_decide_conv(i, (char *)new_name, new_len);
 end1:
 	if (old_dir != new_dir)
-		up(&hpfs_i(new_dir)->i_sem);
-	up(&hpfs_i(old_dir)->i_sem);
-	up(&hpfs_i(i)->i_parent);
+		mutex_unlock(&hpfs_i(new_dir)->i_mutex);
+	mutex_unlock(&hpfs_i(old_dir)->i_mutex);
+	mutex_unlock(&hpfs_i(i)->i_parent_mutex);
 	if (new_inode)
-		up(&hpfs_i(new_inode)->i_parent);
+		mutex_unlock(&hpfs_i(new_inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c
index 63e88d7e2c3b..9488a794076e 100644
--- a/fs/hpfs/super.c
+++ b/fs/hpfs/super.c
@@ -181,8 +181,8 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
 	    SLAB_CTOR_CONSTRUCTOR) {
-		init_MUTEX(&ei->i_sem);
-		init_MUTEX(&ei->i_parent);
+		mutex_init(&ei->i_mutex);
+		mutex_init(&ei->i_parent_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index b35195289945..25fa8bba8cb5 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -56,48 +56,10 @@ static void huge_pagevec_release(struct pagevec *pvec)
 	pagevec_reinit(pvec);
 }
 
-/*
- * huge_pages_needed tries to determine the number of new huge pages that
- * will be required to fully populate this VMA.  This will be equal to
- * the size of the VMA in huge pages minus the number of huge pages
- * (covered by this VMA) that are found in the page cache.
- *
- * Result is in bytes to be compatible with is_hugepage_mem_enough()
- */
-static unsigned long
-huge_pages_needed(struct address_space *mapping, struct vm_area_struct *vma)
-{
-	int i;
-	struct pagevec pvec;
-	unsigned long start = vma->vm_start;
-	unsigned long end = vma->vm_end;
-	unsigned long hugepages = (end - start) >> HPAGE_SHIFT;
-	pgoff_t next = vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT);
-	pgoff_t endpg = next + hugepages;
-
-	pagevec_init(&pvec, 0);
-	while (next < endpg) {
-		if (!pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE))
-			break;
-		for (i = 0; i < pagevec_count(&pvec); i++) {
-			struct page *page = pvec.pages[i];
-			if (page->index > next)
-				next = page->index;
-			if (page->index >= endpg)
-				break;
-			next++;
-			hugepages--;
-		}
-		huge_pagevec_release(&pvec);
-	}
-	return hugepages << HPAGE_SHIFT;
-}
-
 static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	struct inode *inode = file->f_dentry->d_inode;
-	struct address_space *mapping = inode->i_mapping;
-	unsigned long bytes;
+	struct hugetlbfs_inode_info *info = HUGETLBFS_I(inode);
 	loff_t len, vma_len;
 	int ret;
 
@@ -113,10 +75,6 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	if (vma->vm_end - vma->vm_start < HPAGE_SIZE)
 		return -EINVAL;
 
-	bytes = huge_pages_needed(mapping, vma);
-	if (!is_hugepage_mem_enough(bytes))
-		return -ENOMEM;
-
 	vma_len = (loff_t)(vma->vm_end - vma->vm_start);
 
 	mutex_lock(&inode->i_mutex);
@@ -129,6 +87,10 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	if (!(vma->vm_flags & VM_WRITE) && len > inode->i_size)
 		goto out;
 
+	if (vma->vm_flags & VM_MAYSHARE)
+		if (hugetlb_extend_reservation(info, len >> HPAGE_SHIFT) != 0)
+			goto out;
+
 	ret = 0;
 	hugetlb_prefault_arch_hook(vma->vm_mm);
 	if (inode->i_size < len)
@@ -227,13 +189,18 @@ static void truncate_huge_page(struct page *page)
 	put_page(page);
 }
 
-static void truncate_hugepages(struct address_space *mapping, loff_t lstart)
+static void truncate_hugepages(struct inode *inode, loff_t lstart)
 {
+	struct address_space *mapping = &inode->i_data;
 	const pgoff_t start = lstart >> HPAGE_SHIFT;
 	struct pagevec pvec;
 	pgoff_t next;
 	int i;
 
+	hugetlb_truncate_reservation(HUGETLBFS_I(inode),
+				     lstart >> HPAGE_SHIFT);
+	if (!mapping->nrpages)
+		return;
 	pagevec_init(&pvec, 0);
 	next = start;
 	while (1) {
@@ -262,8 +229,7 @@ static void truncate_hugepages(struct address_space *mapping, loff_t lstart)
 
 static void hugetlbfs_delete_inode(struct inode *inode)
 {
-	if (inode->i_data.nrpages)
-		truncate_hugepages(&inode->i_data, 0);
+	truncate_hugepages(inode, 0);
 	clear_inode(inode);
 }
 
@@ -296,8 +262,7 @@ static void hugetlbfs_forget_inode(struct inode *inode)
 	inode->i_state |= I_FREEING;
 	inodes_stat.nr_inodes--;
 	spin_unlock(&inode_lock);
-	if (inode->i_data.nrpages)
-		truncate_hugepages(&inode->i_data, 0);
+	truncate_hugepages(inode, 0);
 	clear_inode(inode);
 	destroy_inode(inode);
 }
@@ -356,7 +321,7 @@ static int hugetlb_vmtruncate(struct inode *inode, loff_t offset)
 	if (!prio_tree_empty(&mapping->i_mmap))
 		hugetlb_vmtruncate_list(&mapping->i_mmap, pgoff);
 	spin_unlock(&mapping->i_mmap_lock);
-	truncate_hugepages(mapping, offset);
+	truncate_hugepages(inode, offset);
 	return 0;
 }
 
@@ -573,6 +538,7 @@ static struct inode *hugetlbfs_alloc_inode(struct super_block *sb)
 		hugetlbfs_inc_free_inodes(sbinfo);
 		return NULL;
 	}
+	p->prereserved_hpages = 0;
 	return &p->vfs_inode;
 }
 
@@ -771,21 +737,6 @@ static struct file_system_type hugetlbfs_fs_type = {
 
 static struct vfsmount *hugetlbfs_vfsmount;
 
-/*
- * Return the next identifier for a shm file
- */
-static unsigned long hugetlbfs_counter(void)
-{
-	static DEFINE_SPINLOCK(lock);
-	static unsigned long counter;
-	unsigned long ret;
-
-	spin_lock(&lock);
-	ret = ++counter;
-	spin_unlock(&lock);
-	return ret;
-}
-
 static int can_do_hugetlb_shm(void)
 {
 	return likely(capable(CAP_IPC_LOCK) ||
@@ -801,18 +752,16 @@ struct file *hugetlb_zero_setup(size_t size)
 	struct dentry *dentry, *root;
 	struct qstr quick_string;
 	char buf[16];
+	static atomic_t counter;
 
 	if (!can_do_hugetlb_shm())
 		return ERR_PTR(-EPERM);
 
-	if (!is_hugepage_mem_enough(size))
-		return ERR_PTR(-ENOMEM);
-
 	if (!user_shm_lock(size, current->user))
 		return ERR_PTR(-ENOMEM);
 
 	root = hugetlbfs_vfsmount->mnt_root;
-	snprintf(buf, 16, "%lu", hugetlbfs_counter());
+	snprintf(buf, 16, "%u", atomic_inc_return(&counter));
 	quick_string.name = buf;
 	quick_string.len = strlen(quick_string.name);
 	quick_string.hash = 0;
@@ -831,6 +780,11 @@ struct file *hugetlb_zero_setup(size_t size)
 	if (!inode)
 		goto out_file;
 
+	error = -ENOMEM;
+	if (hugetlb_extend_reservation(HUGETLBFS_I(inode),
+				       size >> HPAGE_SHIFT) != 0)
+		goto out_inode;
+
 	d_instantiate(dentry, inode);
 	inode->i_size = size;
 	inode->i_nlink = 0;
@@ -841,6 +795,8 @@ struct file *hugetlb_zero_setup(size_t size)
 	file->f_mode = FMODE_WRITE | FMODE_READ;
 	return file;
 
+out_inode:
+	iput(inode);
 out_file:
 	put_filp(file);
 out_dentry:
diff --git a/fs/inode.c b/fs/inode.c
index d0be6159eb7f..25967b67903d 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -84,14 +84,14 @@ static struct hlist_head *inode_hashtable;
 DEFINE_SPINLOCK(inode_lock);
 
 /*
- * iprune_sem provides exclusion between the kswapd or try_to_free_pages
+ * iprune_mutex provides exclusion between the kswapd or try_to_free_pages
  * icache shrinking path, and the umount path.  Without this exclusion,
  * by the time prune_icache calls iput for the inode whose pages it has
  * been invalidating, or by the time it calls clear_inode & destroy_inode
  * from its final dispose_list, the struct super_block they refer to
  * (for inode->i_sb->s_op) may already have been freed and reused.
  */
-DECLARE_MUTEX(iprune_sem);
+DEFINE_MUTEX(iprune_mutex);
 
 /*
  * Statistics gathering..
@@ -206,7 +206,7 @@ void inode_init_once(struct inode *inode)
 	i_size_ordered_init(inode);
 #ifdef CONFIG_INOTIFY
 	INIT_LIST_HEAD(&inode->inotify_watches);
-	sema_init(&inode->inotify_sem, 1);
+	mutex_init(&inode->inotify_mutex);
 #endif
 }
 
@@ -319,7 +319,7 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose)
 		/*
 		 * We can reschedule here without worrying about the list's
 		 * consistency because the per-sb list of inodes must not
-		 * change during umount anymore, and because iprune_sem keeps
+		 * change during umount anymore, and because iprune_mutex keeps
 		 * shrink_icache_memory() away.
 		 */
 		cond_resched_lock(&inode_lock);
@@ -355,14 +355,14 @@ int invalidate_inodes(struct super_block * sb)
 	int busy;
 	LIST_HEAD(throw_away);
 
-	down(&iprune_sem);
+	mutex_lock(&iprune_mutex);
 	spin_lock(&inode_lock);
 	inotify_unmount_inodes(&sb->s_inodes);
 	busy = invalidate_list(&sb->s_inodes, &throw_away);
 	spin_unlock(&inode_lock);
 
 	dispose_list(&throw_away);
-	up(&iprune_sem);
+	mutex_unlock(&iprune_mutex);
 
 	return busy;
 }
@@ -377,7 +377,7 @@ int __invalidate_device(struct block_device *bdev)
 	if (sb) {
 		/*
 		 * no need to lock the super, get_super holds the
-		 * read semaphore so the filesystem cannot go away
+		 * read mutex so the filesystem cannot go away
 		 * under us (->put_super runs with the write lock
 		 * hold).
 		 */
@@ -423,7 +423,7 @@ static void prune_icache(int nr_to_scan)
 	int nr_scanned;
 	unsigned long reap = 0;
 
-	down(&iprune_sem);
+	mutex_lock(&iprune_mutex);
 	spin_lock(&inode_lock);
 	for (nr_scanned = 0; nr_scanned < nr_to_scan; nr_scanned++) {
 		struct inode *inode;
@@ -459,7 +459,7 @@ static void prune_icache(int nr_to_scan)
 	spin_unlock(&inode_lock);
 
 	dispose_list(&freeable);
-	up(&iprune_sem);
+	mutex_unlock(&iprune_mutex);
 
 	if (current_is_kswapd())
 		mod_page_state(kswapd_inodesteal, reap);
diff --git a/fs/inotify.c b/fs/inotify.c
index 3041503bde02..0ee39ef591c6 100644
--- a/fs/inotify.c
+++ b/fs/inotify.c
@@ -54,10 +54,10 @@ int inotify_max_queued_events;
  * Lock ordering:
  *
  * dentry->d_lock (used to keep d_move() away from dentry->d_parent)
- * iprune_sem (synchronize shrink_icache_memory())
+ * iprune_mutex (synchronize shrink_icache_memory())
  * 	inode_lock (protects the super_block->s_inodes list)
- * 	inode->inotify_sem (protects inode->inotify_watches and watches->i_list)
- * 		inotify_dev->sem (protects inotify_device and watches->d_list)
+ * 	inode->inotify_mutex (protects inode->inotify_watches and watches->i_list)
+ * 		inotify_dev->mutex (protects inotify_device and watches->d_list)
  */
 
 /*
@@ -79,12 +79,12 @@ int inotify_max_queued_events;
 /*
  * struct inotify_device - represents an inotify instance
  *
- * This structure is protected by the semaphore 'sem'.
+ * This structure is protected by the mutex 'mutex'.
  */
 struct inotify_device {
 	wait_queue_head_t 	wq;		/* wait queue for i/o */
 	struct idr		idr;		/* idr mapping wd -> watch */
-	struct semaphore	sem;		/* protects this bad boy */
+	struct mutex		mutex;		/* protects this bad boy */
 	struct list_head 	events;		/* list of queued events */
 	struct list_head	watches;	/* list of watches */
 	atomic_t		count;		/* reference count */
@@ -101,7 +101,7 @@ struct inotify_device {
  * device.  In read(), this list is walked and all events that can fit in the
  * buffer are returned.
  *
- * Protected by dev->sem of the device in which we are queued.
+ * Protected by dev->mutex of the device in which we are queued.
  */
 struct inotify_kernel_event {
 	struct inotify_event	event;	/* the user-space event */
@@ -112,8 +112,8 @@ struct inotify_kernel_event {
 /*
  * struct inotify_watch - represents a watch request on a specific inode
  *
- * d_list is protected by dev->sem of the associated watch->dev.
- * i_list and mask are protected by inode->inotify_sem of the associated inode.
+ * d_list is protected by dev->mutex of the associated watch->dev.
+ * i_list and mask are protected by inode->inotify_mutex of the associated inode.
  * dev, inode, and wd are never written to once the watch is created.
  */
 struct inotify_watch {
@@ -261,7 +261,7 @@ static struct inotify_kernel_event * kernel_event(s32 wd, u32 mask, u32 cookie,
 /*
  * inotify_dev_get_event - return the next event in the given dev's queue
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static inline struct inotify_kernel_event *
 inotify_dev_get_event(struct inotify_device *dev)
@@ -272,7 +272,7 @@ inotify_dev_get_event(struct inotify_device *dev)
 /*
  * inotify_dev_queue_event - add a new event to the given device
  *
- * Caller must hold dev->sem.  Can sleep (calls kernel_event()).
+ * Caller must hold dev->mutex.  Can sleep (calls kernel_event()).
  */
 static void inotify_dev_queue_event(struct inotify_device *dev,
 				    struct inotify_watch *watch, u32 mask,
@@ -315,7 +315,7 @@ static void inotify_dev_queue_event(struct inotify_device *dev,
 /*
  * remove_kevent - cleans up and ultimately frees the given kevent
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static void remove_kevent(struct inotify_device *dev,
 			  struct inotify_kernel_event *kevent)
@@ -332,7 +332,7 @@ static void remove_kevent(struct inotify_device *dev,
 /*
  * inotify_dev_event_dequeue - destroy an event on the given device
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static void inotify_dev_event_dequeue(struct inotify_device *dev)
 {
@@ -346,7 +346,7 @@ static void inotify_dev_event_dequeue(struct inotify_device *dev)
 /*
  * inotify_dev_get_wd - returns the next WD for use by the given dev
  *
- * Callers must hold dev->sem.  This function can sleep.
+ * Callers must hold dev->mutex.  This function can sleep.
  */
 static int inotify_dev_get_wd(struct inotify_device *dev,
 			      struct inotify_watch *watch)
@@ -383,7 +383,7 @@ static int find_inode(const char __user *dirname, struct nameidata *nd,
 /*
  * create_watch - creates a watch on the given device.
  *
- * Callers must hold dev->sem.  Calls inotify_dev_get_wd() so may sleep.
+ * Callers must hold dev->mutex.  Calls inotify_dev_get_wd() so may sleep.
  * Both 'dev' and 'inode' (by way of nameidata) need to be pinned.
  */
 static struct inotify_watch *create_watch(struct inotify_device *dev,
@@ -434,7 +434,7 @@ static struct inotify_watch *create_watch(struct inotify_device *dev,
 /*
  * inotify_find_dev - find the watch associated with the given inode and dev
  *
- * Callers must hold inode->inotify_sem.
+ * Callers must hold inode->inotify_mutex.
  */
 static struct inotify_watch *inode_find_dev(struct inode *inode,
 					    struct inotify_device *dev)
@@ -469,7 +469,7 @@ static void remove_watch_no_event(struct inotify_watch *watch,
  * the IN_IGNORED event to the given device signifying that the inode is no
  * longer watched.
  *
- * Callers must hold both inode->inotify_sem and dev->sem.  We drop a
+ * Callers must hold both inode->inotify_mutex and dev->mutex.  We drop a
  * reference to the inode before returning.
  *
  * The inode is not iput() so as to remain atomic.  If the inode needs to be
@@ -507,21 +507,21 @@ void inotify_inode_queue_event(struct inode *inode, u32 mask, u32 cookie,
 	if (!inotify_inode_watched(inode))
 		return;
 
-	down(&inode->inotify_sem);
+	mutex_lock(&inode->inotify_mutex);
 	list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
 		u32 watch_mask = watch->mask;
 		if (watch_mask & mask) {
 			struct inotify_device *dev = watch->dev;
 			get_inotify_watch(watch);
-			down(&dev->sem);
+			mutex_lock(&dev->mutex);
 			inotify_dev_queue_event(dev, watch, mask, cookie, name);
 			if (watch_mask & IN_ONESHOT)
 				remove_watch_no_event(watch, dev);
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 			put_inotify_watch(watch);
 		}
 	}
-	up(&inode->inotify_sem);
+	mutex_unlock(&inode->inotify_mutex);
 }
 EXPORT_SYMBOL_GPL(inotify_inode_queue_event);
 
@@ -569,7 +569,7 @@ EXPORT_SYMBOL_GPL(inotify_get_cookie);
  * @list: list of inodes being unmounted (sb->s_inodes)
  *
  * Called with inode_lock held, protecting the unmounting super block's list
- * of inodes, and with iprune_sem held, keeping shrink_icache_memory() at bay.
+ * of inodes, and with iprune_mutex held, keeping shrink_icache_memory() at bay.
  * We temporarily drop inode_lock, however, and CAN block.
  */
 void inotify_unmount_inodes(struct list_head *list)
@@ -618,7 +618,7 @@ void inotify_unmount_inodes(struct list_head *list)
 		 * We can safely drop inode_lock here because we hold
 		 * references on both inode and next_i.  Also no new inodes
 		 * will be added since the umount has begun.  Finally,
-		 * iprune_sem keeps shrink_icache_memory() away.
+		 * iprune_mutex keeps shrink_icache_memory() away.
 		 */
 		spin_unlock(&inode_lock);
 
@@ -626,16 +626,16 @@ void inotify_unmount_inodes(struct list_head *list)
 			iput(need_iput_tmp);
 
 		/* for each watch, send IN_UNMOUNT and then remove it */
-		down(&inode->inotify_sem);
+		mutex_lock(&inode->inotify_mutex);
 		watches = &inode->inotify_watches;
 		list_for_each_entry_safe(watch, next_w, watches, i_list) {
 			struct inotify_device *dev = watch->dev;
-			down(&dev->sem);
+			mutex_lock(&dev->mutex);
 			inotify_dev_queue_event(dev, watch, IN_UNMOUNT,0,NULL);
 			remove_watch(watch, dev);
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 		}
-		up(&inode->inotify_sem);
+		mutex_unlock(&inode->inotify_mutex);
 		iput(inode);		
 
 		spin_lock(&inode_lock);
@@ -651,14 +651,14 @@ void inotify_inode_is_dead(struct inode *inode)
 {
 	struct inotify_watch *watch, *next;
 
-	down(&inode->inotify_sem);
+	mutex_lock(&inode->inotify_mutex);
 	list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
 		struct inotify_device *dev = watch->dev;
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		remove_watch(watch, dev);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 	}
-	up(&inode->inotify_sem);
+	mutex_unlock(&inode->inotify_mutex);
 }
 EXPORT_SYMBOL_GPL(inotify_inode_is_dead);
 
@@ -670,10 +670,10 @@ static unsigned int inotify_poll(struct file *file, poll_table *wait)
 	int ret = 0;
 
 	poll_wait(file, &dev->wq, wait);
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	if (!list_empty(&dev->events))
 		ret = POLLIN | POLLRDNORM;
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	return ret;
 }
@@ -695,9 +695,9 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 
 		prepare_to_wait(&dev->wq, &wait, TASK_INTERRUPTIBLE);
 
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		events = !list_empty(&dev->events);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 		if (events) {
 			ret = 0;
 			break;
@@ -720,7 +720,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 	if (ret)
 		return ret;
 
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	while (1) {
 		struct inotify_kernel_event *kevent;
 
@@ -750,7 +750,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 
 		remove_kevent(dev, kevent);
 	}
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	return ret;
 }
@@ -763,37 +763,37 @@ static int inotify_release(struct inode *ignored, struct file *file)
 	 * Destroy all of the watches on this device.  Unfortunately, not very
 	 * pretty.  We cannot do a simple iteration over the list, because we
 	 * do not know the inode until we iterate to the watch.  But we need to
-	 * hold inode->inotify_sem before dev->sem.  The following works.
+	 * hold inode->inotify_mutex before dev->mutex.  The following works.
 	 */
 	while (1) {
 		struct inotify_watch *watch;
 		struct list_head *watches;
 		struct inode *inode;
 
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		watches = &dev->watches;
 		if (list_empty(watches)) {
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 			break;
 		}
 		watch = list_entry(watches->next, struct inotify_watch, d_list);
 		get_inotify_watch(watch);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 
 		inode = watch->inode;
-		down(&inode->inotify_sem);
-		down(&dev->sem);
+		mutex_lock(&inode->inotify_mutex);
+		mutex_lock(&dev->mutex);
 		remove_watch_no_event(watch, dev);
-		up(&dev->sem);
-		up(&inode->inotify_sem);
+		mutex_unlock(&dev->mutex);
+		mutex_unlock(&inode->inotify_mutex);
 		put_inotify_watch(watch);
 	}
 
 	/* destroy all of the events on this device */
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	while (!list_empty(&dev->events))
 		inotify_dev_event_dequeue(dev);
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	/* free this device: the put matching the get in inotify_init() */
 	put_inotify_dev(dev);
@@ -811,26 +811,26 @@ static int inotify_ignore(struct inotify_device *dev, s32 wd)
 	struct inotify_watch *watch;
 	struct inode *inode;
 
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	watch = idr_find(&dev->idr, wd);
 	if (unlikely(!watch)) {
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 		return -EINVAL;
 	}
 	get_inotify_watch(watch);
 	inode = watch->inode;
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
-	down(&inode->inotify_sem);
-	down(&dev->sem);
+	mutex_lock(&inode->inotify_mutex);
+	mutex_lock(&dev->mutex);
 
 	/* make sure that we did not race */
 	watch = idr_find(&dev->idr, wd);
 	if (likely(watch))
 		remove_watch(watch, dev);
 
-	up(&dev->sem);
-	up(&inode->inotify_sem);
+	mutex_unlock(&dev->mutex);
+	mutex_unlock(&inode->inotify_mutex);
 	put_inotify_watch(watch);
 
 	return 0;
@@ -905,7 +905,7 @@ asmlinkage long sys_inotify_init(void)
 	INIT_LIST_HEAD(&dev->events);
 	INIT_LIST_HEAD(&dev->watches);
 	init_waitqueue_head(&dev->wq);
-	sema_init(&dev->sem, 1);
+	mutex_init(&dev->mutex);
 	dev->event_count = 0;
 	dev->queue_size = 0;
 	dev->max_events = inotify_max_queued_events;
@@ -960,8 +960,8 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
 	inode = nd.dentry->d_inode;
 	dev = filp->private_data;
 
-	down(&inode->inotify_sem);
-	down(&dev->sem);
+	mutex_lock(&inode->inotify_mutex);
+	mutex_lock(&dev->mutex);
 
 	if (mask & IN_MASK_ADD)
 		mask_add = 1;
@@ -998,8 +998,8 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
 	list_add(&watch->i_list, &inode->inotify_watches);
 	ret = watch->wd;
 out:
-	up(&dev->sem);
-	up(&inode->inotify_sem);
+	mutex_unlock(&dev->mutex);
+	mutex_unlock(&inode->inotify_mutex);
 	path_release(&nd);
 fput_and_out:
 	fput_light(filp, fput_needed);
diff --git a/fs/isofs/joliet.c b/fs/isofs/joliet.c
index 2931de7f1a6a..81a90e170ac3 100644
--- a/fs/isofs/joliet.c
+++ b/fs/isofs/joliet.c
@@ -11,7 +11,7 @@
 #include "isofs.h"
 
 /*
- * Convert Unicode 16 to UTF8 or ASCII.
+ * Convert Unicode 16 to UTF-8 or ASCII.
  */
 static int
 uni16_to_x8(unsigned char *ascii, u16 *uni, int len, struct nls_table *nls)
diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c
index 543ed543d1e5..3f5102b069db 100644
--- a/fs/jbd/checkpoint.c
+++ b/fs/jbd/checkpoint.c
@@ -85,7 +85,7 @@ void __log_wait_for_space(journal_t *journal)
 		if (journal->j_flags & JFS_ABORT)
 			return;
 		spin_unlock(&journal->j_state_lock);
-		down(&journal->j_checkpoint_sem);
+		mutex_lock(&journal->j_checkpoint_mutex);
 
 		/*
 		 * Test again, another process may have checkpointed while we
@@ -98,7 +98,7 @@ void __log_wait_for_space(journal_t *journal)
 			log_do_checkpoint(journal);
 			spin_lock(&journal->j_state_lock);
 		}
-		up(&journal->j_checkpoint_sem);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 	}
 }
 
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index e4b516ac4989..95a628d8cac8 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -659,8 +659,8 @@ static journal_t * journal_init_common (void)
 	init_waitqueue_head(&journal->j_wait_checkpoint);
 	init_waitqueue_head(&journal->j_wait_commit);
 	init_waitqueue_head(&journal->j_wait_updates);
-	init_MUTEX(&journal->j_barrier);
-	init_MUTEX(&journal->j_checkpoint_sem);
+	mutex_init(&journal->j_barrier);
+	mutex_init(&journal->j_checkpoint_mutex);
 	spin_lock_init(&journal->j_revoke_lock);
 	spin_lock_init(&journal->j_list_lock);
 	spin_lock_init(&journal->j_state_lock);
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index ca917973c2c0..5fc40888f4cf 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -455,7 +455,7 @@ void journal_lock_updates(journal_t *journal)
 	 * to make sure that we serialise special journal-locked operations
 	 * too.
 	 */
-	down(&journal->j_barrier);
+	mutex_lock(&journal->j_barrier);
 }
 
 /**
@@ -470,7 +470,7 @@ void journal_unlock_updates (journal_t *journal)
 {
 	J_ASSERT(journal->j_barrier_count != 0);
 
-	up(&journal->j_barrier);
+	mutex_unlock(&journal->j_barrier);
 	spin_lock(&journal->j_state_lock);
 	--journal->j_barrier_count;
 	spin_unlock(&journal->j_state_lock);
diff --git a/fs/jffs/inode-v23.c b/fs/jffs/inode-v23.c
index fc3855a1aef3..890d7ff7456d 100644
--- a/fs/jffs/inode-v23.c
+++ b/fs/jffs/inode-v23.c
@@ -42,7 +42,7 @@
 #include <linux/quotaops.h>
 #include <linux/highmem.h>
 #include <linux/vfs.h>
-#include <asm/semaphore.h>
+#include <linux/mutex.h>
 #include <asm/byteorder.h>
 #include <asm/uaccess.h>
 
@@ -203,7 +203,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 	fmc = c->fmc;
 
 	D3(printk (KERN_NOTICE "notify_change(): down biglock\n"));
-	down(&fmc->biglock);
+	mutex_lock(&fmc->biglock);
 
 	f = jffs_find_file(c, inode->i_ino);
 
@@ -211,7 +211,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 		printk("jffs_setattr(): Invalid inode number: %lu\n",
 		       inode->i_ino);
 		D3(printk (KERN_NOTICE "notify_change(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 		res = -EINVAL;
 		goto out;
 	});
@@ -232,7 +232,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 	if (!(new_node = jffs_alloc_node())) {
 		D(printk("jffs_setattr(): Allocation failed!\n"));
 		D3(printk (KERN_NOTICE "notify_change(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 		res = -ENOMEM;
 		goto out;
 	}
@@ -319,7 +319,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 		D(printk("jffs_notify_change(): The write failed!\n"));
 		jffs_free_node(new_node);
 		D3(printk (KERN_NOTICE "n_c(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		goto out;
 	}
 
@@ -327,7 +327,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 
 	mark_inode_dirty(inode);
 	D3(printk (KERN_NOTICE "n_c(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 out:
 	unlock_kernel();
 	return res;
@@ -461,7 +461,7 @@ jffs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		goto jffs_rename_end;
 	}
 	D3(printk (KERN_NOTICE "rename(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	/* Create a node and initialize as much as needed.  */
 	result = -ENOMEM;
 	if (!(node = jffs_alloc_node())) {
@@ -555,7 +555,7 @@ jffs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 jffs_rename_end:
 	D3(printk (KERN_NOTICE "rename(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_rename()  */
@@ -574,14 +574,14 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	int ddino;
 	lock_kernel();
 	D3(printk (KERN_NOTICE "readdir(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	D2(printk("jffs_readdir(): inode: 0x%p, filp: 0x%p\n", inode, filp));
 	if (filp->f_pos == 0) {
 		D3(printk("jffs_readdir(): \".\" %lu\n", inode->i_ino));
 		if (filldir(dirent, ".", 1, filp->f_pos, inode->i_ino, DT_DIR) < 0) {
 			D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -598,7 +598,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		D3(printk("jffs_readdir(): \"..\" %u\n", ddino));
 		if (filldir(dirent, "..", 2, filp->f_pos, ddino, DT_DIR) < 0) {
 			D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -617,7 +617,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		if (filldir(dirent, f->name, f->nsize,
 			    filp->f_pos , f->ino, DT_UNKNOWN) < 0) {
 		        D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -627,7 +627,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		} while(f && f->deleted);
 	}
 	D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return filp->f_pos;
 } /* jffs_readdir()  */
@@ -660,7 +660,7 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 	});
 
 	D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	r = -ENAMETOOLONG;
 	if (len > JFFS_MAX_NAME_LEN) {
@@ -683,31 +683,31 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 
 	if ((len == 1) && (name[0] == '.')) {
 		D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		if (!(inode = iget(dir->i_sb, d->ino))) {
 			D(printk("jffs_lookup(): . iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else if ((len == 2) && (name[0] == '.') && (name[1] == '.')) {
 	        D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
  		if (!(inode = iget(dir->i_sb, d->pino))) {
 			D(printk("jffs_lookup(): .. iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else if ((f = jffs_find_child(d, name, len))) {
 	        D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		if (!(inode = iget(dir->i_sb, f->ino))) {
 			D(printk("jffs_lookup(): iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else {
 		D3(printk("jffs_lookup(): Couldn't find the file. "
 			  "f = 0x%p, name = \"%s\", d = 0x%p, d->ino = %u\n",
@@ -717,13 +717,13 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 
 	d_add(dentry, inode);
 	D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return NULL;
 
 jffs_lookup_end:
 	D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 jffs_lookup_end_no_biglock:
 	unlock_kernel();
@@ -753,7 +753,7 @@ jffs_do_readpage_nolock(struct file *file, struct page *page)
 	ClearPageError(page);
 
 	D3(printk (KERN_NOTICE "readpage(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	read_len = 0;
 	result = 0;
@@ -782,7 +782,7 @@ jffs_do_readpage_nolock(struct file *file, struct page *page)
 	kunmap(page);
 
 	D3(printk (KERN_NOTICE "readpage(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 	if (result) {
 	        SetPageError(page);
@@ -839,7 +839,7 @@ jffs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
 	c = dir_f->c;
 	D3(printk (KERN_NOTICE "mkdir(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	dir_mode = S_IFDIR | (mode & (S_IRWXUGO|S_ISVTX)
 			      & ~current->fs->umask);
@@ -906,7 +906,7 @@ jffs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	result = 0;
 jffs_mkdir_end:
 	D3(printk (KERN_NOTICE "mkdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_mkdir()  */
@@ -921,10 +921,10 @@ jffs_rmdir(struct inode *dir, struct dentry *dentry)
 	D3(printk("***jffs_rmdir()\n"));
 	D3(printk (KERN_NOTICE "rmdir(): down biglock\n"));
 	lock_kernel();
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	ret = jffs_remove(dir, dentry, S_IFDIR);
 	D3(printk (KERN_NOTICE "rmdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return ret;
 }
@@ -940,10 +940,10 @@ jffs_unlink(struct inode *dir, struct dentry *dentry)
 	lock_kernel();
 	D3(printk("***jffs_unlink()\n"));
 	D3(printk (KERN_NOTICE "unlink(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	ret = jffs_remove(dir, dentry, 0);
 	D3(printk (KERN_NOTICE "unlink(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return ret;
 }
@@ -1086,7 +1086,7 @@ jffs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 	c = dir_f->c;
 
 	D3(printk (KERN_NOTICE "mknod(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	/* Create and initialize a new node.  */
 	if (!(node = jffs_alloc_node())) {
@@ -1152,7 +1152,7 @@ jffs_mknod_err:
 
 jffs_mknod_end:
 	D3(printk (KERN_NOTICE "mknod(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_mknod()  */
@@ -1203,7 +1203,7 @@ jffs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
 		return -ENOMEM;
 	}
 	D3(printk (KERN_NOTICE "symlink(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	node->data_offset = 0;
 	node->removed_size = 0;
@@ -1253,7 +1253,7 @@ jffs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
 	d_instantiate(dentry, inode);
  jffs_symlink_end:
 	D3(printk (KERN_NOTICE "symlink(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return err;
 } /* jffs_symlink()  */
@@ -1306,7 +1306,7 @@ jffs_create(struct inode *dir, struct dentry *dentry, int mode,
 		return -ENOMEM;
 	}
 	D3(printk (KERN_NOTICE "create(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	node->data_offset = 0;
 	node->removed_size = 0;
@@ -1359,7 +1359,7 @@ jffs_create(struct inode *dir, struct dentry *dentry, int mode,
 	d_instantiate(dentry, inode);
  jffs_create_end:
 	D3(printk (KERN_NOTICE "create(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return err;
 } /* jffs_create()  */
@@ -1423,7 +1423,7 @@ jffs_file_write(struct file *filp, const char *buf, size_t count,
 	thiscount = min(c->fmc->max_chunk_size - sizeof(struct jffs_raw_inode), count);
 
 	D3(printk (KERN_NOTICE "file_write(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	/* Urgh. POSIX says we can do short writes if we feel like it. 
 	 * In practice, we can't. Nothing will cope. So we loop until
@@ -1511,7 +1511,7 @@ jffs_file_write(struct file *filp, const char *buf, size_t count,
 	}
  out:
 	D3(printk (KERN_NOTICE "file_write(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 	/* Fix things in the real inode.  */
 	if (pos > inode->i_size) {
@@ -1567,7 +1567,7 @@ jffs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 		return -EIO;
 	}
 	D3(printk (KERN_NOTICE "ioctl(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	switch (cmd) {
 	case JFFS_PRINT_HASH:
@@ -1609,7 +1609,7 @@ jffs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 		ret = -ENOTTY;
 	}
 	D3(printk (KERN_NOTICE "ioctl(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	return ret;
 } /* jffs_ioctl()  */
 
@@ -1685,12 +1685,12 @@ jffs_read_inode(struct inode *inode)
 	}
 	c = (struct jffs_control *)inode->i_sb->s_fs_info;
 	D3(printk (KERN_NOTICE "read_inode(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	if (!(f = jffs_find_file(c, inode->i_ino))) {
 		D(printk("jffs_read_inode(): No such inode (%lu).\n",
 			 inode->i_ino));
 		D3(printk (KERN_NOTICE "read_inode(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		return;
 	}
 	inode->u.generic_ip = (void *)f;
@@ -1732,7 +1732,7 @@ jffs_read_inode(struct inode *inode)
 	}
 
 	D3(printk (KERN_NOTICE "read_inode(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 }
 
 
diff --git a/fs/jffs/intrep.c b/fs/jffs/intrep.c
index ce7b54b0b2b7..0ef207dfaf6f 100644
--- a/fs/jffs/intrep.c
+++ b/fs/jffs/intrep.c
@@ -62,7 +62,7 @@
 #include <linux/fs.h>
 #include <linux/stat.h>
 #include <linux/pagemap.h>
-#include <asm/semaphore.h>
+#include <linux/mutex.h>
 #include <asm/byteorder.h>
 #include <linux/smp_lock.h>
 #include <linux/time.h>
@@ -3416,7 +3416,7 @@ jffs_garbage_collect_thread(void *ptr)
 		D1(printk (KERN_NOTICE "jffs_garbage_collect_thread(): collecting.\n"));
 
 		D3(printk (KERN_NOTICE "g_c_thread(): down biglock\n"));
-		down(&fmc->biglock);
+		mutex_lock(&fmc->biglock);
 		
 		D1(printk("***jffs_garbage_collect_thread(): round #%u, "
 			  "fmc->dirty_size = %u\n", i++, fmc->dirty_size));
@@ -3447,6 +3447,6 @@ jffs_garbage_collect_thread(void *ptr)
 		
 	gc_end:
 		D3(printk (KERN_NOTICE "g_c_thread(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 	} /* for (;;) */
 } /* jffs_garbage_collect_thread() */
diff --git a/fs/jffs/jffs_fm.c b/fs/jffs/jffs_fm.c
index 6da13b309bd1..7d8ca1aeace2 100644
--- a/fs/jffs/jffs_fm.c
+++ b/fs/jffs/jffs_fm.c
@@ -139,7 +139,7 @@ jffs_build_begin(struct jffs_control *c, int unit)
 	fmc->tail = NULL;
 	fmc->head_extra = NULL;
 	fmc->tail_extra = NULL;
-	init_MUTEX(&fmc->biglock);
+	mutex_init(&fmc->biglock);
 	return fmc;
 }
 
diff --git a/fs/jffs/jffs_fm.h b/fs/jffs/jffs_fm.h
index f64151e74122..c794d923df2a 100644
--- a/fs/jffs/jffs_fm.h
+++ b/fs/jffs/jffs_fm.h
@@ -20,10 +20,11 @@
 #ifndef __LINUX_JFFS_FM_H__
 #define __LINUX_JFFS_FM_H__
 
+#include <linux/config.h>
 #include <linux/types.h>
 #include <linux/jffs.h>
 #include <linux/mtd/mtd.h>
-#include <linux/config.h>
+#include <linux/mutex.h>
 
 /* The alignment between two nodes in the flash memory.  */
 #define JFFS_ALIGN_SIZE 4
@@ -97,7 +98,7 @@ struct jffs_fmcontrol
 	struct jffs_fm *tail;
 	struct jffs_fm *head_extra;
 	struct jffs_fm *tail_extra;
-	struct semaphore biglock;
+	struct mutex biglock;
 };
 
 /* Notice the two members head_extra and tail_extra in the jffs_control
diff --git a/fs/libfs.c b/fs/libfs.c
index 71fd08fa4103..4fdeaceb892c 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -7,6 +7,8 @@
 #include <linux/pagemap.h>
 #include <linux/mount.h>
 #include <linux/vfs.h>
+#include <linux/mutex.h>
+
 #include <asm/uaccess.h>
 
 int simple_getattr(struct vfsmount *mnt, struct dentry *dentry,
@@ -530,7 +532,7 @@ struct simple_attr {
 	char set_buf[24];
 	void *data;
 	const char *fmt;	/* format for read operation */
-	struct semaphore sem;	/* protects access to these buffers */
+	struct mutex mutex;	/* protects access to these buffers */
 };
 
 /* simple_attr_open is called by an actual attribute open file operation
@@ -549,7 +551,7 @@ int simple_attr_open(struct inode *inode, struct file *file,
 	attr->set = set;
 	attr->data = inode->u.generic_ip;
 	attr->fmt = fmt;
-	init_MUTEX(&attr->sem);
+	mutex_init(&attr->mutex);
 
 	file->private_data = attr;
 
@@ -575,7 +577,7 @@ ssize_t simple_attr_read(struct file *file, char __user *buf,
 	if (!attr->get)
 		return -EACCES;
 
-	down(&attr->sem);
+	mutex_lock(&attr->mutex);
 	if (*ppos) /* continued read */
 		size = strlen(attr->get_buf);
 	else	  /* first read */
@@ -584,7 +586,7 @@ ssize_t simple_attr_read(struct file *file, char __user *buf,
 				 (unsigned long long)attr->get(attr->data));
 
 	ret = simple_read_from_buffer(buf, len, ppos, attr->get_buf, size);
-	up(&attr->sem);
+	mutex_unlock(&attr->mutex);
 	return ret;
 }
 
@@ -602,7 +604,7 @@ ssize_t simple_attr_write(struct file *file, const char __user *buf,
 	if (!attr->set)
 		return -EACCES;
 
-	down(&attr->sem);
+	mutex_lock(&attr->mutex);
 	ret = -EFAULT;
 	size = min(sizeof(attr->set_buf) - 1, len);
 	if (copy_from_user(attr->set_buf, buf, size))
@@ -613,7 +615,7 @@ ssize_t simple_attr_write(struct file *file, const char __user *buf,
 	val = simple_strtol(attr->set_buf, NULL, 0);
 	attr->set(attr->data, val);
 out:
-	up(&attr->sem);
+	mutex_unlock(&attr->mutex);
 	return ret;
 }
 
diff --git a/fs/minix/namei.c b/fs/minix/namei.c
index b25bca5bdb57..5b6a4540a05b 100644
--- a/fs/minix/namei.c
+++ b/fs/minix/namei.c
@@ -6,18 +6,6 @@
 
 #include "minix.h"
 
-static inline void inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static int add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = minix_add_link(dentry, inode);
@@ -25,7 +13,7 @@ static int add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -125,7 +113,7 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -139,7 +127,7 @@ static int minix_link(struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 	return add_nondir(dentry, inode);
 }
@@ -152,7 +140,7 @@ static int minix_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 	if (dir->i_nlink >= minix_sb(dir->i_sb)->s_link_max)
 		goto out;
 
-	inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = minix_new_inode(dir, &err);
 	if (!inode)
@@ -163,7 +151,7 @@ static int minix_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 		inode->i_mode |= S_ISGID;
 	minix_set_inode(inode, 0);
 
-	inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = minix_make_empty(inode, dir);
 	if (err)
@@ -178,11 +166,11 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
-	dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -202,7 +190,7 @@ static int minix_unlink(struct inode * dir, struct dentry *dentry)
 		goto end_unlink;
 
 	inode->i_ctime = dir->i_ctime;
-	dec_count(inode);
+	inode_dec_link_count(inode);
 end_unlink:
 	return err;
 }
@@ -215,8 +203,8 @@ static int minix_rmdir(struct inode * dir, struct dentry *dentry)
 	if (minix_empty_dir(inode)) {
 		err = minix_unlink(dir, dentry);
 		if (!err) {
-			dec_count(dir);
-			dec_count(inode);
+			inode_dec_link_count(dir);
+			inode_dec_link_count(inode);
 		}
 	}
 	return err;
@@ -257,34 +245,34 @@ static int minix_rename(struct inode * old_dir, struct dentry *old_dentry,
 		new_de = minix_find_entry(new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		minix_set_link(new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= info->s_link_max)
 				goto out_dir;
 		}
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = minix_add_link(new_dentry, old_inode);
 		if (err) {
-			dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	minix_delete_entry(old_de, old_page);
-	dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		minix_set_link(dir_de, dir_page, new_dir);
-		dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/namei.c b/fs/namei.c
index 8dc2b038d5d9..c72b940797fc 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -104,7 +104,7 @@
  */
 /*
  * [Sep 2001 AV] Single-semaphore locking scheme (kudos to David Holland)
- * implemented.  Let's see if raised priority of ->s_vfs_rename_sem gives
+ * implemented.  Let's see if raised priority of ->s_vfs_rename_mutex gives
  * any extra contention...
  */
 
@@ -1422,7 +1422,7 @@ struct dentry *lock_rename(struct dentry *p1, struct dentry *p2)
 		return NULL;
 	}
 
-	down(&p1->d_inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
 
 	for (p = p1; p->d_parent != p; p = p->d_parent) {
 		if (p->d_parent == p2) {
@@ -1450,7 +1450,7 @@ void unlock_rename(struct dentry *p1, struct dentry *p2)
 	mutex_unlock(&p1->d_inode->i_mutex);
 	if (p1 != p2) {
 		mutex_unlock(&p2->d_inode->i_mutex);
-		up(&p1->d_inode->i_sb->s_vfs_rename_sem);
+		mutex_unlock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
 	}
 }
 
@@ -2277,17 +2277,17 @@ asmlinkage long sys_link(const char __user *oldname, const char __user *newname)
  *	a) we can get into loop creation. Check is done in is_subdir().
  *	b) race potential - two innocent renames can create a loop together.
  *	   That's where 4.4 screws up. Current fix: serialization on
- *	   sb->s_vfs_rename_sem. We might be more accurate, but that's another
+ *	   sb->s_vfs_rename_mutex. We might be more accurate, but that's another
  *	   story.
  *	c) we have to lock _three_ objects - parents and victim (if it exists).
  *	   And that - after we got ->i_mutex on parents (until then we don't know
  *	   whether the target exists).  Solution: try to be smart with locking
  *	   order for inodes.  We rely on the fact that tree topology may change
- *	   only under ->s_vfs_rename_sem _and_ that parent of the object we
+ *	   only under ->s_vfs_rename_mutex _and_ that parent of the object we
  *	   move will be locked.  Thus we can rank directories by the tree
  *	   (ancestors first) and rank all non-directories after them.
  *	   That works since everybody except rename does "lock parent, lookup,
- *	   lock child" and rename is under ->s_vfs_rename_sem.
+ *	   lock child" and rename is under ->s_vfs_rename_mutex.
  *	   HOWEVER, it relies on the assumption that any object with ->lookup()
  *	   has no more than 1 dentry.  If "hybrid" objects will ever appear,
  *	   we'd better make sure that there's no link(2) for them.
diff --git a/fs/ncpfs/file.c b/fs/ncpfs/file.c
index 973b444d6914..ebdad8f6398f 100644
--- a/fs/ncpfs/file.c
+++ b/fs/ncpfs/file.c
@@ -46,7 +46,7 @@ int ncp_make_open(struct inode *inode, int right)
 		NCP_FINFO(inode)->volNumber, 
 		NCP_FINFO(inode)->dirEntNum);
 	error = -EACCES;
-	down(&NCP_FINFO(inode)->open_sem);
+	mutex_lock(&NCP_FINFO(inode)->open_mutex);
 	if (!atomic_read(&NCP_FINFO(inode)->opened)) {
 		struct ncp_entry_info finfo;
 		int result;
@@ -93,7 +93,7 @@ int ncp_make_open(struct inode *inode, int right)
 	}
 
 out_unlock:
-	up(&NCP_FINFO(inode)->open_sem);
+	mutex_unlock(&NCP_FINFO(inode)->open_mutex);
 out:
 	return error;
 }
diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c
index d277a58bd128..0b521d3d97ce 100644
--- a/fs/ncpfs/inode.c
+++ b/fs/ncpfs/inode.c
@@ -63,7 +63,7 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
 	    SLAB_CTOR_CONSTRUCTOR) {
-		init_MUTEX(&ei->open_sem);
+		mutex_init(&ei->open_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -520,7 +520,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent)
 	}
 
 /*	server->lock = 0;	*/
-	init_MUTEX(&server->sem);
+	mutex_init(&server->mutex);
 	server->packet = NULL;
 /*	server->buffer_size = 0;	*/
 /*	server->conn_status = 0;	*/
@@ -557,7 +557,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent)
 	server->dentry_ttl = 0;	/* no caching */
 
 	INIT_LIST_HEAD(&server->tx.requests);
-	init_MUTEX(&server->rcv.creq_sem);
+	mutex_init(&server->rcv.creq_mutex);
 	server->tx.creq		= NULL;
 	server->rcv.creq	= NULL;
 	server->data_ready	= sock->sk->sk_data_ready;
diff --git a/fs/ncpfs/ncplib_kernel.c b/fs/ncpfs/ncplib_kernel.c
index c755e1848a42..d9ebf6439f59 100644
--- a/fs/ncpfs/ncplib_kernel.c
+++ b/fs/ncpfs/ncplib_kernel.c
@@ -291,7 +291,7 @@ ncp_make_closed(struct inode *inode)
 	int err;
 
 	err = 0;
-	down(&NCP_FINFO(inode)->open_sem);	
+	mutex_lock(&NCP_FINFO(inode)->open_mutex);
 	if (atomic_read(&NCP_FINFO(inode)->opened) == 1) {
 		atomic_set(&NCP_FINFO(inode)->opened, 0);
 		err = ncp_close_file(NCP_SERVER(inode), NCP_FINFO(inode)->file_handle);
@@ -301,7 +301,7 @@ ncp_make_closed(struct inode *inode)
 				NCP_FINFO(inode)->volNumber,
 				NCP_FINFO(inode)->dirEntNum, err);
 	}
-	up(&NCP_FINFO(inode)->open_sem);
+	mutex_unlock(&NCP_FINFO(inode)->open_mutex);
 	return err;
 }
 
diff --git a/fs/ncpfs/sock.c b/fs/ncpfs/sock.c
index 6593a5ca88ba..8783eb7ec641 100644
--- a/fs/ncpfs/sock.c
+++ b/fs/ncpfs/sock.c
@@ -171,9 +171,9 @@ static inline void __ncp_abort_request(struct ncp_server *server, struct ncp_req
 
 static inline void ncp_abort_request(struct ncp_server *server, struct ncp_request_reply *req, int err)
 {
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncp_abort_request(server, req, err);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static inline void __ncptcp_abort(struct ncp_server *server)
@@ -303,20 +303,20 @@ static inline void __ncp_start_request(struct ncp_server *server, struct ncp_req
 
 static int ncp_add_request(struct ncp_server *server, struct ncp_request_reply *req)
 {
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	if (!ncp_conn_valid(server)) {
-		up(&server->rcv.creq_sem);
+		mutex_unlock(&server->rcv.creq_mutex);
 		printk(KERN_ERR "ncpfs: tcp: Server died\n");
 		return -EIO;
 	}
 	if (server->tx.creq || server->rcv.creq) {
 		req->status = RQ_QUEUED;
 		list_add_tail(&req->req, &server->tx.requests);
-		up(&server->rcv.creq_sem);
+		mutex_unlock(&server->rcv.creq_mutex);
 		return 0;
 	}
 	__ncp_start_request(server, req);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 	return 0;
 }
 
@@ -400,7 +400,7 @@ void ncpdgram_rcv_proc(void *s)
 				info_server(server, 0, server->unexpected_packet.data, result);
 				continue;
 			}
-			down(&server->rcv.creq_sem);		
+			mutex_lock(&server->rcv.creq_mutex);
 			req = server->rcv.creq;
 			if (req && (req->tx_type == NCP_ALLOC_SLOT_REQUEST || (server->sequence == reply.sequence && 
 					server->connection == get_conn_number(&reply)))) {
@@ -430,11 +430,11 @@ void ncpdgram_rcv_proc(void *s)
 				     	server->rcv.creq = NULL;
 					ncp_finish_request(req, result);
 					__ncp_next_request(server);
-					up(&server->rcv.creq_sem);
+					mutex_unlock(&server->rcv.creq_mutex);
 					continue;
 				}
 			}
-			up(&server->rcv.creq_sem);
+			mutex_unlock(&server->rcv.creq_mutex);
 		}
 drop:;		
 		_recv(sock, &reply, sizeof(reply), MSG_DONTWAIT);
@@ -472,9 +472,9 @@ static void __ncpdgram_timeout_proc(struct ncp_server *server)
 void ncpdgram_timeout_proc(void *s)
 {
 	struct ncp_server *server = s;
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncpdgram_timeout_proc(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static inline void ncp_init_req(struct ncp_request_reply* req)
@@ -657,18 +657,18 @@ void ncp_tcp_rcv_proc(void *s)
 {
 	struct ncp_server *server = s;
 
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncptcp_rcv_proc(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 void ncp_tcp_tx_proc(void *s)
 {
 	struct ncp_server *server = s;
 	
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncptcp_try_send(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static int do_ncp_rpc_call(struct ncp_server *server, int size,
@@ -833,7 +833,7 @@ int ncp_disconnect(struct ncp_server *server)
 
 void ncp_lock_server(struct ncp_server *server)
 {
-	down(&server->sem);
+	mutex_lock(&server->mutex);
 	if (server->lock)
 		printk(KERN_WARNING "ncp_lock_server: was locked!\n");
 	server->lock = 1;
@@ -846,5 +846,5 @@ void ncp_unlock_server(struct ncp_server *server)
 		return;
 	}
 	server->lock = 0;
-	up(&server->sem);
+	mutex_unlock(&server->mutex);
 }
diff --git a/fs/nls/Kconfig b/fs/nls/Kconfig
index 0ab8f00bdbb2..976ecccd6f56 100644
--- a/fs/nls/Kconfig
+++ b/fs/nls/Kconfig
@@ -491,7 +491,7 @@ config NLS_KOI8_U
 	  (koi8-u) and Belarusian (koi8-ru) character sets.
 
 config NLS_UTF8
-	tristate "NLS UTF8"
+	tristate "NLS UTF-8"
 	depends on NLS
 	help
 	  If you want to display filenames with native language characters
diff --git a/fs/ntfs/ChangeLog b/fs/ntfs/ChangeLog
index 9d8ffa89e2c2..35cc4b1d60f7 100644
--- a/fs/ntfs/ChangeLog
+++ b/fs/ntfs/ChangeLog
@@ -16,8 +16,34 @@ ToDo/Notes:
 	  inode having been discarded already.  Whether this can actually ever
 	  happen is unclear however so it is worth waiting until someone hits
 	  the problem.
-	- Enable the code for setting the NT4 compatibility flag when we start
-	  making NTFS 1.2 specific modifications.
+
+2.1.27 - Various bug fixes and cleanups.
+
+	- Fix two compiler warnings on Alpha.  Thanks to Andrew Morton for
+	  reporting them.
+	- Fix an (innocent) off-by-one error in the runlist code.
+	- Fix a buggette in an "should be impossible" case handling where we
+	  continued the attribute lookup loop instead of aborting it.
+	- Use buffer_migrate_page() for the ->migratepage function of all ntfs
+	  address space operations.
+	- Fix comparison of $MFT and $MFTMirr to not bail out when there are
+	  unused, invalid mft records which are the same in both $MFT and
+	  $MFTMirr.
+	- Add support for sparse files which have a compression unit of 0.
+	- Remove all the make_bad_inode() calls.  This should only be called
+	  from read inode and new inode code paths.
+	- Limit name length in fs/ntfs/unistr.c::ntfs_nlstoucs() to maximum
+	  allowed by NTFS, i.e. 255 Unicode characters, not including the
+	  terminating NULL (which is not stored on disk).
+	- Improve comments on file attribute flags in fs/ntfs/layout.h.
+	- Fix a bug in fs/ntfs/inode.c::ntfs_read_locked_index_inode() where we
+	  forgot to update a temporary variable so loading index inodes which
+	  have an index allocation attribute failed.
+	- Add a missing call to flush_dcache_mft_record_page() in
+	  fs/ntfs/inode.c::ntfs_write_inode().
+	- Handle the recently introduced -ENAMETOOLONG return value from
+	  fs/ntfs/unistr.c::ntfs_nlstoucs() in fs/ntfs/namei.c::ntfs_lookup().
+	- Semaphore to mutex conversion.  (Ingo Molnar)
 
 2.1.26 - Minor bug fixes and updates.
 
diff --git a/fs/ntfs/Makefile b/fs/ntfs/Makefile
index d95fac7fdeb6..e27b4eacffbf 100644
--- a/fs/ntfs/Makefile
+++ b/fs/ntfs/Makefile
@@ -6,7 +6,7 @@ ntfs-objs := aops.o attrib.o collate.o compress.o debug.o dir.o file.o \
 	     index.o inode.o mft.o mst.o namei.o runlist.o super.o sysctl.o \
 	     unistr.o upcase.o
 
-EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.26\"
+EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.27\"
 
 ifeq ($(CONFIG_NTFS_DEBUG),y)
 EXTRA_CFLAGS += -DDEBUG
diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index 7e361da770b3..580412d330cb 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -22,6 +22,7 @@
  */
 
 #include <linux/errno.h>
+#include <linux/fs.h>
 #include <linux/mm.h>
 #include <linux/pagemap.h>
 #include <linux/swap.h>
@@ -1277,18 +1278,18 @@ unm_done:
 		
 		tni = locked_nis[nr_locked_nis];
 		/* Get the base inode. */
-		down(&tni->extent_lock);
+		mutex_lock(&tni->extent_lock);
 		if (tni->nr_extents >= 0)
 			base_tni = tni;
 		else {
 			base_tni = tni->ext.base_ntfs_ino;
 			BUG_ON(!base_tni);
 		}
-		up(&tni->extent_lock);
+		mutex_unlock(&tni->extent_lock);
 		ntfs_debug("Unlocking %s inode 0x%lx.",
 				tni == base_tni ? "base" : "extent",
 				tni->mft_no);
-		up(&tni->mrec_lock);
+		mutex_unlock(&tni->mrec_lock);
 		atomic_dec(&tni->count);
 		iput(VFS_I(base_tni));
 	}
@@ -1529,7 +1530,6 @@ err_out:
 				"error %i.", err);
 		SetPageError(page);
 		NVolSetErrors(ni->vol);
-		make_bad_inode(vi);
 	}
 	unlock_page(page);
 	if (ctx)
@@ -1551,6 +1551,9 @@ struct address_space_operations ntfs_aops = {
 #ifdef NTFS_RW
 	.writepage	= ntfs_writepage,	/* Write dirty page to disk. */
 #endif /* NTFS_RW */
+	.migratepage	= buffer_migrate_page,	/* Move a page cache page from
+						   one physical page to an
+						   other. */
 };
 
 /**
@@ -1567,6 +1570,9 @@ struct address_space_operations ntfs_mst_aops = {
 						   without touching the buffers
 						   belonging to the page. */
 #endif /* NTFS_RW */
+	.migratepage	= buffer_migrate_page,	/* Move a page cache page from
+						   one physical page to an
+						   other. */
 };
 
 #ifdef NTFS_RW
diff --git a/fs/ntfs/attrib.c b/fs/ntfs/attrib.c
index 9480a0526cd3..1663f5c3c6aa 100644
--- a/fs/ntfs/attrib.c
+++ b/fs/ntfs/attrib.c
@@ -1,7 +1,7 @@
 /**
  * attrib.c - NTFS attribute operations.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -1048,7 +1048,7 @@ do_next_attr_loop:
 				le32_to_cpu(ctx->mrec->bytes_allocated))
 			break;
 		if (a->type == AT_END)
-			continue;
+			break;
 		if (!a->length)
 			break;
 		if (al_entry->instance != a->instance)
@@ -1695,7 +1695,9 @@ int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size)
 			a->data.non_resident.initialized_size =
 			cpu_to_sle64(attr_size);
 	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		a->data.non_resident.compression_unit = 4;
+		a->data.non_resident.compression_unit = 0;
+		if (NInoCompressed(ni) || vol->major_ver < 3)
+			a->data.non_resident.compression_unit = 4;
 		a->data.non_resident.compressed_size =
 				a->data.non_resident.allocated_size;
 	} else
@@ -1714,13 +1716,20 @@ int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size)
 	ni->allocated_size = new_size;
 	if (NInoSparse(ni) || NInoCompressed(ni)) {
 		ni->itype.compressed.size = ni->allocated_size;
-		ni->itype.compressed.block_size = 1U <<
-				(a->data.non_resident.compression_unit +
-				vol->cluster_size_bits);
-		ni->itype.compressed.block_size_bits =
-				ffs(ni->itype.compressed.block_size) - 1;
-		ni->itype.compressed.block_clusters = 1U <<
-				a->data.non_resident.compression_unit;
+		if (a->data.non_resident.compression_unit) {
+			ni->itype.compressed.block_size = 1U << (a->data.
+					non_resident.compression_unit +
+					vol->cluster_size_bits);
+			ni->itype.compressed.block_size_bits =
+					ffs(ni->itype.compressed.block_size) -
+					1;
+			ni->itype.compressed.block_clusters = 1U <<
+					a->data.non_resident.compression_unit;
+		} else {
+			ni->itype.compressed.block_size = 0;
+			ni->itype.compressed.block_size_bits = 0;
+			ni->itype.compressed.block_clusters = 0;
+		}
 		vi->i_blocks = ni->itype.compressed.size >> 9;
 	} else
 		vi->i_blocks = ni->allocated_size >> 9;
@@ -2429,16 +2438,12 @@ undo_alloc:
 				"chkdsk to recover.", IS_ERR(m) ?
 				"restore attribute search context" :
 				"truncate attribute runlist");
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
 		NVolSetErrors(vol);
 	} else if (mp_rebuilt) {
 		if (ntfs_attr_record_resize(m, a, attr_len)) {
 			ntfs_error(vol->sb, "Failed to restore attribute "
 					"record in error code path.  Run "
 					"chkdsk to recover.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			if (ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
@@ -2451,8 +2456,6 @@ undo_alloc:
 						"mapping pairs array in error "
 						"code path.  Run chkdsk to "
 						"recover.");
-				make_bad_inode(vi);
-				make_bad_inode(VFS_I(base_ni));
 				NVolSetErrors(vol);
 			}
 			flush_dcache_mft_record_page(ctx->ntfs_ino);
diff --git a/fs/ntfs/compress.c b/fs/ntfs/compress.c
index 25d24106f893..68a607ff9fd3 100644
--- a/fs/ntfs/compress.c
+++ b/fs/ntfs/compress.c
@@ -67,7 +67,7 @@ static DEFINE_SPINLOCK(ntfs_cb_lock);
 /**
  * allocate_compression_buffers - allocate the decompression buffers
  *
- * Caller has to hold the ntfs_lock semaphore.
+ * Caller has to hold the ntfs_lock mutex.
  *
  * Return 0 on success or -ENOMEM if the allocations failed.
  */
@@ -84,7 +84,7 @@ int allocate_compression_buffers(void)
 /**
  * free_compression_buffers - free the decompression buffers
  *
- * Caller has to hold the ntfs_lock semaphore.
+ * Caller has to hold the ntfs_lock mutex.
  */
 void free_compression_buffers(void)
 {
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
index b0690d4c8906..9d9ed3fe371d 100644
--- a/fs/ntfs/dir.c
+++ b/fs/ntfs/dir.c
@@ -1136,7 +1136,7 @@ static int ntfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	if (fpos == 1) {
 		ntfs_debug("Calling filldir for .. with len 2, fpos 0x1, "
 				"inode 0x%lx, DT_DIR.",
-				parent_ino(filp->f_dentry));
+				(unsigned long)parent_ino(filp->f_dentry));
 		rc = filldir(dirent, "..", 2, fpos,
 				parent_ino(filp->f_dentry), DT_DIR);
 		if (rc)
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index 5027d3d1b3fe..f5d057e4acc2 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -943,7 +943,8 @@ rl_not_mapped_enoent:
 		}
 		ni->runlist.rl = rl;
 		status.runlist_merged = 1;
-		ntfs_debug("Allocated cluster, lcn 0x%llx.", lcn);
+		ntfs_debug("Allocated cluster, lcn 0x%llx.",
+				(unsigned long long)lcn);
 		/* Map and lock the mft record and get the attribute record. */
 		if (!NInoAttr(ni))
 			base_ni = ni;
@@ -1206,8 +1207,6 @@ rl_not_mapped_enoent:
 					"attribute runlist in error code "
 					"path.  Run chkdsk to recover the "
 					"lost cluster.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			status.runlist_merged = 0;
@@ -1238,8 +1237,6 @@ rl_not_mapped_enoent:
 			ntfs_error(vol->sb, "Failed to restore attribute "
 					"record in error code path.  Run "
 					"chkdsk to recover.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			if (ntfs_mapping_pairs_build(vol, (u8*)a +
@@ -1252,8 +1249,6 @@ rl_not_mapped_enoent:
 						"mapping pairs array in error "
 						"code path.  Run chkdsk to "
 						"recover.");
-				make_bad_inode(vi);
-				make_bad_inode(VFS_I(base_ni));
 				NVolSetErrors(vol);
 			}
 			flush_dcache_mft_record_page(ctx->ntfs_ino);
@@ -1622,11 +1617,8 @@ err_out:
 		unmap_mft_record(base_ni);
 	ntfs_error(vi->i_sb, "Failed to update initialized_size/i_size (error "
 			"code %i).", err);
-	if (err != -ENOMEM) {
+	if (err != -ENOMEM)
 		NVolSetErrors(ni->vol);
-		make_bad_inode(VFS_I(base_ni));
-		make_bad_inode(vi);
-	}
 	return err;
 }
 
@@ -1801,8 +1793,6 @@ err_out:
 		ntfs_error(vi->i_sb, "Resident attribute commit write failed "
 				"with error %i.", err);
 		NVolSetErrors(ni->vol);
-		make_bad_inode(VFS_I(base_ni));
-		make_bad_inode(vi);
 	}
 	if (ctx)
 		ntfs_attr_put_search_ctx(ctx);
diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c
index 55263b7de9c0..4c86b7e1d1eb 100644
--- a/fs/ntfs/inode.c
+++ b/fs/ntfs/inode.c
@@ -1,7 +1,7 @@
 /**
  * inode.c - NTFS kernel inode handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -19,13 +19,19 @@
  * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
-#include <linux/pagemap.h>
 #include <linux/buffer_head.h>
-#include <linux/smp_lock.h>
-#include <linux/quotaops.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
 #include <linux/mount.h>
+#include <linux/mutex.h>
+#include <linux/pagemap.h>
+#include <linux/quotaops.h>
+#include <linux/slab.h>
+#include <linux/smp_lock.h>
 
 #include "aops.h"
+#include "attrib.h"
+#include "bitmap.h"
 #include "dir.h"
 #include "debug.h"
 #include "inode.h"
@@ -382,7 +388,7 @@ void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
 	atomic_set(&ni->count, 1);
 	ni->vol = NTFS_SB(sb);
 	ntfs_init_runlist(&ni->runlist);
-	init_MUTEX(&ni->mrec_lock);
+	mutex_init(&ni->mrec_lock);
 	ni->page = NULL;
 	ni->page_ofs = 0;
 	ni->attr_list_size = 0;
@@ -394,7 +400,7 @@ void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
 	ni->itype.index.collation_rule = 0;
 	ni->itype.index.block_size_bits = 0;
 	ni->itype.index.vcn_size_bits = 0;
-	init_MUTEX(&ni->extent_lock);
+	mutex_init(&ni->extent_lock);
 	ni->nr_extents = 0;
 	ni->ext.base_ntfs_ino = NULL;
 }
@@ -1064,10 +1070,10 @@ skip_large_dir_stuff:
 		if (a->non_resident) {
 			NInoSetNonResident(ni);
 			if (NInoCompressed(ni) || NInoSparse(ni)) {
-				if (a->data.non_resident.compression_unit !=
-						4) {
+				if (NInoCompressed(ni) && a->data.non_resident.
+						compression_unit != 4) {
 					ntfs_error(vi->i_sb, "Found "
-							"nonstandard "
+							"non-standard "
 							"compression unit (%u "
 							"instead of 4).  "
 							"Cannot handle this.",
@@ -1076,16 +1082,26 @@ skip_large_dir_stuff:
 					err = -EOPNOTSUPP;
 					goto unm_err_out;
 				}
-				ni->itype.compressed.block_clusters = 1U <<
-						a->data.non_resident.
-						compression_unit;
-				ni->itype.compressed.block_size = 1U << (
-						a->data.non_resident.
-						compression_unit +
-						vol->cluster_size_bits);
-				ni->itype.compressed.block_size_bits = ffs(
-						ni->itype.compressed.
-						block_size) - 1;
+				if (a->data.non_resident.compression_unit) {
+					ni->itype.compressed.block_size = 1U <<
+							(a->data.non_resident.
+							compression_unit +
+							vol->cluster_size_bits);
+					ni->itype.compressed.block_size_bits =
+							ffs(ni->itype.
+							compressed.
+							block_size) - 1;
+					ni->itype.compressed.block_clusters =
+							1U << a->data.
+							non_resident.
+							compression_unit;
+				} else {
+					ni->itype.compressed.block_size = 0;
+					ni->itype.compressed.block_size_bits =
+							0;
+					ni->itype.compressed.block_clusters =
+							0;
+				}
 				ni->itype.compressed.size = sle64_to_cpu(
 						a->data.non_resident.
 						compressed_size);
@@ -1338,8 +1354,9 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 			goto unm_err_out;
 		}
 		if (NInoCompressed(ni) || NInoSparse(ni)) {
-			if (a->data.non_resident.compression_unit != 4) {
-				ntfs_error(vi->i_sb, "Found nonstandard "
+			if (NInoCompressed(ni) && a->data.non_resident.
+					compression_unit != 4) {
+				ntfs_error(vi->i_sb, "Found non-standard "
 						"compression unit (%u instead "
 						"of 4).  Cannot handle this.",
 						a->data.non_resident.
@@ -1347,13 +1364,22 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 				err = -EOPNOTSUPP;
 				goto unm_err_out;
 			}
-			ni->itype.compressed.block_clusters = 1U <<
-					a->data.non_resident.compression_unit;
-			ni->itype.compressed.block_size = 1U << (
-					a->data.non_resident.compression_unit +
-					vol->cluster_size_bits);
-			ni->itype.compressed.block_size_bits = ffs(
-					ni->itype.compressed.block_size) - 1;
+			if (a->data.non_resident.compression_unit) {
+				ni->itype.compressed.block_size = 1U <<
+						(a->data.non_resident.
+						compression_unit +
+						vol->cluster_size_bits);
+				ni->itype.compressed.block_size_bits =
+						ffs(ni->itype.compressed.
+						block_size) - 1;
+				ni->itype.compressed.block_clusters = 1U <<
+						a->data.non_resident.
+						compression_unit;
+			} else {
+				ni->itype.compressed.block_size = 0;
+				ni->itype.compressed.block_size_bits = 0;
+				ni->itype.compressed.block_clusters = 0;
+			}
 			ni->itype.compressed.size = sle64_to_cpu(
 					a->data.non_resident.compressed_size);
 		}
@@ -1406,7 +1432,6 @@ err_out:
 			"Run chkdsk.", err, vi->i_ino, ni->type, ni->name_len,
 			base_vi->i_ino);
 	make_bad_inode(vi);
-	make_bad_inode(base_vi);
 	if (err != -ENOMEM)
 		NVolSetErrors(vol);
 	return err;
@@ -1591,6 +1616,7 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 					"$INDEX_ALLOCATION attribute.");
 		goto unm_err_out;
 	}
+	a = ctx->attr;
 	if (!a->non_resident) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
 				"resident.");
@@ -2823,11 +2849,8 @@ done:
 old_bad_out:
 	old_size = -1;
 bad_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP) {
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
+	if (err != -ENOMEM && err != -EOPNOTSUPP)
 		NVolSetErrors(vol);
-	}
 	if (err != -EOPNOTSUPP)
 		NInoSetTruncateFailed(ni);
 	else if (old_size >= 0)
@@ -2842,11 +2865,8 @@ out:
 	ntfs_debug("Failed.  Returning error code %i.", err);
 	return err;
 conv_err_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP) {
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
+	if (err != -ENOMEM && err != -EOPNOTSUPP)
 		NVolSetErrors(vol);
-	}
 	if (err != -EOPNOTSUPP)
 		NInoSetTruncateFailed(ni);
 	else
@@ -3044,15 +3064,18 @@ int ntfs_write_inode(struct inode *vi, int sync)
 	 * record will be cleaned and written out to disk below, i.e. before
 	 * this function returns.
 	 */
-	if (modified && !NInoTestSetDirty(ctx->ntfs_ino))
-		mark_ntfs_record_dirty(ctx->ntfs_ino->page,
-				ctx->ntfs_ino->page_ofs);
+	if (modified) {
+		flush_dcache_mft_record_page(ctx->ntfs_ino);
+		if (!NInoTestSetDirty(ctx->ntfs_ino))
+			mark_ntfs_record_dirty(ctx->ntfs_ino->page,
+					ctx->ntfs_ino->page_ofs);
+	}
 	ntfs_attr_put_search_ctx(ctx);
 	/* Now the access times are updated, write the base mft record. */
 	if (NInoDirty(ni))
 		err = write_mft_record(ni, m, sync);
 	/* Write all attached extent mft records. */
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (ni->nr_extents > 0) {
 		ntfs_inode **extent_nis = ni->ext.extent_ntfs_inos;
 		int i;
@@ -3079,7 +3102,7 @@ int ntfs_write_inode(struct inode *vi, int sync)
 			}
 		}
 	}
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	unmap_mft_record(ni);
 	if (unlikely(err))
 		goto err_out;
@@ -3094,9 +3117,7 @@ err_out:
 				"retries later.");
 		mark_inode_dirty(vi);
 	} else {
-		ntfs_error(vi->i_sb, "Failed (error code %i):  Marking inode "
-				"as bad.  You should run chkdsk.", -err);
-		make_bad_inode(vi);
+		ntfs_error(vi->i_sb, "Failed (error %i):  Run chkdsk.", -err);
 		NVolSetErrors(ni->vol);
 	}
 	return err;
diff --git a/fs/ntfs/inode.h b/fs/ntfs/inode.h
index 3de5c0231966..f088291e017c 100644
--- a/fs/ntfs/inode.h
+++ b/fs/ntfs/inode.h
@@ -24,12 +24,13 @@
 #ifndef _LINUX_NTFS_INODE_H
 #define _LINUX_NTFS_INODE_H
 
-#include <linux/mm.h>
+#include <asm/atomic.h>
+
 #include <linux/fs.h>
-#include <linux/seq_file.h>
 #include <linux/list.h>
-#include <asm/atomic.h>
-#include <asm/semaphore.h>
+#include <linux/mm.h>
+#include <linux/mutex.h>
+#include <linux/seq_file.h>
 
 #include "layout.h"
 #include "volume.h"
@@ -81,7 +82,7 @@ struct _ntfs_inode {
 	 * The following fields are only valid for real inodes and extent
 	 * inodes.
 	 */
-	struct semaphore mrec_lock; /* Lock for serializing access to the
+	struct mutex mrec_lock;	/* Lock for serializing access to the
 				   mft record belonging to this inode. */
 	struct page *page;	/* The page containing the mft record of the
 				   inode. This should only be touched by the
@@ -119,7 +120,7 @@ struct _ntfs_inode {
 			u8 block_clusters;	/* Number of clusters per cb. */
 		} compressed;
 	} itype;
-	struct semaphore extent_lock;	/* Lock for accessing/modifying the
+	struct mutex extent_lock;	/* Lock for accessing/modifying the
 					   below . */
 	s32 nr_extents;	/* For a base mft record, the number of attached extent
 			   inodes (0 if none), for extent records and for fake
diff --git a/fs/ntfs/layout.h b/fs/ntfs/layout.h
index bb408d4dcbb0..d34b93cb8b48 100644
--- a/fs/ntfs/layout.h
+++ b/fs/ntfs/layout.h
@@ -769,7 +769,7 @@ typedef struct {
 				compressed.  (This effectively limits the
 				compression unit size to be a power of two
 				clusters.)  WinNT4 only uses a value of 4.
-				Sparse files also have this set to 4. */
+				Sparse files have this set to 0 on XPSP2. */
 /* 35*/			u8 reserved[5];		/* Align to 8-byte boundary. */
 /* The sizes below are only used when lowest_vcn is zero, as otherwise it would
    be difficult to keep them up-to-date.*/
@@ -801,13 +801,16 @@ typedef struct {
 typedef ATTR_RECORD ATTR_REC;
 
 /*
- * File attribute flags (32-bit).
+ * File attribute flags (32-bit) appearing in the file_attributes fields of the
+ * STANDARD_INFORMATION attribute of MFT_RECORDs and the FILENAME_ATTR
+ * attributes of MFT_RECORDs and directory index entries.
+ *
+ * All of the below flags appear in the directory index entries but only some
+ * appear in the STANDARD_INFORMATION attribute whilst only some others appear
+ * in the FILENAME_ATTR attribute of MFT_RECORDs.  Unless otherwise stated the
+ * flags appear in all of the above.
  */
 enum {
-	/*
-	 * The following flags are only present in the STANDARD_INFORMATION
-	 * attribute (in the field file_attributes).
-	 */
 	FILE_ATTR_READONLY		= const_cpu_to_le32(0x00000001),
 	FILE_ATTR_HIDDEN		= const_cpu_to_le32(0x00000002),
 	FILE_ATTR_SYSTEM		= const_cpu_to_le32(0x00000004),
@@ -839,18 +842,14 @@ enum {
 	   F_A_COMPRESSED, and F_A_ENCRYPTED and preserves the rest.  This mask
 	   is used to to obtain all flags that are valid for setting. */
 	/*
-	 * The following flag is only present in the FILE_NAME attribute (in
-	 * the field file_attributes).
+	 * The flag FILE_ATTR_DUP_FILENAME_INDEX_PRESENT is present in all
+	 * FILENAME_ATTR attributes but not in the STANDARD_INFORMATION
+	 * attribute of an mft record.
 	 */
 	FILE_ATTR_DUP_FILE_NAME_INDEX_PRESENT	= const_cpu_to_le32(0x10000000),
 	/* Note, this is a copy of the corresponding bit from the mft record,
 	   telling us whether this is a directory or not, i.e. whether it has
 	   an index root attribute or not. */
-	/*
-	 * The following flag is present both in the STANDARD_INFORMATION
-	 * attribute and in the FILE_NAME attribute (in the field
-	 * file_attributes).
-	 */
 	FILE_ATTR_DUP_VIEW_INDEX_PRESENT	= const_cpu_to_le32(0x20000000),
 	/* Note, this is a copy of the corresponding bit from the mft record,
 	   telling us whether this file has a view index present (eg. object id
@@ -891,7 +890,7 @@ typedef struct {
 					   Windows this is only updated when
 					   accessed if some time delta has
 					   passed since the last update. Also,
-					   last access times updates can be
+					   last access time updates can be
 					   disabled altogether for speed. */
 /* 32*/	FILE_ATTR_FLAGS file_attributes; /* Flags describing the file. */
 /* 36*/	union {
@@ -1076,16 +1075,21 @@ typedef struct {
 /* 20*/	sle64 last_access_time;		/* Time this mft record was last
 					   accessed. */
 /* 28*/	sle64 allocated_size;		/* Byte size of on-disk allocated space
-					   for the data attribute.  So for
-					   normal $DATA, this is the
+					   for the unnamed data attribute.  So
+					   for normal $DATA, this is the
 					   allocated_size from the unnamed
 					   $DATA attribute and for compressed
 					   and/or sparse $DATA, this is the
 					   compressed_size from the unnamed
-					   $DATA attribute.  NOTE: This is a
-					   multiple of the cluster size. */
-/* 30*/	sle64 data_size;		/* Byte size of actual data in data
-					   attribute. */
+					   $DATA attribute.  For a directory or
+					   other inode without an unnamed $DATA
+					   attribute, this is always 0.  NOTE:
+					   This is a multiple of the cluster
+					   size. */
+/* 30*/	sle64 data_size;		/* Byte size of actual data in unnamed
+					   data attribute.  For a directory or
+					   other inode without an unnamed $DATA
+					   attribute, this is always 0. */
 /* 38*/	FILE_ATTR_FLAGS file_attributes;	/* Flags describing the file. */
 /* 3c*/	union {
 	/* 3c*/	struct {
diff --git a/fs/ntfs/mft.c b/fs/ntfs/mft.c
index 6499aafc2258..4e72bc7afdf9 100644
--- a/fs/ntfs/mft.c
+++ b/fs/ntfs/mft.c
@@ -93,6 +93,7 @@ static inline MFT_RECORD *map_mft_record_page(ntfs_inode *ni)
 				"Run chkdsk.", ni->mft_no);
 		ntfs_unmap_page(page);
 		page = ERR_PTR(-EIO);
+		NVolSetErrors(vol);
 	}
 err_out:
 	ni->page = NULL;
@@ -104,8 +105,8 @@ err_out:
  * map_mft_record - map, pin and lock an mft record
  * @ni:		ntfs inode whose MFT record to map
  *
- * First, take the mrec_lock semaphore. We might now be sleeping, while waiting
- * for the semaphore if it was already locked by someone else.
+ * First, take the mrec_lock mutex.  We might now be sleeping, while waiting
+ * for the mutex if it was already locked by someone else.
  *
  * The page of the record is mapped using map_mft_record_page() before being
  * returned to the caller.
@@ -135,9 +136,9 @@ err_out:
  * So that code will end up having to own the mrec_lock of all mft
  * records/inodes present in the page before I/O can proceed. In that case we
  * wouldn't need to bother with PG_locked and PG_uptodate as nobody will be
- * accessing anything without owning the mrec_lock semaphore. But we do need
- * to use them because of the read_cache_page() invocation and the code becomes
- * so much simpler this way that it is well worth it.
+ * accessing anything without owning the mrec_lock mutex.  But we do need to
+ * use them because of the read_cache_page() invocation and the code becomes so
+ * much simpler this way that it is well worth it.
  *
  * The mft record is now ours and we return a pointer to it. You need to check
  * the returned pointer with IS_ERR() and if that is true, PTR_ERR() will return
@@ -160,13 +161,13 @@ MFT_RECORD *map_mft_record(ntfs_inode *ni)
 	atomic_inc(&ni->count);
 
 	/* Serialize access to this mft record. */
-	down(&ni->mrec_lock);
+	mutex_lock(&ni->mrec_lock);
 
 	m = map_mft_record_page(ni);
 	if (likely(!IS_ERR(m)))
 		return m;
 
-	up(&ni->mrec_lock);
+	mutex_unlock(&ni->mrec_lock);
 	atomic_dec(&ni->count);
 	ntfs_error(ni->vol->sb, "Failed with error code %lu.", -PTR_ERR(m));
 	return m;
@@ -217,7 +218,7 @@ void unmap_mft_record(ntfs_inode *ni)
 	ntfs_debug("Entering for mft_no 0x%lx.", ni->mft_no);
 
 	unmap_mft_record_page(ni);
-	up(&ni->mrec_lock);
+	mutex_unlock(&ni->mrec_lock);
 	atomic_dec(&ni->count);
 	/*
 	 * If pure ntfs_inode, i.e. no vfs inode attached, we leave it to
@@ -261,7 +262,7 @@ MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
 	 * in which case just return it. If not found, add it to the base
 	 * inode before returning it.
 	 */
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 	if (base_ni->nr_extents > 0) {
 		extent_nis = base_ni->ext.extent_ntfs_inos;
 		for (i = 0; i < base_ni->nr_extents; i++) {
@@ -274,7 +275,7 @@ MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
 		}
 	}
 	if (likely(ni != NULL)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		/* We found the record; just have to map and return it. */
 		m = map_mft_record(ni);
@@ -301,7 +302,7 @@ map_err_out:
 	/* Record wasn't there. Get a new ntfs inode and initialize it. */
 	ni = ntfs_new_extent_inode(base_ni->vol->sb, mft_no);
 	if (unlikely(!ni)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		return ERR_PTR(-ENOMEM);
 	}
@@ -312,7 +313,7 @@ map_err_out:
 	/* Now map the record. */
 	m = map_mft_record(ni);
 	if (IS_ERR(m)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		ntfs_clear_extent_inode(ni);
 		goto map_err_out;
@@ -347,14 +348,14 @@ map_err_out:
 		base_ni->ext.extent_ntfs_inos = tmp;
 	}
 	base_ni->ext.extent_ntfs_inos[base_ni->nr_extents++] = ni;
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	atomic_dec(&base_ni->count);
 	ntfs_debug("Done 2.");
 	*ntfs_ino = ni;
 	return m;
 unm_err_out:
 	unmap_mft_record(ni);
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	atomic_dec(&base_ni->count);
 	/*
 	 * If the extent inode was not attached to the base inode we need to
@@ -399,12 +400,12 @@ void __mark_mft_record_dirty(ntfs_inode *ni)
 	BUG_ON(NInoAttr(ni));
 	mark_ntfs_record_dirty(ni->page, ni->page_ofs);
 	/* Determine the base vfs inode and mark it dirty, too. */
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (likely(ni->nr_extents >= 0))
 		base_ni = ni;
 	else
 		base_ni = ni->ext.base_ntfs_ino;
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	__mark_inode_dirty(VFS_I(base_ni), I_DIRTY_SYNC | I_DIRTY_DATASYNC);
 }
 
@@ -650,10 +651,7 @@ err_out:
  * fs/ntfs/aops.c::mark_ntfs_record_dirty().
  *
  * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.  The caller should call make_bad_inode() on
- * the base inode to ensure no more access happens to this inode.  We do not do
- * it here as the caller may want to finish writing other extent mft records
- * first to minimize on-disk metadata inconsistencies.
+ * record dirty and return -errno.
  *
  * NOTE:  We always perform synchronous i/o and ignore the @sync parameter.
  * However, if the mft record has a counterpart in the mft mirror and @sync is
@@ -983,7 +981,7 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 		}
 		ntfs_debug("Inode 0x%lx is not dirty.", mft_no);
 		/* The inode is not dirty, try to take the mft record lock. */
-		if (unlikely(down_trylock(&ni->mrec_lock))) {
+		if (unlikely(!mutex_trylock(&ni->mrec_lock))) {
 			ntfs_debug("Mft record 0x%lx is already locked, do "
 					"not write it.", mft_no);
 			atomic_dec(&ni->count);
@@ -1043,13 +1041,13 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	 * corresponding to this extent mft record attached.
 	 */
 	ni = NTFS_I(vi);
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (ni->nr_extents <= 0) {
 		/*
 		 * The base inode has no attached extent inodes, write this
 		 * extent mft record.
 		 */
-		up(&ni->extent_lock);
+		mutex_unlock(&ni->extent_lock);
 		iput(vi);
 		ntfs_debug("Base inode 0x%lx has no attached extent inodes, "
 				"write the extent record.", na.mft_no);
@@ -1072,7 +1070,7 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	 * extent mft record.
 	 */
 	if (!eni) {
-		up(&ni->extent_lock);
+		mutex_unlock(&ni->extent_lock);
 		iput(vi);
 		ntfs_debug("Extent inode 0x%lx is not attached to its base "
 				"inode 0x%lx, write the extent record.",
@@ -1083,12 +1081,12 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 			mft_no, na.mft_no);
 	/* Take a reference to the extent ntfs inode. */
 	atomic_inc(&eni->count);
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	/*
 	 * Found the extent inode coresponding to this extent mft record.
 	 * Try to take the mft record lock.
 	 */
-	if (unlikely(down_trylock(&eni->mrec_lock))) {
+	if (unlikely(!mutex_trylock(&eni->mrec_lock))) {
 		atomic_dec(&eni->count);
 		iput(vi);
 		ntfs_debug("Extent mft record 0x%lx is already locked, do "
@@ -2711,7 +2709,7 @@ mft_rec_already_initialized:
 		 * have its page mapped and it is very easy to do.
 		 */
 		atomic_inc(&ni->count);
-		down(&ni->mrec_lock);
+		mutex_lock(&ni->mrec_lock);
 		ni->page = page;
 		ni->page_ofs = ofs;
 		/*
@@ -2798,22 +2796,22 @@ int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m)
 	BUG_ON(NInoAttr(ni));
 	BUG_ON(ni->nr_extents != -1);
 
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	base_ni = ni->ext.base_ntfs_ino;
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 
 	BUG_ON(base_ni->nr_extents <= 0);
 
 	ntfs_debug("Entering for extent inode 0x%lx, base inode 0x%lx.\n",
 			mft_no, base_ni->mft_no);
 
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 
 	/* Make sure we are holding the only reference to the extent inode. */
 	if (atomic_read(&ni->count) > 2) {
 		ntfs_error(vol->sb, "Tried to free busy extent inode 0x%lx, "
 				"not freeing.", base_ni->mft_no);
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		return -EBUSY;
 	}
 
@@ -2831,7 +2829,7 @@ int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m)
 		break;
 	}
 
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 
 	if (unlikely(err)) {
 		ntfs_error(vol->sb, "Extent inode 0x%lx is not attached to "
@@ -2890,7 +2888,7 @@ rollback_error:
 	return 0;
 rollback:
 	/* Rollback what we did... */
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 	extent_nis = base_ni->ext.extent_ntfs_inos;
 	if (!(base_ni->nr_extents & 3)) {
 		int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode*);
@@ -2899,7 +2897,7 @@ rollback:
 		if (unlikely(!extent_nis)) {
 			ntfs_error(vol->sb, "Failed to allocate internal "
 					"buffer during rollback.%s", es);
-			up(&base_ni->extent_lock);
+			mutex_unlock(&base_ni->extent_lock);
 			NVolSetErrors(vol);
 			goto rollback_error;
 		}
@@ -2914,7 +2912,7 @@ rollback:
 	m->flags |= MFT_RECORD_IN_USE;
 	m->sequence_number = old_seq_no;
 	extent_nis[base_ni->nr_extents++] = ni;
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	mark_mft_record_dirty(ni);
 	return err;
 }
diff --git a/fs/ntfs/mft.h b/fs/ntfs/mft.h
index 407de2cef1d6..639cd1bab08b 100644
--- a/fs/ntfs/mft.h
+++ b/fs/ntfs/mft.h
@@ -97,10 +97,7 @@ extern int write_mft_record_nolock(ntfs_inode *ni, MFT_RECORD *m, int sync);
  * uptodate.
  *
  * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.  The caller should call make_bad_inode() on
- * the base inode to ensure no more access happens to this inode.  We do not do
- * it here as the caller may want to finish writing other extent mft records
- * first to minimize on-disk metadata inconsistencies.
+ * record dirty and return -errno.
  */
 static inline int write_mft_record(ntfs_inode *ni, MFT_RECORD *m, int sync)
 {
diff --git a/fs/ntfs/namei.c b/fs/ntfs/namei.c
index 5ea9eb93af62..eddb2247cec5 100644
--- a/fs/ntfs/namei.c
+++ b/fs/ntfs/namei.c
@@ -2,7 +2,7 @@
  * namei.c - NTFS kernel directory inode operations. Part of the Linux-NTFS
  *	     project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -115,7 +115,9 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 	uname_len = ntfs_nlstoucs(vol, dent->d_name.name, dent->d_name.len,
 			&uname);
 	if (uname_len < 0) {
-		ntfs_error(vol->sb, "Failed to convert name to Unicode.");
+		if (uname_len != -ENAMETOOLONG)
+			ntfs_error(vol->sb, "Failed to convert name to "
+					"Unicode.");
 		return ERR_PTR(uname_len);
 	}
 	mref = ntfs_lookup_inode_by_name(NTFS_I(dir_ino), uname, uname_len,
@@ -157,7 +159,7 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 		/* Return the error code. */
 		return (struct dentry *)dent_inode;
 	}
-	/* It is guaranteed that name is no longer allocated at this point. */
+	/* It is guaranteed that @name is no longer allocated at this point. */
 	if (MREF_ERR(mref) == -ENOENT) {
 		ntfs_debug("Entry was not found, adding negative dentry.");
 		/* The dcache will handle negative entries. */
@@ -168,7 +170,6 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 	ntfs_error(vol->sb, "ntfs_lookup_ino_by_name() failed with error "
 			"code %i.", -MREF_ERR(mref));
 	return ERR_PTR(MREF_ERR(mref));
-
 	// TODO: Consider moving this lot to a separate function! (AIA)
 handle_name:
    {
diff --git a/fs/ntfs/ntfs.h b/fs/ntfs/ntfs.h
index 653d2a5c4899..0624c8ef4d9c 100644
--- a/fs/ntfs/ntfs.h
+++ b/fs/ntfs/ntfs.h
@@ -91,7 +91,7 @@ extern void free_compression_buffers(void);
 
 /* From fs/ntfs/super.c */
 #define default_upcase_len 0x10000
-extern struct semaphore ntfs_lock;
+extern struct mutex ntfs_lock;
 
 typedef struct {
 	int val;
diff --git a/fs/ntfs/runlist.c b/fs/ntfs/runlist.c
index 061b5ff6b73c..eb52b801512b 100644
--- a/fs/ntfs/runlist.c
+++ b/fs/ntfs/runlist.c
@@ -381,6 +381,7 @@ static inline runlist_element *ntfs_rl_insert(runlist_element *dst,
 static inline runlist_element *ntfs_rl_replace(runlist_element *dst,
 		int dsize, runlist_element *src, int ssize, int loc)
 {
+	signed delta;
 	BOOL left = FALSE;	/* Left end of @src needs merging. */
 	BOOL right = FALSE;	/* Right end of @src needs merging. */
 	int tail;		/* Start of tail of @dst. */
@@ -396,11 +397,14 @@ static inline runlist_element *ntfs_rl_replace(runlist_element *dst,
 		left = ntfs_are_rl_mergeable(dst + loc - 1, src);
 	/*
 	 * Allocate some space.  We will need less if the left, right, or both
-	 * ends get merged.
+	 * ends get merged.  The -1 accounts for the run being replaced.
 	 */
-	dst = ntfs_rl_realloc(dst, dsize, dsize + ssize - left - right);
-	if (IS_ERR(dst))
-		return dst;
+	delta = ssize - 1 - left - right;
+	if (delta > 0) {
+		dst = ntfs_rl_realloc(dst, dsize, dsize + delta);
+		if (IS_ERR(dst))
+			return dst;
+	}
 	/*
 	 * We are guaranteed to succeed from here so can start modifying the
 	 * original runlists.
diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c
index 368a8ec10668..7646b5059389 100644
--- a/fs/ntfs/super.c
+++ b/fs/ntfs/super.c
@@ -1099,26 +1099,38 @@ static BOOL check_mft_mirror(ntfs_volume *vol)
 			kmirr = page_address(mirr_page);
 			++index;
 		}
-		/* Make sure the record is ok. */
-		if (ntfs_is_baad_recordp((le32*)kmft)) {
-			ntfs_error(sb, "Incomplete multi sector transfer "
-					"detected in mft record %i.", i);
+		/* Do not check the record if it is not in use. */
+		if (((MFT_RECORD*)kmft)->flags & MFT_RECORD_IN_USE) {
+			/* Make sure the record is ok. */
+			if (ntfs_is_baad_recordp((le32*)kmft)) {
+				ntfs_error(sb, "Incomplete multi sector "
+						"transfer detected in mft "
+						"record %i.", i);
 mm_unmap_out:
-			ntfs_unmap_page(mirr_page);
+				ntfs_unmap_page(mirr_page);
 mft_unmap_out:
-			ntfs_unmap_page(mft_page);
-			return FALSE;
+				ntfs_unmap_page(mft_page);
+				return FALSE;
+			}
 		}
-		if (ntfs_is_baad_recordp((le32*)kmirr)) {
-			ntfs_error(sb, "Incomplete multi sector transfer "
-					"detected in mft mirror record %i.", i);
-			goto mm_unmap_out;
+		/* Do not check the mirror record if it is not in use. */
+		if (((MFT_RECORD*)kmirr)->flags & MFT_RECORD_IN_USE) {
+			if (ntfs_is_baad_recordp((le32*)kmirr)) {
+				ntfs_error(sb, "Incomplete multi sector "
+						"transfer detected in mft "
+						"mirror record %i.", i);
+				goto mm_unmap_out;
+			}
 		}
 		/* Get the amount of data in the current record. */
 		bytes = le32_to_cpu(((MFT_RECORD*)kmft)->bytes_in_use);
-		if (!bytes || bytes > vol->mft_record_size) {
+		if (bytes < sizeof(MFT_RECORD_OLD) ||
+				bytes > vol->mft_record_size ||
+				ntfs_is_baad_recordp((le32*)kmft)) {
 			bytes = le32_to_cpu(((MFT_RECORD*)kmirr)->bytes_in_use);
-			if (!bytes || bytes > vol->mft_record_size)
+			if (bytes < sizeof(MFT_RECORD_OLD) ||
+					bytes > vol->mft_record_size ||
+					ntfs_is_baad_recordp((le32*)kmirr))
 				bytes = vol->mft_record_size;
 		}
 		/* Compare the two records. */
@@ -1665,11 +1677,11 @@ read_partial_upcase_page:
 	ntfs_debug("Read %llu bytes from $UpCase (expected %zu bytes).",
 			i_size, 64 * 1024 * sizeof(ntfschar));
 	iput(ino);
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (!default_upcase) {
 		ntfs_debug("Using volume specified $UpCase since default is "
 				"not present.");
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		return TRUE;
 	}
 	max = default_upcase_len;
@@ -1683,12 +1695,12 @@ read_partial_upcase_page:
 		vol->upcase = default_upcase;
 		vol->upcase_len = max;
 		ntfs_nr_upcase_users++;
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		ntfs_debug("Volume specified $UpCase matches default. Using "
 				"default.");
 		return TRUE;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	ntfs_debug("Using volume specified $UpCase since it does not match "
 			"the default.");
 	return TRUE;
@@ -1697,17 +1709,17 @@ iput_upcase_failed:
 	ntfs_free(vol->upcase);
 	vol->upcase = NULL;
 upcase_failed:
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (default_upcase) {
 		vol->upcase = default_upcase;
 		vol->upcase_len = default_upcase_len;
 		ntfs_nr_upcase_users++;
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		ntfs_error(sb, "Failed to load $UpCase from the volume. Using "
 				"default.");
 		return TRUE;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	ntfs_error(sb, "Failed to initialize upcase table.");
 	return FALSE;
 }
@@ -2183,12 +2195,12 @@ iput_attrdef_err_out:
 iput_upcase_err_out:
 #endif /* NTFS_RW */
 	vol->upcase_len = 0;
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -2393,7 +2405,7 @@ static void ntfs_put_super(struct super_block *sb)
 	 * Destroy the global default upcase table if necessary.  Also decrease
 	 * the number of upcase users if we are a user.
 	 */
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
@@ -2404,7 +2416,7 @@ static void ntfs_put_super(struct super_block *sb)
 	}
 	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
 		free_compression_buffers();
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -2878,7 +2890,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 			ntfs_error(sb, "Failed to load essential metadata.");
 		goto iput_tmp_ino_err_out_now;
 	}
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	/*
 	 * The current mount is a compression user if the cluster size is
 	 * less than or equal 4kiB.
@@ -2889,7 +2901,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 			ntfs_error(NULL, "Failed to allocate buffers "
 					"for compression engine.");
 			ntfs_nr_compression_users--;
-			up(&ntfs_lock);
+			mutex_unlock(&ntfs_lock);
 			goto iput_tmp_ino_err_out_now;
 		}
 	}
@@ -2901,7 +2913,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 	if (!default_upcase)
 		default_upcase = generate_default_upcase();
 	ntfs_nr_upcase_users++;
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	/*
 	 * From now on, ignore @silent parameter. If we fail below this line,
 	 * it will be due to a corrupt fs or a system error, so we report it.
@@ -2919,12 +2931,12 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 		atomic_inc(&vol->root_ino->i_count);
 		ntfs_debug("Exiting, status successful.");
 		/* Release the default upcase if it has no users. */
-		down(&ntfs_lock);
+		mutex_lock(&ntfs_lock);
 		if (!--ntfs_nr_upcase_users && default_upcase) {
 			ntfs_free(default_upcase);
 			default_upcase = NULL;
 		}
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		sb->s_export_op = &ntfs_export_ops;
 		lock_kernel();
 		return 0;
@@ -2992,12 +3004,12 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 		vol->attrdef = NULL;
 	}
 	vol->upcase_len = 0;
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -3012,14 +3024,14 @@ unl_upcase_iput_tmp_ino_err_out_now:
 	 * Decrease the number of upcase users and destroy the global default
 	 * upcase table if necessary.
 	 */
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (!--ntfs_nr_upcase_users && default_upcase) {
 		ntfs_free(default_upcase);
 		default_upcase = NULL;
 	}
 	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
 		free_compression_buffers();
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 iput_tmp_ino_err_out_now:
 	iput(tmp_ino);
 	if (vol->mft_ino && vol->mft_ino != tmp_ino)
@@ -3078,8 +3090,8 @@ static void ntfs_big_inode_init_once(void *foo, struct kmem_cache *cachep,
 struct kmem_cache *ntfs_attr_ctx_cache;
 struct kmem_cache *ntfs_index_ctx_cache;
 
-/* Driver wide semaphore. */
-DECLARE_MUTEX(ntfs_lock);
+/* Driver wide mutex. */
+DEFINE_MUTEX(ntfs_lock);
 
 static struct super_block *ntfs_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *data)
@@ -3234,7 +3246,7 @@ static void __exit exit_ntfs_fs(void)
 }
 
 MODULE_AUTHOR("Anton Altaparmakov <aia21@cantab.net>");
-MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2005 Anton Altaparmakov");
+MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2006 Anton Altaparmakov");
 MODULE_VERSION(NTFS_VERSION);
 MODULE_LICENSE("GPL");
 #ifdef DEBUG
diff --git a/fs/ntfs/unistr.c b/fs/ntfs/unistr.c
index 0ea887fc859c..b123c0fa6bf6 100644
--- a/fs/ntfs/unistr.c
+++ b/fs/ntfs/unistr.c
@@ -1,7 +1,7 @@
 /*
  * unistr.c - NTFS Unicode string handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -19,6 +19,8 @@
  * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
+#include <linux/slab.h>
+
 #include "types.h"
 #include "debug.h"
 #include "ntfs.h"
@@ -242,7 +244,7 @@ int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
  * map dictates, into a little endian, 2-byte Unicode string.
  *
  * This function allocates the string and the caller is responsible for
- * calling kmem_cache_free(ntfs_name_cache, @outs); when finished with it.
+ * calling kmem_cache_free(ntfs_name_cache, *@outs); when finished with it.
  *
  * On success the function returns the number of Unicode characters written to
  * the output string *@outs (>= 0), not counting the terminating Unicode NULL
@@ -262,37 +264,48 @@ int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
 	wchar_t wc;
 	int i, o, wc_len;
 
-	/* We don't trust outside sources. */
-	if (ins) {
+	/* We do not trust outside sources. */
+	if (likely(ins)) {
 		ucs = kmem_cache_alloc(ntfs_name_cache, SLAB_NOFS);
-		if (ucs) {
+		if (likely(ucs)) {
 			for (i = o = 0; i < ins_len; i += wc_len) {
 				wc_len = nls->char2uni(ins + i, ins_len - i,
 						&wc);
-				if (wc_len >= 0) {
-					if (wc) {
+				if (likely(wc_len >= 0 &&
+						o < NTFS_MAX_NAME_LEN)) {
+					if (likely(wc)) {
 						ucs[o++] = cpu_to_le16(wc);
 						continue;
-					} /* else (!wc) */
+					} /* else if (!wc) */
 					break;
-				} /* else (wc_len < 0) */
-				goto conversion_err;
+				} /* else if (wc_len < 0 ||
+						o >= NTFS_MAX_NAME_LEN) */
+				goto name_err;
 			}
 			ucs[o] = 0;
 			*outs = ucs;
 			return o;
-		} /* else (!ucs) */
-		ntfs_error(vol->sb, "Failed to allocate name from "
-				"ntfs_name_cache!");
+		} /* else if (!ucs) */
+		ntfs_error(vol->sb, "Failed to allocate buffer for converted "
+				"name from ntfs_name_cache.");
 		return -ENOMEM;
-	} /* else (!ins) */
-	ntfs_error(NULL, "Received NULL pointer.");
+	} /* else if (!ins) */
+	ntfs_error(vol->sb, "Received NULL pointer.");
 	return -EINVAL;
-conversion_err:
-	ntfs_error(vol->sb, "Name using character set %s contains characters "
-			"that cannot be converted to Unicode.", nls->charset);
+name_err:
 	kmem_cache_free(ntfs_name_cache, ucs);
-	return -EILSEQ;
+	if (wc_len < 0) {
+		ntfs_error(vol->sb, "Name using character set %s contains "
+				"characters that cannot be converted to "
+				"Unicode.", nls->charset);
+		i = -EILSEQ;
+	} else /* if (o >= NTFS_MAX_NAME_LEN) */ {
+		ntfs_error(vol->sb, "Name is too long (maximum length for a "
+				"name on NTFS is %d Unicode characters.",
+				NTFS_MAX_NAME_LEN);
+		i = -ENAMETOOLONG;
+	}
+	return i;
 }
 
 /**
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 8dd3aafec499..09e1c57a86a0 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -959,7 +959,7 @@ static int ocfs2_initialize_mem_caches(void)
 	ocfs2_lock_cache = kmem_cache_create("ocfs2_lock",
 					     sizeof(struct ocfs2_journal_lock),
 					     0,
-					     SLAB_NO_REAP|SLAB_HWCACHE_ALIGN,
+					     SLAB_HWCACHE_ALIGN,
 					     NULL, NULL);
 	if (!ocfs2_lock_cache)
 		return -ENOMEM;
diff --git a/fs/open.c b/fs/open.c
index 70e0230d8e77..1091dadd6c38 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -973,7 +973,7 @@ repeat:
 	fdt = files_fdtable(files);
  	fd = find_next_zero_bit(fdt->open_fds->fds_bits,
 				fdt->max_fdset,
-				fdt->next_fd);
+				files->next_fd);
 
 	/*
 	 * N.B. For clone tasks sharing a files structure, this test
@@ -998,7 +998,7 @@ repeat:
 
 	FD_SET(fd, fdt->open_fds);
 	FD_CLR(fd, fdt->close_on_exec);
-	fdt->next_fd = fd + 1;
+	files->next_fd = fd + 1;
 #if 1
 	/* Sanity check */
 	if (fdt->fd[fd] != NULL) {
@@ -1019,8 +1019,8 @@ static void __put_unused_fd(struct files_struct *files, unsigned int fd)
 {
 	struct fdtable *fdt = files_fdtable(files);
 	__FD_CLR(fd, fdt->open_fds);
-	if (fd < fdt->next_fd)
-		fdt->next_fd = fd;
+	if (fd < files->next_fd)
+		files->next_fd = fd;
 }
 
 void fastcall put_unused_fd(unsigned int fd)
diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c
index 1d24fead51a6..826c131994c3 100644
--- a/fs/proc/proc_misc.c
+++ b/fs/proc/proc_misc.c
@@ -312,7 +312,7 @@ static void *devinfo_next(struct seq_file *f, void *v, loff_t *pos)
 		case BLK_HDR:
 			info->state = BLK_LIST;
 			(*pos)++;
-			break;
+			/*fallthrough*/
 		case BLK_LIST:
 			if (get_blkdev_info(info->blkdev,&idummy,&ndummy)) {
 				/*
diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c
index b471315e24ef..c33963fded9e 100644
--- a/fs/qnx4/file.c
+++ b/fs/qnx4/file.c
@@ -12,10 +12,7 @@
  * 27-06-1998 by Frank Denis : file overwriting.
  */
 
-#include <linux/config.h>
-#include <linux/types.h>
 #include <linux/fs.h>
-#include <linux/time.h>
 #include <linux/qnx4_fs.h>
 
 /*
diff --git a/fs/quota.c b/fs/quota.c
index ba9e0bf32f67..d6a2be826e29 100644
--- a/fs/quota.c
+++ b/fs/quota.c
@@ -170,10 +170,10 @@ static void quota_sync_sb(struct super_block *sb, int type)
 
 	/* Now when everything is written we can discard the pagecache so
 	 * that userspace sees the changes. We need i_mutex and so we could
-	 * not do it inside dqonoff_sem. Moreover we need to be carefull
+	 * not do it inside dqonoff_mutex. Moreover we need to be carefull
 	 * about races with quotaoff() (that is the reason why we have own
 	 * reference to inode). */
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		discard[cnt] = NULL;
 		if (type != -1 && cnt != type)
@@ -182,7 +182,7 @@ static void quota_sync_sb(struct super_block *sb, int type)
 			continue;
 		discard[cnt] = igrab(sb_dqopt(sb)->files[cnt]);
 	}
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (discard[cnt]) {
 			mutex_lock(&discard[cnt]->i_mutex);
diff --git a/fs/quota_v2.c b/fs/quota_v2.c
index b4199ec3ece4..c519a583e681 100644
--- a/fs/quota_v2.c
+++ b/fs/quota_v2.c
@@ -394,7 +394,7 @@ static int v2_write_dquot(struct dquot *dquot)
 	ssize_t ret;
 	struct v2_disk_dqblk ddquot, empty;
 
-	/* dq_off is guarded by dqio_sem */
+	/* dq_off is guarded by dqio_mutex */
 	if (!dquot->dq_off)
 		if ((ret = dq_insert_tree(dquot)) < 0) {
 			printk(KERN_ERR "VFS: Error %zd occurred while creating quota.\n", ret);
diff --git a/fs/ramfs/file-mmu.c b/fs/ramfs/file-mmu.c
index 2115383dcc8d..6ada2095b9ac 100644
--- a/fs/ramfs/file-mmu.c
+++ b/fs/ramfs/file-mmu.c
@@ -24,18 +24,7 @@
  * caches is sufficient.
  */
 
-#include <linux/module.h>
 #include <linux/fs.h>
-#include <linux/pagemap.h>
-#include <linux/highmem.h>
-#include <linux/init.h>
-#include <linux/string.h>
-#include <linux/smp_lock.h>
-#include <linux/backing-dev.h>
-#include <linux/ramfs.h>
-
-#include <asm/uaccess.h>
-#include "internal.h"
 
 struct address_space_operations ramfs_aops = {
 	.readpage	= simple_readpage,
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index 3f810acd0bfa..b1ca234068f6 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -87,8 +87,7 @@ static int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize)
 	xpages = 1UL << order;
 	npages = (newsize + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
-	for (loop = 0; loop < npages; loop++)
-		set_page_count(pages + loop, 1);
+	split_page(pages, order);
 
 	/* trim off any pages we don't actually require */
 	for (loop = npages; loop < xpages; loop++)
diff --git a/fs/relayfs/Makefile b/fs/relayfs/Makefile
deleted file mode 100644
index e76e182cdb38..000000000000
--- a/fs/relayfs/Makefile
+++ /dev/null
@@ -1,4 +0,0 @@
-obj-$(CONFIG_RELAYFS_FS) += relayfs.o
-
-relayfs-y := relay.o inode.o buffers.o
-
diff --git a/fs/relayfs/buffers.c b/fs/relayfs/buffers.c
deleted file mode 100644
index 10187812771e..000000000000
--- a/fs/relayfs/buffers.c
+++ /dev/null
@@ -1,190 +0,0 @@
-/*
- * RelayFS buffer management code.
- *
- * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
- *
- * This file is released under the GPL.
- */
-
-#include <linux/module.h>
-#include <linux/vmalloc.h>
-#include <linux/mm.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-/*
- * close() vm_op implementation for relayfs file mapping.
- */
-static void relay_file_mmap_close(struct vm_area_struct *vma)
-{
-	struct rchan_buf *buf = vma->vm_private_data;
-	buf->chan->cb->buf_unmapped(buf, vma->vm_file);
-}
-
-/*
- * nopage() vm_op implementation for relayfs file mapping.
- */
-static struct page *relay_buf_nopage(struct vm_area_struct *vma,
-				     unsigned long address,
-				     int *type)
-{
-	struct page *page;
-	struct rchan_buf *buf = vma->vm_private_data;
-	unsigned long offset = address - vma->vm_start;
-
-	if (address > vma->vm_end)
-		return NOPAGE_SIGBUS; /* Disallow mremap */
-	if (!buf)
-		return NOPAGE_OOM;
-
-	page = vmalloc_to_page(buf->start + offset);
-	if (!page)
-		return NOPAGE_OOM;
-	get_page(page);
-
-	if (type)
-		*type = VM_FAULT_MINOR;
-
-	return page;
-}
-
-/*
- * vm_ops for relay file mappings.
- */
-static struct vm_operations_struct relay_file_mmap_ops = {
-	.nopage = relay_buf_nopage,
-	.close = relay_file_mmap_close,
-};
-
-/**
- *	relay_mmap_buf: - mmap channel buffer to process address space
- *	@buf: relay channel buffer
- *	@vma: vm_area_struct describing memory to be mapped
- *
- *	Returns 0 if ok, negative on error
- *
- *	Caller should already have grabbed mmap_sem.
- */
-int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma)
-{
-	unsigned long length = vma->vm_end - vma->vm_start;
-	struct file *filp = vma->vm_file;
-
-	if (!buf)
-		return -EBADF;
-
-	if (length != (unsigned long)buf->chan->alloc_size)
-		return -EINVAL;
-
-	vma->vm_ops = &relay_file_mmap_ops;
-	vma->vm_private_data = buf;
-	buf->chan->cb->buf_mapped(buf, filp);
-
-	return 0;
-}
-
-/**
- *	relay_alloc_buf - allocate a channel buffer
- *	@buf: the buffer struct
- *	@size: total size of the buffer
- *
- *	Returns a pointer to the resulting buffer, NULL if unsuccessful
- */
-static void *relay_alloc_buf(struct rchan_buf *buf, unsigned long size)
-{
-	void *mem;
-	unsigned int i, j, n_pages;
-
-	size = PAGE_ALIGN(size);
-	n_pages = size >> PAGE_SHIFT;
-
-	buf->page_array = kcalloc(n_pages, sizeof(struct page *), GFP_KERNEL);
-	if (!buf->page_array)
-		return NULL;
-
-	for (i = 0; i < n_pages; i++) {
-		buf->page_array[i] = alloc_page(GFP_KERNEL);
-		if (unlikely(!buf->page_array[i]))
-			goto depopulate;
-	}
-	mem = vmap(buf->page_array, n_pages, VM_MAP, PAGE_KERNEL);
-	if (!mem)
-		goto depopulate;
-
-	memset(mem, 0, size);
-	buf->page_count = n_pages;
-	return mem;
-
-depopulate:
-	for (j = 0; j < i; j++)
-		__free_page(buf->page_array[j]);
-	kfree(buf->page_array);
-	return NULL;
-}
-
-/**
- *	relay_create_buf - allocate and initialize a channel buffer
- *	@alloc_size: size of the buffer to allocate
- *	@n_subbufs: number of sub-buffers in the channel
- *
- *	Returns channel buffer if successful, NULL otherwise
- */
-struct rchan_buf *relay_create_buf(struct rchan *chan)
-{
-	struct rchan_buf *buf = kcalloc(1, sizeof(struct rchan_buf), GFP_KERNEL);
-	if (!buf)
-		return NULL;
-
-	buf->padding = kmalloc(chan->n_subbufs * sizeof(size_t *), GFP_KERNEL);
-	if (!buf->padding)
-		goto free_buf;
-
-	buf->start = relay_alloc_buf(buf, chan->alloc_size);
-	if (!buf->start)
-		goto free_buf;
-
-	buf->chan = chan;
-	kref_get(&buf->chan->kref);
-	return buf;
-
-free_buf:
-	kfree(buf->padding);
-	kfree(buf);
-	return NULL;
-}
-
-/**
- *	relay_destroy_buf - destroy an rchan_buf struct and associated buffer
- *	@buf: the buffer struct
- */
-void relay_destroy_buf(struct rchan_buf *buf)
-{
-	struct rchan *chan = buf->chan;
-	unsigned int i;
-
-	if (likely(buf->start)) {
-		vunmap(buf->start);
-		for (i = 0; i < buf->page_count; i++)
-			__free_page(buf->page_array[i]);
-		kfree(buf->page_array);
-	}
-	kfree(buf->padding);
-	kfree(buf);
-	kref_put(&chan->kref, relay_destroy_channel);
-}
-
-/**
- *	relay_remove_buf - remove a channel buffer
- *
- *	Removes the file from the relayfs fileystem, which also frees the
- *	rchan_buf_struct and the channel buffer.  Should only be called from
- *	kref_put().
- */
-void relay_remove_buf(struct kref *kref)
-{
-	struct rchan_buf *buf = container_of(kref, struct rchan_buf, kref);
-	buf->chan->cb->remove_buf_file(buf->dentry);
-	relay_destroy_buf(buf);
-}
diff --git a/fs/relayfs/buffers.h b/fs/relayfs/buffers.h
deleted file mode 100644
index 37a12493f641..000000000000
--- a/fs/relayfs/buffers.h
+++ /dev/null
@@ -1,12 +0,0 @@
-#ifndef _BUFFERS_H
-#define _BUFFERS_H
-
-/* This inspired by rtai/shmem */
-#define FIX_SIZE(x) (((x) - 1) & PAGE_MASK) + PAGE_SIZE
-
-extern int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma);
-extern struct rchan_buf *relay_create_buf(struct rchan *chan);
-extern void relay_destroy_buf(struct rchan_buf *buf);
-extern void relay_remove_buf(struct kref *kref);
-
-#endif/* _BUFFERS_H */
diff --git a/fs/relayfs/inode.c b/fs/relayfs/inode.c
deleted file mode 100644
index 383523011aad..000000000000
--- a/fs/relayfs/inode.c
+++ /dev/null
@@ -1,581 +0,0 @@
-/*
- * VFS-related code for RelayFS, a high-speed data relay filesystem.
- *
- * Copyright (C) 2003-2005 - Tom Zanussi <zanussi@us.ibm.com>, IBM Corp
- * Copyright (C) 2003-2005 - Karim Yaghmour <karim@opersys.com>
- *
- * Based on ramfs, Copyright (C) 2002 - Linus Torvalds
- *
- * This file is released under the GPL.
- */
-
-#include <linux/module.h>
-#include <linux/fs.h>
-#include <linux/mount.h>
-#include <linux/pagemap.h>
-#include <linux/init.h>
-#include <linux/string.h>
-#include <linux/backing-dev.h>
-#include <linux/namei.h>
-#include <linux/poll.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-#define RELAYFS_MAGIC			0xF0B4A981
-
-static struct vfsmount *		relayfs_mount;
-static int				relayfs_mount_count;
-
-static struct backing_dev_info		relayfs_backing_dev_info = {
-	.ra_pages	= 0,	/* No readahead */
-	.capabilities	= BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK,
-};
-
-static struct inode *relayfs_get_inode(struct super_block *sb,
-				       int mode,
- 				       struct file_operations *fops,
-				       void *data)
-{
-	struct inode *inode;
-
-	inode = new_inode(sb);
-	if (!inode)
-		return NULL;
-
-	inode->i_mode = mode;
-	inode->i_uid = 0;
-	inode->i_gid = 0;
-	inode->i_blksize = PAGE_CACHE_SIZE;
-	inode->i_blocks = 0;
-	inode->i_mapping->backing_dev_info = &relayfs_backing_dev_info;
-	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-	switch (mode & S_IFMT) {
-	case S_IFREG:
-		inode->i_fop = fops;
-		if (data)
-			inode->u.generic_ip = data;
-		break;
-	case S_IFDIR:
-		inode->i_op = &simple_dir_inode_operations;
-		inode->i_fop = &simple_dir_operations;
-
-		/* directory inodes start off with i_nlink == 2 (for "." entry) */
-		inode->i_nlink++;
-		break;
-	default:
-		break;
-	}
-
-	return inode;
-}
-
-/**
- *	relayfs_create_entry - create a relayfs directory or file
- *	@name: the name of the file to create
- *	@parent: parent directory
- *	@mode: mode
- *	@fops: file operations to use for the file
- *	@data: user-associated data for this file
- *
- *	Returns the new dentry, NULL on failure
- *
- *	Creates a file or directory with the specifed permissions.
- */
-static struct dentry *relayfs_create_entry(const char *name,
-					   struct dentry *parent,
-					   int mode,
-					   struct file_operations *fops,
-					   void *data)
-{
-	struct dentry *d;
-	struct inode *inode;
-	int error = 0;
-
-	BUG_ON(!name || !(S_ISREG(mode) || S_ISDIR(mode)));
-
-	error = simple_pin_fs("relayfs", &relayfs_mount, &relayfs_mount_count);
-	if (error) {
-		printk(KERN_ERR "Couldn't mount relayfs: errcode %d\n", error);
-		return NULL;
-	}
-
-	if (!parent && relayfs_mount && relayfs_mount->mnt_sb)
-		parent = relayfs_mount->mnt_sb->s_root;
-
-	if (!parent) {
-		simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-		return NULL;
-	}
-
-	parent = dget(parent);
-	mutex_lock(&parent->d_inode->i_mutex);
-	d = lookup_one_len(name, parent, strlen(name));
-	if (IS_ERR(d)) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	if (d->d_inode) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	inode = relayfs_get_inode(parent->d_inode->i_sb, mode, fops, data);
-	if (!inode) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	d_instantiate(d, inode);
-	dget(d);	/* Extra count - pin the dentry in core */
-
-	if (S_ISDIR(mode))
-		parent->d_inode->i_nlink++;
-
-	goto exit;
-
-release_mount:
-	simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-
-exit:
-	mutex_unlock(&parent->d_inode->i_mutex);
-	dput(parent);
-	return d;
-}
-
-/**
- *	relayfs_create_file - create a file in the relay filesystem
- *	@name: the name of the file to create
- *	@parent: parent directory
- *	@mode: mode, if not specied the default perms are used
- *	@fops: file operations to use for the file
- *	@data: user-associated data for this file
- *
- *	Returns file dentry if successful, NULL otherwise.
- *
- *	The file will be created user r on behalf of current user.
- */
-struct dentry *relayfs_create_file(const char *name,
-				   struct dentry *parent,
-				   int mode,
-				   struct file_operations *fops,
-				   void *data)
-{
-	BUG_ON(!fops);
-
-	if (!mode)
-		mode = S_IRUSR;
-	mode = (mode & S_IALLUGO) | S_IFREG;
-
-	return relayfs_create_entry(name, parent, mode, fops, data);
-}
-
-/**
- *	relayfs_create_dir - create a directory in the relay filesystem
- *	@name: the name of the directory to create
- *	@parent: parent directory, NULL if parent should be fs root
- *
- *	Returns directory dentry if successful, NULL otherwise.
- *
- *	The directory will be created world rwx on behalf of current user.
- */
-struct dentry *relayfs_create_dir(const char *name, struct dentry *parent)
-{
-	int mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
-	return relayfs_create_entry(name, parent, mode, NULL, NULL);
-}
-
-/**
- *	relayfs_remove - remove a file or directory in the relay filesystem
- *	@dentry: file or directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove(struct dentry *dentry)
-{
-	struct dentry *parent;
-	int error = 0;
-
-	if (!dentry)
-		return -EINVAL;
-	parent = dentry->d_parent;
-	if (!parent)
-		return -EINVAL;
-
-	parent = dget(parent);
-	mutex_lock(&parent->d_inode->i_mutex);
-	if (dentry->d_inode) {
-		if (S_ISDIR(dentry->d_inode->i_mode))
-			error = simple_rmdir(parent->d_inode, dentry);
-		else
-			error = simple_unlink(parent->d_inode, dentry);
-		if (!error)
-			d_delete(dentry);
-	}
-	if (!error)
-		dput(dentry);
-	mutex_unlock(&parent->d_inode->i_mutex);
-	dput(parent);
-
-	if (!error)
-		simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-
-	return error;
-}
-
-/**
- *	relayfs_remove_file - remove a file from relay filesystem
- *	@dentry: directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove_file(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/**
- *	relayfs_remove_dir - remove a directory in the relay filesystem
- *	@dentry: directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove_dir(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/**
- *	relay_file_open - open file op for relay files
- *	@inode: the inode
- *	@filp: the file
- *
- *	Increments the channel buffer refcount.
- */
-static int relay_file_open(struct inode *inode, struct file *filp)
-{
-	struct rchan_buf *buf = inode->u.generic_ip;
-	kref_get(&buf->kref);
-	filp->private_data = buf;
-
-	return 0;
-}
-
-/**
- *	relay_file_mmap - mmap file op for relay files
- *	@filp: the file
- *	@vma: the vma describing what to map
- *
- *	Calls upon relay_mmap_buf to map the file into user space.
- */
-static int relay_file_mmap(struct file *filp, struct vm_area_struct *vma)
-{
-	struct rchan_buf *buf = filp->private_data;
-	return relay_mmap_buf(buf, vma);
-}
-
-/**
- *	relay_file_poll - poll file op for relay files
- *	@filp: the file
- *	@wait: poll table
- *
- *	Poll implemention.
- */
-static unsigned int relay_file_poll(struct file *filp, poll_table *wait)
-{
-	unsigned int mask = 0;
-	struct rchan_buf *buf = filp->private_data;
-
-	if (buf->finalized)
-		return POLLERR;
-
-	if (filp->f_mode & FMODE_READ) {
-		poll_wait(filp, &buf->read_wait, wait);
-		if (!relay_buf_empty(buf))
-			mask |= POLLIN | POLLRDNORM;
-	}
-
-	return mask;
-}
-
-/**
- *	relay_file_release - release file op for relay files
- *	@inode: the inode
- *	@filp: the file
- *
- *	Decrements the channel refcount, as the filesystem is
- *	no longer using it.
- */
-static int relay_file_release(struct inode *inode, struct file *filp)
-{
-	struct rchan_buf *buf = filp->private_data;
-	kref_put(&buf->kref, relay_remove_buf);
-
-	return 0;
-}
-
-/**
- *	relay_file_read_consume - update the consumed count for the buffer
- */
-static void relay_file_read_consume(struct rchan_buf *buf,
-				    size_t read_pos,
-				    size_t bytes_consumed)
-{
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-	size_t read_subbuf;
-
-	if (buf->bytes_consumed + bytes_consumed > subbuf_size) {
-		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-		buf->bytes_consumed = 0;
-	}
-
-	buf->bytes_consumed += bytes_consumed;
-	read_subbuf = read_pos / buf->chan->subbuf_size;
-	if (buf->bytes_consumed + buf->padding[read_subbuf] == subbuf_size) {
-		if ((read_subbuf == buf->subbufs_produced % n_subbufs) &&
-		    (buf->offset == subbuf_size))
-			return;
-		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-		buf->bytes_consumed = 0;
-	}
-}
-
-/**
- *	relay_file_read_avail - boolean, are there unconsumed bytes available?
- */
-static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos)
-{
-	size_t bytes_produced, bytes_consumed, write_offset;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-	size_t produced = buf->subbufs_produced % n_subbufs;
-	size_t consumed = buf->subbufs_consumed % n_subbufs;
-
-	write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-
-	if (consumed > produced) {
-		if ((produced > n_subbufs) &&
-		    (produced + n_subbufs - consumed <= n_subbufs))
-			produced += n_subbufs;
-	} else if (consumed == produced) {
-		if (buf->offset > subbuf_size) {
-			produced += n_subbufs;
-			if (buf->subbufs_produced == buf->subbufs_consumed)
-				consumed += n_subbufs;
-		}
-	}
-
-	if (buf->offset > subbuf_size)
-		bytes_produced = (produced - 1) * subbuf_size + write_offset;
-	else
-		bytes_produced = produced * subbuf_size + write_offset;
-	bytes_consumed = consumed * subbuf_size + buf->bytes_consumed;
-
-	if (bytes_produced == bytes_consumed)
-		return 0;
-
-	relay_file_read_consume(buf, read_pos, 0);
-
-	return 1;
-}
-
-/**
- *	relay_file_read_subbuf_avail - return bytes available in sub-buffer
- */
-static size_t relay_file_read_subbuf_avail(size_t read_pos,
-					   struct rchan_buf *buf)
-{
-	size_t padding, avail = 0;
-	size_t read_subbuf, read_offset, write_subbuf, write_offset;
-	size_t subbuf_size = buf->chan->subbuf_size;
-
-	write_subbuf = (buf->data - buf->start) / subbuf_size;
-	write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-	read_subbuf = read_pos / subbuf_size;
-	read_offset = read_pos % subbuf_size;
-	padding = buf->padding[read_subbuf];
-
-	if (read_subbuf == write_subbuf) {
-		if (read_offset + padding < write_offset)
-			avail = write_offset - (read_offset + padding);
-	} else
-		avail = (subbuf_size - padding) - read_offset;
-
-	return avail;
-}
-
-/**
- *	relay_file_read_start_pos - find the first available byte to read
- *
- *	If the read_pos is in the middle of padding, return the
- *	position of the first actually available byte, otherwise
- *	return the original value.
- */
-static size_t relay_file_read_start_pos(size_t read_pos,
-					struct rchan_buf *buf)
-{
-	size_t read_subbuf, padding, padding_start, padding_end;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-
-	read_subbuf = read_pos / subbuf_size;
-	padding = buf->padding[read_subbuf];
-	padding_start = (read_subbuf + 1) * subbuf_size - padding;
-	padding_end = (read_subbuf + 1) * subbuf_size;
-	if (read_pos >= padding_start && read_pos < padding_end) {
-		read_subbuf = (read_subbuf + 1) % n_subbufs;
-		read_pos = read_subbuf * subbuf_size;
-	}
-
-	return read_pos;
-}
-
-/**
- *	relay_file_read_end_pos - return the new read position
- */
-static size_t relay_file_read_end_pos(struct rchan_buf *buf,
-				      size_t read_pos,
-				      size_t count)
-{
-	size_t read_subbuf, padding, end_pos;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-
-	read_subbuf = read_pos / subbuf_size;
-	padding = buf->padding[read_subbuf];
-	if (read_pos % subbuf_size + count + padding == subbuf_size)
-		end_pos = (read_subbuf + 1) * subbuf_size;
-	else
-		end_pos = read_pos + count;
-	if (end_pos >= subbuf_size * n_subbufs)
-		end_pos = 0;
-
-	return end_pos;
-}
-
-/**
- *	relay_file_read - read file op for relay files
- *	@filp: the file
- *	@buffer: the userspace buffer
- *	@count: number of bytes to read
- *	@ppos: position to read from
- *
- *	Reads count bytes or the number of bytes available in the
- *	current sub-buffer being read, whichever is smaller.
- */
-static ssize_t relay_file_read(struct file *filp,
-			       char __user *buffer,
-			       size_t count,
-			       loff_t *ppos)
-{
-	struct rchan_buf *buf = filp->private_data;
-	struct inode *inode = filp->f_dentry->d_inode;
-	size_t read_start, avail;
-	ssize_t ret = 0;
-	void *from;
-
-	mutex_lock(&inode->i_mutex);
-	if(!relay_file_read_avail(buf, *ppos))
-		goto out;
-
-	read_start = relay_file_read_start_pos(*ppos, buf);
-	avail = relay_file_read_subbuf_avail(read_start, buf);
-	if (!avail)
-		goto out;
-
-	from = buf->start + read_start;
-	ret = count = min(count, avail);
-	if (copy_to_user(buffer, from, count)) {
-		ret = -EFAULT;
-		goto out;
-	}
-	relay_file_read_consume(buf, read_start, count);
-	*ppos = relay_file_read_end_pos(buf, read_start, count);
-out:
-	mutex_unlock(&inode->i_mutex);
-	return ret;
-}
-
-struct file_operations relay_file_operations = {
-	.open		= relay_file_open,
-	.poll		= relay_file_poll,
-	.mmap		= relay_file_mmap,
-	.read		= relay_file_read,
-	.llseek		= no_llseek,
-	.release	= relay_file_release,
-};
-
-static struct super_operations relayfs_ops = {
-	.statfs		= simple_statfs,
-	.drop_inode	= generic_delete_inode,
-};
-
-static int relayfs_fill_super(struct super_block * sb, void * data, int silent)
-{
-	struct inode *inode;
-	struct dentry *root;
-	int mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
-
-	sb->s_blocksize = PAGE_CACHE_SIZE;
-	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
-	sb->s_magic = RELAYFS_MAGIC;
-	sb->s_op = &relayfs_ops;
-	inode = relayfs_get_inode(sb, mode, NULL, NULL);
-
-	if (!inode)
-		return -ENOMEM;
-
-	root = d_alloc_root(inode);
-	if (!root) {
-		iput(inode);
-		return -ENOMEM;
-	}
-	sb->s_root = root;
-
-	return 0;
-}
-
-static struct super_block * relayfs_get_sb(struct file_system_type *fs_type,
-					   int flags, const char *dev_name,
-					   void *data)
-{
-	return get_sb_single(fs_type, flags, data, relayfs_fill_super);
-}
-
-static struct file_system_type relayfs_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "relayfs",
-	.get_sb		= relayfs_get_sb,
-	.kill_sb	= kill_litter_super,
-};
-
-static int __init init_relayfs_fs(void)
-{
-	return register_filesystem(&relayfs_fs_type);
-}
-
-static void __exit exit_relayfs_fs(void)
-{
-
-
-
-
-
-	unregister_filesystem(&relayfs_fs_type);
-}
-
-module_init(init_relayfs_fs)
-module_exit(exit_relayfs_fs)
-
-EXPORT_SYMBOL_GPL(relay_file_operations);
-EXPORT_SYMBOL_GPL(relayfs_create_dir);
-EXPORT_SYMBOL_GPL(relayfs_remove_dir);
-EXPORT_SYMBOL_GPL(relayfs_create_file);
-EXPORT_SYMBOL_GPL(relayfs_remove_file);
-
-MODULE_AUTHOR("Tom Zanussi <zanussi@us.ibm.com> and Karim Yaghmour <karim@opersys.com>");
-MODULE_DESCRIPTION("Relay Filesystem");
-MODULE_LICENSE("GPL");
-
diff --git a/fs/relayfs/relay.c b/fs/relayfs/relay.c
deleted file mode 100644
index abf3ceaace49..000000000000
--- a/fs/relayfs/relay.c
+++ /dev/null
@@ -1,482 +0,0 @@
-/*
- * Public API and common code for RelayFS.
- *
- * See Documentation/filesystems/relayfs.txt for an overview of relayfs.
- *
- * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
- *
- * This file is released under the GPL.
- */
-
-#include <linux/errno.h>
-#include <linux/stddef.h>
-#include <linux/slab.h>
-#include <linux/module.h>
-#include <linux/string.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-/**
- *	relay_buf_empty - boolean, is the channel buffer empty?
- *	@buf: channel buffer
- *
- *	Returns 1 if the buffer is empty, 0 otherwise.
- */
-int relay_buf_empty(struct rchan_buf *buf)
-{
-	return (buf->subbufs_produced - buf->subbufs_consumed) ? 0 : 1;
-}
-
-/**
- *	relay_buf_full - boolean, is the channel buffer full?
- *	@buf: channel buffer
- *
- *	Returns 1 if the buffer is full, 0 otherwise.
- */
-int relay_buf_full(struct rchan_buf *buf)
-{
-	size_t ready = buf->subbufs_produced - buf->subbufs_consumed;
-	return (ready >= buf->chan->n_subbufs) ? 1 : 0;
-}
-
-/*
- * High-level relayfs kernel API and associated functions.
- */
-
-/*
- * rchan_callback implementations defining default channel behavior.  Used
- * in place of corresponding NULL values in client callback struct.
- */
-
-/*
- * subbuf_start() default callback.  Does nothing.
- */
-static int subbuf_start_default_callback (struct rchan_buf *buf,
-					  void *subbuf,
-					  void *prev_subbuf,
-					  size_t prev_padding)
-{
-	if (relay_buf_full(buf))
-		return 0;
-
-	return 1;
-}
-
-/*
- * buf_mapped() default callback.  Does nothing.
- */
-static void buf_mapped_default_callback(struct rchan_buf *buf,
-					struct file *filp)
-{
-}
-
-/*
- * buf_unmapped() default callback.  Does nothing.
- */
-static void buf_unmapped_default_callback(struct rchan_buf *buf,
-					  struct file *filp)
-{
-}
-
-/*
- * create_buf_file_create() default callback.  Creates file to represent buf.
- */
-static struct dentry *create_buf_file_default_callback(const char *filename,
-						       struct dentry *parent,
-						       int mode,
-						       struct rchan_buf *buf,
-						       int *is_global)
-{
-	return relayfs_create_file(filename, parent, mode,
-				   &relay_file_operations, buf);
-}
-
-/*
- * remove_buf_file() default callback.  Removes file representing relay buffer.
- */
-static int remove_buf_file_default_callback(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/* relay channel default callbacks */
-static struct rchan_callbacks default_channel_callbacks = {
-	.subbuf_start = subbuf_start_default_callback,
-	.buf_mapped = buf_mapped_default_callback,
-	.buf_unmapped = buf_unmapped_default_callback,
-	.create_buf_file = create_buf_file_default_callback,
-	.remove_buf_file = remove_buf_file_default_callback,
-};
-
-/**
- *	wakeup_readers - wake up readers waiting on a channel
- *	@private: the channel buffer
- *
- *	This is the work function used to defer reader waking.  The
- *	reason waking is deferred is that calling directly from write
- *	causes problems if you're writing from say the scheduler.
- */
-static void wakeup_readers(void *private)
-{
-	struct rchan_buf *buf = private;
-	wake_up_interruptible(&buf->read_wait);
-}
-
-/**
- *	__relay_reset - reset a channel buffer
- *	@buf: the channel buffer
- *	@init: 1 if this is a first-time initialization
- *
- *	See relay_reset for description of effect.
- */
-static inline void __relay_reset(struct rchan_buf *buf, unsigned int init)
-{
-	size_t i;
-
-	if (init) {
-		init_waitqueue_head(&buf->read_wait);
-		kref_init(&buf->kref);
-		INIT_WORK(&buf->wake_readers, NULL, NULL);
-	} else {
-		cancel_delayed_work(&buf->wake_readers);
-		flush_scheduled_work();
-	}
-
-	buf->subbufs_produced = 0;
-	buf->subbufs_consumed = 0;
-	buf->bytes_consumed = 0;
-	buf->finalized = 0;
-	buf->data = buf->start;
-	buf->offset = 0;
-
-	for (i = 0; i < buf->chan->n_subbufs; i++)
-		buf->padding[i] = 0;
-
-	buf->chan->cb->subbuf_start(buf, buf->data, NULL, 0);
-}
-
-/**
- *	relay_reset - reset the channel
- *	@chan: the channel
- *
- *	This has the effect of erasing all data from all channel buffers
- *	and restarting the channel in its initial state.  The buffers
- *	are not freed, so any mappings are still in effect.
- *
- *	NOTE: Care should be taken that the channel isn't actually
- *	being used by anything when this call is made.
- */
-void relay_reset(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		__relay_reset(chan->buf[i], 0);
-		prev = chan->buf[i];
-	}
-}
-
-/**
- *	relay_open_buf - create a new channel buffer in relayfs
- *
- *	Internal - used by relay_open().
- */
-static struct rchan_buf *relay_open_buf(struct rchan *chan,
-					const char *filename,
-					struct dentry *parent,
-					int *is_global)
-{
-	struct rchan_buf *buf;
-	struct dentry *dentry;
-
-	if (*is_global)
-		return chan->buf[0];
-
- 	buf = relay_create_buf(chan);
- 	if (!buf)
- 		return NULL;
-
-	/* Create file in fs */
- 	dentry = chan->cb->create_buf_file(filename, parent, S_IRUSR,
- 					   buf, is_global);
- 	if (!dentry) {
- 		relay_destroy_buf(buf);
-		return NULL;
- 	}
-
-	buf->dentry = dentry;
-	__relay_reset(buf, 1);
-
-	return buf;
-}
-
-/**
- *	relay_close_buf - close a channel buffer
- *	@buf: channel buffer
- *
- *	Marks the buffer finalized and restores the default callbacks.
- *	The channel buffer and channel buffer data structure are then freed
- *	automatically when the last reference is given up.
- */
-static inline void relay_close_buf(struct rchan_buf *buf)
-{
-	buf->finalized = 1;
-	buf->chan->cb = &default_channel_callbacks;
-	cancel_delayed_work(&buf->wake_readers);
-	flush_scheduled_work();
-	kref_put(&buf->kref, relay_remove_buf);
-}
-
-static inline void setup_callbacks(struct rchan *chan,
-				   struct rchan_callbacks *cb)
-{
-	if (!cb) {
-		chan->cb = &default_channel_callbacks;
-		return;
-	}
-
-	if (!cb->subbuf_start)
-		cb->subbuf_start = subbuf_start_default_callback;
-	if (!cb->buf_mapped)
-		cb->buf_mapped = buf_mapped_default_callback;
-	if (!cb->buf_unmapped)
-		cb->buf_unmapped = buf_unmapped_default_callback;
-	if (!cb->create_buf_file)
-		cb->create_buf_file = create_buf_file_default_callback;
-	if (!cb->remove_buf_file)
-		cb->remove_buf_file = remove_buf_file_default_callback;
-	chan->cb = cb;
-}
-
-/**
- *	relay_open - create a new relayfs channel
- *	@base_filename: base name of files to create
- *	@parent: dentry of parent directory, NULL for root directory
- *	@subbuf_size: size of sub-buffers
- *	@n_subbufs: number of sub-buffers
- *	@cb: client callback functions
- *
- *	Returns channel pointer if successful, NULL otherwise.
- *
- *	Creates a channel buffer for each cpu using the sizes and
- *	attributes specified.  The created channel buffer files
- *	will be named base_filename0...base_filenameN-1.  File
- *	permissions will be S_IRUSR.
- */
-struct rchan *relay_open(const char *base_filename,
-			 struct dentry *parent,
-			 size_t subbuf_size,
-			 size_t n_subbufs,
-			 struct rchan_callbacks *cb)
-{
-	unsigned int i;
-	struct rchan *chan;
-	char *tmpname;
-	int is_global = 0;
-
-	if (!base_filename)
-		return NULL;
-
-	if (!(subbuf_size && n_subbufs))
-		return NULL;
-
-	chan = kcalloc(1, sizeof(struct rchan), GFP_KERNEL);
-	if (!chan)
-		return NULL;
-
-	chan->version = RELAYFS_CHANNEL_VERSION;
-	chan->n_subbufs = n_subbufs;
-	chan->subbuf_size = subbuf_size;
-	chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs);
-	setup_callbacks(chan, cb);
-	kref_init(&chan->kref);
-
-	tmpname = kmalloc(NAME_MAX + 1, GFP_KERNEL);
-	if (!tmpname)
-		goto free_chan;
-
-	for_each_online_cpu(i) {
-		sprintf(tmpname, "%s%d", base_filename, i);
-		chan->buf[i] = relay_open_buf(chan, tmpname, parent,
-					      &is_global);
-		chan->buf[i]->cpu = i;
-		if (!chan->buf[i])
-			goto free_bufs;
-	}
-
-	kfree(tmpname);
-	return chan;
-
-free_bufs:
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i])
-			break;
-		relay_close_buf(chan->buf[i]);
-		if (is_global)
-			break;
-	}
-	kfree(tmpname);
-
-free_chan:
-	kref_put(&chan->kref, relay_destroy_channel);
-	return NULL;
-}
-
-/**
- *	relay_switch_subbuf - switch to a new sub-buffer
- *	@buf: channel buffer
- *	@length: size of current event
- *
- *	Returns either the length passed in or 0 if full.
-
- *	Performs sub-buffer-switch tasks such as invoking callbacks,
- *	updating padding counts, waking up readers, etc.
- */
-size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
-{
-	void *old, *new;
-	size_t old_subbuf, new_subbuf;
-
-	if (unlikely(length > buf->chan->subbuf_size))
-		goto toobig;
-
-	if (buf->offset != buf->chan->subbuf_size + 1) {
-		buf->prev_padding = buf->chan->subbuf_size - buf->offset;
-		old_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
-		buf->padding[old_subbuf] = buf->prev_padding;
-		buf->subbufs_produced++;
-		if (waitqueue_active(&buf->read_wait)) {
-			PREPARE_WORK(&buf->wake_readers, wakeup_readers, buf);
-			schedule_delayed_work(&buf->wake_readers, 1);
-		}
-	}
-
-	old = buf->data;
-	new_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
-	new = buf->start + new_subbuf * buf->chan->subbuf_size;
-	buf->offset = 0;
-	if (!buf->chan->cb->subbuf_start(buf, new, old, buf->prev_padding)) {
-		buf->offset = buf->chan->subbuf_size + 1;
-		return 0;
-	}
-	buf->data = new;
-	buf->padding[new_subbuf] = 0;
-
-	if (unlikely(length + buf->offset > buf->chan->subbuf_size))
-		goto toobig;
-
-	return length;
-
-toobig:
-	buf->chan->last_toobig = length;
-	return 0;
-}
-
-/**
- *	relay_subbufs_consumed - update the buffer's sub-buffers-consumed count
- *	@chan: the channel
- *	@cpu: the cpu associated with the channel buffer to update
- *	@subbufs_consumed: number of sub-buffers to add to current buf's count
- *
- *	Adds to the channel buffer's consumed sub-buffer count.
- *	subbufs_consumed should be the number of sub-buffers newly consumed,
- *	not the total consumed.
- *
- *	NOTE: kernel clients don't need to call this function if the channel
- *	mode is 'overwrite'.
- */
-void relay_subbufs_consumed(struct rchan *chan,
-			    unsigned int cpu,
-			    size_t subbufs_consumed)
-{
-	struct rchan_buf *buf;
-
-	if (!chan)
-		return;
-
-	if (cpu >= NR_CPUS || !chan->buf[cpu])
-		return;
-
-	buf = chan->buf[cpu];
-	buf->subbufs_consumed += subbufs_consumed;
-	if (buf->subbufs_consumed > buf->subbufs_produced)
-		buf->subbufs_consumed = buf->subbufs_produced;
-}
-
-/**
- *	relay_destroy_channel - free the channel struct
- *
- *	Should only be called from kref_put().
- */
-void relay_destroy_channel(struct kref *kref)
-{
-	struct rchan *chan = container_of(kref, struct rchan, kref);
-	kfree(chan);
-}
-
-/**
- *	relay_close - close the channel
- *	@chan: the channel
- *
- *	Closes all channel buffers and frees the channel.
- */
-void relay_close(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_close_buf(chan->buf[i]);
-		prev = chan->buf[i];
-	}
-
-	if (chan->last_toobig)
-		printk(KERN_WARNING "relayfs: one or more items not logged "
-		       "[item size (%Zd) > sub-buffer size (%Zd)]\n",
-		       chan->last_toobig, chan->subbuf_size);
-
-	kref_put(&chan->kref, relay_destroy_channel);
-}
-
-/**
- *	relay_flush - close the channel
- *	@chan: the channel
- *
- *	Flushes all channel buffers i.e. forces buffer switch.
- */
-void relay_flush(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_switch_subbuf(chan->buf[i], 0);
-		prev = chan->buf[i];
-	}
-}
-
-EXPORT_SYMBOL_GPL(relay_open);
-EXPORT_SYMBOL_GPL(relay_close);
-EXPORT_SYMBOL_GPL(relay_flush);
-EXPORT_SYMBOL_GPL(relay_reset);
-EXPORT_SYMBOL_GPL(relay_subbufs_consumed);
-EXPORT_SYMBOL_GPL(relay_switch_subbuf);
-EXPORT_SYMBOL_GPL(relay_buf_full);
diff --git a/fs/relayfs/relay.h b/fs/relayfs/relay.h
deleted file mode 100644
index 0993d3e5753b..000000000000
--- a/fs/relayfs/relay.h
+++ /dev/null
@@ -1,8 +0,0 @@
-#ifndef _RELAY_H
-#define _RELAY_H
-
-extern int relayfs_remove(struct dentry *dentry);
-extern int relay_buf_empty(struct rchan_buf *buf);
-extern void relay_destroy_channel(struct kref *kref);
-
-#endif /* _RELAY_H */
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 7c40570b71dc..555b9ac04c25 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -37,7 +37,7 @@ int seq_open(struct file *file, struct seq_operations *op)
 		file->private_data = p;
 	}
 	memset(p, 0, sizeof(*p));
-	sema_init(&p->sem, 1);
+	mutex_init(&p->lock);
 	p->op = op;
 
 	/*
@@ -71,7 +71,7 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
 	void *p;
 	int err = 0;
 
-	down(&m->sem);
+	mutex_lock(&m->lock);
 	/*
 	 * seq_file->op->..m_start/m_stop/m_next may do special actions
 	 * or optimisations based on the file->f_version, so we want to
@@ -164,7 +164,7 @@ Done:
 	else
 		*ppos += copied;
 	file->f_version = m->version;
-	up(&m->sem);
+	mutex_unlock(&m->lock);
 	return copied;
 Enomem:
 	err = -ENOMEM;
@@ -237,7 +237,7 @@ loff_t seq_lseek(struct file *file, loff_t offset, int origin)
 	struct seq_file *m = (struct seq_file *)file->private_data;
 	long long retval = -EINVAL;
 
-	down(&m->sem);
+	mutex_lock(&m->lock);
 	m->version = file->f_version;
 	switch (origin) {
 		case 1:
@@ -260,7 +260,7 @@ loff_t seq_lseek(struct file *file, loff_t offset, int origin)
 				}
 			}
 	}
-	up(&m->sem);
+	mutex_unlock(&m->lock);
 	file->f_version = m->version;
 	return retval;
 }
diff --git a/fs/super.c b/fs/super.c
index e20b5580afd5..425861cb1caa 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -76,9 +76,9 @@ static struct super_block *alloc_super(void)
 		down_write(&s->s_umount);
 		s->s_count = S_BIAS;
 		atomic_set(&s->s_active, 1);
-		sema_init(&s->s_vfs_rename_sem,1);
-		sema_init(&s->s_dquot.dqio_sem, 1);
-		sema_init(&s->s_dquot.dqonoff_sem, 1);
+		mutex_init(&s->s_vfs_rename_mutex);
+		mutex_init(&s->s_dquot.dqio_mutex);
+		mutex_init(&s->s_dquot.dqonoff_mutex);
 		init_rwsem(&s->s_dquot.dqptr_sem);
 		init_waitqueue_head(&s->s_wait_unfrozen);
 		s->s_maxbytes = MAX_NON_LFS;
@@ -693,9 +693,9 @@ struct super_block *get_sb_bdev(struct file_system_type *fs_type,
 	 * will protect the lockfs code from trying to start a snapshot
 	 * while we are mounting
 	 */
-	down(&bdev->bd_mount_sem);
+	mutex_lock(&bdev->bd_mount_mutex);
 	s = sget(fs_type, test_bdev_super, set_bdev_super, bdev);
-	up(&bdev->bd_mount_sem);
+	mutex_unlock(&bdev->bd_mount_mutex);
 	if (IS_ERR(s))
 		goto out;
 
diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c
index 49bd219275db..9ee956864445 100644
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -50,6 +50,32 @@ static struct sysfs_dirent * sysfs_new_dirent(struct sysfs_dirent * parent_sd,
 	return sd;
 }
 
+/**
+ *
+ * Return -EEXIST if there is already a sysfs element with the same name for
+ * the same parent.
+ *
+ * called with parent inode's i_mutex held
+ */
+int sysfs_dirent_exist(struct sysfs_dirent *parent_sd,
+			  const unsigned char *new)
+{
+	struct sysfs_dirent * sd;
+
+	list_for_each_entry(sd, &parent_sd->s_children, s_sibling) {
+		if (sd->s_element) {
+			const unsigned char *existing = sysfs_get_name(sd);
+			if (strcmp(existing, new))
+				continue;
+			else
+				return -EEXIST;
+		}
+	}
+
+	return 0;
+}
+
+
 int sysfs_make_dirent(struct sysfs_dirent * parent_sd, struct dentry * dentry,
 			void * element, umode_t mode, int type)
 {
@@ -102,7 +128,11 @@ static int create_dir(struct kobject * k, struct dentry * p,
 	mutex_lock(&p->d_inode->i_mutex);
 	*d = lookup_one_len(n, p, strlen(n));
 	if (!IS_ERR(*d)) {
-		error = sysfs_make_dirent(p->d_fsdata, *d, k, mode, SYSFS_DIR);
+ 		if (sysfs_dirent_exist(p->d_fsdata, n))
+  			error = -EEXIST;
+  		else
+			error = sysfs_make_dirent(p->d_fsdata, *d, k, mode,
+								SYSFS_DIR);
 		if (!error) {
 			error = sysfs_create(*d, mode, init_dir);
 			if (!error) {
@@ -302,6 +332,7 @@ void sysfs_remove_dir(struct kobject * kobj)
 	 * Drop reference from dget() on entrance.
 	 */
 	dput(dentry);
+	kobj->dentry = NULL;
 }
 
 int sysfs_rename_dir(struct kobject * kobj, const char *new_name)
@@ -479,7 +510,3 @@ struct file_operations sysfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= sysfs_readdir,
 };
-
-EXPORT_SYMBOL_GPL(sysfs_create_dir);
-EXPORT_SYMBOL_GPL(sysfs_remove_dir);
-EXPORT_SYMBOL_GPL(sysfs_rename_dir);
diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index d0e3d8495165..5e83e7246788 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -301,9 +301,8 @@ static int check_perm(struct inode * inode, struct file * file)
 	/* No error? Great, allocate a buffer for the file, and store it
 	 * it in file->private_data for easy access.
 	 */
-	buffer = kmalloc(sizeof(struct sysfs_buffer),GFP_KERNEL);
+	buffer = kzalloc(sizeof(struct sysfs_buffer), GFP_KERNEL);
 	if (buffer) {
-		memset(buffer,0,sizeof(struct sysfs_buffer));
 		init_MUTEX(&buffer->sem);
 		buffer->needs_read_fill = 1;
 		buffer->ops = ops;
@@ -362,10 +361,12 @@ int sysfs_add_file(struct dentry * dir, const struct attribute * attr, int type)
 {
 	struct sysfs_dirent * parent_sd = dir->d_fsdata;
 	umode_t mode = (attr->mode & S_IALLUGO) | S_IFREG;
-	int error = 0;
+	int error = -EEXIST;
 
 	mutex_lock(&dir->d_inode->i_mutex);
-	error = sysfs_make_dirent(parent_sd, NULL, (void *) attr, mode, type);
+	if (!sysfs_dirent_exist(parent_sd, attr->name))
+		error = sysfs_make_dirent(parent_sd, NULL, (void *)attr,
+					  mode, type);
 	mutex_unlock(&dir->d_inode->i_mutex);
 
 	return error;
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index 689f7bcfaf30..4c29ac41ac3e 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -54,11 +54,10 @@ int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 
 	if (!sd_iattr) {
 		/* setting attributes for the first time, allocate now */
-		sd_iattr = kmalloc(sizeof(struct iattr), GFP_KERNEL);
+		sd_iattr = kzalloc(sizeof(struct iattr), GFP_KERNEL);
 		if (!sd_iattr)
 			return -ENOMEM;
 		/* assign default attributes */
-		memset(sd_iattr, 0, sizeof(struct iattr));
 		sd_iattr->ia_mode = sd->s_mode;
 		sd_iattr->ia_uid = 0;
 		sd_iattr->ia_gid = 0;
@@ -227,12 +226,16 @@ void sysfs_drop_dentry(struct sysfs_dirent * sd, struct dentry * parent)
 void sysfs_hash_and_remove(struct dentry * dir, const char * name)
 {
 	struct sysfs_dirent * sd;
-	struct sysfs_dirent * parent_sd = dir->d_fsdata;
+	struct sysfs_dirent * parent_sd;
+
+	if (!dir)
+		return;
 
 	if (dir->d_inode == NULL)
 		/* no inode means this hasn't been made visible yet */
 		return;
 
+	parent_sd = dir->d_fsdata;
 	mutex_lock(&dir->d_inode->i_mutex);
 	list_for_each_entry(sd, &parent_sd->s_children, s_sibling) {
 		if (!sd->s_element)
diff --git a/fs/sysfs/symlink.c b/fs/sysfs/symlink.c
index e38d6338a20d..d2eac3ceed5f 100644
--- a/fs/sysfs/symlink.c
+++ b/fs/sysfs/symlink.c
@@ -66,6 +66,7 @@ static int sysfs_add_link(struct dentry * parent, const char * name, struct kobj
 	if (!error)
 		return 0;
 
+	kobject_put(target);
 	kfree(sl->link_name);
 exit2:
 	kfree(sl);
@@ -82,12 +83,13 @@ exit1:
 int sysfs_create_link(struct kobject * kobj, struct kobject * target, const char * name)
 {
 	struct dentry * dentry = kobj->dentry;
-	int error = 0;
+	int error = -EEXIST;
 
 	BUG_ON(!kobj || !kobj->dentry || !name);
 
 	mutex_lock(&dentry->d_inode->i_mutex);
-	error = sysfs_add_link(dentry, name, target);
+	if (!sysfs_dirent_exist(dentry->d_fsdata, name))
+		error = sysfs_add_link(dentry, name, target);
 	mutex_unlock(&dentry->d_inode->i_mutex);
 	return error;
 }
diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h
index 3f8953e0e5d0..cf11d5b789d9 100644
--- a/fs/sysfs/sysfs.h
+++ b/fs/sysfs/sysfs.h
@@ -5,6 +5,7 @@ extern kmem_cache_t *sysfs_dir_cachep;
 extern struct inode * sysfs_new_inode(mode_t mode, struct sysfs_dirent *);
 extern int sysfs_create(struct dentry *, int mode, int (*init)(struct inode *));
 
+extern int sysfs_dirent_exist(struct sysfs_dirent *, const unsigned char *);
 extern int sysfs_make_dirent(struct sysfs_dirent *, struct dentry *, void *,
 				umode_t, int);
 
diff --git a/fs/sysv/namei.c b/fs/sysv/namei.c
index 7f0e4b53085e..b8a73f716fbe 100644
--- a/fs/sysv/namei.c
+++ b/fs/sysv/namei.c
@@ -16,18 +16,6 @@
 #include <linux/smp_lock.h>
 #include "sysv.h"
 
-static inline void inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static int add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = sysv_add_link(dentry, inode);
@@ -35,7 +23,7 @@ static int add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -124,7 +112,7 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -138,7 +126,7 @@ static int sysv_link(struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	return add_nondir(dentry, inode);
@@ -151,7 +139,7 @@ static int sysv_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 
 	if (dir->i_nlink >= SYSV_SB(dir->i_sb)->s_link_max) 
 		goto out;
-	inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = sysv_new_inode(dir, S_IFDIR|mode);
 	err = PTR_ERR(inode);
@@ -160,7 +148,7 @@ static int sysv_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 
 	sysv_set_inode(inode, 0);
 
-	inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = sysv_make_empty(inode, dir);
 	if (err)
@@ -175,11 +163,11 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
-	dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -199,7 +187,7 @@ static int sysv_unlink(struct inode * dir, struct dentry * dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	dec_count(inode);
+	inode_dec_link_count(inode);
 out:
 	return err;
 }
@@ -213,8 +201,8 @@ static int sysv_rmdir(struct inode * dir, struct dentry * dentry)
 		err = sysv_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			dec_count(inode);
-			dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	return err;
@@ -258,34 +246,34 @@ static int sysv_rename(struct inode * old_dir, struct dentry * old_dentry,
 		new_de = sysv_find_entry(new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		sysv_set_link(new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= SYSV_SB(new_dir->i_sb)->s_link_max)
 				goto out_dir;
 		}
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = sysv_add_link(new_dentry, old_inode);
 		if (err) {
-			dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	sysv_delete_entry(old_de, old_page);
-	dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		sysv_set_link(dir_de, dir_page, new_dir);
-		dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c
index 201049ac8a96..ea521f846d97 100644
--- a/fs/udf/balloc.c
+++ b/fs/udf/balloc.c
@@ -152,7 +152,7 @@ static void udf_bitmap_free_blocks(struct super_block * sb,
 	int bitmap_nr;
 	unsigned long overflow;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (bloc.logicalBlockNum < 0 ||
 		(bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))
 	{
@@ -211,7 +211,7 @@ error_return:
 	sb->s_dirt = 1;
 	if (UDF_SB_LVIDBH(sb))
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return;
 }
 
@@ -226,7 +226,7 @@ static int udf_bitmap_prealloc_blocks(struct super_block * sb,
 	int nr_groups, bitmap_nr;
 	struct buffer_head *bh;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (first_block < 0 || first_block >= UDF_SB_PARTLEN(sb, partition))
 		goto out;
 
@@ -275,7 +275,7 @@ out:
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 	}
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return alloc_count;
 }
 
@@ -291,7 +291,7 @@ static int udf_bitmap_new_block(struct super_block * sb,
 	int newblock = 0;
 
 	*err = -ENOSPC;
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 
 repeat:
 	if (goal < 0 || goal >= UDF_SB_PARTLEN(sb, partition))
@@ -364,7 +364,7 @@ repeat:
 	}
 	if (i >= (nr_groups*2))
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return newblock;
 	}
 	if (bit < sb->s_blocksize << 3)
@@ -373,7 +373,7 @@ repeat:
 		bit = udf_find_next_one_bit(bh->b_data, sb->s_blocksize << 3, group_start << 3);
 	if (bit >= sb->s_blocksize << 3)
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return 0;
 	}
 
@@ -387,7 +387,7 @@ got_block:
 	 */
 	if (inode && DQUOT_ALLOC_BLOCK(inode, 1))
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		*err = -EDQUOT;
 		return 0;
 	}
@@ -410,13 +410,13 @@ got_block:
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 	}
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	*err = 0;
 	return newblock;
 
 error_return:
 	*err = -EIO;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return 0;
 }
 
@@ -433,7 +433,7 @@ static void udf_table_free_blocks(struct super_block * sb,
 	int8_t etype;
 	int i;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (bloc.logicalBlockNum < 0 ||
 		(bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))
 	{
@@ -666,7 +666,7 @@ static void udf_table_free_blocks(struct super_block * sb,
 
 error_return:
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return;
 }
 
@@ -692,7 +692,7 @@ static int udf_table_prealloc_blocks(struct super_block * sb,
 	else
 		return 0;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	extoffset = sizeof(struct unallocSpaceEntry);
 	bloc = UDF_I_LOCATION(table);
 
@@ -736,7 +736,7 @@ static int udf_table_prealloc_blocks(struct super_block * sb,
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 		sb->s_dirt = 1;
 	}
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return alloc_count;
 }
 
@@ -761,7 +761,7 @@ static int udf_table_new_block(struct super_block * sb,
 	else
 		return newblock;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (goal < 0 || goal >= UDF_SB_PARTLEN(sb, partition))
 		goal = 0;
 
@@ -811,7 +811,7 @@ static int udf_table_new_block(struct super_block * sb,
 	if (spread == 0xFFFFFFFF)
 	{
 		udf_release_data(goal_bh);
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return 0;
 	}
 
@@ -827,7 +827,7 @@ static int udf_table_new_block(struct super_block * sb,
 	if (inode && DQUOT_ALLOC_BLOCK(inode, 1))
 	{
 		udf_release_data(goal_bh);
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		*err = -EDQUOT;
 		return 0;
 	}
@@ -846,7 +846,7 @@ static int udf_table_new_block(struct super_block * sb,
 	}
 
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	*err = 0;
 	return newblock;
 }
diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
index c9b707b470ca..3873c672cb4c 100644
--- a/fs/udf/ialloc.c
+++ b/fs/udf/ialloc.c
@@ -42,7 +42,7 @@ void udf_free_inode(struct inode * inode)
 
 	clear_inode(inode);
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (sbi->s_lvidbh) {
 		if (S_ISDIR(inode->i_mode))
 			UDF_SB_LVIDIU(sb)->numDirs =
@@ -53,7 +53,7 @@ void udf_free_inode(struct inode * inode)
 		
 		mark_buffer_dirty(sbi->s_lvidbh);
 	}
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 
 	udf_free_blocks(sb, NULL, UDF_I_LOCATION(inode), 0, 1);
 }
@@ -83,7 +83,7 @@ struct inode * udf_new_inode (struct inode *dir, int mode, int * err)
 		return NULL;
 	}
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	UDF_I_UNIQUE(inode) = 0;
 	UDF_I_LENEXTENTS(inode) = 0;
 	UDF_I_NEXT_ALLOC_BLOCK(inode) = 0;
@@ -148,7 +148,7 @@ struct inode * udf_new_inode (struct inode *dir, int mode, int * err)
 		UDF_I_CRTIME(inode) = current_fs_time(inode->i_sb);
 	insert_inode_hash(inode);
 	mark_inode_dirty(inode);
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 
 	if (DQUOT_ALLOC_INODE(inode))
 	{
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 368d8f81fe54..9303c50c5d55 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -1515,7 +1515,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 	sb->s_fs_info = sbi;
 	memset(UDF_SB(sb), 0x00, sizeof(struct udf_sb_info));
 
-	init_MUTEX(&sbi->s_alloc_sem);
+	mutex_init(&sbi->s_alloc_mutex);
 
 	if (!udf_parse_options((char *)options, &uopt))
 		goto error_out;
diff --git a/fs/ufs/file.c b/fs/ufs/file.c
index ed69d7fe1b5d..62ad481810ef 100644
--- a/fs/ufs/file.c
+++ b/fs/ufs/file.c
@@ -23,18 +23,8 @@
  *  ext2 fs regular file handling primitives
  */
 
-#include <asm/uaccess.h>
-#include <asm/system.h>
-
-#include <linux/errno.h>
 #include <linux/fs.h>
 #include <linux/ufs_fs.h>
-#include <linux/fcntl.h>
-#include <linux/time.h>
-#include <linux/stat.h>
-#include <linux/mm.h>
-#include <linux/pagemap.h>
-#include <linux/smp_lock.h>
 
 /*
  * We have mostly NULL's here: the current defaults are ok for
diff --git a/fs/ufs/namei.c b/fs/ufs/namei.c
index 2958cde7d3d6..8d5f98a01c74 100644
--- a/fs/ufs/namei.c
+++ b/fs/ufs/namei.c
@@ -43,18 +43,6 @@
 #define UFSD(x)
 #endif
 
-static inline void ufs_inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void ufs_dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static inline int ufs_add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = ufs_add_link(dentry, inode);
@@ -62,7 +50,7 @@ static inline int ufs_add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -173,7 +161,7 @@ out:
 	return err;
 
 out_fail:
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -191,7 +179,7 @@ static int ufs_link (struct dentry * old_dentry, struct inode * dir,
 	}
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	ufs_inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	error = ufs_add_nondir(dentry, inode);
@@ -208,7 +196,7 @@ static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 		goto out;
 
 	lock_kernel();
-	ufs_inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = ufs_new_inode(dir, S_IFDIR|mode);
 	err = PTR_ERR(inode);
@@ -218,7 +206,7 @@ static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	inode->i_op = &ufs_dir_inode_operations;
 	inode->i_fop = &ufs_dir_operations;
 
-	ufs_inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = ufs_make_empty(inode, dir);
 	if (err)
@@ -234,11 +222,11 @@ out:
 	return err;
 
 out_fail:
-	ufs_dec_count(inode);
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput (inode);
 out_dir:
-	ufs_dec_count(dir);
+	inode_dec_link_count(dir);
 	unlock_kernel();
 	goto out;
 }
@@ -260,7 +248,7 @@ static int ufs_unlink(struct inode * dir, struct dentry *dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	err = 0;
 out:
 	unlock_kernel();
@@ -277,8 +265,8 @@ static int ufs_rmdir (struct inode * dir, struct dentry *dentry)
 		err = ufs_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			ufs_dec_count(inode);
-			ufs_dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	unlock_kernel();
@@ -319,35 +307,35 @@ static int ufs_rename (struct inode * old_dir, struct dentry * old_dentry,
 		new_de = ufs_find_entry (new_dentry, &new_bh);
 		if (!new_de)
 			goto out_dir;
-		ufs_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		ufs_set_link(new_dir, new_de, new_bh, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		ufs_dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= UFS_LINK_MAX)
 				goto out_dir;
 		}
-		ufs_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = ufs_add_link(new_dentry, old_inode);
 		if (err) {
-			ufs_dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			ufs_inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	ufs_delete_entry (old_dir, old_de, old_bh);
 
-	ufs_dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		ufs_set_link(old_inode, dir_de, dir_bh, new_dir);
-		ufs_dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	unlock_kernel();
 	return 0;
diff --git a/fs/xfs/Makefile-linux-2.6 b/fs/xfs/Makefile-linux-2.6
index 97bd4743b461..5d73eaa1971f 100644
--- a/fs/xfs/Makefile-linux-2.6
+++ b/fs/xfs/Makefile-linux-2.6
@@ -1,33 +1,19 @@
 #
-# Copyright (c) 2000-2005 Silicon Graphics, Inc.  All Rights Reserved.
+# 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 version 2 of the GNU General Public License as
+# 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.
+# 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.
 #
-# Further, this software is distributed without any warranty that it is
-# free of the rightful claim of any third person regarding infringement
-# or the like.  Any license provided herein, whether implied or
-# otherwise, applies only to this software file.  Patent licenses, if
-# any, provided herein do not apply to combinations of this program with
-# other software, or any other product whatsoever.
-#
-# You should have received a copy of the GNU General Public License along
-# with this program; if not, write the Free Software Foundation, Inc., 59
-# Temple Place - Suite 330, Boston MA 02111-1307, USA.
-#
-# Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
-# Mountain View, CA  94043, or:
-#
-# http://www.sgi.com
-#
-# For further information regarding this notice, see:
-#
-# http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
+# 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
 #
 
 EXTRA_CFLAGS +=	 -Ifs/xfs -Ifs/xfs/linux-2.6 -funsigned-char
@@ -36,7 +22,7 @@ XFS_LINUX := linux-2.6
 
 ifeq ($(CONFIG_XFS_DEBUG),y)
 	EXTRA_CFLAGS += -g -DSTATIC="" -DDEBUG
-	EXTRA_CFLAGS += -DPAGEBUF_LOCK_TRACKING
+	EXTRA_CFLAGS += -DXFS_BUF_LOCK_TRACKING
 endif
 ifeq ($(CONFIG_XFS_TRACE),y)
 	EXTRA_CFLAGS += -DXFS_ALLOC_TRACE
@@ -50,7 +36,7 @@ ifeq ($(CONFIG_XFS_TRACE),y)
 	EXTRA_CFLAGS += -DXFS_ILOCK_TRACE
 	EXTRA_CFLAGS += -DXFS_LOG_TRACE
 	EXTRA_CFLAGS += -DXFS_RW_TRACE
-	EXTRA_CFLAGS += -DPAGEBUF_TRACE
+	EXTRA_CFLAGS += -DXFS_BUF_TRACE
 	EXTRA_CFLAGS += -DXFS_VNODE_TRACE
 endif
 
diff --git a/fs/xfs/linux-2.6/kmem.h b/fs/xfs/linux-2.6/kmem.h
index c64a29cdfff3..f0268a84e6fd 100644
--- a/fs/xfs/linux-2.6/kmem.h
+++ b/fs/xfs/linux-2.6/kmem.h
@@ -23,17 +23,8 @@
 #include <linux/mm.h>
 
 /*
- * memory management routines
+ * Process flags handling
  */
-#define KM_SLEEP	0x0001u
-#define KM_NOSLEEP	0x0002u
-#define KM_NOFS		0x0004u
-#define KM_MAYFAIL	0x0008u
-
-#define	kmem_zone	kmem_cache
-#define kmem_zone_t	struct kmem_cache
-
-typedef unsigned long xfs_pflags_t;
 
 #define PFLAGS_TEST_NOIO()              (current->flags & PF_NOIO)
 #define PFLAGS_TEST_FSTRANS()           (current->flags & PF_FSTRANS)
@@ -67,74 +58,102 @@ typedef unsigned long xfs_pflags_t;
 	*(NSTATEP) = *(OSTATEP);	\
 } while (0)
 
-static __inline gfp_t kmem_flags_convert(unsigned int __nocast flags)
+/*
+ * 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 = __GFP_NOWARN;	/* we'll report problems, if need be */
+	gfp_t	lflags;
 
-#ifdef DEBUG
-	if (unlikely(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL))) {
-		printk(KERN_WARNING
-		    "XFS: memory allocation with wrong flags (%x)\n", flags);
-		BUG();
-	}
-#endif
+	BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL));
 
 	if (flags & KM_NOSLEEP) {
-		lflags |= GFP_ATOMIC;
+		lflags = GFP_ATOMIC | __GFP_NOWARN;
 	} else {
-		lflags |= GFP_KERNEL;
-
-		/* avoid recusive callbacks to filesystem during transactions */
+		lflags = GFP_KERNEL | __GFP_NOWARN;
 		if (PFLAGS_TEST_FSTRANS() || (flags & KM_NOFS))
 			lflags &= ~__GFP_FS;
 	}
-
-        return lflags;
+	return lflags;
 }
 
-static __inline kmem_zone_t *
+extern void *kmem_alloc(size_t, unsigned int __nocast);
+extern void *kmem_realloc(void *, size_t, size_t, unsigned int __nocast);
+extern void *kmem_zalloc(size_t, unsigned int __nocast);
+extern void  kmem_free(void *, size_t);
+
+/*
+ * Zone interfaces
+ */
+
+#define KM_ZONE_HWALIGN	SLAB_HWCACHE_ALIGN
+#define KM_ZONE_RECLAIM	SLAB_RECLAIM_ACCOUNT
+#define KM_ZONE_SPREAD	0
+
+#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, NULL);
 }
 
-static __inline void
+static inline kmem_zone_t *
+kmem_zone_init_flags(int size, char *zone_name, unsigned long flags,
+		     void (*construct)(void *, kmem_zone_t *, unsigned long))
+{
+	return kmem_cache_create(zone_name, size, 0, flags, construct, NULL);
+}
+
+static inline void
 kmem_zone_free(kmem_zone_t *zone, void *ptr)
 {
 	kmem_cache_free(zone, ptr);
 }
 
-static __inline void
+static inline void
 kmem_zone_destroy(kmem_zone_t *zone)
 {
 	if (zone && kmem_cache_destroy(zone))
 		BUG();
 }
 
-extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
 extern void *kmem_zone_alloc(kmem_zone_t *, unsigned int __nocast);
+extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
 
-extern void *kmem_alloc(size_t, unsigned int __nocast);
-extern void *kmem_realloc(void *, size_t, size_t, unsigned int __nocast);
-extern void *kmem_zalloc(size_t, unsigned int __nocast);
-extern void  kmem_free(void *, size_t);
+/*
+ * Low memory cache shrinkers
+ */
 
 typedef struct shrinker *kmem_shaker_t;
 typedef int (*kmem_shake_func_t)(int, gfp_t);
 
-static __inline kmem_shaker_t
+static inline kmem_shaker_t
 kmem_shake_register(kmem_shake_func_t sfunc)
 {
 	return set_shrinker(DEFAULT_SEEKS, sfunc);
 }
 
-static __inline void
+static inline void
 kmem_shake_deregister(kmem_shaker_t shrinker)
 {
 	remove_shrinker(shrinker);
 }
 
-static __inline int
+static inline int
 kmem_shake_allow(gfp_t gfp_mask)
 {
 	return (gfp_mask & __GFP_WAIT);
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index 74d8be87f983..97fc056130eb 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -43,7 +43,29 @@
 #include <linux/pagevec.h>
 #include <linux/writeback.h>
 
-STATIC void xfs_count_page_state(struct page *, int *, int *, int *);
+STATIC void
+xfs_count_page_state(
+	struct page		*page,
+	int			*delalloc,
+	int			*unmapped,
+	int			*unwritten)
+{
+	struct buffer_head	*bh, *head;
+
+	*delalloc = *unmapped = *unwritten = 0;
+
+	bh = head = page_buffers(page);
+	do {
+		if (buffer_uptodate(bh) && !buffer_mapped(bh))
+			(*unmapped) = 1;
+		else if (buffer_unwritten(bh) && !buffer_delay(bh))
+			clear_buffer_unwritten(bh);
+		else if (buffer_unwritten(bh))
+			(*unwritten) = 1;
+		else if (buffer_delay(bh))
+			(*delalloc) = 1;
+	} while ((bh = bh->b_this_page) != head);
+}
 
 #if defined(XFS_RW_TRACE)
 void
@@ -54,7 +76,7 @@ xfs_page_trace(
 	int		mask)
 {
 	xfs_inode_t	*ip;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	loff_t		isize = i_size_read(inode);
 	loff_t		offset = page_offset(page);
 	int		delalloc = -1, unmapped = -1, unwritten = -1;
@@ -81,7 +103,7 @@ xfs_page_trace(
 		(void *)((unsigned long)delalloc),
 		(void *)((unsigned long)unmapped),
 		(void *)((unsigned long)unwritten),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL);
 }
 #else
@@ -192,7 +214,7 @@ xfs_alloc_ioend(
 	ioend->io_uptodate = 1; /* cleared if any I/O fails */
 	ioend->io_list = NULL;
 	ioend->io_type = type;
-	ioend->io_vnode = LINVFS_GET_VP(inode);
+	ioend->io_vnode = vn_from_inode(inode);
 	ioend->io_buffer_head = NULL;
 	ioend->io_buffer_tail = NULL;
 	atomic_inc(&ioend->io_vnode->v_iocount);
@@ -217,7 +239,7 @@ xfs_map_blocks(
 	xfs_iomap_t		*mapp,
 	int			flags)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error, nmaps = 1;
 
 	VOP_BMAP(vp, offset, count, flags, mapp, &nmaps, error);
@@ -462,28 +484,37 @@ xfs_add_to_ioend(
 }
 
 STATIC void
+xfs_map_buffer(
+	struct buffer_head	*bh,
+	xfs_iomap_t		*mp,
+	xfs_off_t		offset,
+	uint			block_bits)
+{
+	sector_t		bn;
+
+	ASSERT(mp->iomap_bn != IOMAP_DADDR_NULL);
+
+	bn = (mp->iomap_bn >> (block_bits - BBSHIFT)) +
+	      ((offset - mp->iomap_offset) >> block_bits);
+
+	ASSERT(bn || (mp->iomap_flags & IOMAP_REALTIME));
+
+	bh->b_blocknr = bn;
+	set_buffer_mapped(bh);
+}
+
+STATIC void
 xfs_map_at_offset(
 	struct buffer_head	*bh,
 	loff_t			offset,
 	int			block_bits,
 	xfs_iomap_t		*iomapp)
 {
-	xfs_daddr_t		bn;
-	int			sector_shift;
-
 	ASSERT(!(iomapp->iomap_flags & IOMAP_HOLE));
 	ASSERT(!(iomapp->iomap_flags & IOMAP_DELAY));
-	ASSERT(iomapp->iomap_bn != IOMAP_DADDR_NULL);
-
-	sector_shift = block_bits - BBSHIFT;
-	bn = (iomapp->iomap_bn >> sector_shift) +
-	      ((offset - iomapp->iomap_offset) >> block_bits);
-
-	ASSERT(bn || (iomapp->iomap_flags & IOMAP_REALTIME));
-	ASSERT((bn << sector_shift) >= iomapp->iomap_bn);
 
 	lock_buffer(bh);
-	bh->b_blocknr = bn;
+	xfs_map_buffer(bh, iomapp, offset, block_bits);
 	bh->b_bdev = iomapp->iomap_target->bt_bdev;
 	set_buffer_mapped(bh);
 	clear_buffer_delay(bh);
@@ -616,7 +647,7 @@ xfs_is_delayed_page(
 				acceptable = (type == IOMAP_UNWRITTEN);
 			else if (buffer_delay(bh))
 				acceptable = (type == IOMAP_DELAY);
-			else if (buffer_mapped(bh))
+			else if (buffer_dirty(bh) && buffer_mapped(bh))
 				acceptable = (type == 0);
 			else
 				break;
@@ -1040,8 +1071,159 @@ error:
 	return err;
 }
 
+/*
+ * writepage: Called from one of two places:
+ *
+ * 1. we are flushing a delalloc buffer head.
+ *
+ * 2. we are writing out a dirty page. Typically the page dirty
+ *    state is cleared before we get here. In this case is it
+ *    conceivable we have no buffer heads.
+ *
+ * For delalloc space on the page we need to allocate space and
+ * flush it. For unmapped buffer heads on the page we should
+ * allocate space if the page is uptodate. For any other dirty
+ * buffer heads on the page we should flush them.
+ *
+ * If we detect that a transaction would be required to flush
+ * the page, we have to check the process flags first, if we
+ * are already in a transaction or disk I/O during allocations
+ * is off, we need to fail the writepage and redirty the page.
+ */
+
+STATIC int
+xfs_vm_writepage(
+	struct page		*page,
+	struct writeback_control *wbc)
+{
+	int			error;
+	int			need_trans;
+	int			delalloc, unmapped, unwritten;
+	struct inode		*inode = page->mapping->host;
+
+	xfs_page_trace(XFS_WRITEPAGE_ENTER, inode, page, 0);
+
+	/*
+	 * We need a transaction if:
+	 *  1. There are delalloc buffers on the page
+	 *  2. The page is uptodate and we have unmapped buffers
+	 *  3. The page is uptodate and we have no buffers
+	 *  4. There are unwritten buffers on the page
+	 */
+
+	if (!page_has_buffers(page)) {
+		unmapped = 1;
+		need_trans = 1;
+	} else {
+		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+		if (!PageUptodate(page))
+			unmapped = 0;
+		need_trans = delalloc + unmapped + unwritten;
+	}
+
+	/*
+	 * If we need a transaction and the process flags say
+	 * we are already in a transaction, or no IO is allowed
+	 * then mark the page dirty again and leave the page
+	 * as is.
+	 */
+	if (PFLAGS_TEST_FSTRANS() && need_trans)
+		goto out_fail;
+
+	/*
+	 * Delay hooking up buffer heads until we have
+	 * made our go/no-go decision.
+	 */
+	if (!page_has_buffers(page))
+		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
+
+	/*
+	 * Convert delayed allocate, unwritten or unmapped space
+	 * to real space and flush out to disk.
+	 */
+	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
+	if (error == -EAGAIN)
+		goto out_fail;
+	if (unlikely(error < 0))
+		goto out_unlock;
+
+	return 0;
+
+out_fail:
+	redirty_page_for_writepage(wbc, page);
+	unlock_page(page);
+	return 0;
+out_unlock:
+	unlock_page(page);
+	return error;
+}
+
+/*
+ * Called to move a page into cleanable state - and from there
+ * to be released. Possibly the page is already clean. We always
+ * have buffer heads in this call.
+ *
+ * Returns 0 if the page is ok to release, 1 otherwise.
+ *
+ * Possible scenarios are:
+ *
+ * 1. We are being called to release a page which has been written
+ *    to via regular I/O. buffer heads will be dirty and possibly
+ *    delalloc. If no delalloc buffer heads in this case then we
+ *    can just return zero.
+ *
+ * 2. We are called to release a page which has been written via
+ *    mmap, all we need to do is ensure there is no delalloc
+ *    state in the buffer heads, if not we can let the caller
+ *    free them and we should come back later via writepage.
+ */
 STATIC int
-__linvfs_get_block(
+xfs_vm_releasepage(
+	struct page		*page,
+	gfp_t			gfp_mask)
+{
+	struct inode		*inode = page->mapping->host;
+	int			dirty, delalloc, unmapped, unwritten;
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_ALL,
+		.nr_to_write = 1,
+	};
+
+	xfs_page_trace(XFS_RELEASEPAGE_ENTER, inode, page, gfp_mask);
+
+	if (!page_has_buffers(page))
+		return 0;
+
+	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+	if (!delalloc && !unwritten)
+		goto free_buffers;
+
+	if (!(gfp_mask & __GFP_FS))
+		return 0;
+
+	/* If we are already inside a transaction or the thread cannot
+	 * do I/O, we cannot release this page.
+	 */
+	if (PFLAGS_TEST_FSTRANS())
+		return 0;
+
+	/*
+	 * Convert delalloc space to real space, do not flush the
+	 * data out to disk, that will be done by the caller.
+	 * Never need to allocate space here - we will always
+	 * come back to writepage in that case.
+	 */
+	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
+	if (dirty == 0 && !unwritten)
+		goto free_buffers;
+	return 0;
+
+free_buffers:
+	return try_to_free_buffers(page);
+}
+
+STATIC int
+__xfs_get_block(
 	struct inode		*inode,
 	sector_t		iblock,
 	unsigned long		blocks,
@@ -1050,7 +1232,7 @@ __linvfs_get_block(
 	int			direct,
 	bmapi_flags_t		flags)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	xfs_iomap_t		iomap;
 	xfs_off_t		offset;
 	ssize_t			size;
@@ -1073,21 +1255,13 @@ __linvfs_get_block(
 		return 0;
 
 	if (iomap.iomap_bn != IOMAP_DADDR_NULL) {
-		xfs_daddr_t	bn;
-		xfs_off_t	delta;
-
-		/* For unwritten extents do not report a disk address on
+		/*
+		 * For unwritten extents do not report a disk address on
 		 * the read case (treat as if we're reading into a hole).
 		 */
 		if (create || !(iomap.iomap_flags & IOMAP_UNWRITTEN)) {
-			delta = offset - iomap.iomap_offset;
-			delta >>= inode->i_blkbits;
-
-			bn = iomap.iomap_bn >> (inode->i_blkbits - BBSHIFT);
-			bn += delta;
-			BUG_ON(!bn && !(iomap.iomap_flags & IOMAP_REALTIME));
-			bh_result->b_blocknr = bn;
-			set_buffer_mapped(bh_result);
+			xfs_map_buffer(bh_result, &iomap, offset,
+				       inode->i_blkbits);
 		}
 		if (create && (iomap.iomap_flags & IOMAP_UNWRITTEN)) {
 			if (direct)
@@ -1130,30 +1304,30 @@ __linvfs_get_block(
 }
 
 int
-linvfs_get_block(
+xfs_get_block(
 	struct inode		*inode,
 	sector_t		iblock,
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __linvfs_get_block(inode, iblock, 0, bh_result,
+	return __xfs_get_block(inode, iblock, 0, bh_result,
 					create, 0, BMAPI_WRITE);
 }
 
 STATIC int
-linvfs_get_blocks_direct(
+xfs_get_blocks_direct(
 	struct inode		*inode,
 	sector_t		iblock,
 	unsigned long		max_blocks,
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __linvfs_get_block(inode, iblock, max_blocks, bh_result,
+	return __xfs_get_block(inode, iblock, max_blocks, bh_result,
 					create, 1, BMAPI_WRITE|BMAPI_DIRECT);
 }
 
 STATIC void
-linvfs_end_io_direct(
+xfs_end_io_direct(
 	struct kiocb	*iocb,
 	loff_t		offset,
 	ssize_t		size,
@@ -1191,7 +1365,7 @@ linvfs_end_io_direct(
 }
 
 STATIC ssize_t
-linvfs_direct_IO(
+xfs_vm_direct_IO(
 	int			rw,
 	struct kiocb		*iocb,
 	const struct iovec	*iov,
@@ -1200,7 +1374,7 @@ linvfs_direct_IO(
 {
 	struct file	*file = iocb->ki_filp;
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_iomap_t	iomap;
 	int		maps = 1;
 	int		error;
@@ -1215,164 +1389,61 @@ linvfs_direct_IO(
 	ret = blockdev_direct_IO_own_locking(rw, iocb, inode,
 		iomap.iomap_target->bt_bdev,
 		iov, offset, nr_segs,
-		linvfs_get_blocks_direct,
-		linvfs_end_io_direct);
+		xfs_get_blocks_direct,
+		xfs_end_io_direct);
 
 	if (unlikely(ret <= 0 && iocb->private))
 		xfs_destroy_ioend(iocb->private);
 	return ret;
 }
 
+STATIC int
+xfs_vm_prepare_write(
+	struct file		*file,
+	struct page		*page,
+	unsigned int		from,
+	unsigned int		to)
+{
+	return block_prepare_write(page, from, to, xfs_get_block);
+}
 
 STATIC sector_t
-linvfs_bmap(
+xfs_vm_bmap(
 	struct address_space	*mapping,
 	sector_t		block)
 {
 	struct inode		*inode = (struct inode *)mapping->host;
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error;
 
-	vn_trace_entry(vp, "linvfs_bmap", (inst_t *)__return_address);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
 
 	VOP_RWLOCK(vp, VRWLOCK_READ);
 	VOP_FLUSH_PAGES(vp, (xfs_off_t)0, -1, 0, FI_REMAPF, error);
 	VOP_RWUNLOCK(vp, VRWLOCK_READ);
-	return generic_block_bmap(mapping, block, linvfs_get_block);
+	return generic_block_bmap(mapping, block, xfs_get_block);
 }
 
 STATIC int
-linvfs_readpage(
+xfs_vm_readpage(
 	struct file		*unused,
 	struct page		*page)
 {
-	return mpage_readpage(page, linvfs_get_block);
+	return mpage_readpage(page, xfs_get_block);
 }
 
 STATIC int
-linvfs_readpages(
+xfs_vm_readpages(
 	struct file		*unused,
 	struct address_space	*mapping,
 	struct list_head	*pages,
 	unsigned		nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, linvfs_get_block);
-}
-
-STATIC void
-xfs_count_page_state(
-	struct page		*page,
-	int			*delalloc,
-	int			*unmapped,
-	int			*unwritten)
-{
-	struct buffer_head	*bh, *head;
-
-	*delalloc = *unmapped = *unwritten = 0;
-
-	bh = head = page_buffers(page);
-	do {
-		if (buffer_uptodate(bh) && !buffer_mapped(bh))
-			(*unmapped) = 1;
-		else if (buffer_unwritten(bh) && !buffer_delay(bh))
-			clear_buffer_unwritten(bh);
-		else if (buffer_unwritten(bh))
-			(*unwritten) = 1;
-		else if (buffer_delay(bh))
-			(*delalloc) = 1;
-	} while ((bh = bh->b_this_page) != head);
+	return mpage_readpages(mapping, pages, nr_pages, xfs_get_block);
 }
 
-
-/*
- * writepage: Called from one of two places:
- *
- * 1. we are flushing a delalloc buffer head.
- *
- * 2. we are writing out a dirty page. Typically the page dirty
- *    state is cleared before we get here. In this case is it
- *    conceivable we have no buffer heads.
- *
- * For delalloc space on the page we need to allocate space and
- * flush it. For unmapped buffer heads on the page we should
- * allocate space if the page is uptodate. For any other dirty
- * buffer heads on the page we should flush them.
- *
- * If we detect that a transaction would be required to flush
- * the page, we have to check the process flags first, if we
- * are already in a transaction or disk I/O during allocations
- * is off, we need to fail the writepage and redirty the page.
- */
-
 STATIC int
-linvfs_writepage(
-	struct page		*page,
-	struct writeback_control *wbc)
-{
-	int			error;
-	int			need_trans;
-	int			delalloc, unmapped, unwritten;
-	struct inode		*inode = page->mapping->host;
-
-	xfs_page_trace(XFS_WRITEPAGE_ENTER, inode, page, 0);
-
-	/*
-	 * We need a transaction if:
-	 *  1. There are delalloc buffers on the page
-	 *  2. The page is uptodate and we have unmapped buffers
-	 *  3. The page is uptodate and we have no buffers
-	 *  4. There are unwritten buffers on the page
-	 */
-
-	if (!page_has_buffers(page)) {
-		unmapped = 1;
-		need_trans = 1;
-	} else {
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-		if (!PageUptodate(page))
-			unmapped = 0;
-		need_trans = delalloc + unmapped + unwritten;
-	}
-
-	/*
-	 * If we need a transaction and the process flags say
-	 * we are already in a transaction, or no IO is allowed
-	 * then mark the page dirty again and leave the page
-	 * as is.
-	 */
-	if (PFLAGS_TEST_FSTRANS() && need_trans)
-		goto out_fail;
-
-	/*
-	 * Delay hooking up buffer heads until we have
-	 * made our go/no-go decision.
-	 */
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-
-	/*
-	 * Convert delayed allocate, unwritten or unmapped space
-	 * to real space and flush out to disk.
-	 */
-	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
-	if (error == -EAGAIN)
-		goto out_fail;
-	if (unlikely(error < 0))
-		goto out_unlock;
-
-	return 0;
-
-out_fail:
-	redirty_page_for_writepage(wbc, page);
-	unlock_page(page);
-	return 0;
-out_unlock:
-	unlock_page(page);
-	return error;
-}
-
-STATIC int
-linvfs_invalidate_page(
+xfs_vm_invalidatepage(
 	struct page		*page,
 	unsigned long		offset)
 {
@@ -1381,87 +1452,16 @@ linvfs_invalidate_page(
 	return block_invalidatepage(page, offset);
 }
 
-/*
- * Called to move a page into cleanable state - and from there
- * to be released. Possibly the page is already clean. We always
- * have buffer heads in this call.
- *
- * Returns 0 if the page is ok to release, 1 otherwise.
- *
- * Possible scenarios are:
- *
- * 1. We are being called to release a page which has been written
- *    to via regular I/O. buffer heads will be dirty and possibly
- *    delalloc. If no delalloc buffer heads in this case then we
- *    can just return zero.
- *
- * 2. We are called to release a page which has been written via
- *    mmap, all we need to do is ensure there is no delalloc
- *    state in the buffer heads, if not we can let the caller
- *    free them and we should come back later via writepage.
- */
-STATIC int
-linvfs_release_page(
-	struct page		*page,
-	gfp_t			gfp_mask)
-{
-	struct inode		*inode = page->mapping->host;
-	int			dirty, delalloc, unmapped, unwritten;
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
-		.nr_to_write = 1,
-	};
-
-	xfs_page_trace(XFS_RELEASEPAGE_ENTER, inode, page, gfp_mask);
-
-	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-	if (!delalloc && !unwritten)
-		goto free_buffers;
-
-	if (!(gfp_mask & __GFP_FS))
-		return 0;
-
-	/* If we are already inside a transaction or the thread cannot
-	 * do I/O, we cannot release this page.
-	 */
-	if (PFLAGS_TEST_FSTRANS())
-		return 0;
-
-	/*
-	 * Convert delalloc space to real space, do not flush the
-	 * data out to disk, that will be done by the caller.
-	 * Never need to allocate space here - we will always
-	 * come back to writepage in that case.
-	 */
-	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
-	if (dirty == 0 && !unwritten)
-		goto free_buffers;
-	return 0;
-
-free_buffers:
-	return try_to_free_buffers(page);
-}
-
-STATIC int
-linvfs_prepare_write(
-	struct file		*file,
-	struct page		*page,
-	unsigned int		from,
-	unsigned int		to)
-{
-	return block_prepare_write(page, from, to, linvfs_get_block);
-}
-
-struct address_space_operations linvfs_aops = {
-	.readpage		= linvfs_readpage,
-	.readpages		= linvfs_readpages,
-	.writepage		= linvfs_writepage,
+struct address_space_operations xfs_address_space_operations = {
+	.readpage		= xfs_vm_readpage,
+	.readpages		= xfs_vm_readpages,
+	.writepage		= xfs_vm_writepage,
 	.sync_page		= block_sync_page,
-	.releasepage		= linvfs_release_page,
-	.invalidatepage		= linvfs_invalidate_page,
-	.prepare_write		= linvfs_prepare_write,
+	.releasepage		= xfs_vm_releasepage,
+	.invalidatepage		= xfs_vm_invalidatepage,
+	.prepare_write		= xfs_vm_prepare_write,
 	.commit_write		= generic_commit_write,
-	.bmap			= linvfs_bmap,
-	.direct_IO		= linvfs_direct_IO,
+	.bmap			= xfs_vm_bmap,
+	.direct_IO		= xfs_vm_direct_IO,
 	.migratepage		= buffer_migrate_page,
 };
diff --git a/fs/xfs/linux-2.6/xfs_aops.h b/fs/xfs/linux-2.6/xfs_aops.h
index 55339dd5a30d..795699f121d2 100644
--- a/fs/xfs/linux-2.6/xfs_aops.h
+++ b/fs/xfs/linux-2.6/xfs_aops.h
@@ -40,7 +40,7 @@ typedef struct xfs_ioend {
 	struct work_struct	io_work;	/* xfsdatad work queue */
 } xfs_ioend_t;
 
-extern struct address_space_operations linvfs_aops;
-extern int linvfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
+extern struct address_space_operations xfs_address_space_operations;
+extern int xfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
 
 #endif /* __XFS_IOPS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index bfb4f2917bb6..9fb0312665ca 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -29,6 +29,7 @@
 #include <linux/blkdev.h>
 #include <linux/hash.h>
 #include <linux/kthread.h>
+#include <linux/migrate.h>
 #include "xfs_linux.h"
 
 STATIC kmem_zone_t *xfs_buf_zone;
@@ -1805,13 +1806,12 @@ xfs_flush_buftarg(
 int __init
 xfs_buf_init(void)
 {
-	int		error = -ENOMEM;
-
 #ifdef XFS_BUF_TRACE
 	xfs_buf_trace_buf = ktrace_alloc(XFS_BUF_TRACE_SIZE, KM_SLEEP);
 #endif
 
-	xfs_buf_zone = kmem_zone_init(sizeof(xfs_buf_t), "xfs_buf");
+	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",
+						KM_ZONE_HWALIGN, NULL);
 	if (!xfs_buf_zone)
 		goto out_free_trace_buf;
 
@@ -1839,7 +1839,7 @@ xfs_buf_init(void)
 #ifdef XFS_BUF_TRACE
 	ktrace_free(xfs_buf_trace_buf);
 #endif
-	return error;
+	return -ENOMEM;
 }
 
 void
diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c
index 80eb249f2fa0..b768ea910bbe 100644
--- a/fs/xfs/linux-2.6/xfs_export.c
+++ b/fs/xfs/linux-2.6/xfs_export.c
@@ -25,6 +25,8 @@
 #include "xfs_mount.h"
 #include "xfs_export.h"
 
+STATIC struct dentry dotdot = { .d_name.name = "..", .d_name.len = 2, };
+
 /*
  * XFS encodes and decodes the fileid portion of NFS filehandles
  * itself instead of letting the generic NFS code do it.  This
@@ -37,7 +39,7 @@
  */
 
 STATIC struct dentry *
-linvfs_decode_fh(
+xfs_fs_decode_fh(
 	struct super_block	*sb,
 	__u32			*fh,
 	int			fh_len,
@@ -78,12 +80,12 @@ linvfs_decode_fh(
 	}
 
 	fh = (__u32 *)&ifid;
-	return find_exported_dentry(sb, fh, parent, acceptable, context);
+	return sb->s_export_op->find_exported_dentry(sb, fh, parent, acceptable, context);
 }
 
 
 STATIC int
-linvfs_encode_fh(
+xfs_fs_encode_fh(
 	struct dentry		*dentry,
 	__u32			*fh,
 	int			*max_len,
@@ -95,7 +97,7 @@ linvfs_encode_fh(
 	int			len;
 	int			is64 = 0;
 #if XFS_BIG_INUMS
-	vfs_t			*vfs = LINVFS_GET_VFS(inode->i_sb);
+	vfs_t			*vfs = vfs_from_sb(inode->i_sb);
 
 	if (!(vfs->vfs_flag & VFS_32BITINODES)) {
 		/* filesystem may contain 64bit inode numbers */
@@ -130,21 +132,21 @@ linvfs_encode_fh(
 }
 
 STATIC struct dentry *
-linvfs_get_dentry(
+xfs_fs_get_dentry(
 	struct super_block	*sb,
 	void			*data)
 {
 	vnode_t			*vp;
 	struct inode		*inode;
 	struct dentry		*result;
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_VGET(vfsp, &vp, (fid_t *)data, error);
 	if (error || vp == NULL)
 		return ERR_PTR(-ESTALE) ;
 
-	inode = LINVFS_GET_IP(vp);
+	inode = vn_to_inode(vp);
 	result = d_alloc_anon(inode);
         if (!result) {
 		iput(inode);
@@ -154,25 +156,20 @@ linvfs_get_dentry(
 }
 
 STATIC struct dentry *
-linvfs_get_parent(
+xfs_fs_get_parent(
 	struct dentry		*child)
 {
 	int			error;
 	vnode_t			*vp, *cvp;
 	struct dentry		*parent;
-	struct dentry		dotdot;
-
-	dotdot.d_name.name = "..";
-	dotdot.d_name.len = 2;
-	dotdot.d_inode = NULL;
 
 	cvp = NULL;
-	vp = LINVFS_GET_VP(child->d_inode);
+	vp = vn_from_inode(child->d_inode);
 	VOP_LOOKUP(vp, &dotdot, &cvp, 0, NULL, NULL, error);
 	if (unlikely(error))
 		return ERR_PTR(-error);
 
-	parent = d_alloc_anon(LINVFS_GET_IP(cvp));
+	parent = d_alloc_anon(vn_to_inode(cvp));
 	if (unlikely(!parent)) {
 		VN_RELE(cvp);
 		return ERR_PTR(-ENOMEM);
@@ -180,9 +177,9 @@ linvfs_get_parent(
 	return parent;
 }
 
-struct export_operations linvfs_export_ops = {
-	.decode_fh		= linvfs_decode_fh,
-	.encode_fh		= linvfs_encode_fh,
-	.get_parent		= linvfs_get_parent,
-	.get_dentry		= linvfs_get_dentry,
+struct export_operations xfs_export_operations = {
+	.decode_fh		= xfs_fs_decode_fh,
+	.encode_fh		= xfs_fs_encode_fh,
+	.get_parent		= xfs_fs_get_parent,
+	.get_dentry		= xfs_fs_get_dentry,
 };
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index ced4404339c7..185567a6a561 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -43,13 +43,13 @@
 #include <linux/dcache.h>
 #include <linux/smp_lock.h>
 
-static struct vm_operations_struct linvfs_file_vm_ops;
+static struct vm_operations_struct xfs_file_vm_ops;
 #ifdef CONFIG_XFS_DMAPI
-static struct vm_operations_struct linvfs_dmapi_file_vm_ops;
+static struct vm_operations_struct xfs_dmapi_file_vm_ops;
 #endif
 
 STATIC inline ssize_t
-__linvfs_read(
+__xfs_file_read(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	int			ioflags,
@@ -58,7 +58,7 @@ __linvfs_read(
 {
 	struct iovec		iov = {buf, count};
 	struct file		*file = iocb->ki_filp;
-	vnode_t			*vp = LINVFS_GET_VP(file->f_dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(file->f_dentry->d_inode);
 	ssize_t			rval;
 
 	BUG_ON(iocb->ki_pos != pos);
@@ -71,28 +71,28 @@ __linvfs_read(
 
 
 STATIC ssize_t
-linvfs_aio_read(
+xfs_file_aio_read(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_read(iocb, buf, IO_ISAIO, count, pos);
+	return __xfs_file_read(iocb, buf, IO_ISAIO, count, pos);
 }
 
 STATIC ssize_t
-linvfs_aio_read_invis(
+xfs_file_aio_read_invis(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_read(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
+	return __xfs_file_read(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_write(
+__xfs_file_write(
 	struct kiocb	*iocb,
 	const char	__user *buf,
 	int		ioflags,
@@ -102,7 +102,7 @@ __linvfs_write(
 	struct iovec	iov = {(void __user *)buf, count};
 	struct file	*file = iocb->ki_filp;
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	ssize_t		rval;
 
 	BUG_ON(iocb->ki_pos != pos);
@@ -115,28 +115,28 @@ __linvfs_write(
 
 
 STATIC ssize_t
-linvfs_aio_write(
+xfs_file_aio_write(
 	struct kiocb		*iocb,
 	const char		__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_write(iocb, buf, IO_ISAIO, count, pos);
+	return __xfs_file_write(iocb, buf, IO_ISAIO, count, pos);
 }
 
 STATIC ssize_t
-linvfs_aio_write_invis(
+xfs_file_aio_write_invis(
 	struct kiocb		*iocb,
 	const char		__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_write(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
+	return __xfs_file_write(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_readv(
+__xfs_file_readv(
 	struct file		*file,
 	const struct iovec 	*iov,
 	int			ioflags,
@@ -144,8 +144,8 @@ __linvfs_readv(
 	loff_t			*ppos)
 {
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	struct		kiocb kiocb;
+	vnode_t		*vp = vn_from_inode(inode);
+	struct kiocb	kiocb;
 	ssize_t		rval;
 
 	init_sync_kiocb(&kiocb, file);
@@ -160,28 +160,28 @@ __linvfs_readv(
 }
 
 STATIC ssize_t
-linvfs_readv(
+xfs_file_readv(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_readv(file, iov, 0, nr_segs, ppos);
+	return __xfs_file_readv(file, iov, 0, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_readv_invis(
+xfs_file_readv_invis(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_readv(file, iov, IO_INVIS, nr_segs, ppos);
+	return __xfs_file_readv(file, iov, IO_INVIS, nr_segs, ppos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_writev(
+__xfs_file_writev(
 	struct file		*file,
 	const struct iovec 	*iov,
 	int			ioflags,
@@ -189,8 +189,8 @@ __linvfs_writev(
 	loff_t			*ppos)
 {
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	struct		kiocb kiocb;
+	vnode_t		*vp = vn_from_inode(inode);
+	struct kiocb	kiocb;
 	ssize_t		rval;
 
 	init_sync_kiocb(&kiocb, file);
@@ -206,34 +206,34 @@ __linvfs_writev(
 
 
 STATIC ssize_t
-linvfs_writev(
+xfs_file_writev(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_writev(file, iov, 0, nr_segs, ppos);
+	return __xfs_file_writev(file, iov, 0, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_writev_invis(
+xfs_file_writev_invis(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_writev(file, iov, IO_INVIS, nr_segs, ppos);
+	return __xfs_file_writev(file, iov, IO_INVIS, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_sendfile(
+xfs_file_sendfile(
 	struct file		*filp,
 	loff_t			*ppos,
 	size_t			count,
 	read_actor_t		actor,
 	void			*target)
 {
-	vnode_t			*vp = LINVFS_GET_VP(filp->f_dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(filp->f_dentry->d_inode);
 	ssize_t			rval;
 
 	VOP_SENDFILE(vp, filp, ppos, 0, count, actor, target, NULL, rval);
@@ -242,11 +242,11 @@ linvfs_sendfile(
 
 
 STATIC int
-linvfs_open(
+xfs_file_open(
 	struct inode	*inode,
 	struct file	*filp)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 
 	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
@@ -259,11 +259,11 @@ linvfs_open(
 
 
 STATIC int
-linvfs_release(
+xfs_file_release(
 	struct inode	*inode,
 	struct file	*filp)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error = 0;
 
 	if (vp)
@@ -273,13 +273,13 @@ linvfs_release(
 
 
 STATIC int
-linvfs_fsync(
+xfs_file_fsync(
 	struct file	*filp,
 	struct dentry	*dentry,
 	int		datasync)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 	int		flags = FSYNC_WAIT;
 
@@ -292,7 +292,7 @@ linvfs_fsync(
 }
 
 /*
- * linvfs_readdir maps to VOP_READDIR().
+ * xfs_file_readdir maps to VOP_READDIR().
  * We need to build a uio, cred, ...
  */
 
@@ -301,13 +301,13 @@ linvfs_fsync(
 #ifdef CONFIG_XFS_DMAPI
 
 STATIC struct page *
-linvfs_filemap_nopage(
+xfs_vm_nopage(
 	struct vm_area_struct	*area,
 	unsigned long		address,
 	int			*type)
 {
 	struct inode	*inode = area->vm_file->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
 	int		error;
 
@@ -324,7 +324,7 @@ linvfs_filemap_nopage(
 
 
 STATIC int
-linvfs_readdir(
+xfs_file_readdir(
 	struct file	*filp,
 	void		*dirent,
 	filldir_t	filldir)
@@ -340,7 +340,7 @@ linvfs_readdir(
 	xfs_off_t	start_offset, curr_offset;
 	xfs_dirent_t	*dbp = NULL;
 
-	vp = LINVFS_GET_VP(filp->f_dentry->d_inode);
+	vp = vn_from_inode(filp->f_dentry->d_inode);
 	ASSERT(vp);
 
 	/* Try fairly hard to get memory */
@@ -404,39 +404,40 @@ done:
 
 
 STATIC int
-linvfs_file_mmap(
+xfs_file_mmap(
 	struct file	*filp,
 	struct vm_area_struct *vma)
 {
 	struct inode	*ip = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(ip);
-	vattr_t		va = { .va_mask = XFS_AT_UPDATIME };
+	vnode_t		*vp = vn_from_inode(ip);
+	vattr_t		vattr;
 	int		error;
 
-	vma->vm_ops = &linvfs_file_vm_ops;
+	vma->vm_ops = &xfs_file_vm_ops;
 
 #ifdef CONFIG_XFS_DMAPI
 	if (vp->v_vfsp->vfs_flag & VFS_DMI) {
-		vma->vm_ops = &linvfs_dmapi_file_vm_ops;
+		vma->vm_ops = &xfs_dmapi_file_vm_ops;
 	}
 #endif /* CONFIG_XFS_DMAPI */
 
-	VOP_SETATTR(vp, &va, XFS_AT_UPDATIME, NULL, error);
-	if (!error)
-		vn_revalidate(vp);	/* update Linux inode flags */
+	vattr.va_mask = XFS_AT_UPDATIME;
+	VOP_SETATTR(vp, &vattr, XFS_AT_UPDATIME, NULL, error);
+	if (likely(!error))
+		__vn_revalidate(vp, &vattr);	/* update flags */
 	return 0;
 }
 
 
 STATIC long
-linvfs_ioctl(
+xfs_file_ioctl(
 	struct file	*filp,
 	unsigned int	cmd,
 	unsigned long	arg)
 {
 	int		error;
-	struct inode *inode = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	struct inode	*inode = filp->f_dentry->d_inode;
+	vnode_t		*vp = vn_from_inode(inode);
 
 	VOP_IOCTL(vp, inode, filp, 0, cmd, (void __user *)arg, error);
 	VMODIFY(vp);
@@ -451,14 +452,14 @@ linvfs_ioctl(
 }
 
 STATIC long
-linvfs_ioctl_invis(
+xfs_file_ioctl_invis(
 	struct file	*filp,
 	unsigned int	cmd,
 	unsigned long	arg)
 {
 	int		error;
-	struct inode *inode = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	struct inode	*inode = filp->f_dentry->d_inode;
+	vnode_t		*vp = vn_from_inode(inode);
 
 	ASSERT(vp);
 	VOP_IOCTL(vp, inode, filp, IO_INVIS, cmd, (void __user *)arg, error);
@@ -476,11 +477,11 @@ linvfs_ioctl_invis(
 #ifdef CONFIG_XFS_DMAPI
 #ifdef HAVE_VMOP_MPROTECT
 STATIC int
-linvfs_mprotect(
+xfs_vm_mprotect(
 	struct vm_area_struct *vma,
 	unsigned int	newflags)
 {
-	vnode_t		*vp = LINVFS_GET_VP(vma->vm_file->f_dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(vma->vm_file->f_dentry->d_inode);
 	int		error = 0;
 
 	if (vp->v_vfsp->vfs_flag & VFS_DMI) {
@@ -503,10 +504,10 @@ linvfs_mprotect(
  * it back online.
  */
 STATIC int
-linvfs_open_exec(
+xfs_file_open_exec(
 	struct inode	*inode)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
 	int		error = 0;
 	xfs_inode_t	*ip;
@@ -527,69 +528,69 @@ open_exec_out:
 }
 #endif /* HAVE_FOP_OPEN_EXEC */
 
-struct file_operations linvfs_file_operations = {
+struct file_operations xfs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
-	.readv		= linvfs_readv,
-	.writev		= linvfs_writev,
-	.aio_read	= linvfs_aio_read,
-	.aio_write	= linvfs_aio_write,
-	.sendfile	= linvfs_sendfile,
-	.unlocked_ioctl	= linvfs_ioctl,
+	.readv		= xfs_file_readv,
+	.writev		= xfs_file_writev,
+	.aio_read	= xfs_file_aio_read,
+	.aio_write	= xfs_file_aio_write,
+	.sendfile	= xfs_file_sendfile,
+	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_ioctl,
+	.compat_ioctl	= xfs_file_compat_ioctl,
 #endif
-	.mmap		= linvfs_file_mmap,
-	.open		= linvfs_open,
-	.release	= linvfs_release,
-	.fsync		= linvfs_fsync,
+	.mmap		= xfs_file_mmap,
+	.open		= xfs_file_open,
+	.release	= xfs_file_release,
+	.fsync		= xfs_file_fsync,
 #ifdef HAVE_FOP_OPEN_EXEC
-	.open_exec	= linvfs_open_exec,
+	.open_exec	= xfs_file_open_exec,
 #endif
 };
 
-struct file_operations linvfs_invis_file_operations = {
+struct file_operations xfs_invis_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
-	.readv		= linvfs_readv_invis,
-	.writev		= linvfs_writev_invis,
-	.aio_read	= linvfs_aio_read_invis,
-	.aio_write	= linvfs_aio_write_invis,
-	.sendfile	= linvfs_sendfile,
-	.unlocked_ioctl	= linvfs_ioctl_invis,
+	.readv		= xfs_file_readv_invis,
+	.writev		= xfs_file_writev_invis,
+	.aio_read	= xfs_file_aio_read_invis,
+	.aio_write	= xfs_file_aio_write_invis,
+	.sendfile	= xfs_file_sendfile,
+	.unlocked_ioctl	= xfs_file_ioctl_invis,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_invis_ioctl,
+	.compat_ioctl	= xfs_file_compat_invis_ioctl,
 #endif
-	.mmap		= linvfs_file_mmap,
-	.open		= linvfs_open,
-	.release	= linvfs_release,
-	.fsync		= linvfs_fsync,
+	.mmap		= xfs_file_mmap,
+	.open		= xfs_file_open,
+	.release	= xfs_file_release,
+	.fsync		= xfs_file_fsync,
 };
 
 
-struct file_operations linvfs_dir_operations = {
+struct file_operations xfs_dir_file_operations = {
 	.read		= generic_read_dir,
-	.readdir	= linvfs_readdir,
-	.unlocked_ioctl	= linvfs_ioctl,
+	.readdir	= xfs_file_readdir,
+	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_ioctl,
+	.compat_ioctl	= xfs_file_compat_ioctl,
 #endif
-	.fsync		= linvfs_fsync,
+	.fsync		= xfs_file_fsync,
 };
 
-static struct vm_operations_struct linvfs_file_vm_ops = {
+static struct vm_operations_struct xfs_file_vm_ops = {
 	.nopage		= filemap_nopage,
 	.populate	= filemap_populate,
 };
 
 #ifdef CONFIG_XFS_DMAPI
-static struct vm_operations_struct linvfs_dmapi_file_vm_ops = {
-	.nopage		= linvfs_filemap_nopage,
+static struct vm_operations_struct xfs_dmapi_file_vm_ops = {
+	.nopage		= xfs_vm_nopage,
 	.populate	= filemap_populate,
 #ifdef HAVE_VMOP_MPROTECT
-	.mprotect	= linvfs_mprotect,
+	.mprotect	= xfs_vm_mprotect,
 #endif
 };
 #endif /* CONFIG_XFS_DMAPI */
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c
index 4fa4b1a5187e..575f2a790f31 100644
--- a/fs/xfs/linux-2.6/xfs_fs_subr.c
+++ b/fs/xfs/linux-2.6/xfs_fs_subr.c
@@ -57,7 +57,7 @@ fs_tosspages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp))
 		truncate_inode_pages(ip->i_mapping, first);
@@ -76,7 +76,7 @@ fs_flushinval_pages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp)) {
 		filemap_write_and_wait(ip->i_mapping);
@@ -98,7 +98,7 @@ fs_flush_pages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp)) {
 		filemap_fdatawrite(ip->i_mapping);
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
index 4db47790415c..84478491609b 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl.c
@@ -138,7 +138,7 @@ xfs_find_handle(
 	}
 
 	/* we need the vnode */
-	vp = LINVFS_GET_VP(inode);
+	vp = vn_from_inode(inode);
 
 	/* now we can grab the fsid */
 	memcpy(&handle.ha_fsid, vp->v_vfsp->vfs_altfsid, sizeof(xfs_fsid_t));
@@ -256,7 +256,7 @@ xfs_vget_fsop_handlereq(
 	}
 
 	vpp = XFS_ITOV(ip);
-	inodep = LINVFS_GET_IP(vpp);
+	inodep = vn_to_inode(vpp);
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
 	*vp = vpp;
@@ -344,7 +344,7 @@ xfs_open_by_handle(
 		return -XFS_ERROR(-PTR_ERR(filp));
 	}
 	if (inode->i_mode & S_IFREG)
-		filp->f_op = &linvfs_invis_file_operations;
+		filp->f_op = &xfs_invis_file_operations;
 
 	fd_install(new_fd, filp);
 	return new_fd;
@@ -715,7 +715,7 @@ xfs_ioctl(
 	xfs_inode_t		*ip;
 	xfs_mount_t		*mp;
 
-	vp = LINVFS_GET_VP(inode);
+	vp = vn_from_inode(inode);
 
 	vn_trace_entry(vp, "xfs_ioctl", (inst_t *)__return_address);
 
@@ -1160,105 +1160,129 @@ xfs_ioc_xattr(
 	void			__user *arg)
 {
 	struct fsxattr		fa;
-	vattr_t			va;
-	int			error;
+	struct vattr		*vattr;
+	int			error = 0;
 	int			attr_flags;
 	unsigned int		flags;
 
+	vattr = kmalloc(sizeof(*vattr), GFP_KERNEL);
+	if (unlikely(!vattr))
+		return -ENOMEM;
+
 	switch (cmd) {
 	case XFS_IOC_FSGETXATTR: {
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-			     XFS_AT_NEXTENTS | XFS_AT_PROJID;
-		VOP_GETATTR(vp, &va, 0, NULL, error);
-		if (error)
-			return -error;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
+				 XFS_AT_NEXTENTS | XFS_AT_PROJID;
+		VOP_GETATTR(vp, vattr, 0, NULL, error);
+		if (unlikely(error)) {
+			error = -error;
+			break;
+		}
 
-		fa.fsx_xflags	= va.va_xflags;
-		fa.fsx_extsize	= va.va_extsize;
-		fa.fsx_nextents = va.va_nextents;
-		fa.fsx_projid	= va.va_projid;
+		fa.fsx_xflags	= vattr->va_xflags;
+		fa.fsx_extsize	= vattr->va_extsize;
+		fa.fsx_nextents = vattr->va_nextents;
+		fa.fsx_projid	= vattr->va_projid;
 
-		if (copy_to_user(arg, &fa, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+		if (copy_to_user(arg, &fa, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
+		break;
 	}
 
 	case XFS_IOC_FSSETXATTR: {
-		if (copy_from_user(&fa, arg, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
+		if (copy_from_user(&fa, arg, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
 
 		attr_flags = 0;
 		if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
 			attr_flags |= ATTR_NONBLOCK;
 
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID;
-		va.va_xflags  = fa.fsx_xflags;
-		va.va_extsize = fa.fsx_extsize;
-		va.va_projid  = fa.fsx_projid;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID;
+		vattr->va_xflags  = fa.fsx_xflags;
+		vattr->va_extsize = fa.fsx_extsize;
+		vattr->va_projid  = fa.fsx_projid;
 
-		VOP_SETATTR(vp, &va, attr_flags, NULL, error);
-		if (!error)
-			vn_revalidate(vp);	/* update Linux inode flags */
-		return -error;
+		VOP_SETATTR(vp, vattr, attr_flags, NULL, error);
+		if (likely(!error))
+			__vn_revalidate(vp, vattr);	/* update flags */
+		error = -error;
+		break;
 	}
 
 	case XFS_IOC_FSGETXATTRA: {
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-			     XFS_AT_ANEXTENTS | XFS_AT_PROJID;
-		VOP_GETATTR(vp, &va, 0, NULL, error);
-		if (error)
-			return -error;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
+				 XFS_AT_ANEXTENTS | XFS_AT_PROJID;
+		VOP_GETATTR(vp, vattr, 0, NULL, error);
+		if (unlikely(error)) {
+			error = -error;
+			break;
+		}
 
-		fa.fsx_xflags	= va.va_xflags;
-		fa.fsx_extsize	= va.va_extsize;
-		fa.fsx_nextents = va.va_anextents;
-		fa.fsx_projid	= va.va_projid;
+		fa.fsx_xflags	= vattr->va_xflags;
+		fa.fsx_extsize	= vattr->va_extsize;
+		fa.fsx_nextents = vattr->va_anextents;
+		fa.fsx_projid	= vattr->va_projid;
 
-		if (copy_to_user(arg, &fa, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+		if (copy_to_user(arg, &fa, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
+		break;
 	}
 
 	case XFS_IOC_GETXFLAGS: {
 		flags = xfs_di2lxflags(ip->i_d.di_flags);
 		if (copy_to_user(arg, &flags, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+			error = -EFAULT;
+		break;
 	}
 
 	case XFS_IOC_SETXFLAGS: {
-		if (copy_from_user(&flags, arg, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
+		if (copy_from_user(&flags, arg, sizeof(flags))) {
+			error = -EFAULT;
+			break;
+		}
 
 		if (flags & ~(LINUX_XFLAG_IMMUTABLE | LINUX_XFLAG_APPEND | \
 			      LINUX_XFLAG_NOATIME | LINUX_XFLAG_NODUMP | \
-			      LINUX_XFLAG_SYNC))
-			return -XFS_ERROR(EOPNOTSUPP);
+			      LINUX_XFLAG_SYNC)) {
+			error = -EOPNOTSUPP;
+			break;
+		}
 
 		attr_flags = 0;
 		if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
 			attr_flags |= ATTR_NONBLOCK;
 
-		va.va_mask = XFS_AT_XFLAGS;
-		va.va_xflags = xfs_merge_ioc_xflags(flags,
-				xfs_ip2xflags(ip));
+		vattr->va_mask = XFS_AT_XFLAGS;
+		vattr->va_xflags = xfs_merge_ioc_xflags(flags,
+							xfs_ip2xflags(ip));
 
-		VOP_SETATTR(vp, &va, attr_flags, NULL, error);
-		if (!error)
-			vn_revalidate(vp);	/* update Linux inode flags */
-		return -error;
+		VOP_SETATTR(vp, vattr, attr_flags, NULL, error);
+		if (likely(!error))
+			__vn_revalidate(vp, vattr);	/* update flags */
+		error = -error;
+		break;
 	}
 
 	case XFS_IOC_GETVERSION: {
-		flags = LINVFS_GET_IP(vp)->i_generation;
+		flags = vn_to_inode(vp)->i_generation;
 		if (copy_to_user(arg, &flags, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+			error = -EFAULT;
+		break;
 	}
 
 	default:
-		return -ENOTTY;
+		error = -ENOTTY;
+		break;
 	}
+
+	kfree(vattr);
+	return error;
 }
 
 STATIC int
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
index a7c9ba1a9f7b..b6321abd9a81 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.c
@@ -107,11 +107,11 @@ xfs_ioctl32_bulkstat(
 #endif
 
 STATIC long
-__linvfs_compat_ioctl(int mode, struct file *f, unsigned cmd, unsigned long arg)
+xfs_compat_ioctl(int mode, struct file *f, unsigned cmd, unsigned long arg)
 {
 	int		error;
 	struct		inode *inode = f->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_to_inode(inode);
 
 	switch (cmd) {
 	case XFS_IOC_DIOINFO:
@@ -196,19 +196,19 @@ __linvfs_compat_ioctl(int mode, struct file *f, unsigned cmd, unsigned long arg)
 }
 
 long
-linvfs_compat_ioctl(
+xfs_file_compat_ioctl(
 	struct file		*f,
 	unsigned		cmd,
 	unsigned long		arg)
 {
-	return __linvfs_compat_ioctl(0, f, cmd, arg);
+	return xfs_compat_ioctl(0, f, cmd, arg);
 }
 
 long
-linvfs_compat_invis_ioctl(
+xfs_file_compat_invis_ioctl(
 	struct file		*f,
 	unsigned		cmd,
 	unsigned long		arg)
 {
-	return __linvfs_compat_ioctl(IO_INVIS, f, cmd, arg);
+	return xfs_compat_ioctl(IO_INVIS, f, cmd, arg);
 }
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.h b/fs/xfs/linux-2.6/xfs_ioctl32.h
index 011c273bec50..02de6e62ee37 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.h
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.h
@@ -18,7 +18,7 @@
 #ifndef __XFS_IOCTL32_H__
 #define __XFS_IOCTL32_H__
 
-extern long linvfs_compat_ioctl(struct file *, unsigned, unsigned long);
-extern long linvfs_compat_invis_ioctl(struct file *f, unsigned, unsigned long);
+extern long xfs_file_compat_ioctl(struct file *, unsigned, unsigned long);
+extern long xfs_file_compat_invis_ioctl(struct file *, unsigned, unsigned long);
 
 #endif /* __XFS_IOCTL32_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index d7f6f2d8ac8e..af487437bd7e 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -106,7 +106,7 @@ xfs_ichgtime(
 	xfs_inode_t	*ip,
 	int		flags)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	timespec_t	tv;
 
 	nanotime(&tv);
@@ -198,22 +198,22 @@ xfs_ichgtime_fast(
  * Pull the link count and size up from the xfs inode to the linux inode
  */
 STATIC void
-validate_fields(
-	struct inode	*ip)
+xfs_validate_fields(
+	struct inode	*ip,
+	struct vattr	*vattr)
 {
-	vnode_t		*vp = LINVFS_GET_VP(ip);
-	vattr_t		va;
+	vnode_t		*vp = vn_from_inode(ip);
 	int		error;
 
-	va.va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
-	VOP_GETATTR(vp, &va, ATTR_LAZY, NULL, error);
-	if (likely(!error)) {
-		ip->i_nlink = va.va_nlink;
-		ip->i_blocks = va.va_nblocks;
+	vattr->va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
+	VOP_GETATTR(vp, vattr, ATTR_LAZY, NULL, error);
+  	if (likely(!error)) {
+		ip->i_nlink = vattr->va_nlink;
+		ip->i_blocks = vattr->va_nblocks;
 
-		/* we're under i_mutex so i_size can't change under us */
-		if (i_size_read(ip) != va.va_size)
-			i_size_write(ip, va.va_size);
+		/* we're under i_sem so i_size can't change under us */
+		if (i_size_read(ip) != vattr->va_size)
+			i_size_write(ip, vattr->va_size);
 	}
 }
 
@@ -224,11 +224,11 @@ validate_fields(
  * inode, of course, such that log replay can't cause these to be lost).
  */
 STATIC int
-linvfs_init_security(
+xfs_init_security(
 	struct vnode	*vp,
 	struct inode	*dir)
 {
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 	size_t		length;
 	void		*value;
 	char		*name;
@@ -257,46 +257,46 @@ linvfs_init_security(
  * XXX(hch):  nfsd is broken, better fix it instead.
  */
 STATIC inline int
-has_fs_struct(struct task_struct *task)
+xfs_has_fs_struct(struct task_struct *task)
 {
 	return (task->fs != init_task.fs);
 }
 
 STATIC inline void
-cleanup_inode(
+xfs_cleanup_inode(
 	vnode_t		*dvp,
 	vnode_t		*vp,
-	struct dentry	*dentry,	
+	struct dentry	*dentry,
 	int		mode)
 {
 	struct dentry   teardown = {};
-	int             err2;
+	int             error;
 
 	/* Oh, the horror.
-	 * If we can't add the ACL or we fail in 
-	 * linvfs_init_security we must back out.
+	 * 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.
 	 */
-	teardown.d_inode = LINVFS_GET_IP(vp);
+	teardown.d_inode = vn_to_inode(vp);
 	teardown.d_name = dentry->d_name;
 
 	if (S_ISDIR(mode))
-	  	VOP_RMDIR(dvp, &teardown, NULL, err2);
+	  	VOP_RMDIR(dvp, &teardown, NULL, error);
 	else
-		VOP_REMOVE(dvp, &teardown, NULL, err2);
+		VOP_REMOVE(dvp, &teardown, NULL, error);
 	VN_RELE(vp);
 }
 
 STATIC int
-linvfs_mknod(
+xfs_vn_mknod(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode,
 	dev_t		rdev)
 {
 	struct inode	*ip;
-	vattr_t		va;
-	vnode_t		*vp = NULL, *dvp = LINVFS_GET_VP(dir);
+	vattr_t		vattr = { 0 };
+	vnode_t		*vp = NULL, *dvp = vn_from_inode(dir);
 	xfs_acl_t	*default_acl = NULL;
 	attrexists_t	test_default_acl = _ACL_DEFAULT_EXISTS;
 	int		error;
@@ -305,99 +305,98 @@ linvfs_mknod(
 	 * Irix uses Missed'em'V split, but doesn't want to see
 	 * the upper 5 bits of (14bit) major.
 	 */
-	if (!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff)
+	if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff))
 		return -EINVAL;
 
-	if (test_default_acl && test_default_acl(dvp)) {
-		if (!_ACL_ALLOC(default_acl))
+	if (unlikely(test_default_acl && test_default_acl(dvp))) {
+		if (!_ACL_ALLOC(default_acl)) {
 			return -ENOMEM;
+		}
 		if (!_ACL_GET_DEFAULT(dvp, default_acl)) {
 			_ACL_FREE(default_acl);
 			default_acl = NULL;
 		}
 	}
 
-	if (IS_POSIXACL(dir) && !default_acl && has_fs_struct(current))
+	if (IS_POSIXACL(dir) && !default_acl && xfs_has_fs_struct(current))
 		mode &= ~current->fs->umask;
 
-	memset(&va, 0, sizeof(va));
-	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
-	va.va_mode = mode;
+	vattr.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
+	vattr.va_mode = mode;
 
 	switch (mode & S_IFMT) {
 	case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK:
-		va.va_rdev = sysv_encode_dev(rdev);
-		va.va_mask |= XFS_AT_RDEV;
+		vattr.va_rdev = sysv_encode_dev(rdev);
+		vattr.va_mask |= XFS_AT_RDEV;
 		/*FALLTHROUGH*/
 	case S_IFREG:
-		VOP_CREATE(dvp, dentry, &va, &vp, NULL, error);
+		VOP_CREATE(dvp, dentry, &vattr, &vp, NULL, error);
 		break;
 	case S_IFDIR:
-		VOP_MKDIR(dvp, dentry, &va, &vp, NULL, error);
+		VOP_MKDIR(dvp, dentry, &vattr, &vp, NULL, error);
 		break;
 	default:
 		error = EINVAL;
 		break;
 	}
 
-	if (!error)
-	{
-		error = linvfs_init_security(vp, dir);
+	if (unlikely(!error)) {
+		error = xfs_init_security(vp, dir);
 		if (error)
-			cleanup_inode(dvp, vp, dentry, mode);
+			xfs_cleanup_inode(dvp, vp, dentry, mode);
 	}
 
-	if (default_acl) {
+	if (unlikely(default_acl)) {
 		if (!error) {
-			error = _ACL_INHERIT(vp, &va, default_acl);
-			if (!error) 
+			error = _ACL_INHERIT(vp, &vattr, default_acl);
+			if (!error)
 				VMODIFY(vp);
 			else
-				cleanup_inode(dvp, vp, dentry, mode);
+				xfs_cleanup_inode(dvp, vp, dentry, mode);
 		}
 		_ACL_FREE(default_acl);
 	}
 
-	if (!error) {
+	if (likely(!error)) {
 		ASSERT(vp);
-		ip = LINVFS_GET_IP(vp);
+		ip = vn_to_inode(vp);
 
 		if (S_ISCHR(mode) || S_ISBLK(mode))
 			ip->i_rdev = rdev;
 		else if (S_ISDIR(mode))
-			validate_fields(ip);
+			xfs_validate_fields(ip, &vattr);
 		d_instantiate(dentry, ip);
-		validate_fields(dir);
+		xfs_validate_fields(dir, &vattr);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_create(
+xfs_vn_create(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode,
 	struct nameidata *nd)
 {
-	return linvfs_mknod(dir, dentry, mode, 0);
+	return xfs_vn_mknod(dir, dentry, mode, 0);
 }
 
 STATIC int
-linvfs_mkdir(
+xfs_vn_mkdir(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode)
 {
-	return linvfs_mknod(dir, dentry, mode|S_IFDIR, 0);
+	return xfs_vn_mknod(dir, dentry, mode|S_IFDIR, 0);
 }
 
 STATIC struct dentry *
-linvfs_lookup(
+xfs_vn_lookup(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	struct nameidata *nd)
 {
-	struct vnode	*vp = LINVFS_GET_VP(dir), *cvp;
+	struct vnode	*vp = vn_from_inode(dir), *cvp;
 	int		error;
 
 	if (dentry->d_name.len >= MAXNAMELEN)
@@ -411,11 +410,11 @@ linvfs_lookup(
 		return NULL;
 	}
 
-	return d_splice_alias(LINVFS_GET_IP(cvp), dentry);
+	return d_splice_alias(vn_to_inode(cvp), dentry);
 }
 
 STATIC int
-linvfs_link(
+xfs_vn_link(
 	struct dentry	*old_dentry,
 	struct inode	*dir,
 	struct dentry	*dentry)
@@ -423,99 +422,102 @@ linvfs_link(
 	struct inode	*ip;	/* inode of guy being linked to */
 	vnode_t		*tdvp;	/* target directory for new name/link */
 	vnode_t		*vp;	/* vp of name being linked */
+	vattr_t		vattr;
 	int		error;
 
 	ip = old_dentry->d_inode;	/* inode being linked to */
 	if (S_ISDIR(ip->i_mode))
 		return -EPERM;
 
-	tdvp = LINVFS_GET_VP(dir);
-	vp = LINVFS_GET_VP(ip);
+	tdvp = vn_from_inode(dir);
+	vp = vn_from_inode(ip);
 
 	VOP_LINK(tdvp, vp, dentry, NULL, error);
-	if (!error) {
+	if (likely(!error)) {
 		VMODIFY(tdvp);
 		VN_HOLD(vp);
-		validate_fields(ip);
+		xfs_validate_fields(ip, &vattr);
 		d_instantiate(dentry, ip);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_unlink(
+xfs_vn_unlink(
 	struct inode	*dir,
 	struct dentry	*dentry)
 {
 	struct inode	*inode;
 	vnode_t		*dvp;	/* directory containing name to remove */
+	vattr_t		vattr;
 	int		error;
 
 	inode = dentry->d_inode;
-	dvp = LINVFS_GET_VP(dir);
+	dvp = vn_from_inode(dir);
 
 	VOP_REMOVE(dvp, dentry, NULL, error);
-	if (!error) {
-		validate_fields(dir);	/* For size only */
-		validate_fields(inode);
+	if (likely(!error)) {
+		xfs_validate_fields(dir, &vattr);	/* size needs update */
+		xfs_validate_fields(inode, &vattr);
 	}
-
 	return -error;
 }
 
 STATIC int
-linvfs_symlink(
+xfs_vn_symlink(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	const char	*symname)
 {
 	struct inode	*ip;
-	vattr_t		va;
+	vattr_t		vattr = { 0 };
 	vnode_t		*dvp;	/* directory containing name of symlink */
 	vnode_t		*cvp;	/* used to lookup symlink to put in dentry */
 	int		error;
 
-	dvp = LINVFS_GET_VP(dir);
+	dvp = vn_from_inode(dir);
 	cvp = NULL;
 
-	memset(&va, 0, sizeof(va));
-	va.va_mode = S_IFLNK |
+	vattr.va_mode = S_IFLNK |
 		(irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO);
-	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
+	vattr.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
 
 	error = 0;
-	VOP_SYMLINK(dvp, dentry, &va, (char *)symname, &cvp, NULL, error);
+	VOP_SYMLINK(dvp, dentry, &vattr, (char *)symname, &cvp, NULL, error);
 	if (likely(!error && cvp)) {
-		error = linvfs_init_security(cvp, dir);
+		error = xfs_init_security(cvp, dir);
 		if (likely(!error)) {
-			ip = LINVFS_GET_IP(cvp);
+			ip = vn_to_inode(cvp);
 			d_instantiate(dentry, ip);
-			validate_fields(dir);
-			validate_fields(ip);
+			xfs_validate_fields(dir, &vattr);
+			xfs_validate_fields(ip, &vattr);
+		} else {
+			xfs_cleanup_inode(dvp, cvp, dentry, 0);
 		}
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_rmdir(
+xfs_vn_rmdir(
 	struct inode	*dir,
 	struct dentry	*dentry)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*dvp = LINVFS_GET_VP(dir);
+	vnode_t		*dvp = vn_from_inode(dir);
+	vattr_t		vattr;
 	int		error;
 
 	VOP_RMDIR(dvp, dentry, NULL, error);
-	if (!error) {
-		validate_fields(inode);
-		validate_fields(dir);
+	if (likely(!error)) {
+		xfs_validate_fields(inode, &vattr);
+		xfs_validate_fields(dir, &vattr);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_rename(
+xfs_vn_rename(
 	struct inode	*odir,
 	struct dentry	*odentry,
 	struct inode	*ndir,
@@ -524,22 +526,21 @@ linvfs_rename(
 	struct inode	*new_inode = ndentry->d_inode;
 	vnode_t		*fvp;	/* from directory */
 	vnode_t		*tvp;	/* target directory */
+	vattr_t		vattr;
 	int		error;
 
-	fvp = LINVFS_GET_VP(odir);
-	tvp = LINVFS_GET_VP(ndir);
+	fvp = vn_from_inode(odir);
+	tvp = vn_from_inode(ndir);
 
 	VOP_RENAME(fvp, odentry, tvp, ndentry, NULL, error);
-	if (error)
-		return -error;
-
-	if (new_inode)
-		validate_fields(new_inode);
-
-	validate_fields(odir);
-	if (ndir != odir)
-		validate_fields(ndir);
-	return 0;
+	if (likely(!error)) {
+		if (new_inode)
+			xfs_validate_fields(new_inode, &vattr);
+		xfs_validate_fields(odir, &vattr);
+		if (ndir != odir)
+			xfs_validate_fields(ndir, &vattr);
+	}
+	return -error;
 }
 
 /*
@@ -548,7 +549,7 @@ linvfs_rename(
  * uio is kmalloced for this reason...
  */
 STATIC void *
-linvfs_follow_link(
+xfs_vn_follow_link(
 	struct dentry		*dentry,
 	struct nameidata	*nd)
 {
@@ -574,7 +575,7 @@ linvfs_follow_link(
 		return NULL;
 	}
 
-	vp = LINVFS_GET_VP(dentry->d_inode);
+	vp = vn_from_inode(dentry->d_inode);
 
 	iov.iov_base = link;
 	iov.iov_len = MAXPATHLEN;
@@ -599,7 +600,7 @@ linvfs_follow_link(
 }
 
 STATIC void
-linvfs_put_link(
+xfs_vn_put_link(
 	struct dentry	*dentry,
 	struct nameidata *nd,
 	void		*p)
@@ -612,12 +613,12 @@ linvfs_put_link(
 
 #ifdef CONFIG_XFS_POSIX_ACL
 STATIC int
-linvfs_permission(
+xfs_vn_permission(
 	struct inode	*inode,
 	int		mode,
 	struct nameidata *nd)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 
 	mode <<= 6;		/* convert from linux to vnode access bits */
@@ -625,17 +626,17 @@ linvfs_permission(
 	return -error;
 }
 #else
-#define linvfs_permission NULL
+#define xfs_vn_permission NULL
 #endif
 
 STATIC int
-linvfs_getattr(
+xfs_vn_getattr(
 	struct vfsmount	*mnt,
 	struct dentry	*dentry,
 	struct kstat	*stat)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error = 0;
 
 	if (unlikely(vp->v_flag & VMODIFIED))
@@ -646,18 +647,17 @@ linvfs_getattr(
 }
 
 STATIC int
-linvfs_setattr(
+xfs_vn_setattr(
 	struct dentry	*dentry,
 	struct iattr	*attr)
 {
 	struct inode	*inode = dentry->d_inode;
 	unsigned int	ia_valid = attr->ia_valid;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	vattr_t		vattr;
+	vnode_t		*vp = vn_from_inode(inode);
+	vattr_t		vattr = { 0 };
 	int		flags = 0;
 	int		error;
 
-	memset(&vattr, 0, sizeof(vattr_t));
 	if (ia_valid & ATTR_UID) {
 		vattr.va_mask |= XFS_AT_UID;
 		vattr.va_uid = attr->ia_uid;
@@ -699,28 +699,27 @@ linvfs_setattr(
 #endif
 
 	VOP_SETATTR(vp, &vattr, flags, NULL, error);
-	if (error)
-		return -error;
-	vn_revalidate(vp);
-	return error;
+	if (likely(!error))
+		__vn_revalidate(vp, &vattr);
+	return -error;
 }
 
 STATIC void
-linvfs_truncate(
+xfs_vn_truncate(
 	struct inode	*inode)
 {
-	block_truncate_page(inode->i_mapping, inode->i_size, linvfs_get_block);
+	block_truncate_page(inode->i_mapping, inode->i_size, xfs_get_block);
 }
 
 STATIC int
-linvfs_setxattr(
+xfs_vn_setxattr(
 	struct dentry	*dentry,
 	const char	*name,
 	const void	*data,
 	size_t		size,
 	int		flags)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -744,13 +743,13 @@ linvfs_setxattr(
 }
 
 STATIC ssize_t
-linvfs_getxattr(
+xfs_vn_getxattr(
 	struct dentry	*dentry,
 	const char	*name,
 	void		*data,
 	size_t		size)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -774,12 +773,12 @@ linvfs_getxattr(
 }
 
 STATIC ssize_t
-linvfs_listxattr(
+xfs_vn_listxattr(
 	struct dentry		*dentry,
 	char			*data,
 	size_t			size)
 {
-	vnode_t			*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(dentry->d_inode);
 	int			error, xflags = ATTR_KERNAMELS;
 	ssize_t			result;
 
@@ -794,11 +793,11 @@ linvfs_listxattr(
 }
 
 STATIC int
-linvfs_removexattr(
+xfs_vn_removexattr(
 	struct dentry	*dentry,
 	const char	*name)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -816,45 +815,45 @@ linvfs_removexattr(
 }
 
 
-struct inode_operations linvfs_file_inode_operations = {
-	.permission		= linvfs_permission,
-	.truncate		= linvfs_truncate,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+struct inode_operations xfs_inode_operations = {
+	.permission		= xfs_vn_permission,
+	.truncate		= xfs_vn_truncate,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
 
-struct inode_operations linvfs_dir_inode_operations = {
-	.create			= linvfs_create,
-	.lookup			= linvfs_lookup,
-	.link			= linvfs_link,
-	.unlink			= linvfs_unlink,
-	.symlink		= linvfs_symlink,
-	.mkdir			= linvfs_mkdir,
-	.rmdir			= linvfs_rmdir,
-	.mknod			= linvfs_mknod,
-	.rename			= linvfs_rename,
-	.permission		= linvfs_permission,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+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,
+	.rmdir			= xfs_vn_rmdir,
+	.mknod			= xfs_vn_mknod,
+	.rename			= xfs_vn_rename,
+	.permission		= xfs_vn_permission,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
 
-struct inode_operations linvfs_symlink_inode_operations = {
+struct inode_operations xfs_symlink_inode_operations = {
 	.readlink		= generic_readlink,
-	.follow_link		= linvfs_follow_link,
-	.put_link		= linvfs_put_link,
-	.permission		= linvfs_permission,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+	.follow_link		= xfs_vn_follow_link,
+	.put_link		= xfs_vn_put_link,
+	.permission		= xfs_vn_permission,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
diff --git a/fs/xfs/linux-2.6/xfs_iops.h b/fs/xfs/linux-2.6/xfs_iops.h
index 6899a6b4a50a..a8417d7af5f9 100644
--- a/fs/xfs/linux-2.6/xfs_iops.h
+++ b/fs/xfs/linux-2.6/xfs_iops.h
@@ -18,13 +18,13 @@
 #ifndef __XFS_IOPS_H__
 #define __XFS_IOPS_H__
 
-extern struct inode_operations linvfs_file_inode_operations;
-extern struct inode_operations linvfs_dir_inode_operations;
-extern struct inode_operations linvfs_symlink_inode_operations;
+extern struct inode_operations xfs_inode_operations;
+extern struct inode_operations xfs_dir_inode_operations;
+extern struct inode_operations xfs_symlink_inode_operations;
 
-extern struct file_operations linvfs_file_operations;
-extern struct file_operations linvfs_invis_file_operations;
-extern struct file_operations linvfs_dir_operations;
+extern struct file_operations xfs_file_operations;
+extern struct file_operations xfs_dir_file_operations;
+extern struct file_operations xfs_invis_file_operations;
 
 extern int xfs_ioctl(struct bhv_desc *, struct inode *, struct file *,
                         int, unsigned int, void __user *);
diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h
index 67389b745526..1fe09f2d6519 100644
--- a/fs/xfs/linux-2.6/xfs_linux.h
+++ b/fs/xfs/linux-2.6/xfs_linux.h
@@ -73,6 +73,9 @@
 #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 <asm/page.h>
 #include <asm/div64.h>
@@ -100,6 +103,11 @@
  */
 #undef  HAVE_REFCACHE	/* reference cache not needed for NFS in 2.6 */
 #define HAVE_SENDFILE	/* sendfile(2) exists in 2.6, but not in 2.4 */
+#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
 
 /*
  * State flag for unwritten extent buffers.
@@ -226,7 +234,7 @@ BUFFER_FNS(PrivateStart, unwritten);
 #define xfs_sort(a,n,s,fn)	sort(a,n,s,fn,NULL)
 #define xfs_stack_trace()	dump_stack()
 #define xfs_itruncate_data(ip, off)	\
-	(-vmtruncate(LINVFS_GET_IP(XFS_ITOV(ip)), (off)))
+	(-vmtruncate(vn_to_inode(XFS_ITOV(ip)), (off)))
 #define xfs_statvfs_fsid(statp, mp)	\
 	({ u64 id = huge_encode_dev((mp)->m_ddev_targp->bt_dev); \
 	   __kernel_fsid_t *fsid = &(statp)->f_fsid;	\
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index e0ab45fbfebd..0169360475c4 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -83,7 +83,7 @@ xfs_rw_enter_trace(
 		(void *)((unsigned long)ioflags),
 		(void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(io->io_new_size & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
@@ -113,7 +113,7 @@ xfs_inval_cached_trace(
 		(void *)((unsigned long)(first & 0xffffffff)),
 		(void *)((unsigned long)((last >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(last & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
@@ -249,9 +249,8 @@ xfs_read(
 	if (n < size)
 		size = n;
 
-	if (XFS_FORCED_SHUTDOWN(mp)) {
+	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
-	}
 
 	if (unlikely(ioflags & IO_ISDIRECT))
 		mutex_lock(&inode->i_mutex);
@@ -267,10 +266,14 @@ xfs_read(
 					dmflags, &locktype);
 		if (ret) {
 			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-			goto unlock_isem;
+			goto unlock_mutex;
 		}
 	}
 
+	if (unlikely((ioflags & IO_ISDIRECT) && VN_CACHED(vp)))
+		VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(*offset)),
+						-1, FI_REMAPF_LOCKED);
+
 	xfs_rw_enter_trace(XFS_READ_ENTER, &ip->i_iocore,
 				(void *)iovp, segs, *offset, ioflags);
 	ret = __generic_file_aio_read(iocb, iovp, segs, offset);
@@ -281,7 +284,7 @@ xfs_read(
 
 	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
-unlock_isem:
+unlock_mutex:
 	if (unlikely(ioflags & IO_ISDIRECT))
 		mutex_unlock(&inode->i_mutex);
 	return ret;
@@ -432,7 +435,7 @@ xfs_zero_eof(
 	xfs_fsize_t	isize,		/* current inode size */
 	xfs_fsize_t	end_size)	/* terminal inode size */
 {
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 	xfs_fileoff_t	start_zero_fsb;
 	xfs_fileoff_t	end_zero_fsb;
 	xfs_fileoff_t	zero_count_fsb;
@@ -573,7 +576,7 @@ xfs_write(
 	vrwlock_t		locktype;
 	size_t			ocount = 0, count;
 	loff_t			pos;
-	int			need_isem = 1, need_flush = 0;
+	int			need_i_mutex = 1, need_flush = 0;
 
 	XFS_STATS_INC(xs_write_calls);
 
@@ -622,14 +625,14 @@ xfs_write(
 			return XFS_ERROR(-EINVAL);
 
 		if (!VN_CACHED(vp) && pos < i_size_read(inode))
-			need_isem = 0;
+			need_i_mutex = 0;
 
 		if (VN_CACHED(vp))
 			need_flush = 1;
 	}
 
 relock:
-	if (need_isem) {
+	if (need_i_mutex) {
 		iolock = XFS_IOLOCK_EXCL;
 		locktype = VRWLOCK_WRITE;
 
@@ -651,7 +654,7 @@ start:
 					S_ISBLK(inode->i_mode));
 	if (error) {
 		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-		goto out_unlock_isem;
+		goto out_unlock_mutex;
 	}
 
 	new_size = pos + count;
@@ -663,7 +666,7 @@ start:
 		loff_t		savedsize = pos;
 		int		dmflags = FILP_DELAY_FLAG(file);
 
-		if (need_isem)
+		if (need_i_mutex)
 			dmflags |= DM_FLAGS_IMUX;
 
 		xfs_iunlock(xip, XFS_ILOCK_EXCL);
@@ -672,7 +675,7 @@ start:
 				      dmflags, &locktype);
 		if (error) {
 			xfs_iunlock(xip, iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 		xfs_ilock(xip, XFS_ILOCK_EXCL);
 		eventsent = 1;
@@ -710,7 +713,7 @@ start:
 					isize, pos + count);
 		if (error) {
 			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 	}
 	xfs_iunlock(xip, XFS_ILOCK_EXCL);
@@ -731,7 +734,7 @@ start:
 			error = -remove_suid(file->f_dentry);
 		if (unlikely(error)) {
 			xfs_iunlock(xip, iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 	}
 
@@ -747,14 +750,14 @@ retry:
 					-1, FI_REMAPF_LOCKED);
 		}
 
-		if (need_isem) {
+		if (need_i_mutex) {
 			/* demote the lock now the cached pages are gone */
 			XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
 			mutex_unlock(&inode->i_mutex);
 
 			iolock = XFS_IOLOCK_SHARED;
 			locktype = VRWLOCK_WRITE_DIRECT;
-			need_isem = 0;
+			need_i_mutex = 0;
 		}
 
  		xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
@@ -772,7 +775,7 @@ retry:
 			pos += ret;
 			count -= ret;
 
-			need_isem = 1;
+			need_i_mutex = 1;
 			ioflags &= ~IO_ISDIRECT;
 			xfs_iunlock(xip, iolock);
 			goto relock;
@@ -794,14 +797,14 @@ retry:
 	    !(ioflags & IO_INVIS)) {
 
 		xfs_rwunlock(bdp, locktype);
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
 		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
 				DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
 				0, 0, 0); /* Delay flag intentionally  unused */
 		if (error)
 			goto out_nounlocks;
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
 		xfs_rwlock(bdp, locktype);
 		pos = xip->i_d.di_size;
@@ -905,9 +908,9 @@ retry:
 			if (error)
 				goto out_unlock_internal;
 		}
-	
+
 		xfs_rwunlock(bdp, locktype);
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
 
 		error = sync_page_range(inode, mapping, pos, ret);
@@ -918,8 +921,8 @@ retry:
 
  out_unlock_internal:
 	xfs_rwunlock(bdp, locktype);
- out_unlock_isem:
-	if (need_isem)
+ out_unlock_mutex:
+	if (need_i_mutex)
 		mutex_unlock(&inode->i_mutex);
  out_nounlocks:
 	return -error;
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
index 8955720a2c6b..713e6a7505d0 100644
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ b/fs/xfs/linux-2.6/xfs_stats.c
@@ -62,18 +62,15 @@ xfs_read_xfsstats(
 		while (j < xstats[i].endpoint) {
 			val = 0;
 			/* sum over all cpus */
-			for (c = 0; c < NR_CPUS; c++) {
-				if (!cpu_possible(c)) continue;
+			for_each_cpu(c)
 				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
-			}
 			len += sprintf(buffer + len, " %u", val);
 			j++;
 		}
 		buffer[len++] = '\n';
 	}
 	/* extra precision counters */
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i)) continue;
+	for_each_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;
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index f22e426d9e42..8355faf8ffde 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -59,8 +59,8 @@
 #include <linux/writeback.h>
 #include <linux/kthread.h>
 
-STATIC struct quotactl_ops linvfs_qops;
-STATIC struct super_operations linvfs_sops;
+STATIC struct quotactl_ops xfs_quotactl_operations;
+STATIC struct super_operations xfs_super_operations;
 STATIC kmem_zone_t *xfs_vnode_zone;
 STATIC kmem_zone_t *xfs_ioend_zone;
 mempool_t *xfs_ioend_pool;
@@ -76,8 +76,6 @@ xfs_args_allocate(
 	strncpy(args->fsname, sb->s_id, MAXNAMELEN);
 
 	/* Copy the already-parsed mount(2) flags we're interested in */
-	if (sb->s_flags & MS_NOATIME)
-		args->flags |= XFSMNT_NOATIME;
 	if (sb->s_flags & MS_DIRSYNC)
 		args->flags |= XFSMNT_DIRSYNC;
 	if (sb->s_flags & MS_SYNCHRONOUS)
@@ -129,21 +127,21 @@ xfs_set_inodeops(
 {
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFREG:
-		inode->i_op = &linvfs_file_inode_operations;
-		inode->i_fop = &linvfs_file_operations;
-		inode->i_mapping->a_ops = &linvfs_aops;
+		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:
-		inode->i_op = &linvfs_dir_inode_operations;
-		inode->i_fop = &linvfs_dir_operations;
+		inode->i_op = &xfs_dir_inode_operations;
+		inode->i_fop = &xfs_dir_file_operations;
 		break;
 	case S_IFLNK:
-		inode->i_op = &linvfs_symlink_inode_operations;
+		inode->i_op = &xfs_symlink_inode_operations;
 		if (inode->i_blocks)
-			inode->i_mapping->a_ops = &linvfs_aops;
+			inode->i_mapping->a_ops = &xfs_address_space_operations;
 		break;
 	default:
-		inode->i_op = &linvfs_file_inode_operations;
+		inode->i_op = &xfs_inode_operations;
 		init_special_inode(inode, inode->i_mode, inode->i_rdev);
 		break;
 	}
@@ -155,7 +153,7 @@ xfs_revalidate_inode(
 	vnode_t			*vp,
 	xfs_inode_t		*ip)
 {
-	struct inode		*inode = LINVFS_GET_IP(vp);
+	struct inode		*inode = vn_to_inode(vp);
 
 	inode->i_mode	= ip->i_d.di_mode;
 	inode->i_nlink	= ip->i_d.di_nlink;
@@ -212,7 +210,7 @@ xfs_initialize_vnode(
 	int			unlock)
 {
 	xfs_inode_t		*ip = XFS_BHVTOI(inode_bhv);
-	struct inode		*inode = LINVFS_GET_IP(vp);
+	struct inode		*inode = vn_to_inode(vp);
 
 	if (!inode_bhv->bd_vobj) {
 		vp->v_vfsp = bhvtovfs(bdp);
@@ -230,7 +228,7 @@ xfs_initialize_vnode(
 	if (ip->i_d.di_mode != 0 && unlock && (inode->i_state & I_NEW)) {
 		xfs_revalidate_inode(XFS_BHVTOM(bdp), vp, ip);
 		xfs_set_inodeops(inode);
-	
+
 		ip->i_flags &= ~XFS_INEW;
 		barrier();
 
@@ -334,43 +332,42 @@ xfs_blkdev_issue_flush(
 }
 
 STATIC struct inode *
-linvfs_alloc_inode(
+xfs_fs_alloc_inode(
 	struct super_block	*sb)
 {
 	vnode_t			*vp;
 
-	vp = kmem_cache_alloc(xfs_vnode_zone, kmem_flags_convert(KM_SLEEP));
-	if (!vp)
+	vp = kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
+	if (unlikely(!vp))
 		return NULL;
-	return LINVFS_GET_IP(vp);
+	return vn_to_inode(vp);
 }
 
 STATIC void
-linvfs_destroy_inode(
+xfs_fs_destroy_inode(
 	struct inode		*inode)
 {
-	kmem_zone_free(xfs_vnode_zone, LINVFS_GET_VP(inode));
+	kmem_zone_free(xfs_vnode_zone, vn_from_inode(inode));
 }
 
 STATIC void
-linvfs_inode_init_once(
-	void			*data,
-	kmem_cache_t		*cachep,
+xfs_fs_inode_init_once(
+	void			*vnode,
+	kmem_zone_t		*zonep,
 	unsigned long		flags)
 {
-	vnode_t			*vp = (vnode_t *)data;
-
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
-	    SLAB_CTOR_CONSTRUCTOR)
-		inode_init_once(LINVFS_GET_IP(vp));
+		      SLAB_CTOR_CONSTRUCTOR)
+		inode_init_once(vn_to_inode((vnode_t *)vnode));
 }
 
 STATIC int
-linvfs_init_zones(void)
+xfs_init_zones(void)
 {
-	xfs_vnode_zone = kmem_cache_create("xfs_vnode",
-				sizeof(vnode_t), 0, SLAB_RECLAIM_ACCOUNT,
-				linvfs_inode_init_once, NULL);
+	xfs_vnode_zone = kmem_zone_init_flags(sizeof(vnode_t), "xfs_vnode_t",
+					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
+					KM_ZONE_SPREAD,
+					xfs_fs_inode_init_once);
 	if (!xfs_vnode_zone)
 		goto out;
 
@@ -379,14 +376,12 @@ linvfs_init_zones(void)
 		goto out_destroy_vnode_zone;
 
 	xfs_ioend_pool = mempool_create(4 * MAX_BUF_PER_PAGE,
-			mempool_alloc_slab, mempool_free_slab,
-			xfs_ioend_zone);
+					mempool_alloc_slab, mempool_free_slab,
+					xfs_ioend_zone);
 	if (!xfs_ioend_pool)
 		goto out_free_ioend_zone;
-
 	return 0;
 
-
  out_free_ioend_zone:
 	kmem_zone_destroy(xfs_ioend_zone);
  out_destroy_vnode_zone:
@@ -396,7 +391,7 @@ linvfs_init_zones(void)
 }
 
 STATIC void
-linvfs_destroy_zones(void)
+xfs_destroy_zones(void)
 {
 	mempool_destroy(xfs_ioend_pool);
 	kmem_zone_destroy(xfs_vnode_zone);
@@ -407,14 +402,14 @@ linvfs_destroy_zones(void)
  * Attempt to flush the inode, this will actually fail
  * if the inode is pinned, but we dirty the inode again
  * at the point when it is unpinned after a log write,
- * since this is when the inode itself becomes flushable. 
+ * since this is when the inode itself becomes flushable.
  */
 STATIC int
-linvfs_write_inode(
+xfs_fs_write_inode(
 	struct inode		*inode,
 	int			sync)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error = 0, flags = FLUSH_INODE;
 
 	if (vp) {
@@ -434,13 +429,13 @@ linvfs_write_inode(
 }
 
 STATIC void
-linvfs_clear_inode(
+xfs_fs_clear_inode(
 	struct inode		*inode)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error, cache;
 
-	vn_trace_entry(vp, "clear_inode", (inst_t *)__return_address);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
 
 	XFS_STATS_INC(vn_rele);
 	XFS_STATS_INC(vn_remove);
@@ -516,7 +511,7 @@ void
 xfs_flush_inode(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	struct vfs	*vfs = XFS_MTOVFS(ip->i_mount);
 
 	igrab(inode);
@@ -541,7 +536,7 @@ void
 xfs_flush_device(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	struct vfs	*vfs = XFS_MTOVFS(ip->i_mount);
 
 	igrab(inode);
@@ -550,7 +545,7 @@ xfs_flush_device(
 	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
 }
 
-#define SYNCD_FLAGS	(SYNC_FSDATA|SYNC_BDFLUSH|SYNC_ATTR)
+#define SYNCD_FLAGS	(SYNC_FSDATA|SYNC_BDFLUSH|SYNC_ATTR|SYNC_REFCACHE)
 STATIC void
 vfs_sync_worker(
 	vfs_t		*vfsp,
@@ -613,7 +608,7 @@ xfssyncd(
 }
 
 STATIC int
-linvfs_start_syncd(
+xfs_fs_start_syncd(
 	vfs_t			*vfsp)
 {
 	vfsp->vfs_sync_work.w_syncer = vfs_sync_worker;
@@ -625,20 +620,20 @@ linvfs_start_syncd(
 }
 
 STATIC void
-linvfs_stop_syncd(
+xfs_fs_stop_syncd(
 	vfs_t			*vfsp)
 {
 	kthread_stop(vfsp->vfs_sync_task);
 }
 
 STATIC void
-linvfs_put_super(
+xfs_fs_put_super(
 	struct super_block	*sb)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
-	linvfs_stop_syncd(vfsp);
+	xfs_fs_stop_syncd(vfsp);
 	VFS_SYNC(vfsp, SYNC_ATTR|SYNC_DELWRI, NULL, error);
 	if (!error)
 		VFS_UNMOUNT(vfsp, 0, NULL, error);
@@ -652,10 +647,10 @@ linvfs_put_super(
 }
 
 STATIC void
-linvfs_write_super(
+xfs_fs_write_super(
 	struct super_block	*sb)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	if (sb->s_flags & MS_RDONLY) {
@@ -668,11 +663,11 @@ linvfs_write_super(
 }
 
 STATIC int
-linvfs_sync_super(
+xfs_fs_sync_super(
 	struct super_block	*sb,
 	int			wait)
 {
-	vfs_t		*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t		*vfsp = vfs_from_sb(sb);
 	int		error;
 	int		flags = SYNC_FSDATA;
 
@@ -707,11 +702,11 @@ linvfs_sync_super(
 }
 
 STATIC int
-linvfs_statfs(
+xfs_fs_statfs(
 	struct super_block	*sb,
 	struct kstatfs		*statp)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_STATVFS(vfsp, statp, NULL, error);
@@ -719,12 +714,12 @@ linvfs_statfs(
 }
 
 STATIC int
-linvfs_remount(
+xfs_fs_remount(
 	struct super_block	*sb,
 	int			*flags,
 	char			*options)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	struct xfs_mount_args	*args = xfs_args_allocate(sb);
 	int			error;
 
@@ -736,18 +731,18 @@ linvfs_remount(
 }
 
 STATIC void
-linvfs_freeze_fs(
+xfs_fs_lockfs(
 	struct super_block	*sb)
 {
-	VFS_FREEZE(LINVFS_GET_VFS(sb));
+	VFS_FREEZE(vfs_from_sb(sb));
 }
 
 STATIC int
-linvfs_show_options(
+xfs_fs_show_options(
 	struct seq_file		*m,
 	struct vfsmount		*mnt)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(mnt->mnt_sb);
+	struct vfs		*vfsp = vfs_from_sb(mnt->mnt_sb);
 	int			error;
 
 	VFS_SHOWARGS(vfsp, m, error);
@@ -755,11 +750,11 @@ linvfs_show_options(
 }
 
 STATIC int
-linvfs_quotasync(
+xfs_fs_quotasync(
 	struct super_block	*sb,
 	int			type)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, Q_XQUOTASYNC, 0, (caddr_t)NULL, error);
@@ -767,11 +762,11 @@ linvfs_quotasync(
 }
 
 STATIC int
-linvfs_getxstate(
+xfs_fs_getxstate(
 	struct super_block	*sb,
 	struct fs_quota_stat	*fqs)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, Q_XGETQSTAT, 0, (caddr_t)fqs, error);
@@ -779,12 +774,12 @@ linvfs_getxstate(
 }
 
 STATIC int
-linvfs_setxstate(
+xfs_fs_setxstate(
 	struct super_block	*sb,
 	unsigned int		flags,
 	int			op)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, op, 0, (caddr_t)&flags, error);
@@ -792,13 +787,13 @@ linvfs_setxstate(
 }
 
 STATIC int
-linvfs_getxquota(
+xfs_fs_getxquota(
 	struct super_block	*sb,
 	int			type,
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error, getmode;
 
 	getmode = (type == USRQUOTA) ? Q_XGETQUOTA :
@@ -808,13 +803,13 @@ linvfs_getxquota(
 }
 
 STATIC int
-linvfs_setxquota(
+xfs_fs_setxquota(
 	struct super_block	*sb,
 	int			type,
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error, setmode;
 
 	setmode = (type == USRQUOTA) ? Q_XSETQLIM :
@@ -824,21 +819,17 @@ linvfs_setxquota(
 }
 
 STATIC int
-linvfs_fill_super(
+xfs_fs_fill_super(
 	struct super_block	*sb,
 	void			*data,
 	int			silent)
 {
 	vnode_t			*rootvp;
-	struct vfs		*vfsp = vfs_allocate();
+	struct vfs		*vfsp = vfs_allocate(sb);
 	struct xfs_mount_args	*args = xfs_args_allocate(sb);
 	struct kstatfs		statvfs;
 	int			error, error2;
 
-	vfsp->vfs_super = sb;
-	LINVFS_SET_VFS(sb, vfsp);
-	if (sb->s_flags & MS_RDONLY)
-		vfsp->vfs_flag |= VFS_RDONLY;
 	bhv_insert_all_vfsops(vfsp);
 
 	VFS_PARSEARGS(vfsp, (char *)data, args, 0, error);
@@ -849,10 +840,10 @@ linvfs_fill_super(
 
 	sb_min_blocksize(sb, BBSIZE);
 #ifdef CONFIG_XFS_EXPORT
-	sb->s_export_op = &linvfs_export_ops;
+	sb->s_export_op = &xfs_export_operations;
 #endif
-	sb->s_qcop = &linvfs_qops;
-	sb->s_op = &linvfs_sops;
+	sb->s_qcop = &xfs_quotactl_operations;
+	sb->s_op = &xfs_super_operations;
 
 	VFS_MOUNT(vfsp, args, NULL, error);
 	if (error) {
@@ -876,7 +867,7 @@ linvfs_fill_super(
 	if (error)
 		goto fail_unmount;
 
-	sb->s_root = d_alloc_root(LINVFS_GET_IP(rootvp));
+	sb->s_root = d_alloc_root(vn_to_inode(rootvp));
 	if (!sb->s_root) {
 		error = ENOMEM;
 		goto fail_vnrele;
@@ -885,7 +876,7 @@ linvfs_fill_super(
 		error = EINVAL;
 		goto fail_vnrele;
 	}
-	if ((error = linvfs_start_syncd(vfsp)))
+	if ((error = xfs_fs_start_syncd(vfsp)))
 		goto fail_vnrele;
 	vn_trace_exit(rootvp, __FUNCTION__, (inst_t *)__return_address);
 
@@ -910,41 +901,41 @@ fail_vfsop:
 }
 
 STATIC struct super_block *
-linvfs_get_sb(
+xfs_fs_get_sb(
 	struct file_system_type	*fs_type,
 	int			flags,
 	const char		*dev_name,
 	void			*data)
 {
-	return get_sb_bdev(fs_type, flags, dev_name, data, linvfs_fill_super);
-}
-
-STATIC struct super_operations linvfs_sops = {
-	.alloc_inode		= linvfs_alloc_inode,
-	.destroy_inode		= linvfs_destroy_inode,
-	.write_inode		= linvfs_write_inode,
-	.clear_inode		= linvfs_clear_inode,
-	.put_super		= linvfs_put_super,
-	.write_super		= linvfs_write_super,
-	.sync_fs		= linvfs_sync_super,
-	.write_super_lockfs	= linvfs_freeze_fs,
-	.statfs			= linvfs_statfs,
-	.remount_fs		= linvfs_remount,
-	.show_options		= linvfs_show_options,
+	return get_sb_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super);
+}
+
+STATIC struct super_operations xfs_super_operations = {
+	.alloc_inode		= xfs_fs_alloc_inode,
+	.destroy_inode		= xfs_fs_destroy_inode,
+	.write_inode		= xfs_fs_write_inode,
+	.clear_inode		= xfs_fs_clear_inode,
+	.put_super		= xfs_fs_put_super,
+	.write_super		= xfs_fs_write_super,
+	.sync_fs		= xfs_fs_sync_super,
+	.write_super_lockfs	= xfs_fs_lockfs,
+	.statfs			= xfs_fs_statfs,
+	.remount_fs		= xfs_fs_remount,
+	.show_options		= xfs_fs_show_options,
 };
 
-STATIC struct quotactl_ops linvfs_qops = {
-	.quota_sync		= linvfs_quotasync,
-	.get_xstate		= linvfs_getxstate,
-	.set_xstate		= linvfs_setxstate,
-	.get_xquota		= linvfs_getxquota,
-	.set_xquota		= linvfs_setxquota,
+STATIC struct quotactl_ops xfs_quotactl_operations = {
+	.quota_sync		= xfs_fs_quotasync,
+	.get_xstate		= xfs_fs_getxstate,
+	.set_xstate		= xfs_fs_setxstate,
+	.get_xquota		= xfs_fs_getxquota,
+	.set_xquota		= xfs_fs_setxquota,
 };
 
 STATIC struct file_system_type xfs_fs_type = {
 	.owner			= THIS_MODULE,
 	.name			= "xfs",
-	.get_sb			= linvfs_get_sb,
+	.get_sb			= xfs_fs_get_sb,
 	.kill_sb		= kill_block_super,
 	.fs_flags		= FS_REQUIRES_DEV,
 };
@@ -965,7 +956,7 @@ init_xfs_fs( void )
 
 	ktrace_init(64);
 
-	error = linvfs_init_zones();
+	error = xfs_init_zones();
 	if (error < 0)
 		goto undo_zones;
 
@@ -981,14 +972,13 @@ init_xfs_fs( void )
 	error = register_filesystem(&xfs_fs_type);
 	if (error)
 		goto undo_register;
-	XFS_DM_INIT(&xfs_fs_type);
 	return 0;
 
 undo_register:
 	xfs_buf_terminate();
 
 undo_buffers:
-	linvfs_destroy_zones();
+	xfs_destroy_zones();
 
 undo_zones:
 	return error;
@@ -998,11 +988,10 @@ STATIC void __exit
 exit_xfs_fs( void )
 {
 	vfs_exitquota();
-	XFS_DM_EXIT(&xfs_fs_type);
 	unregister_filesystem(&xfs_fs_type);
 	xfs_cleanup();
 	xfs_buf_terminate();
-	linvfs_destroy_zones();
+	xfs_destroy_zones();
 	ktrace_uninit();
 }
 
diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h
index df59408dca06..376b96cb513a 100644
--- a/fs/xfs/linux-2.6/xfs_super.h
+++ b/fs/xfs/linux-2.6/xfs_super.h
@@ -98,11 +98,6 @@ extern void xfs_qm_exit(void);
 				XFS_DMAPI_STRING \
 				XFS_DBG_STRING /* DBG must be last */
 
-#define LINVFS_GET_VFS(s) \
-	(vfs_t *)((s)->s_fs_info)
-#define LINVFS_SET_VFS(s, vfsp) \
-	((s)->s_fs_info = vfsp)
-
 struct xfs_inode;
 struct xfs_mount;
 struct xfs_buftarg;
@@ -120,6 +115,6 @@ extern int  xfs_blkdev_get(struct xfs_mount *, const char *,
 extern void xfs_blkdev_put(struct block_device *);
 extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
 
-extern struct export_operations linvfs_export_ops;
+extern struct export_operations xfs_export_operations;
 
 #endif	/* __XFS_SUPER_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_sysctl.c b/fs/xfs/linux-2.6/xfs_sysctl.c
index a02564972420..7079cc837210 100644
--- a/fs/xfs/linux-2.6/xfs_sysctl.c
+++ b/fs/xfs/linux-2.6/xfs_sysctl.c
@@ -38,8 +38,7 @@ xfs_stats_clear_proc_handler(
 
 	if (!ret && write && *valp) {
 		printk("XFS Clearing xfsstats\n");
-		for (c = 0; c < NR_CPUS; c++) {
-			if (!cpu_possible(c)) continue;
+		for_each_cpu(c) {
 			preempt_disable();
 			/* save vn_active, it's a universal truth! */
 			vn_active = per_cpu(xfsstats, c).vn_active;
diff --git a/fs/xfs/linux-2.6/xfs_vfs.c b/fs/xfs/linux-2.6/xfs_vfs.c
index c855d62e5344..6f7c9f7a8624 100644
--- a/fs/xfs/linux-2.6/xfs_vfs.c
+++ b/fs/xfs/linux-2.6/xfs_vfs.c
@@ -227,7 +227,8 @@ vfs_freeze(
 }
 
 vfs_t *
-vfs_allocate( void )
+vfs_allocate(
+	struct super_block	*sb)
 {
 	struct vfs		*vfsp;
 
@@ -236,9 +237,23 @@ vfs_allocate( void )
 	INIT_LIST_HEAD(&vfsp->vfs_sync_list);
 	spin_lock_init(&vfsp->vfs_sync_lock);
 	init_waitqueue_head(&vfsp->vfs_wait_single_sync_task);
+
+	vfsp->vfs_super = sb;
+	sb->s_fs_info = vfsp;
+
+	if (sb->s_flags & MS_RDONLY)
+		vfsp->vfs_flag |= VFS_RDONLY;
+
 	return vfsp;
 }
 
+vfs_t *
+vfs_from_sb(
+	struct super_block	*sb)
+{
+	return (vfs_t *)sb->s_fs_info;
+}
+
 void
 vfs_deallocate(
 	struct vfs		*vfsp)
@@ -295,7 +310,7 @@ bhv_remove_all_vfsops(
 	bhv_remove_vfsops(vfsp, VFS_POSITION_DM);
 	if (!freebase)
 		return;
-	mp = XFS_BHVTOM(bhv_lookup(VFS_BHVHEAD(vfsp), &xfs_vfsops));
+	mp = XFS_VFSTOM(vfsp);
 	VFS_REMOVEBHV(vfsp, &mp->m_bhv);
 	xfs_mount_free(mp, 0);
 }
diff --git a/fs/xfs/linux-2.6/xfs_vfs.h b/fs/xfs/linux-2.6/xfs_vfs.h
index 57caf9eddee0..8fed356db055 100644
--- a/fs/xfs/linux-2.6/xfs_vfs.h
+++ b/fs/xfs/linux-2.6/xfs_vfs.h
@@ -193,7 +193,8 @@ typedef struct bhv_vfsops {
 #define vfs_bhv_set_custom(b,o)	( (b)->bhv_custom = (void *)(o))
 #define vfs_bhv_clr_custom(b)	( (b)->bhv_custom = NULL )
 
-extern vfs_t *vfs_allocate(void);
+extern vfs_t *vfs_allocate(struct super_block *);
+extern vfs_t *vfs_from_sb(struct super_block *);
 extern void vfs_deallocate(vfs_t *);
 extern void vfs_insertops(vfs_t *, bhv_vfsops_t *);
 extern void vfs_insertbhv(vfs_t *, bhv_desc_t *, vfsops_t *, void *);
diff --git a/fs/xfs/linux-2.6/xfs_vnode.c b/fs/xfs/linux-2.6/xfs_vnode.c
index 260dd8415dd7..d27c25b27ccd 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.c
+++ b/fs/xfs/linux-2.6/xfs_vnode.c
@@ -58,7 +58,7 @@ struct vnode *
 vn_initialize(
 	struct inode	*inode)
 {
-	struct vnode	*vp = LINVFS_GET_VP(inode);
+	struct vnode	*vp = vn_from_inode(inode);
 
 	XFS_STATS_INC(vn_active);
 	XFS_STATS_INC(vn_alloc);
@@ -83,7 +83,7 @@ vn_initialize(
 	vp->v_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
 #endif	/* XFS_VNODE_TRACE */
 
-	vn_trace_exit(vp, "vn_initialize", (inst_t *)__return_address);
+	vn_trace_exit(vp, __FUNCTION__, (inst_t *)__return_address);
 	return vp;
 }
 
@@ -97,7 +97,7 @@ vn_revalidate_core(
 	struct vnode	*vp,
 	vattr_t		*vap)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	inode->i_mode	    = vap->va_mode;
 	inode->i_nlink	    = vap->va_nlink;
@@ -129,24 +129,31 @@ vn_revalidate_core(
  * Revalidate the Linux inode from the vnode.
  */
 int
-vn_revalidate(
-	struct vnode	*vp)
+__vn_revalidate(
+	struct vnode	*vp,
+	struct vattr	*vattr)
 {
-	vattr_t		va;
 	int		error;
 
-	vn_trace_entry(vp, "vn_revalidate", (inst_t *)__return_address);
-	ASSERT(vp->v_fbhv != NULL);
-
-	va.va_mask = XFS_AT_STAT|XFS_AT_XFLAGS;
-	VOP_GETATTR(vp, &va, 0, NULL, error);
-	if (!error) {
-		vn_revalidate_core(vp, &va);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
+	vattr->va_mask = XFS_AT_STAT | XFS_AT_XFLAGS;
+	VOP_GETATTR(vp, vattr, 0, NULL, error);
+	if (likely(!error)) {
+		vn_revalidate_core(vp, vattr);
 		VUNMODIFY(vp);
 	}
 	return -error;
 }
 
+int
+vn_revalidate(
+	struct vnode	*vp)
+{
+	vattr_t		vattr;
+
+	return __vn_revalidate(vp, &vattr);
+}
+
 /*
  * Add a reference to a referenced vnode.
  */
@@ -159,7 +166,7 @@ vn_hold(
 	XFS_STATS_INC(vn_hold);
 
 	VN_LOCK(vp);
-	inode = igrab(LINVFS_GET_IP(vp));
+	inode = igrab(vn_to_inode(vp));
 	ASSERT(inode);
 	VN_UNLOCK(vp, 0);
 
diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h
index 0fe2419461d6..06f5845e9568 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.h
+++ b/fs/xfs/linux-2.6/xfs_vnode.h
@@ -116,8 +116,14 @@ typedef enum {
 /*
  * Vnode to Linux inode mapping.
  */
-#define LINVFS_GET_VP(inode)	((vnode_t *)list_entry(inode, vnode_t, v_inode))
-#define LINVFS_GET_IP(vp)	(&(vp)->v_inode)
+static inline struct vnode *vn_from_inode(struct inode *inode)
+{
+	return (vnode_t *)list_entry(inode, vnode_t, v_inode);
+}
+static inline struct inode *vn_to_inode(struct vnode *vnode)
+{
+	return &vnode->v_inode;
+}
 
 /*
  * Vnode flags.
@@ -490,6 +496,7 @@ typedef struct vnode_map {
 			 (vmap).v_ino	 = (vp)->v_inode.i_ino; }
 
 extern int	vn_revalidate(struct vnode *);
+extern int	__vn_revalidate(struct vnode *, vattr_t *);
 extern void	vn_revalidate_core(struct vnode *, vattr_t *);
 
 extern void	vn_iowait(struct vnode *vp);
@@ -497,7 +504,7 @@ extern void	vn_iowake(struct vnode *vp);
 
 static inline int vn_count(struct vnode *vp)
 {
-	return atomic_read(&LINVFS_GET_IP(vp)->i_count);
+	return atomic_read(&vn_to_inode(vp)->i_count);
 }
 
 /*
@@ -511,16 +518,16 @@ extern vnode_t	*vn_hold(struct vnode *);
 	  vn_trace_hold(vp, __FILE__, __LINE__, (inst_t *)__return_address))
 #define VN_RELE(vp)		\
 	  (vn_trace_rele(vp, __FILE__, __LINE__, (inst_t *)__return_address), \
-	   iput(LINVFS_GET_IP(vp)))
+	   iput(vn_to_inode(vp)))
 #else
 #define VN_HOLD(vp)		((void)vn_hold(vp))
-#define VN_RELE(vp)		(iput(LINVFS_GET_IP(vp)))
+#define VN_RELE(vp)		(iput(vn_to_inode(vp)))
 #endif
 
 static inline struct vnode *vn_grab(struct vnode *vp)
 {
-	struct inode *inode = igrab(LINVFS_GET_IP(vp));
-	return inode ? LINVFS_GET_VP(inode) : NULL;
+	struct inode *inode = igrab(vn_to_inode(vp));
+	return inode ? vn_from_inode(inode) : NULL;
 }
 
 /*
@@ -528,7 +535,7 @@ static inline struct vnode *vn_grab(struct vnode *vp)
  */
 #define VNAME(dentry)		((char *) (dentry)->d_name.name)
 #define VNAMELEN(dentry)	((dentry)->d_name.len)
-#define VNAME_TO_VNODE(dentry)	(LINVFS_GET_VP((dentry)->d_inode))
+#define VNAME_TO_VNODE(dentry)	(vn_from_inode((dentry)->d_inode))
 
 /*
  * Vnode spinlock manipulation.
@@ -557,12 +564,12 @@ static __inline__ void vn_flagclr(struct vnode *vp, uint flag)
  */
 static inline void vn_mark_bad(struct vnode *vp)
 {
-	make_bad_inode(LINVFS_GET_IP(vp));
+	make_bad_inode(vn_to_inode(vp));
 }
 
 static inline int VN_BAD(struct vnode *vp)
 {
-	return is_bad_inode(LINVFS_GET_IP(vp));
+	return is_bad_inode(vn_to_inode(vp));
 }
 
 /*
@@ -587,9 +594,9 @@ static inline void vn_atime_to_time_t(struct vnode *vp, time_t *tt)
 /*
  * Some useful predicates.
  */
-#define VN_MAPPED(vp)	mapping_mapped(LINVFS_GET_IP(vp)->i_mapping)
-#define VN_CACHED(vp)	(LINVFS_GET_IP(vp)->i_mapping->nrpages)
-#define VN_DIRTY(vp)	mapping_tagged(LINVFS_GET_IP(vp)->i_mapping, \
+#define VN_MAPPED(vp)	mapping_mapped(vn_to_inode(vp)->i_mapping)
+#define VN_CACHED(vp)	(vn_to_inode(vp)->i_mapping->nrpages)
+#define VN_DIRTY(vp)	mapping_tagged(vn_to_inode(vp)->i_mapping, \
 					PAGECACHE_TAG_DIRTY)
 #define VMODIFY(vp)	VN_FLAGSET(vp, VMODIFIED)
 #define VUNMODIFY(vp)	VN_FLAGCLR(vp, VMODIFIED)
diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c
index 2ec6b441849c..e4e5f05b841b 100644
--- a/fs/xfs/quota/xfs_dquot_item.c
+++ b/fs/xfs/quota/xfs_dquot_item.c
@@ -79,9 +79,11 @@ xfs_qm_dquot_logitem_format(
 
 	logvec->i_addr = (xfs_caddr_t)&logitem->qli_format;
 	logvec->i_len  = sizeof(xfs_dq_logformat_t);
+	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_QFORMAT);
 	logvec++;
 	logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core;
 	logvec->i_len  = sizeof(xfs_disk_dquot_t);
+	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_DQUOT);
 
 	ASSERT(2 == logitem->qli_item.li_desc->lid_size);
 	logitem->qli_format.qlf_size = 2;
diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c
index 7c0e39dc6189..1fb757ef3f41 100644
--- a/fs/xfs/quota/xfs_qm.c
+++ b/fs/xfs/quota/xfs_qm.c
@@ -1704,9 +1704,9 @@ xfs_qm_get_rtblks(
 	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 */
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;
 
@@ -1717,10 +1717,11 @@ xfs_qm_get_rtblks(
 			return error;
 	}
 	rtblks = 0;
-	nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
-	for (ep = base; ep < &base[nextents]; ep++)
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	for (idx = 0; idx < nextents; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		rtblks += xfs_bmbt_get_blockcount(ep);
+	}
 	*O_rtblks = (xfs_qcnt_t)rtblks;
 	return 0;
 }
@@ -2788,9 +2789,7 @@ xfs_qm_freelist_destroy(xfs_frlist_t *ql)
 		xfs_qm_dqdestroy(dqp);
 		dqp = nextdqp;
 	}
-	/*
-	 * Don't bother about unlocking.
-	 */
+	mutex_unlock(&ql->qh_lock);
 	mutex_destroy(&ql->qh_lock);
 
 	ASSERT(ql->qh_nelems == 0);
diff --git a/fs/xfs/quota/xfs_qm_bhv.c b/fs/xfs/quota/xfs_qm_bhv.c
index 90402a1c3983..6838b36d95a9 100644
--- a/fs/xfs/quota/xfs_qm_bhv.c
+++ b/fs/xfs/quota/xfs_qm_bhv.c
@@ -374,7 +374,7 @@ xfs_qm_exit(void)
 	vfs_bhv_clr_custom(&xfs_qmops);
 	xfs_qm_cleanup_procfs();
 	if (qm_dqzone)
-		kmem_cache_destroy(qm_dqzone);
+		kmem_zone_destroy(qm_dqzone);
 	if (qm_dqtrxzone)
-		kmem_cache_destroy(qm_dqtrxzone);
+		kmem_zone_destroy(qm_dqtrxzone);
 }
diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c
index 841aa4c15b8a..addf5a7ea06c 100644
--- a/fs/xfs/support/ktrace.c
+++ b/fs/xfs/support/ktrace.c
@@ -39,8 +39,8 @@ ktrace_init(int zentries)
 void
 ktrace_uninit(void)
 {
-	kmem_cache_destroy(ktrace_hdr_zone);
-	kmem_cache_destroy(ktrace_ent_zone);
+	kmem_zone_destroy(ktrace_hdr_zone);
+	kmem_zone_destroy(ktrace_ent_zone);
 }
 
 /*
diff --git a/fs/xfs/support/uuid.c b/fs/xfs/support/uuid.c
index a3d565a67734..e157015c70ff 100644
--- a/fs/xfs/support/uuid.c
+++ b/fs/xfs/support/uuid.c
@@ -21,13 +21,6 @@ static mutex_t	uuid_monitor;
 static int	uuid_table_size;
 static uuid_t	*uuid_table;
 
-void
-uuid_init(void)
-{
-	mutex_init(&uuid_monitor);
-}
-
-
 /* IRIX interpretation of an uuid_t */
 typedef struct {
 	__be32	uu_timelow;
@@ -50,7 +43,7 @@ uuid_getnodeuniq(uuid_t *uuid, int fsid [2])
 
 	fsid[0] = (be16_to_cpu(uup->uu_clockseq) << 16) |
 		   be16_to_cpu(uup->uu_timemid);
-	fsid[1] = be16_to_cpu(uup->uu_timelow);
+	fsid[1] = be32_to_cpu(uup->uu_timelow);
 }
 
 void
@@ -139,3 +132,9 @@ uuid_table_remove(uuid_t *uuid)
 	ASSERT(i < uuid_table_size);
 	mutex_unlock(&uuid_monitor);
 }
+
+void
+uuid_init(void)
+{
+	mutex_init(&uuid_monitor);
+}
diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h
index f9315bc960cb..538d0d65b04c 100644
--- a/fs/xfs/xfs_acl.h
+++ b/fs/xfs/xfs_acl.h
@@ -55,8 +55,8 @@ struct xfs_inode;
 
 extern struct kmem_zone *xfs_acl_zone;
 #define xfs_acl_zone_init(zone, name)	\
-		(zone) = kmem_zone_init(sizeof(xfs_acl_t), name)
-#define xfs_acl_zone_destroy(zone)	kmem_cache_destroy(zone)
+		(zone) = kmem_zone_init(sizeof(xfs_acl_t), (name))
+#define xfs_acl_zone_destroy(zone)	kmem_zone_destroy(zone)
 
 extern int xfs_acl_inherit(struct vnode *, struct vattr *, xfs_acl_t *);
 extern int xfs_acl_iaccess(struct xfs_inode *, mode_t, cred_t *);
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index e5e91e9c7e89..093fac476bda 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -1127,8 +1127,7 @@ xfs_attr_leaf_list(xfs_attr_list_context_t *context)
 		return(error);
 	ASSERT(bp != NULL);
 	leaf = bp->data;
-	if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						!= XFS_ATTR_LEAF_MAGIC)) {
+	if (unlikely(be16_to_cpu(leaf->hdr.info.magic) != XFS_ATTR_LEAF_MAGIC)) {
 		XFS_CORRUPTION_ERROR("xfs_attr_leaf_list", XFS_ERRLEVEL_LOW,
 				     context->dp->i_mount, leaf);
 		xfs_da_brelse(NULL, bp);
@@ -1541,8 +1540,8 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 						     XFS_ATTR_FORK);
 		if (error)
 			goto out;
-		ASSERT(INT_GET(((xfs_attr_leafblock_t *)
-				      bp->data)->hdr.info.magic, ARCH_CONVERT)
+		ASSERT(be16_to_cpu(((xfs_attr_leafblock_t *)
+				      bp->data)->hdr.info.magic)
 						       == XFS_ATTR_LEAF_MAGIC);
 
 		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
@@ -1763,7 +1762,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 			return(error);
 		if (bp) {
 			node = bp->data;
-			switch (INT_GET(node->hdr.info.magic, ARCH_CONVERT)) {
+			switch (be16_to_cpu(node->hdr.info.magic)) {
 			case XFS_DA_NODE_MAGIC:
 				xfs_attr_trace_l_cn("wrong blk", context, node);
 				xfs_da_brelse(NULL, bp);
@@ -1771,18 +1770,14 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				break;
 			case XFS_ATTR_LEAF_MAGIC:
 				leaf = bp->data;
-				if (cursor->hashval >
-				    INT_GET(leaf->entries[
-					 INT_GET(leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-							ARCH_CONVERT)) {
+				if (cursor->hashval > be32_to_cpu(leaf->entries[
+				    be16_to_cpu(leaf->hdr.count)-1].hashval)) {
 					xfs_attr_trace_l_cl("wrong blk",
 							   context, leaf);
 					xfs_da_brelse(NULL, bp);
 					bp = NULL;
 				} else if (cursor->hashval <=
-					     INT_GET(leaf->entries[0].hashval,
-							ARCH_CONVERT)) {
+					     be32_to_cpu(leaf->entries[0].hashval)) {
 					xfs_attr_trace_l_cl("maybe wrong blk",
 							   context, leaf);
 					xfs_da_brelse(NULL, bp);
@@ -1817,10 +1812,10 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				return(XFS_ERROR(EFSCORRUPTED));
 			}
 			node = bp->data;
-			if (INT_GET(node->hdr.info.magic, ARCH_CONVERT)
+			if (be16_to_cpu(node->hdr.info.magic)
 							== XFS_ATTR_LEAF_MAGIC)
 				break;
-			if (unlikely(INT_GET(node->hdr.info.magic, ARCH_CONVERT)
+			if (unlikely(be16_to_cpu(node->hdr.info.magic)
 							!= XFS_DA_NODE_MAGIC)) {
 				XFS_CORRUPTION_ERROR("xfs_attr_node_list(3)",
 						     XFS_ERRLEVEL_LOW,
@@ -1830,19 +1825,17 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				return(XFS_ERROR(EFSCORRUPTED));
 			}
 			btree = node->btree;
-			for (i = 0;
-				i < INT_GET(node->hdr.count, ARCH_CONVERT);
+			for (i = 0; i < be16_to_cpu(node->hdr.count);
 								btree++, i++) {
 				if (cursor->hashval
-						<= INT_GET(btree->hashval,
-							    ARCH_CONVERT)) {
-					cursor->blkno = INT_GET(btree->before, ARCH_CONVERT);
+						<= be32_to_cpu(btree->hashval)) {
+					cursor->blkno = be32_to_cpu(btree->before);
 					xfs_attr_trace_l_cb("descending",
 							    context, btree);
 					break;
 				}
 			}
-			if (i == INT_GET(node->hdr.count, ARCH_CONVERT)) {
+			if (i == be16_to_cpu(node->hdr.count)) {
 				xfs_da_brelse(NULL, bp);
 				return(0);
 			}
@@ -1858,7 +1851,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 	 */
 	for (;;) {
 		leaf = bp->data;
-		if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
+		if (unlikely(be16_to_cpu(leaf->hdr.info.magic)
 						!= XFS_ATTR_LEAF_MAGIC)) {
 			XFS_CORRUPTION_ERROR("xfs_attr_node_list(4)",
 					     XFS_ERRLEVEL_LOW,
@@ -1869,7 +1862,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 		error = xfs_attr_leaf_list_int(bp, context);
 		if (error || !leaf->hdr.info.forw)
 			break;	/* not really an error, buffer full or EOF */
-		cursor->blkno = INT_GET(leaf->hdr.info.forw, ARCH_CONVERT);
+		cursor->blkno = be32_to_cpu(leaf->hdr.info.forw);
 		xfs_da_brelse(NULL, bp);
 		error = xfs_da_read_buf(NULL, context->dp, cursor->blkno, -1,
 					      &bp, XFS_ATTR_FORK);
@@ -2232,9 +2225,10 @@ xfs_attr_trace_l_cn(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(node->hdr.count, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(node->btree[0].hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(node->btree[INT_GET(node->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+		(__psunsigned_t)be16_to_cpu(node->hdr.count),
+		(__psunsigned_t)be32_to_cpu(node->btree[0].hashval),
+		(__psunsigned_t)be32_to_cpu(node->btree[
+				    be16_to_cpu(node->hdr.count)-1].hashval));
 }
 
 /*
@@ -2261,8 +2255,8 @@ xfs_attr_trace_l_cb(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(btree->hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(btree->before, ARCH_CONVERT),
+		(__psunsigned_t)be32_to_cpu(btree->hashval),
+		(__psunsigned_t)be32_to_cpu(btree->before),
 		(__psunsigned_t)NULL);
 }
 
@@ -2290,9 +2284,10 @@ xfs_attr_trace_l_cl(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(leaf->hdr.count, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(leaf->entries[0].hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(leaf->entries[INT_GET(leaf->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+		(__psunsigned_t)be16_to_cpu(leaf->hdr.count),
+		(__psunsigned_t)be32_to_cpu(leaf->entries[0].hashval),
+		(__psunsigned_t)be32_to_cpu(leaf->entries[
+				be16_to_cpu(leaf->hdr.count)-1].hashval));
 }
 
 /*
@@ -2522,7 +2517,7 @@ attr_user_capable(
 	struct vnode	*vp,
 	cred_t		*cred)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
@@ -2540,7 +2535,7 @@ attr_trusted_capable(
 	struct vnode	*vp,
 	cred_t		*cred)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c
index fe91eac4e2a7..717682747bd2 100644
--- a/fs/xfs/xfs_attr_leaf.c
+++ b/fs/xfs/xfs_attr_leaf.c
@@ -194,7 +194,7 @@ xfs_attr_shortform_create(xfs_da_args_t *args)
 	xfs_idata_realloc(dp, sizeof(*hdr), XFS_ATTR_FORK);
 	hdr = (xfs_attr_sf_hdr_t *)ifp->if_u1.if_data;
 	hdr->count = 0;
-	INT_SET(hdr->totsize, ARCH_CONVERT, sizeof(*hdr));
+	hdr->totsize = cpu_to_be16(sizeof(*hdr));
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_ADATA);
 }
 
@@ -224,8 +224,7 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
-				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
+	for (i = 0; i < sf->hdr.count; sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 #ifdef DEBUG
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -248,13 +247,13 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	sfe = (xfs_attr_sf_entry_t *)((char *)sf + offset);
 
 	sfe->namelen = args->namelen;
-	INT_SET(sfe->valuelen, ARCH_CONVERT, args->valuelen);
+	sfe->valuelen = args->valuelen;
 	sfe->flags = (args->flags & ATTR_SECURE) ? XFS_ATTR_SECURE :
 			((args->flags & ATTR_ROOT) ? XFS_ATTR_ROOT : 0);
 	memcpy(sfe->nameval, args->name, args->namelen);
 	memcpy(&sfe->nameval[args->namelen], args->value, args->valuelen);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, 1);
-	INT_MOD(sf->hdr.totsize, ARCH_CONVERT, size);
+	sf->hdr.count++;
+	be16_add(&sf->hdr.totsize, size);
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_ADATA);
 
 	xfs_sbversion_add_attr2(mp, args->trans);
@@ -277,7 +276,7 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	base = sizeof(xfs_attr_sf_hdr_t);
 	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
 	sfe = &sf->list[0];
-	end = INT_GET(sf->hdr.count, ARCH_CONVERT);
+	end = sf->hdr.count;
 	for (i = 0; i < end; sfe = XFS_ATTR_SF_NEXTENTRY(sfe),
 					base += size, i++) {
 		size = XFS_ATTR_SF_ENTSIZE(sfe);
@@ -300,11 +299,11 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	 * Fix up the attribute fork data, covering the hole
 	 */
 	end = base + size;
-	totsize = INT_GET(sf->hdr.totsize, ARCH_CONVERT);
+	totsize = be16_to_cpu(sf->hdr.totsize);
 	if (end != totsize)
 		memmove(&((char *)sf)[base], &((char *)sf)[end], totsize - end);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, -1);
-	INT_MOD(sf->hdr.totsize, ARCH_CONVERT, -size);
+	sf->hdr.count--;
+	be16_add(&sf->hdr.totsize, -size);
 
 	/*
 	 * Fix up the start offset of the attribute fork
@@ -360,7 +359,7 @@ xfs_attr_shortform_lookup(xfs_da_args_t *args)
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
+	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -391,7 +390,7 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 	ASSERT(args->dp->i_d.di_aformat == XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)args->dp->i_afp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
+	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -404,14 +403,14 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 		    ((sfe->flags & XFS_ATTR_ROOT) != 0))
 			continue;
 		if (args->flags & ATTR_KERNOVAL) {
-			args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+			args->valuelen = sfe->valuelen;
 			return(XFS_ERROR(EEXIST));
 		}
-		if (args->valuelen < INT_GET(sfe->valuelen, ARCH_CONVERT)) {
-			args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		if (args->valuelen < sfe->valuelen) {
+			args->valuelen = sfe->valuelen;
 			return(XFS_ERROR(ERANGE));
 		}
-		args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		args->valuelen = sfe->valuelen;
 		memcpy(args->value, &sfe->nameval[args->namelen],
 						    args->valuelen);
 		return(XFS_ERROR(EEXIST));
@@ -438,7 +437,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	dp = args->dp;
 	ifp = dp->i_afp;
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
-	size = INT_GET(sf->hdr.totsize, ARCH_CONVERT);
+	size = be16_to_cpu(sf->hdr.totsize);
 	tmpbuffer = kmem_alloc(size, KM_SLEEP);
 	ASSERT(tmpbuffer != NULL);
 	memcpy(tmpbuffer, ifp->if_u1.if_data, size);
@@ -481,11 +480,11 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	nargs.oknoent = 1;
 
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0; i < sf->hdr.count; i++) {
 		nargs.name = (char *)sfe->nameval;
 		nargs.namelen = sfe->namelen;
 		nargs.value = (char *)&sfe->nameval[nargs.namelen];
-		nargs.valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		nargs.valuelen = sfe->valuelen;
 		nargs.hashval = xfs_da_hashname((char *)sfe->nameval,
 						sfe->namelen);
 		nargs.flags = (sfe->flags & XFS_ATTR_SECURE) ? ATTR_SECURE :
@@ -514,11 +513,9 @@ xfs_attr_shortform_compare(const void *a, const void *b)
 
 	sa = (xfs_attr_sf_sort_t *)a;
 	sb = (xfs_attr_sf_sort_t *)b;
-	if (INT_GET(sa->hash, ARCH_CONVERT)
-				< INT_GET(sb->hash, ARCH_CONVERT)) {
+	if (sa->hash < sb->hash) {
 		return(-1);
-	} else if (INT_GET(sa->hash, ARCH_CONVERT)
-				> INT_GET(sb->hash, ARCH_CONVERT)) {
+	} else if (sa->hash > sb->hash) {
 		return(1);
 	} else {
 		return(sa->entno - sb->entno);
@@ -560,10 +557,8 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	 * If the buffer is large enough, do not bother with sorting.
 	 * Note the generous fudge factor of 16 overhead bytes per entry.
 	 */
-	if ((dp->i_afp->if_bytes + INT_GET(sf->hdr.count, ARCH_CONVERT) * 16)
-							< context->bufsize) {
-		for (i = 0, sfe = &sf->list[0];
-				i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	if ((dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize) {
+		for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 			attrnames_t	*namesp;
 
 			if (((context->flags & ATTR_SECURE) != 0) !=
@@ -584,14 +579,13 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			if (context->flags & ATTR_KERNOVAL) {
 				ASSERT(context->flags & ATTR_KERNAMELS);
 				context->count += namesp->attr_namelen +
-					INT_GET(sfe->namelen, ARCH_CONVERT) + 1;
+					sfe->namelen + 1;
 			}
 			else {
 				if (xfs_attr_put_listent(context, namesp,
 						   (char *)sfe->nameval,
 						   (int)sfe->namelen,
-						   (int)INT_GET(sfe->valuelen,
-								ARCH_CONVERT)))
+						   (int)sfe->valuelen))
 					break;
 			}
 			sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
@@ -603,7 +597,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	/*
 	 * It didn't all fit, so we have to sort everything on hashval.
 	 */
-	sbsize = INT_GET(sf->hdr.count, ARCH_CONVERT) * sizeof(*sbuf);
+	sbsize = sf->hdr.count * sizeof(*sbuf);
 	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP);
 
 	/*
@@ -611,8 +605,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	 * the relevant info from only those that match into a buffer.
 	 */
 	nsbuf = 0;
-	for (i = 0, sfe = &sf->list[0];
-			i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 		if (unlikely(
 		    ((char *)sfe < (char *)sf) ||
 		    ((char *)sfe >= ((char *)sf + dp->i_afp->if_bytes)))) {
@@ -636,8 +629,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			continue;
 		}
 		sbp->entno = i;
-		INT_SET(sbp->hash, ARCH_CONVERT,
-			xfs_da_hashname((char *)sfe->nameval, sfe->namelen));
+		sbp->hash = xfs_da_hashname((char *)sfe->nameval, sfe->namelen);
 		sbp->name = (char *)sfe->nameval;
 		sbp->namelen = sfe->namelen;
 		/* These are bytes, and both on-disk, don't endian-flip */
@@ -660,12 +652,12 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	cursor->initted = 1;
 	cursor->blkno = 0;
 	for (sbp = sbuf, i = 0; i < nsbuf; i++, sbp++) {
-		if (INT_GET(sbp->hash, ARCH_CONVERT) == cursor->hashval) {
+		if (sbp->hash == cursor->hashval) {
 			if (cursor->offset == count) {
 				break;
 			}
 			count++;
-		} else if (INT_GET(sbp->hash, ARCH_CONVERT) > cursor->hashval) {
+		} else if (sbp->hash > cursor->hashval) {
 			break;
 		}
 	}
@@ -685,8 +677,8 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			((sbp->flags & XFS_ATTR_ROOT) ? &attr_trusted :
 			  &attr_user);
 
-		if (cursor->hashval != INT_GET(sbp->hash, ARCH_CONVERT)) {
-			cursor->hashval = INT_GET(sbp->hash, ARCH_CONVERT);
+		if (cursor->hashval != sbp->hash) {
+			cursor->hashval = sbp->hash;
 			cursor->offset = 0;
 		}
 		if (context->flags & ATTR_KERNOVAL) {
@@ -696,7 +688,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 		} else {
 			if (xfs_attr_put_listent(context, namesp,
 					sbp->name, sbp->namelen,
-					INT_GET(sbp->valuelen, ARCH_CONVERT)))
+					sbp->valuelen))
 				break;
 		}
 		cursor->offset++;
@@ -720,12 +712,11 @@ xfs_attr_shortform_allfit(xfs_dabuf_t *bp, xfs_inode_t *dp)
 	int bytes, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 
 	entry = &leaf->entries[0];
 	bytes = sizeof(struct xfs_attr_sf_hdr);
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;		/* don't copy partial entries */
 		if (!(entry->flags & XFS_ATTR_LOCAL))
@@ -733,11 +724,11 @@ xfs_attr_shortform_allfit(xfs_dabuf_t *bp, xfs_inode_t *dp)
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, i);
 		if (name_loc->namelen >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
-		if (INT_GET(name_loc->valuelen, ARCH_CONVERT) >= XFS_ATTR_SF_ENTSIZE_MAX)
+		if (be16_to_cpu(name_loc->valuelen) >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
 		bytes += sizeof(struct xfs_attr_sf_entry)-1
 				+ name_loc->namelen
-				+ INT_GET(name_loc->valuelen, ARCH_CONVERT);
+				+ be16_to_cpu(name_loc->valuelen);
 	}
 	if ((dp->i_mount->m_flags & XFS_MOUNT_ATTR2) &&
 	    (bytes == sizeof(struct xfs_attr_sf_hdr)))
@@ -766,8 +757,7 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	ASSERT(bp != NULL);
 	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(dp->i_mount));
 	leaf = (xfs_attr_leafblock_t *)tmpbuffer;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	memset(bp->data, 0, XFS_LBSIZE(dp->i_mount));
 
 	/*
@@ -810,7 +800,7 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	nargs.trans = args->trans;
 	nargs.oknoent = 1;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;	/* don't copy partial entries */
 		if (!entry->nameidx)
@@ -820,8 +810,8 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 		nargs.name = (char *)name_loc->nameval;
 		nargs.namelen = name_loc->namelen;
 		nargs.value = (char *)&name_loc->nameval[nargs.namelen];
-		nargs.valuelen = INT_GET(name_loc->valuelen, ARCH_CONVERT);
-		nargs.hashval = INT_GET(entry->hashval, ARCH_CONVERT);
+		nargs.valuelen = be16_to_cpu(name_loc->valuelen);
+		nargs.hashval = be32_to_cpu(entry->hashval);
 		nargs.flags = (entry->flags & XFS_ATTR_SECURE) ? ATTR_SECURE :
 			      ((entry->flags & XFS_ATTR_ROOT) ? ATTR_ROOT : 0);
 		xfs_attr_shortform_add(&nargs, forkoff);
@@ -875,13 +865,12 @@ xfs_attr_leaf_to_node(xfs_da_args_t *args)
 		goto out;
 	node = bp1->data;
 	leaf = bp2->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	/* both on-disk, don't endian-flip twice */
 	node->btree[0].hashval =
-		leaf->entries[INT_GET(leaf->hdr.count, ARCH_CONVERT)-1 ].hashval;
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 1);
+		leaf->entries[be16_to_cpu(leaf->hdr.count)-1 ].hashval;
+	node->btree[0].before = cpu_to_be32(blkno);
+	node->hdr.count = cpu_to_be16(1);
 	xfs_da_log_buf(args->trans, bp1, 0, XFS_LBSIZE(dp->i_mount) - 1);
 	error = 0;
 out:
@@ -920,19 +909,16 @@ xfs_attr_leaf_create(xfs_da_args_t *args, xfs_dablk_t blkno, xfs_dabuf_t **bpp)
 	leaf = bp->data;
 	memset((char *)leaf, 0, XFS_LBSIZE(dp->i_mount));
 	hdr = &leaf->hdr;
-	INT_SET(hdr->info.magic, ARCH_CONVERT, XFS_ATTR_LEAF_MAGIC);
-	INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount));
+	hdr->info.magic = cpu_to_be16(XFS_ATTR_LEAF_MAGIC);
+	hdr->firstused = cpu_to_be16(XFS_LBSIZE(dp->i_mount));
 	if (!hdr->firstused) {
-		INT_SET(hdr->firstused, ARCH_CONVERT,
+		hdr->firstused = cpu_to_be16(
 			XFS_LBSIZE(dp->i_mount) - XFS_ATTR_LEAF_NAME_ALIGN);
 	}
 
-	INT_SET(hdr->freemap[0].base, ARCH_CONVERT,
-						sizeof(xfs_attr_leaf_hdr_t));
-	INT_SET(hdr->freemap[0].size, ARCH_CONVERT,
-					  INT_GET(hdr->firstused, ARCH_CONVERT)
-					- INT_GET(hdr->freemap[0].base,
-								ARCH_CONVERT));
+	hdr->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	hdr->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr->firstused) -
+					   sizeof(xfs_attr_leaf_hdr_t));
 
 	xfs_da_log_buf(args->trans, bp, 0, XFS_LBSIZE(dp->i_mount) - 1);
 
@@ -1004,10 +990,9 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	int tablesize, entsize, sum, tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	ASSERT((args->index >= 0)
-		&& (args->index <= INT_GET(leaf->hdr.count, ARCH_CONVERT)));
+		&& (args->index <= be16_to_cpu(leaf->hdr.count)));
 	hdr = &leaf->hdr;
 	entsize = xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 			   args->trans->t_mountp->m_sb.sb_blocksize, NULL);
@@ -1016,26 +1001,25 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * Search through freemap for first-fit on new name length.
 	 * (may need to figure in size of entry struct too)
 	 */
-	tablesize = (INT_GET(hdr->count, ARCH_CONVERT) + 1)
+	tablesize = (be16_to_cpu(hdr->count) + 1)
 					* sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[XFS_ATTR_LEAF_MAPSIZE-1];
 	for (sum = 0, i = XFS_ATTR_LEAF_MAPSIZE-1; i >= 0; map--, i--) {
-		if (tablesize > INT_GET(hdr->firstused, ARCH_CONVERT)) {
-			sum += INT_GET(map->size, ARCH_CONVERT);
+		if (tablesize > be16_to_cpu(hdr->firstused)) {
+			sum += be16_to_cpu(map->size);
 			continue;
 		}
 		if (!map->size)
 			continue;	/* no space in this map */
 		tmp = entsize;
-		if (INT_GET(map->base, ARCH_CONVERT)
-				< INT_GET(hdr->firstused, ARCH_CONVERT))
+		if (be16_to_cpu(map->base) < be16_to_cpu(hdr->firstused))
 			tmp += sizeof(xfs_attr_leaf_entry_t);
-		if (INT_GET(map->size, ARCH_CONVERT) >= tmp) {
+		if (be16_to_cpu(map->size) >= tmp) {
 			tmp = xfs_attr_leaf_add_work(bp, args, i);
 			return(tmp);
 		}
-		sum += INT_GET(map->size, ARCH_CONVERT);
+		sum += be16_to_cpu(map->size);
 	}
 
 	/*
@@ -1056,7 +1040,7 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * After compaction, the block is guaranteed to have only one
 	 * free region, in freemap[0].  If it is not big enough, give up.
 	 */
-	if (INT_GET(hdr->freemap[0].size, ARCH_CONVERT)
+	if (be16_to_cpu(hdr->freemap[0].size)
 				< (entsize + sizeof(xfs_attr_leaf_entry_t)))
 		return(XFS_ERROR(ENOSPC));
 
@@ -1079,45 +1063,42 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	int tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	ASSERT((mapindex >= 0) && (mapindex < XFS_ATTR_LEAF_MAPSIZE));
-	ASSERT((args->index >= 0)
-		&& (args->index <= INT_GET(hdr->count, ARCH_CONVERT)));
+	ASSERT((args->index >= 0) && (args->index <= be16_to_cpu(hdr->count)));
 
 	/*
 	 * Force open some space in the entry array and fill it in.
 	 */
 	entry = &leaf->entries[args->index];
-	if (args->index < INT_GET(hdr->count, ARCH_CONVERT)) {
-		tmp  = INT_GET(hdr->count, ARCH_CONVERT) - args->index;
+	if (args->index < be16_to_cpu(hdr->count)) {
+		tmp  = be16_to_cpu(hdr->count) - args->index;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		memmove((char *)(entry+1), (char *)entry, tmp);
 		xfs_da_log_buf(args->trans, bp,
 		    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
 	}
-	INT_MOD(hdr->count, ARCH_CONVERT, 1);
+	be16_add(&hdr->count, 1);
 
 	/*
 	 * Allocate space for the new string (at the end of the run).
 	 */
 	map = &hdr->freemap[mapindex];
 	mp = args->trans->t_mountp;
-	ASSERT(INT_GET(map->base, ARCH_CONVERT) < XFS_LBSIZE(mp));
-	ASSERT((INT_GET(map->base, ARCH_CONVERT) & 0x3) == 0);
-	ASSERT(INT_GET(map->size, ARCH_CONVERT) >=
+	ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
+	ASSERT((be16_to_cpu(map->base) & 0x3) == 0);
+	ASSERT(be16_to_cpu(map->size) >=
 		xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 					 mp->m_sb.sb_blocksize, NULL));
-	ASSERT(INT_GET(map->size, ARCH_CONVERT) < XFS_LBSIZE(mp));
-	ASSERT((INT_GET(map->size, ARCH_CONVERT) & 0x3) == 0);
-	INT_MOD(map->size, ARCH_CONVERT,
+	ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
+	ASSERT((be16_to_cpu(map->size) & 0x3) == 0);
+	be16_add(&map->size,
 		-xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 					  mp->m_sb.sb_blocksize, &tmp));
-	INT_SET(entry->nameidx, ARCH_CONVERT,
-					INT_GET(map->base, ARCH_CONVERT)
-				      + INT_GET(map->size, ARCH_CONVERT));
-	INT_SET(entry->hashval, ARCH_CONVERT, args->hashval);
+	entry->nameidx = cpu_to_be16(be16_to_cpu(map->base) +
+				     be16_to_cpu(map->size));
+	entry->hashval = cpu_to_be32(args->hashval);
 	entry->flags = tmp ? XFS_ATTR_LOCAL : 0;
 	entry->flags |= (args->flags & ATTR_SECURE) ? XFS_ATTR_SECURE :
 			((args->flags & ATTR_ROOT) ? XFS_ATTR_ROOT : 0);
@@ -1130,12 +1111,10 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	}
 	xfs_da_log_buf(args->trans, bp,
 			  XFS_DA_LOGRANGE(leaf, entry, sizeof(*entry)));
-	ASSERT((args->index == 0) || (INT_GET(entry->hashval, ARCH_CONVERT)
-						>= INT_GET((entry-1)->hashval,
-							    ARCH_CONVERT)));
-	ASSERT((args->index == INT_GET(hdr->count, ARCH_CONVERT)-1) ||
-	       (INT_GET(entry->hashval, ARCH_CONVERT)
-			    <= (INT_GET((entry+1)->hashval, ARCH_CONVERT))));
+	ASSERT((args->index == 0) ||
+	       (be32_to_cpu(entry->hashval) >= be32_to_cpu((entry-1)->hashval)));
+	ASSERT((args->index == be16_to_cpu(hdr->count)-1) ||
+	       (be32_to_cpu(entry->hashval) <= be32_to_cpu((entry+1)->hashval)));
 
 	/*
 	 * Copy the attribute name and value into the new space.
@@ -1149,10 +1128,10 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	if (entry->flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
 		name_loc->namelen = args->namelen;
-		INT_SET(name_loc->valuelen, ARCH_CONVERT, args->valuelen);
+		name_loc->valuelen = cpu_to_be16(args->valuelen);
 		memcpy((char *)name_loc->nameval, args->name, args->namelen);
 		memcpy((char *)&name_loc->nameval[args->namelen], args->value,
-				   INT_GET(name_loc->valuelen, ARCH_CONVERT));
+				   be16_to_cpu(name_loc->valuelen));
 	} else {
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
 		name_rmt->namelen = args->namelen;
@@ -1171,28 +1150,23 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	/*
 	 * Update the control info for this leaf node
 	 */
-	if (INT_GET(entry->nameidx, ARCH_CONVERT)
-				< INT_GET(hdr->firstused, ARCH_CONVERT)) {
+	if (be16_to_cpu(entry->nameidx) < be16_to_cpu(hdr->firstused)) {
 		/* both on-disk, don't endian-flip twice */
 		hdr->firstused = entry->nameidx;
 	}
-	ASSERT(INT_GET(hdr->firstused, ARCH_CONVERT)
-				>= ((INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(*entry))+sizeof(*hdr)));
-	tmp = (INT_GET(hdr->count, ARCH_CONVERT)-1)
-					* sizeof(xfs_attr_leaf_entry_t)
+	ASSERT(be16_to_cpu(hdr->firstused) >=
+	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
+	tmp = (be16_to_cpu(hdr->count)-1) * sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[0];
 	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		if (INT_GET(map->base, ARCH_CONVERT) == tmp) {
-			INT_MOD(map->base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_entry_t));
-			INT_MOD(map->size, ARCH_CONVERT,
-					-sizeof(xfs_attr_leaf_entry_t));
+		if (be16_to_cpu(map->base) == tmp) {
+			be16_add(&map->base, sizeof(xfs_attr_leaf_entry_t));
+			be16_add(&map->size,
+				 -((int)sizeof(xfs_attr_leaf_entry_t)));
 		}
 	}
-	INT_MOD(hdr->usedbytes, ARCH_CONVERT,
-				xfs_attr_leaf_entsize(leaf, args->index));
+	be16_add(&hdr->usedbytes, xfs_attr_leaf_entsize(leaf, args->index));
 	xfs_da_log_buf(args->trans, bp,
 		XFS_DA_LOGRANGE(leaf, hdr, sizeof(*hdr)));
 	return(0);
@@ -1223,28 +1197,25 @@ xfs_attr_leaf_compact(xfs_trans_t *trans, xfs_dabuf_t *bp)
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
 	hdr_d->info = hdr_s->info;	/* struct copy */
-	INT_SET(hdr_d->firstused, ARCH_CONVERT, XFS_LBSIZE(mp));
+	hdr_d->firstused = cpu_to_be16(XFS_LBSIZE(mp));
 	/* handle truncation gracefully */
 	if (!hdr_d->firstused) {
-		INT_SET(hdr_d->firstused, ARCH_CONVERT,
+		hdr_d->firstused = cpu_to_be16(
 				XFS_LBSIZE(mp) - XFS_ATTR_LEAF_NAME_ALIGN);
 	}
 	hdr_d->usedbytes = 0;
 	hdr_d->count = 0;
 	hdr_d->holes = 0;
-	INT_SET(hdr_d->freemap[0].base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_hdr_t));
-	INT_SET(hdr_d->freemap[0].size, ARCH_CONVERT,
-				INT_GET(hdr_d->firstused, ARCH_CONVERT)
-			      - INT_GET(hdr_d->freemap[0].base, ARCH_CONVERT));
+	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused) -
+					     sizeof(xfs_attr_leaf_hdr_t));
 
 	/*
 	 * Copy all entry's in the same (sorted) order,
 	 * but allocate name/value pairs packed and in sequence.
 	 */
 	xfs_attr_leaf_moveents(leaf_s, 0, leaf_d, 0,
-				(int)INT_GET(hdr_s->count, ARCH_CONVERT), mp);
-
+				be16_to_cpu(hdr_s->count), mp);
 	xfs_da_log_buf(trans, bp, 0, XFS_LBSIZE(mp) - 1);
 
 	kmem_free(tmpbuffer, XFS_LBSIZE(mp));
@@ -1279,10 +1250,8 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(blk2->magic == XFS_ATTR_LEAF_MAGIC);
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	args = state->args;
 
 	/*
@@ -1319,22 +1288,21 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	/*
 	 * Move any entries required from leaf to leaf:
 	 */
-	if (count < INT_GET(hdr1->count, ARCH_CONVERT)) {
+	if (count < be16_to_cpu(hdr1->count)) {
 		/*
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count = INT_GET(hdr1->count, ARCH_CONVERT) - count;
-		space  = INT_GET(hdr1->usedbytes, ARCH_CONVERT) - totallen;
+		count = be16_to_cpu(hdr1->count) - count;
+		space  = be16_to_cpu(hdr1->usedbytes) - totallen;
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf2 is the destination, compact it if it looks tight.
 		 */
-		max  = INT_GET(hdr2->firstused, ARCH_CONVERT)
+		max  = be16_to_cpu(hdr2->firstused)
 						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= INT_GET(hdr2->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
+		max -= be16_to_cpu(hdr2->count) * sizeof(xfs_attr_leaf_entry_t);
 		if (space > max) {
 			xfs_attr_leaf_compact(args->trans, blk2->bp);
 		}
@@ -1342,13 +1310,12 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		/*
 		 * Move high entries from leaf1 to low end of leaf2.
 		 */
-		xfs_attr_leaf_moveents(leaf1,
-				INT_GET(hdr1->count, ARCH_CONVERT)-count,
+		xfs_attr_leaf_moveents(leaf1, be16_to_cpu(hdr1->count) - count,
 				leaf2, 0, count, state->mp);
 
 		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
 		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
-	} else if (count > INT_GET(hdr1->count, ARCH_CONVERT)) {
+	} else if (count > be16_to_cpu(hdr1->count)) {
 		/*
 		 * I assert that since all callers pass in an empty
 		 * second buffer, this code should never execute.
@@ -1358,17 +1325,16 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count -= INT_GET(hdr1->count, ARCH_CONVERT);
-		space  = totallen - INT_GET(hdr1->usedbytes, ARCH_CONVERT);
+		count -= be16_to_cpu(hdr1->count);
+		space  = totallen - be16_to_cpu(hdr1->usedbytes);
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf1 is the destination, compact it if it looks tight.
 		 */
-		max  = INT_GET(hdr1->firstused, ARCH_CONVERT)
+		max  = be16_to_cpu(hdr1->firstused)
 						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= INT_GET(hdr1->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
+		max -= be16_to_cpu(hdr1->count) * sizeof(xfs_attr_leaf_entry_t);
 		if (space > max) {
 			xfs_attr_leaf_compact(args->trans, blk1->bp);
 		}
@@ -1377,8 +1343,7 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Move low entries from leaf2 to high end of leaf1.
 		 */
 		xfs_attr_leaf_moveents(leaf2, 0, leaf1,
-				(int)INT_GET(hdr1->count, ARCH_CONVERT), count,
-				state->mp);
+				be16_to_cpu(hdr1->count), count, state->mp);
 
 		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
 		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
@@ -1387,12 +1352,10 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	blk1->hashval =
-	    INT_GET(leaf1->entries[INT_GET(leaf1->hdr.count,
-				    ARCH_CONVERT)-1].hashval, ARCH_CONVERT);
-	blk2->hashval =
-	    INT_GET(leaf2->entries[INT_GET(leaf2->hdr.count,
-				    ARCH_CONVERT)-1].hashval, ARCH_CONVERT);
+	blk1->hashval = be32_to_cpu(
+		leaf1->entries[be16_to_cpu(leaf1->hdr.count)-1].hashval);
+	blk2->hashval = be32_to_cpu(
+		leaf2->entries[be16_to_cpu(leaf2->hdr.count)-1].hashval);
 
 	/*
 	 * Adjust the expected index for insertion.
@@ -1406,13 +1369,12 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * inserting.  The index/blkno fields refer to the "old" entry,
 	 * while the index2/blkno2 fields refer to the "new" entry.
 	 */
-	if (blk1->index > INT_GET(leaf1->hdr.count, ARCH_CONVERT)) {
+	if (blk1->index > be16_to_cpu(leaf1->hdr.count)) {
 		ASSERT(state->inleaf == 0);
-		blk2->index = blk1->index
-				- INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+		blk2->index = blk1->index - be16_to_cpu(leaf1->hdr.count);
 		args->index = args->index2 = blk2->index;
 		args->blkno = args->blkno2 = blk2->blkno;
-	} else if (blk1->index == INT_GET(leaf1->hdr.count, ARCH_CONVERT)) {
+	} else if (blk1->index == be16_to_cpu(leaf1->hdr.count)) {
 		if (state->inleaf) {
 			args->index = blk1->index;
 			args->blkno = blk1->blkno;
@@ -1420,7 +1382,7 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 			args->blkno2 = blk2->blkno;
 		} else {
 			blk2->index = blk1->index
-				    - INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+				    - be16_to_cpu(leaf1->hdr.count);
 			args->index = args->index2 = blk2->index;
 			args->blkno = args->blkno2 = blk2->blkno;
 		}
@@ -1464,15 +1426,14 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 	 * Examine entries until we reduce the absolute difference in
 	 * byte usage between the two blocks to a minimum.
 	 */
-	max = INT_GET(hdr1->count, ARCH_CONVERT)
-			+ INT_GET(hdr2->count, ARCH_CONVERT);
+	max = be16_to_cpu(hdr1->count) + be16_to_cpu(hdr2->count);
 	half  = (max+1) * sizeof(*entry);
-	half += INT_GET(hdr1->usedbytes, ARCH_CONVERT)
-				+ INT_GET(hdr2->usedbytes, ARCH_CONVERT)
-				+ xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+	half += be16_to_cpu(hdr1->usedbytes) +
+		be16_to_cpu(hdr2->usedbytes) +
+		xfs_attr_leaf_newentsize(
+				state->args->namelen,
+				state->args->valuelen,
+				state->blocksize, NULL);
 	half /= 2;
 	lastdelta = state->blocksize;
 	entry = &leaf1->entries[0];
@@ -1498,7 +1459,7 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 		/*
 		 * Wrap around into the second block if necessary.
 		 */
-		if (count == INT_GET(hdr1->count, ARCH_CONVERT)) {
+		if (count == be16_to_cpu(hdr1->count)) {
 			leaf1 = leaf2;
 			entry = &leaf1->entries[0];
 			index = 0;
@@ -1566,12 +1527,12 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC);
 	leaf = (xfs_attr_leafblock_t *)info;
-	count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(leaf->hdr.count);
 	bytes = sizeof(xfs_attr_leaf_hdr_t) +
 		count * sizeof(xfs_attr_leaf_entry_t) +
-		INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		be16_to_cpu(leaf->hdr.usedbytes);
 	if (bytes > (state->blocksize >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);
@@ -1588,7 +1549,7 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -1610,13 +1571,12 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * to shrink an attribute list over time.
 	 */
 	/* start with smaller blk num */
-	forward = (INT_GET(info->forw, ARCH_CONVERT)
-					< INT_GET(info->back, ARCH_CONVERT));
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -1626,14 +1586,13 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 		ASSERT(bp != NULL);
 
 		leaf = (xfs_attr_leafblock_t *)info;
-		count  = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		count  = be16_to_cpu(leaf->hdr.count);
 		bytes  = state->blocksize - (state->blocksize>>2);
-		bytes -= INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-		count += INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		bytes -= INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+		count += be16_to_cpu(leaf->hdr.count);
+		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
 		bytes -= count * sizeof(xfs_attr_leaf_entry_t);
 		bytes -= sizeof(xfs_attr_leaf_hdr_t);
 		xfs_da_brelse(state->args->trans, bp);
@@ -1685,21 +1644,18 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_mount_t *mp;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	mp = args->trans->t_mountp;
-	ASSERT((INT_GET(hdr->count, ARCH_CONVERT) > 0)
-		&& (INT_GET(hdr->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
+	ASSERT((be16_to_cpu(hdr->count) > 0)
+		&& (be16_to_cpu(hdr->count) < (XFS_LBSIZE(mp)/8)));
 	ASSERT((args->index >= 0)
-		&& (args->index < INT_GET(hdr->count, ARCH_CONVERT)));
-	ASSERT(INT_GET(hdr->firstused, ARCH_CONVERT)
-				>= ((INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(*entry))+sizeof(*hdr)));
+		&& (args->index < be16_to_cpu(hdr->count)));
+	ASSERT(be16_to_cpu(hdr->firstused) >=
+	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
 	entry = &leaf->entries[args->index];
-	ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr->firstused, ARCH_CONVERT));
-	ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT) < XFS_LBSIZE(mp));
+	ASSERT(be16_to_cpu(entry->nameidx) >= be16_to_cpu(hdr->firstused));
+	ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
 
 	/*
 	 * Scan through free region table:
@@ -1707,33 +1663,30 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 *    find smallest free region in case we need to replace it,
 	 *    adjust any map that borders the entry table,
 	 */
-	tablesize = INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t)
+	tablesize = be16_to_cpu(hdr->count) * sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[0];
-	tmp = INT_GET(map->size, ARCH_CONVERT);
+	tmp = be16_to_cpu(map->size);
 	before = after = -1;
 	smallest = XFS_ATTR_LEAF_MAPSIZE - 1;
 	entsize = xfs_attr_leaf_entsize(leaf, args->index);
 	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		ASSERT(INT_GET(map->base, ARCH_CONVERT) < XFS_LBSIZE(mp));
-		ASSERT(INT_GET(map->size, ARCH_CONVERT) < XFS_LBSIZE(mp));
-		if (INT_GET(map->base, ARCH_CONVERT) == tablesize) {
-			INT_MOD(map->base, ARCH_CONVERT,
-					-sizeof(xfs_attr_leaf_entry_t));
-			INT_MOD(map->size, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_entry_t));
+		ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
+		ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
+		if (be16_to_cpu(map->base) == tablesize) {
+			be16_add(&map->base,
+				 -((int)sizeof(xfs_attr_leaf_entry_t)));
+			be16_add(&map->size, sizeof(xfs_attr_leaf_entry_t));
 		}
 
-		if ((INT_GET(map->base, ARCH_CONVERT)
-					+ INT_GET(map->size, ARCH_CONVERT))
-				== INT_GET(entry->nameidx, ARCH_CONVERT)) {
+		if ((be16_to_cpu(map->base) + be16_to_cpu(map->size))
+				== be16_to_cpu(entry->nameidx)) {
 			before = i;
-		} else if (INT_GET(map->base, ARCH_CONVERT)
-			== (INT_GET(entry->nameidx, ARCH_CONVERT) + entsize)) {
+		} else if (be16_to_cpu(map->base)
+			== (be16_to_cpu(entry->nameidx) + entsize)) {
 			after = i;
-		} else if (INT_GET(map->size, ARCH_CONVERT) < tmp) {
-			tmp = INT_GET(map->size, ARCH_CONVERT);
+		} else if (be16_to_cpu(map->size) < tmp) {
+			tmp = be16_to_cpu(map->size);
 			smallest = i;
 		}
 	}
@@ -1745,38 +1698,35 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	if ((before >= 0) || (after >= 0)) {
 		if ((before >= 0) && (after >= 0)) {
 			map = &hdr->freemap[before];
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
-			INT_MOD(map->size, ARCH_CONVERT,
-				INT_GET(hdr->freemap[after].size,
-							ARCH_CONVERT));
+			be16_add(&map->size, entsize);
+			be16_add(&map->size,
+				 be16_to_cpu(hdr->freemap[after].size));
 			hdr->freemap[after].base = 0;
 			hdr->freemap[after].size = 0;
 		} else if (before >= 0) {
 			map = &hdr->freemap[before];
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
+			be16_add(&map->size, entsize);
 		} else {
 			map = &hdr->freemap[after];
 			/* both on-disk, don't endian flip twice */
 			map->base = entry->nameidx;
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
+			be16_add(&map->size, entsize);
 		}
 	} else {
 		/*
 		 * Replace smallest region (if it is smaller than free'd entry)
 		 */
 		map = &hdr->freemap[smallest];
-		if (INT_GET(map->size, ARCH_CONVERT) < entsize) {
-			INT_SET(map->base, ARCH_CONVERT,
-					INT_GET(entry->nameidx, ARCH_CONVERT));
-			INT_SET(map->size, ARCH_CONVERT, entsize);
+		if (be16_to_cpu(map->size) < entsize) {
+			map->base = cpu_to_be16(be16_to_cpu(entry->nameidx));
+			map->size = cpu_to_be16(entsize);
 		}
 	}
 
 	/*
 	 * Did we remove the first entry?
 	 */
-	if (INT_GET(entry->nameidx, ARCH_CONVERT)
-				== INT_GET(hdr->firstused, ARCH_CONVERT))
+	if (be16_to_cpu(entry->nameidx) == be16_to_cpu(hdr->firstused))
 		smallest = 1;
 	else
 		smallest = 0;
@@ -1785,18 +1735,18 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * Compress the remaining entries and zero out the removed stuff.
 	 */
 	memset(XFS_ATTR_LEAF_NAME(leaf, args->index), 0, entsize);
-	INT_MOD(hdr->usedbytes, ARCH_CONVERT, -entsize);
+	be16_add(&hdr->usedbytes, -entsize);
 	xfs_da_log_buf(args->trans, bp,
 	     XFS_DA_LOGRANGE(leaf, XFS_ATTR_LEAF_NAME(leaf, args->index),
 				   entsize));
 
-	tmp = (INT_GET(hdr->count, ARCH_CONVERT) - args->index)
+	tmp = (be16_to_cpu(hdr->count) - args->index)
 					* sizeof(xfs_attr_leaf_entry_t);
 	memmove((char *)entry, (char *)(entry+1), tmp);
-	INT_MOD(hdr->count, ARCH_CONVERT, -1);
+	be16_add(&hdr->count, -1);
 	xfs_da_log_buf(args->trans, bp,
 	    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
-	entry = &leaf->entries[INT_GET(hdr->count, ARCH_CONVERT)];
+	entry = &leaf->entries[be16_to_cpu(hdr->count)];
 	memset((char *)entry, 0, sizeof(xfs_attr_leaf_entry_t));
 
 	/*
@@ -1808,18 +1758,17 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	if (smallest) {
 		tmp = XFS_LBSIZE(mp);
 		entry = &leaf->entries[0];
-		for (i = INT_GET(hdr->count, ARCH_CONVERT)-1;
-						i >= 0; entry++, i--) {
-			ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr->firstused, ARCH_CONVERT));
-			ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-							< XFS_LBSIZE(mp));
-			if (INT_GET(entry->nameidx, ARCH_CONVERT) < tmp)
-				tmp = INT_GET(entry->nameidx, ARCH_CONVERT);
+		for (i = be16_to_cpu(hdr->count)-1; i >= 0; entry++, i--) {
+			ASSERT(be16_to_cpu(entry->nameidx) >=
+			       be16_to_cpu(hdr->firstused));
+			ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+
+			if (be16_to_cpu(entry->nameidx) < tmp)
+				tmp = be16_to_cpu(entry->nameidx);
 		}
-		INT_SET(hdr->firstused, ARCH_CONVERT, tmp);
+		hdr->firstused = cpu_to_be16(tmp);
 		if (!hdr->firstused) {
-			INT_SET(hdr->firstused, ARCH_CONVERT,
+			hdr->firstused = cpu_to_be16(
 					tmp - XFS_ATTR_LEAF_NAME_ALIGN);
 		}
 	} else {
@@ -1833,9 +1782,8 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * "join" the leaf with a sibling if so.
 	 */
 	tmp  = sizeof(xfs_attr_leaf_hdr_t);
-	tmp += INT_GET(leaf->hdr.count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
-	tmp += INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+	tmp += be16_to_cpu(leaf->hdr.count) * sizeof(xfs_attr_leaf_entry_t);
+	tmp += be16_to_cpu(leaf->hdr.usedbytes);
 	return(tmp < mp->m_attr_magicpct); /* leaf is < 37% full */
 }
 
@@ -1859,20 +1807,16 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_ATTR_LEAF_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	drop_hdr = &drop_leaf->hdr;
 	save_hdr = &save_leaf->hdr;
 
 	/*
 	 * Save last hashval from dying block for later Btree fixup.
 	 */
-	drop_blk->hashval =
-		INT_GET(drop_leaf->entries[INT_GET(drop_leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-								ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(
+		drop_leaf->entries[be16_to_cpu(drop_leaf->hdr.count)-1].hashval);
 
 	/*
 	 * Check if we need a temp buffer, or can we do it in place.
@@ -1886,12 +1830,11 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 		 */
 		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
 			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf, 0,
-			     (int)INT_GET(drop_hdr->count, ARCH_CONVERT), mp);
+			     be16_to_cpu(drop_hdr->count), mp);
 		} else {
 			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf,
-				  INT_GET(save_hdr->count, ARCH_CONVERT),
-				  (int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				  mp);
+				  be16_to_cpu(save_hdr->count),
+				  be16_to_cpu(drop_hdr->count), mp);
 		}
 	} else {
 		/*
@@ -1905,28 +1848,24 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 		tmp_hdr = &tmp_leaf->hdr;
 		tmp_hdr->info = save_hdr->info;	/* struct copy */
 		tmp_hdr->count = 0;
-		INT_SET(tmp_hdr->firstused, ARCH_CONVERT, state->blocksize);
+		tmp_hdr->firstused = cpu_to_be16(state->blocksize);
 		if (!tmp_hdr->firstused) {
-			INT_SET(tmp_hdr->firstused, ARCH_CONVERT,
+			tmp_hdr->firstused = cpu_to_be16(
 				state->blocksize - XFS_ATTR_LEAF_NAME_ALIGN);
 		}
 		tmp_hdr->usedbytes = 0;
 		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
 			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf, 0,
-				(int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(drop_hdr->count), mp);
 			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf,
-				  INT_GET(tmp_leaf->hdr.count, ARCH_CONVERT),
-				 (int)INT_GET(save_hdr->count, ARCH_CONVERT),
-				 mp);
+				  be16_to_cpu(tmp_leaf->hdr.count),
+				  be16_to_cpu(save_hdr->count), mp);
 		} else {
 			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf, 0,
-				(int)INT_GET(save_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(save_hdr->count), mp);
 			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf,
-				INT_GET(tmp_leaf->hdr.count, ARCH_CONVERT),
-				(int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(tmp_leaf->hdr.count),
+				be16_to_cpu(drop_hdr->count), mp);
 		}
 		memcpy((char *)save_leaf, (char *)tmp_leaf, state->blocksize);
 		kmem_free(tmpbuffer, state->blocksize);
@@ -1938,10 +1877,8 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	save_blk->hashval =
-		INT_GET(save_leaf->entries[INT_GET(save_leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-								ARCH_CONVERT);
+	save_blk->hashval = be32_to_cpu(
+		save_leaf->entries[be16_to_cpu(save_leaf->hdr.count)-1].hashval);
 }
 
 /*========================================================================
@@ -1972,48 +1909,45 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_dahash_t hashval;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count)
 					< (XFS_LBSIZE(args->dp->i_mount)/8));
 
 	/*
 	 * Binary search.  (note: small blocks will skip this loop)
 	 */
 	hashval = args->hashval;
-	probe = span = INT_GET(leaf->hdr.count, ARCH_CONVERT) / 2;
+	probe = span = be16_to_cpu(leaf->hdr.count) / 2;
 	for (entry = &leaf->entries[probe]; span > 4;
 		   entry = &leaf->entries[probe]) {
 		span /= 2;
-		if (INT_GET(entry->hashval, ARCH_CONVERT) < hashval)
+		if (be32_to_cpu(entry->hashval) < hashval)
 			probe += span;
-		else if (INT_GET(entry->hashval, ARCH_CONVERT) > hashval)
+		else if (be32_to_cpu(entry->hashval) > hashval)
 			probe -= span;
 		else
 			break;
 	}
 	ASSERT((probe >= 0) && 
 	       (!leaf->hdr.count
-	       || (probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))));
-	ASSERT((span <= 4) || (INT_GET(entry->hashval, ARCH_CONVERT)
-							== hashval));
+	       || (probe < be16_to_cpu(leaf->hdr.count))));
+	ASSERT((span <= 4) || (be32_to_cpu(entry->hashval) == hashval));
 
 	/*
 	 * Since we may have duplicate hashval's, find the first matching
 	 * hashval in the leaf.
 	 */
-	while ((probe > 0) && (INT_GET(entry->hashval, ARCH_CONVERT)
-							>= hashval)) {
+	while ((probe > 0) && (be32_to_cpu(entry->hashval) >= hashval)) {
 		entry--;
 		probe--;
 	}
-	while ((probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-		&& (INT_GET(entry->hashval, ARCH_CONVERT) < hashval)) {
+	while ((probe < be16_to_cpu(leaf->hdr.count)) &&
+	       (be32_to_cpu(entry->hashval) < hashval)) {
 		entry++;
 		probe++;
 	}
-	if ((probe == INT_GET(leaf->hdr.count, ARCH_CONVERT))
-		    || (INT_GET(entry->hashval, ARCH_CONVERT) != hashval)) {
+	if ((probe == be16_to_cpu(leaf->hdr.count)) ||
+	    (be32_to_cpu(entry->hashval) != hashval)) {
 		args->index = probe;
 		return(XFS_ERROR(ENOATTR));
 	}
@@ -2021,8 +1955,8 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	/*
 	 * Duplicate keys may be present, so search all of them for a match.
 	 */
-	for (  ; (probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-			&& (INT_GET(entry->hashval, ARCH_CONVERT) == hashval);
+	for (  ; (probe < be16_to_cpu(leaf->hdr.count)) &&
+			(be32_to_cpu(entry->hashval) == hashval);
 			entry++, probe++) {
 /*
  * GROT: Add code to remove incomplete entries.
@@ -2064,11 +1998,9 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 			    ((entry->flags & XFS_ATTR_ROOT) != 0))
 				continue;
 			args->index = probe;
-			args->rmtblkno
-				  = INT_GET(name_rmt->valueblk, ARCH_CONVERT);
+			args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
 			args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount,
-						   INT_GET(name_rmt->valuelen,
-								ARCH_CONVERT));
+						   be32_to_cpu(name_rmt->valuelen));
 			return(XFS_ERROR(EEXIST));
 		}
 	}
@@ -2090,18 +2022,17 @@ xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_attr_leaf_name_remote_t *name_rmt;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count)
 					< (XFS_LBSIZE(args->dp->i_mount)/8));
-	ASSERT(args->index < ((int)INT_GET(leaf->hdr.count, ARCH_CONVERT)));
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 
 	entry = &leaf->entries[args->index];
 	if (entry->flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
 		ASSERT(name_loc->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_loc->nameval, args->namelen) == 0);
-		valuelen = INT_GET(name_loc->valuelen, ARCH_CONVERT);
+		valuelen = be16_to_cpu(name_loc->valuelen);
 		if (args->flags & ATTR_KERNOVAL) {
 			args->valuelen = valuelen;
 			return(0);
@@ -2116,8 +2047,8 @@ xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
 		ASSERT(name_rmt->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_rmt->name, args->namelen) == 0);
-		valuelen = INT_GET(name_rmt->valuelen, ARCH_CONVERT);
-		args->rmtblkno = INT_GET(name_rmt->valueblk, ARCH_CONVERT);
+		valuelen = be32_to_cpu(name_rmt->valuelen);
+		args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
 		args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, valuelen);
 		if (args->flags & ATTR_KERNOVAL) {
 			args->valuelen = valuelen;
@@ -2159,32 +2090,29 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Set up environment.
 	 */
-	ASSERT(INT_GET(leaf_s->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf_d->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_s->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_d->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
-	ASSERT((INT_GET(hdr_s->count, ARCH_CONVERT) > 0)
-				&& (INT_GET(hdr_s->count, ARCH_CONVERT)
-						< (XFS_LBSIZE(mp)/8)));
-	ASSERT(INT_GET(hdr_s->firstused, ARCH_CONVERT) >=
-		((INT_GET(hdr_s->count, ARCH_CONVERT)
+	ASSERT((be16_to_cpu(hdr_s->count) > 0) &&
+	       (be16_to_cpu(hdr_s->count) < (XFS_LBSIZE(mp)/8)));
+	ASSERT(be16_to_cpu(hdr_s->firstused) >=
+		((be16_to_cpu(hdr_s->count)
 					* sizeof(*entry_s))+sizeof(*hdr_s)));
-	ASSERT(INT_GET(hdr_d->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8));
-	ASSERT(INT_GET(hdr_d->firstused, ARCH_CONVERT) >=
-		((INT_GET(hdr_d->count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(hdr_d->count) < (XFS_LBSIZE(mp)/8));
+	ASSERT(be16_to_cpu(hdr_d->firstused) >=
+		((be16_to_cpu(hdr_d->count)
 					* sizeof(*entry_d))+sizeof(*hdr_d)));
 
-	ASSERT(start_s < INT_GET(hdr_s->count, ARCH_CONVERT));
-	ASSERT(start_d <= INT_GET(hdr_d->count, ARCH_CONVERT));
-	ASSERT(count <= INT_GET(hdr_s->count, ARCH_CONVERT));
+	ASSERT(start_s < be16_to_cpu(hdr_s->count));
+	ASSERT(start_d <= be16_to_cpu(hdr_d->count));
+	ASSERT(count <= be16_to_cpu(hdr_s->count));
 
 	/*
 	 * Move the entries in the destination leaf up to make a hole?
 	 */
-	if (start_d < INT_GET(hdr_d->count, ARCH_CONVERT)) {
-		tmp  = INT_GET(hdr_d->count, ARCH_CONVERT) - start_d;
+	if (start_d < be16_to_cpu(hdr_d->count)) {
+		tmp  = be16_to_cpu(hdr_d->count) - start_d;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_d->entries[start_d];
 		entry_d = &leaf_d->entries[start_d + count];
@@ -2199,8 +2127,8 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	entry_d = &leaf_d->entries[start_d];
 	desti = start_d;
 	for (i = 0; i < count; entry_s++, entry_d++, desti++, i++) {
-		ASSERT(INT_GET(entry_s->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr_s->firstused, ARCH_CONVERT));
+		ASSERT(be16_to_cpu(entry_s->nameidx)
+				>= be16_to_cpu(hdr_s->firstused));
 		tmp = xfs_attr_leaf_entsize(leaf_s, start_s + i);
 #ifdef GROT
 		/*
@@ -2210,35 +2138,35 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 		 */
 		if (entry_s->flags & XFS_ATTR_INCOMPLETE) { /* skip partials? */
 			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			INT_MOD(hdr_s->usedbytes, ARCH_CONVERT, -tmp);
-			INT_MOD(hdr_s->count, ARCH_CONVERT, -1);
+			be16_add(&hdr_s->usedbytes, -tmp);
+			be16_add(&hdr_s->count, -1);
 			entry_d--;	/* to compensate for ++ in loop hdr */
 			desti--;
 			if ((start_s + i) < offset)
 				result++;	/* insertion index adjustment */
 		} else {
 #endif /* GROT */
-			INT_MOD(hdr_d->firstused, ARCH_CONVERT, -tmp);
+			be16_add(&hdr_d->firstused, -tmp);
 			/* both on-disk, don't endian flip twice */
 			entry_d->hashval = entry_s->hashval;
 			/* both on-disk, don't endian flip twice */
 			entry_d->nameidx = hdr_d->firstused;
 			entry_d->flags = entry_s->flags;
-			ASSERT(INT_GET(entry_d->nameidx, ARCH_CONVERT) + tmp
+			ASSERT(be16_to_cpu(entry_d->nameidx) + tmp
 							<= XFS_LBSIZE(mp));
 			memmove(XFS_ATTR_LEAF_NAME(leaf_d, desti),
 				XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), tmp);
-			ASSERT(INT_GET(entry_s->nameidx, ARCH_CONVERT) + tmp
+			ASSERT(be16_to_cpu(entry_s->nameidx) + tmp
 							<= XFS_LBSIZE(mp));
 			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			INT_MOD(hdr_s->usedbytes, ARCH_CONVERT, -tmp);
-			INT_MOD(hdr_d->usedbytes, ARCH_CONVERT, tmp);
-			INT_MOD(hdr_s->count, ARCH_CONVERT, -1);
-			INT_MOD(hdr_d->count, ARCH_CONVERT, 1);
-			tmp = INT_GET(hdr_d->count, ARCH_CONVERT)
+			be16_add(&hdr_s->usedbytes, -tmp);
+			be16_add(&hdr_d->usedbytes, tmp);
+			be16_add(&hdr_s->count, -1);
+			be16_add(&hdr_d->count, 1);
+			tmp = be16_to_cpu(hdr_d->count)
 						* sizeof(xfs_attr_leaf_entry_t)
 						+ sizeof(xfs_attr_leaf_hdr_t);
-			ASSERT(INT_GET(hdr_d->firstused, ARCH_CONVERT) >= tmp);
+			ASSERT(be16_to_cpu(hdr_d->firstused) >= tmp);
 #ifdef GROT
 		}
 #endif /* GROT */
@@ -2247,7 +2175,7 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Zero out the entries we just copied.
 	 */
-	if (start_s == INT_GET(hdr_s->count, ARCH_CONVERT)) {
+	if (start_s == be16_to_cpu(hdr_s->count)) {
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_s->entries[start_s];
 		ASSERT(((char *)entry_s + tmp) <=
@@ -2258,15 +2186,14 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 		 * Move the remaining entries down to fill the hole,
 		 * then zero the entries at the top.
 		 */
-		tmp  = INT_GET(hdr_s->count, ARCH_CONVERT) - count;
+		tmp  = be16_to_cpu(hdr_s->count) - count;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_s->entries[start_s + count];
 		entry_d = &leaf_s->entries[start_s];
 		memmove((char *)entry_d, (char *)entry_s, tmp);
 
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_s->entries[INT_GET(hdr_s->count,
-							ARCH_CONVERT)];
+		entry_s = &leaf_s->entries[be16_to_cpu(hdr_s->count)];
 		ASSERT(((char *)entry_s + tmp) <=
 		       ((char *)leaf_s + XFS_LBSIZE(mp)));
 		memset((char *)entry_s, 0, tmp);
@@ -2275,14 +2202,11 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Fill in the freemap information
 	 */
-	INT_SET(hdr_d->freemap[0].base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_hdr_t));
-	INT_MOD(hdr_d->freemap[0].base, ARCH_CONVERT,
-				INT_GET(hdr_d->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t));
-	INT_SET(hdr_d->freemap[0].size, ARCH_CONVERT,
-				INT_GET(hdr_d->firstused, ARCH_CONVERT)
-			      - INT_GET(hdr_d->freemap[0].base, ARCH_CONVERT));
+	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	be16_add(&hdr_d->freemap[0].base, be16_to_cpu(hdr_d->count) *
+			sizeof(xfs_attr_leaf_entry_t));
+	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused)
+			      - be16_to_cpu(hdr_d->freemap[0].base));
 	hdr_d->freemap[1].base = 0;
 	hdr_d->freemap[2].base = 0;
 	hdr_d->freemap[1].size = 0;
@@ -2301,18 +2225,16 @@ xfs_attr_leaf_order(xfs_dabuf_t *leaf1_bp, xfs_dabuf_t *leaf2_bp)
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT((INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC) &&
-	       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC));
-	if (   (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0)
-	    && (INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0)
-	    && (   (INT_GET(leaf2->entries[ 0 ].hashval, ARCH_CONVERT) <
-		      INT_GET(leaf1->entries[ 0 ].hashval, ARCH_CONVERT))
-		|| (INT_GET(leaf2->entries[INT_GET(leaf2->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT) <
-		      INT_GET(leaf1->entries[INT_GET(leaf1->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT))) ) {
+	ASSERT((be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC) &&
+	       (be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC));
+	if ((be16_to_cpu(leaf1->hdr.count) > 0) &&
+	    (be16_to_cpu(leaf2->hdr.count) > 0) &&
+	    ((be32_to_cpu(leaf2->entries[0].hashval) <
+	      be32_to_cpu(leaf1->entries[0].hashval)) ||
+	     (be32_to_cpu(leaf2->entries[
+			be16_to_cpu(leaf2->hdr.count)-1].hashval) <
+	      be32_to_cpu(leaf1->entries[
+			be16_to_cpu(leaf1->hdr.count)-1].hashval)))) {
 		return(1);
 	}
 	return(0);
@@ -2327,14 +2249,12 @@ xfs_attr_leaf_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_attr_leafblock_t *leaf;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	if (count)
-		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(leaf->hdr.count);
 	if (!leaf->hdr.count)
 		return(0);
-	return(INT_GET(leaf->entries[INT_GET(leaf->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+	return be32_to_cpu(leaf->entries[be16_to_cpu(leaf->hdr.count)-1].hashval);
 }
 
 /*
@@ -2348,13 +2268,11 @@ xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index)
 	xfs_attr_leaf_name_remote_t *name_rmt;
 	int size;
 
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	if (leaf->entries[index].flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, index);
 		size = XFS_ATTR_LEAF_ENTSIZE_LOCAL(name_loc->namelen,
-						   INT_GET(name_loc->valuelen,
-								ARCH_CONVERT));
+						   be16_to_cpu(name_loc->valuelen));
 	} else {
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, index);
 		size = XFS_ATTR_LEAF_ENTSIZE_REMOTE(name_rmt->namelen);
@@ -2412,22 +2330,20 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 	 */
 	if (context->resynch) {
 		entry = &leaf->entries[0];
-		for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT);
-							entry++, i++) {
-			if (INT_GET(entry->hashval, ARCH_CONVERT)
-							== cursor->hashval) {
+		for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+			if (be32_to_cpu(entry->hashval) == cursor->hashval) {
 				if (cursor->offset == context->dupcnt) {
 					context->dupcnt = 0;
 					break;
 				}
 				context->dupcnt++;
-			} else if (INT_GET(entry->hashval, ARCH_CONVERT)
-							> cursor->hashval) {
+			} else if (be32_to_cpu(entry->hashval) >
+					cursor->hashval) {
 				context->dupcnt = 0;
 				break;
 			}
 		}
-		if (i == INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
+		if (i == be16_to_cpu(leaf->hdr.count)) {
 			xfs_attr_trace_l_c("not found", context);
 			return(0);
 		}
@@ -2441,12 +2357,12 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 	 * We have found our place, start copying out the new attributes.
 	 */
 	retval = 0;
-	for (  ; (i < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+	for (  ; (i < be16_to_cpu(leaf->hdr.count))
 	     && (retval == 0); entry++, i++) {
 		attrnames_t	*namesp;
 
-		if (INT_GET(entry->hashval, ARCH_CONVERT) != cursor->hashval) {
-			cursor->hashval = INT_GET(entry->hashval, ARCH_CONVERT);
+		if (be32_to_cpu(entry->hashval) != cursor->hashval) {
+			cursor->hashval = be32_to_cpu(entry->hashval);
 			cursor->offset = 0;
 		}
 
@@ -2475,8 +2391,7 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 				retval = xfs_attr_put_listent(context, namesp,
 					(char *)name_loc->nameval,
 					(int)name_loc->namelen,
-					(int)INT_GET(name_loc->valuelen,
-								ARCH_CONVERT));
+					be16_to_cpu(name_loc->valuelen));
 			}
 		} else {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
@@ -2488,8 +2403,7 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 				retval = xfs_attr_put_listent(context, namesp,
 					(char *)name_rmt->name,
 					(int)name_rmt->namelen,
-					(int)INT_GET(name_rmt->valuelen,
-								ARCH_CONVERT));
+					be32_to_cpu(name_rmt->valuelen));
 			}
 		}
 		if (retval == 0) {
@@ -2596,9 +2510,8 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 	ASSERT(bp != NULL);
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 	ASSERT(args->index >= 0);
 	entry = &leaf->entries[ args->index ];
 	ASSERT(entry->flags & XFS_ATTR_INCOMPLETE);
@@ -2613,7 +2526,7 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 		namelen = name_rmt->namelen;
 		name = (char *)name_rmt->name;
 	}
-	ASSERT(INT_GET(entry->hashval, ARCH_CONVERT) == args->hashval);
+	ASSERT(be32_to_cpu(entry->hashval) == args->hashval);
 	ASSERT(namelen == args->namelen);
 	ASSERT(memcmp(name, args->name, namelen) == 0);
 #endif /* DEBUG */
@@ -2625,8 +2538,8 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 	if (args->rmtblkno) {
 		ASSERT((entry->flags & XFS_ATTR_LOCAL) == 0);
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
-		INT_SET(name_rmt->valueblk, ARCH_CONVERT, args->rmtblkno);
-		INT_SET(name_rmt->valuelen, ARCH_CONVERT, args->valuelen);
+		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
+		name_rmt->valuelen = cpu_to_be32(args->valuelen);
 		xfs_da_log_buf(args->trans, bp,
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
@@ -2663,9 +2576,8 @@ xfs_attr_leaf_setflag(xfs_da_args_t *args)
 	ASSERT(bp != NULL);
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 	ASSERT(args->index >= 0);
 	entry = &leaf->entries[ args->index ];
 
@@ -2736,16 +2648,14 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	}
 
 	leaf1 = bp1->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf1->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf1->hdr.count));
 	ASSERT(args->index >= 0);
 	entry1 = &leaf1->entries[ args->index ];
 
 	leaf2 = bp2->data;
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index2 < INT_GET(leaf2->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index2 < be16_to_cpu(leaf2->hdr.count));
 	ASSERT(args->index2 >= 0);
 	entry2 = &leaf2->entries[ args->index2 ];
 
@@ -2768,7 +2678,7 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 		namelen2 = name_rmt->namelen;
 		name2 = (char *)name_rmt->name;
 	}
-	ASSERT(INT_GET(entry1->hashval, ARCH_CONVERT) == INT_GET(entry2->hashval, ARCH_CONVERT));
+	ASSERT(be32_to_cpu(entry1->hashval) == be32_to_cpu(entry2->hashval));
 	ASSERT(namelen1 == namelen2);
 	ASSERT(memcmp(name1, name2, namelen1) == 0);
 #endif /* DEBUG */
@@ -2782,8 +2692,8 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	if (args->rmtblkno) {
 		ASSERT((entry1->flags & XFS_ATTR_LOCAL) == 0);
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf1, args->index);
-		INT_SET(name_rmt->valueblk, ARCH_CONVERT, args->rmtblkno);
-		INT_SET(name_rmt->valuelen, ARCH_CONVERT, args->valuelen);
+		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
+		name_rmt->valuelen = cpu_to_be32(args->valuelen);
 		xfs_da_log_buf(args->trans, bp1,
 			 XFS_DA_LOGRANGE(leaf1, name_rmt, sizeof(*name_rmt)));
 	}
@@ -2842,9 +2752,9 @@ xfs_attr_root_inactive(xfs_trans_t **trans, xfs_inode_t *dp)
 	 * This is a depth-first traversal!
 	 */
 	info = bp->data;
-	if (INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+	if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
 		error = xfs_attr_node_inactive(trans, dp, bp, 1);
-	} else if (INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC) {
+	} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
 		error = xfs_attr_leaf_inactive(trans, dp, bp);
 	} else {
 		error = XFS_ERROR(EIO);
@@ -2892,15 +2802,14 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 	}
 
 	node = bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT)
-						== XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	parent_blkno = xfs_da_blkno(bp);	/* save for re-read later */
-	count = INT_GET(node->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(node->hdr.count);
 	if (!count) {
 		xfs_da_brelse(*trans, bp);
 		return(0);
 	}
-	child_fsb = INT_GET(node->btree[0].before, ARCH_CONVERT);
+	child_fsb = be32_to_cpu(node->btree[0].before);
 	xfs_da_brelse(*trans, bp);	/* no locks for later trans */
 
 	/*
@@ -2927,12 +2836,10 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 			 * Invalidate the subtree, however we have to.
 			 */
 			info = child_bp->data;
-			if (INT_GET(info->magic, ARCH_CONVERT)
-							== XFS_DA_NODE_MAGIC) {
+			if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
 				error = xfs_attr_node_inactive(trans, dp,
 						child_bp, level+1);
-			} else if (INT_GET(info->magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC) {
+			} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
 				error = xfs_attr_leaf_inactive(trans, dp,
 						child_bp);
 			} else {
@@ -2962,7 +2869,7 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 				&bp, XFS_ATTR_FORK);
 			if (error)
 				return(error);
-			child_fsb = INT_GET(node->btree[i+1].before, ARCH_CONVERT);
+			child_fsb = be32_to_cpu(node->btree[i+1].before);
 			xfs_da_brelse(*trans, bp);
 		}
 		/*
@@ -2991,17 +2898,16 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	int error, count, size, tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 
 	/*
 	 * Count the number of "remote" value extents.
 	 */
 	count = 0;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
-		if (   INT_GET(entry->nameidx, ARCH_CONVERT)
-		    && ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
 			if (name_rmt->valueblk)
 				count++;
@@ -3027,17 +2933,14 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	 */
 	lp = list;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
-		if (   INT_GET(entry->nameidx, ARCH_CONVERT)
-		    && ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
 			if (name_rmt->valueblk) {
-				/* both on-disk, don't endian flip twice */
-				lp->valueblk = name_rmt->valueblk;
-				INT_SET(lp->valuelen, ARCH_CONVERT,
-						XFS_B_TO_FSB(dp->i_mount,
-						    INT_GET(name_rmt->valuelen,
-							      ARCH_CONVERT)));
+				lp->valueblk = be32_to_cpu(name_rmt->valueblk);
+				lp->valuelen = XFS_B_TO_FSB(dp->i_mount,
+						    be32_to_cpu(name_rmt->valuelen));
 				lp++;
 			}
 		}
@@ -3050,10 +2953,8 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	error = 0;
 	for (lp = list, i = 0; i < count; i++, lp++) {
 		tmp = xfs_attr_leaf_freextent(trans, dp,
-						     INT_GET(lp->valueblk,
-								ARCH_CONVERT),
-						     INT_GET(lp->valuelen,
-								ARCH_CONVERT));
+				lp->valueblk, lp->valuelen);
+
 		if (error == 0)
 			error = tmp;	/* save only the 1st errno */
 	}
diff --git a/fs/xfs/xfs_attr_leaf.h b/fs/xfs/xfs_attr_leaf.h
index 541e34109bb9..51c3ee156b2f 100644
--- a/fs/xfs/xfs_attr_leaf.h
+++ b/fs/xfs/xfs_attr_leaf.h
@@ -73,39 +73,39 @@ struct xfs_trans;
 #define XFS_ATTR_LEAF_MAPSIZE	3	/* how many freespace slots */
 
 typedef struct xfs_attr_leaf_map {	/* RLE map of free bytes */
-	__uint16_t	base;	 	/* base of free region */
-	__uint16_t	size;	  	/* length of free region */
+	__be16	base;			  /* base of free region */
+	__be16	size;			  /* length of free region */
 } xfs_attr_leaf_map_t;
 
 typedef struct xfs_attr_leaf_hdr {	/* constant-structure header block */
 	xfs_da_blkinfo_t info;		/* block type, links, etc. */
-	__uint16_t	count;		/* count of active leaf_entry's */
-	__uint16_t	usedbytes;	/* num bytes of names/values stored */
-	__uint16_t	firstused;	/* first used byte in name area */
-	__uint8_t	holes;		/* != 0 if blk needs compaction */
-	__uint8_t	pad1;
+	__be16	count;			/* count of active leaf_entry's */
+	__be16	usedbytes;		/* num bytes of names/values stored */
+	__be16	firstused;		/* first used byte in name area */
+	__u8	holes;			/* != 0 if blk needs compaction */
+	__u8	pad1;
 	xfs_attr_leaf_map_t freemap[XFS_ATTR_LEAF_MAPSIZE];
 					/* N largest free regions */
 } xfs_attr_leaf_hdr_t;
 
 typedef struct xfs_attr_leaf_entry {	/* sorted on key, not name */
-	xfs_dahash_t	hashval;	/* hash value of name */
-	__uint16_t	nameidx;	/* index into buffer of name/value */
-	__uint8_t	flags;		/* LOCAL/ROOT/SECURE/INCOMPLETE flag */
-	__uint8_t	pad2;		/* unused pad byte */
+	__be32	hashval;		/* hash value of name */
+ 	__be16	nameidx;		/* index into buffer of name/value */
+	__u8	flags;			/* LOCAL/ROOT/SECURE/INCOMPLETE flag */
+	__u8	pad2;			/* unused pad byte */
 } xfs_attr_leaf_entry_t;
 
 typedef struct xfs_attr_leaf_name_local {
-	__uint16_t	valuelen;	/* number of bytes in value */
-	__uint8_t	namelen;	/* length of name bytes */
-	__uint8_t	nameval[1];	/* name/value bytes */
+	__be16	valuelen;		/* number of bytes in value */
+	__u8	namelen;		/* length of name bytes */
+	__u8	nameval[1];		/* name/value bytes */
 } xfs_attr_leaf_name_local_t;
 
 typedef struct xfs_attr_leaf_name_remote {
-	xfs_dablk_t	valueblk;	/* block number of value bytes */
-	__uint32_t	valuelen;	/* number of bytes in value */
-	__uint8_t	namelen;	/* length of name bytes */
-	__uint8_t	name[1];	/* name bytes */
+	__be32	valueblk;		/* block number of value bytes */
+	__be32	valuelen;		/* number of bytes in value */
+	__u8	namelen;		/* length of name bytes */
+	__u8	name[1];		/* name bytes */
 } xfs_attr_leaf_name_remote_t;
 
 typedef struct xfs_attr_leafblock {
@@ -143,8 +143,8 @@ typedef struct xfs_attr_leafblock {
 static inline xfs_attr_leaf_name_remote_t *
 xfs_attr_leaf_name_remote(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (xfs_attr_leaf_name_remote_t *) &((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)];
+	return (xfs_attr_leaf_name_remote_t *)
+		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 #define XFS_ATTR_LEAF_NAME_LOCAL(leafp,idx)	\
@@ -152,16 +152,15 @@ xfs_attr_leaf_name_remote(xfs_attr_leafblock_t *leafp, int idx)
 static inline xfs_attr_leaf_name_local_t *
 xfs_attr_leaf_name_local(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (xfs_attr_leaf_name_local_t *) &((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)];
+	return (xfs_attr_leaf_name_local_t *)
+		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 #define XFS_ATTR_LEAF_NAME(leafp,idx)		\
 	xfs_attr_leaf_name(leafp,idx)
 static inline char *xfs_attr_leaf_name(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (&((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)]);
+	return &((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 /*
diff --git a/fs/xfs/xfs_attr_sf.h b/fs/xfs/xfs_attr_sf.h
index ffed6ca81a52..f67f917803b1 100644
--- a/fs/xfs/xfs_attr_sf.h
+++ b/fs/xfs/xfs_attr_sf.h
@@ -32,8 +32,8 @@ struct xfs_inode;
  */
 typedef struct xfs_attr_shortform {
 	struct xfs_attr_sf_hdr {	/* constant-structure header block */
-		__uint16_t totsize;	/* total bytes in shortform list */
-		__uint8_t count;	/* count of active entries */
+		__be16	totsize;	/* total bytes in shortform list */
+		__u8	count;	/* count of active entries */
 	} hdr;
 	struct xfs_attr_sf_entry {
 		__uint8_t namelen;	/* actual length of name (no NULL) */
@@ -66,8 +66,8 @@ typedef struct xfs_attr_sf_sort {
 #define XFS_ATTR_SF_NEXTENTRY(sfep)		/* next entry in struct */ \
 	((xfs_attr_sf_entry_t *)((char *)(sfep) + XFS_ATTR_SF_ENTSIZE(sfep)))
 #define XFS_ATTR_SF_TOTSIZE(dp)			/* total space in use */ \
-	(INT_GET(((xfs_attr_shortform_t *)	\
-		((dp)->i_afp->if_u1.if_data))->hdr.totsize, ARCH_CONVERT))
+	(be16_to_cpu(((xfs_attr_shortform_t *)	\
+		((dp)->i_afp->if_u1.if_data))->hdr.totsize))
 
 #if defined(XFS_ATTR_TRACE)
 /*
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 70625e577c70..2d702e4a74a3 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -89,7 +89,7 @@ xfs_bmap_add_attrfork_local(
 	int			*flags);	/* inode logging flags */
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after allocating space (or doing a delayed allocation).
  */
 STATIC int				/* error */
@@ -97,7 +97,7 @@ xfs_bmap_add_extent(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
@@ -113,7 +113,7 @@ xfs_bmap_add_extent_delay_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
@@ -129,7 +129,7 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp,/* inode logging flags */
 	int			rsvd);	/* OK to allocate reserved blocks */
 
@@ -142,7 +142,7 @@ xfs_bmap_add_extent_hole_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork); /* data or attr fork */
 
@@ -155,7 +155,7 @@ xfs_bmap_add_extent_unwritten_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp); /* inode logging flags */
 
 /*
@@ -169,7 +169,7 @@ xfs_bmap_alloc(
 /*
  * Transform a btree format file with only one leaf node, where the
  * extents list will fit in the inode, into an extents format file.
- * Since the extent list is already in-core, all we have to do is
+ * Since the file extents are already in-core, all we have to do is
  * give up the space for the btree root and pitch the leaf block.
  */
 STATIC int				/* error */
@@ -191,7 +191,7 @@ xfs_bmap_check_extents(
 #endif
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after removing space (or undoing a delayed allocation).
  */
 STATIC int				/* error */
@@ -201,7 +201,7 @@ xfs_bmap_del_extent(
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp,/* inode logging flags */
 	int			whichfork, /* data or attr fork */
 	int			rsvd);	 /* OK to allocate reserved blocks */
@@ -217,18 +217,6 @@ xfs_bmap_del_free(
 	xfs_bmap_free_item_t	*free);	/* list item to be freed */
 
 /*
- * Remove count entries from the extents array for inode "ip", starting
- * at index "idx".  Copies the remaining items down over the deleted ones,
- * and gives back the excess memory.
- */
-STATIC void
-xfs_bmap_delete_exlist(
-	xfs_inode_t	*ip,		/* incode inode pointer */
-	xfs_extnum_t	idx,		/* starting delete index */
-	xfs_extnum_t	count,		/* count of items to delete */
-	int		whichfork);	/* data or attr fork */
-
-/*
  * Convert an extents-format file into a btree-format file.
  * The new file will have a root block (in the inode) and a single child block.
  */
@@ -244,18 +232,6 @@ xfs_bmap_extents_to_btree(
 	int			whichfork);	/* data or attr fork */
 
 /*
- * Insert new item(s) in the extent list for inode "ip".
- * Count new items are inserted at offset idx.
- */
-STATIC void
-xfs_bmap_insert_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting index of new items */
-	xfs_extnum_t	count,		/* number of inserted items */
-	xfs_bmbt_irec_t	*new,		/* items to insert */
-	int		whichfork);	/* data or attr fork */
-
-/*
  * Convert a local file to an extents file.
  * This code is sort of bogus, since the file data needs to get
  * logged so it won't be lost.  The bmap-level manipulations are ok, though.
@@ -316,7 +292,7 @@ xfs_bmap_trace_addentry(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_delete_exlist.
+ * Add bmap trace entry prior to a call to xfs_iext_remove.
  */
 STATIC void
 xfs_bmap_trace_delete(
@@ -328,7 +304,7 @@ xfs_bmap_trace_delete(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_insert_exlist, or
+ * Add bmap trace entry prior to a call to xfs_iext_insert, or
  * reading in the extents list from the disk (in the btree).
  */
 STATIC void
@@ -343,7 +319,7 @@ xfs_bmap_trace_insert(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry after updating an extent list entry in place.
+ * Add bmap trace entry after updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_post_update(
@@ -354,7 +330,7 @@ xfs_bmap_trace_post_update(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to updating an extent list entry in place.
+ * Add bmap trace entry prior to updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_pre_update(
@@ -413,19 +389,24 @@ STATIC int
 xfs_bmap_count_tree(
 	xfs_mount_t     *mp,
 	xfs_trans_t     *tp,
+	xfs_ifork_t	*ifp,
 	xfs_fsblock_t   blockno,
 	int             levelin,
 	int		*count);
 
 STATIC int
 xfs_bmap_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	int			numrecs,
 	int			*count);
 
 STATIC int
 xfs_bmap_disk_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_mount_t		*mp,
+	xfs_extnum_t		idx,
+	xfs_bmbt_block_t	*block,
 	int			numrecs,
 	int			*count);
 
@@ -537,7 +518,7 @@ xfs_bmap_add_attrfork_local(
 }
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after allocating space (or doing a delayed allocation).
  */
 STATIC int				/* error */
@@ -545,7 +526,7 @@ xfs_bmap_add_extent(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
@@ -578,7 +559,7 @@ xfs_bmap_add_extent(
 	if (nextents == 0) {
 		xfs_bmap_trace_insert(fname, "insert empty", ip, 0, 1, new,
 			NULL, whichfork);
-		xfs_bmap_insert_exlist(ip, 0, 1, new, whichfork);
+		xfs_iext_insert(ifp, 0, 1, new);
 		ASSERT(cur == NULL);
 		ifp->if_lastex = 0;
 		if (!ISNULLSTARTBLOCK(new->br_startblock)) {
@@ -614,7 +595,7 @@ xfs_bmap_add_extent(
 		/*
 		 * Get the record referred to by idx.
 		 */
-		xfs_bmbt_get_all(&ifp->if_u1.if_extents[idx], &prev);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &prev);
 		/*
 		 * If it's a real allocation record, and the new allocation ends
 		 * after the start of the referred to record, then we're filling
@@ -714,14 +695,13 @@ xfs_bmap_add_extent_delay_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
 	int			rsvd)	/* OK to use reserved data block allocation */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	int			diff;	/* temp value */
 	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
@@ -730,6 +710,7 @@ xfs_bmap_add_extent_delay_real(
 	static char		fname[] = "xfs_bmap_add_extent_delay_real";
 #endif
 	int			i;	/* temp state */
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
 	xfs_bmbt_irec_t		r[3];	/* neighbor extent entries */
 					/* left is 0, right is 1, prev is 2 */
@@ -763,8 +744,8 @@ xfs_bmap_add_extent_delay_real(
 	 * Set up a bunch of variables to make the tests simpler.
 	 */
 	cur = *curp;
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	new_endoff = new->br_startoff + new->br_blockcount;
 	ASSERT(PREV.br_startoff <= new->br_startoff);
@@ -781,7 +762,7 @@ xfs_bmap_add_extent_delay_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &LEFT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
 	}
 	STATE_SET(LEFT_CONTIG,
@@ -798,7 +779,7 @@ xfs_bmap_add_extent_delay_real(
 	if (STATE_SET_TEST(RIGHT_VALID,
 			idx <
 			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(ep + 1, &RIGHT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
 		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
 	}
 	STATE_SET(RIGHT_CONTIG,
@@ -825,14 +806,14 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LF|RF|LC|RC", ip, idx, 2,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 2, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 2);
 		ip->i_df.if_lastex = idx - 1;
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
@@ -867,14 +848,14 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
 		xfs_bmap_trace_delete(fname, "LF|RF|LC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -908,7 +889,7 @@ xfs_bmap_add_extent_delay_real(
 		ip->i_df.if_lastex = idx;
 		xfs_bmap_trace_delete(fname, "LF|RF|RC", ip, idx + 1, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx + 1, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx + 1, 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -964,7 +945,7 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
@@ -1010,7 +991,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmbt_set_blockcount(ep, temp);
 		xfs_bmap_trace_insert(fname, "LF", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1039,8 +1020,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 			STARTBLOCKVAL(PREV.br_startblock) -
 			(cur ? cur->bc_private.b.allocated : 0));
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx + 1];
+		ep = xfs_iext_get_ext(ifp, idx + 1);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "LF", ip, idx + 1,
 			XFS_DATA_FORK);
@@ -1058,7 +1038,8 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmap_trace_pre_update(fname, "RF|RC", ip, idx + 1,
 			XFS_DATA_FORK);
 		xfs_bmbt_set_blockcount(ep, temp);
-		xfs_bmbt_set_allf(ep + 1, new->br_startoff, new->br_startblock,
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount,
 			RIGHT.br_state);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx + 1,
@@ -1098,7 +1079,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmbt_set_blockcount(ep, temp);
 		xfs_bmap_trace_insert(fname, "RF", ip, idx + 1, 1,
 			new, NULL, XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 1, new);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1127,8 +1108,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 			STARTBLOCKVAL(PREV.br_startblock) -
 			(cur ? cur->bc_private.b.allocated : 0));
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx];
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "RF", ip, idx, XFS_DATA_FORK);
 		*dnew = temp;
@@ -1149,7 +1129,7 @@ xfs_bmap_add_extent_delay_real(
 		r[1].br_blockcount = temp2;
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 2, &r[0], &r[1],
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 2, &r[0], XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1204,13 +1184,13 @@ xfs_bmap_add_extent_delay_real(
 				}
 			}
 		}
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx];
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "0", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_pre_update(fname, "0", ip, idx + 2,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_startblock(ep + 2, NULLSTARTBLOCK((int)temp2));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx + 2),
+			NULLSTARTBLOCK((int)temp2));
 		xfs_bmap_trace_post_update(fname, "0", ip, idx + 2,
 			XFS_DATA_FORK);
 		*dnew = temp + temp2;
@@ -1254,10 +1234,9 @@ xfs_bmap_add_extent_unwritten_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp) /* inode logging flags */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
 	int			error;	/* error return value */
@@ -1265,6 +1244,7 @@ xfs_bmap_add_extent_unwritten_real(
 	static char		fname[] = "xfs_bmap_add_extent_unwritten_real";
 #endif
 	int			i;	/* temp state */
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
 	xfs_exntst_t		newext;	/* new extent state */
 	xfs_exntst_t		oldext;	/* old extent state */
@@ -1298,8 +1278,8 @@ xfs_bmap_add_extent_unwritten_real(
 	 */
 	error = 0;
 	cur = *curp;
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	newext = new->br_state;
 	oldext = (newext == XFS_EXT_UNWRITTEN) ?
@@ -1320,7 +1300,7 @@ xfs_bmap_add_extent_unwritten_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &LEFT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
 	}
 	STATE_SET(LEFT_CONTIG,
@@ -1337,7 +1317,7 @@ xfs_bmap_add_extent_unwritten_real(
 	if (STATE_SET_TEST(RIGHT_VALID,
 			idx <
 			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(ep + 1, &RIGHT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
 		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
 	}
 	STATE_SET(RIGHT_CONTIG,
@@ -1363,14 +1343,14 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LF|RF|LC|RC", ip, idx, 2,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 2, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 2);
 		ip->i_df.if_lastex = idx - 1;
 		ip->i_d.di_nextents -= 2;
 		if (cur == NULL)
@@ -1409,14 +1389,14 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
 		xfs_bmap_trace_delete(fname, "LF|RF|LC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1456,7 +1436,7 @@ xfs_bmap_add_extent_unwritten_real(
 		ip->i_df.if_lastex = idx;
 		xfs_bmap_trace_delete(fname, "LF|RF|RC", ip, idx + 1, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx + 1, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx + 1, 1);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1516,7 +1496,7 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
@@ -1571,7 +1551,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_bmap_trace_post_update(fname, "LF", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_insert(fname, "LF", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1609,7 +1589,8 @@ xfs_bmap_add_extent_unwritten_real(
 			PREV.br_blockcount - new->br_blockcount);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_allf(ep + 1, new->br_startoff, new->br_startblock,
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount, newext);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx + 1,
 			XFS_DATA_FORK);
@@ -1649,7 +1630,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_bmap_trace_post_update(fname, "RF", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_insert(fname, "RF", ip, idx + 1, 1,
 			new, NULL, XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 1, new);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1696,7 +1677,7 @@ xfs_bmap_add_extent_unwritten_real(
 		r[1].br_state = oldext;
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 2, &r[0], &r[1],
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 2, &r[0], XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents += 2;
 		if (cur == NULL)
@@ -1770,15 +1751,15 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			rsvd)		/* OK to allocate reserved blocks */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
-	xfs_bmbt_rec_t		*ep;	/* extent list entry for idx */
+	xfs_bmbt_rec_t		*ep;	/* extent record for idx */
 #ifdef XFS_BMAP_TRACE
 	static char		fname[] = "xfs_bmap_add_extent_hole_delay";
 #endif
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
 	xfs_filblks_t		newlen=0;	/* new indirect size */
 	xfs_filblks_t		oldlen=0;	/* old indirect size */
@@ -1799,15 +1780,15 @@ xfs_bmap_add_extent_hole_delay(
 				       ((state &= ~MASK(b)), 0))
 #define	SWITCH_STATE		(state & MASK2(LEFT_CONTIG, RIGHT_CONTIG))
 
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	state = 0;
 	ASSERT(ISNULLSTARTBLOCK(new->br_startblock));
 	/*
 	 * Check and set flags if this segment has a left neighbor
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &left);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
 	}
 	/*
@@ -1844,23 +1825,24 @@ xfs_bmap_add_extent_hole_delay(
 		/*
 		 * New allocation is contiguous with delayed allocations
 		 * on the left and on the right.
-		 * Merge all three into a single extent list entry.
+		 * Merge all three into a single extent record.
 		 */
 		temp = left.br_blockcount + new->br_blockcount +
 			right.br_blockcount;
 		xfs_bmap_trace_pre_update(fname, "LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1, temp);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
 		oldlen = STARTBLOCKVAL(left.br_startblock) +
 			STARTBLOCKVAL(new->br_startblock) +
 			STARTBLOCKVAL(right.br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(ep - 1, NULLSTARTBLOCK((int)newlen));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
+			NULLSTARTBLOCK((int)newlen));
 		xfs_bmap_trace_post_update(fname, "LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LC|RC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		ip->i_df.if_lastex = idx - 1;
 		break;
 
@@ -1873,11 +1855,12 @@ xfs_bmap_add_extent_hole_delay(
 		temp = left.br_blockcount + new->br_blockcount;
 		xfs_bmap_trace_pre_update(fname, "LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1, temp);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
 		oldlen = STARTBLOCKVAL(left.br_startblock) +
 			STARTBLOCKVAL(new->br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(ep - 1, NULLSTARTBLOCK((int)newlen));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
+			NULLSTARTBLOCK((int)newlen));
 		xfs_bmap_trace_post_update(fname, "LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
@@ -1909,7 +1892,7 @@ xfs_bmap_add_extent_hole_delay(
 		oldlen = newlen = 0;
 		xfs_bmap_trace_insert(fname, "0", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		break;
 	}
@@ -1940,7 +1923,7 @@ xfs_bmap_add_extent_hole_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork) /* data or attr fork */
 {
@@ -1970,13 +1953,13 @@ xfs_bmap_add_extent_hole_real(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
-	ep = &ifp->if_u1.if_extents[idx];
+	ep = xfs_iext_get_ext(ifp, idx);
 	state = 0;
 	/*
 	 * Check and set flags if this segment has a left neighbor.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &left);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
 	}
 	/*
@@ -2019,18 +2002,18 @@ xfs_bmap_add_extent_hole_real(
 		/*
 		 * New allocation is contiguous with real allocations on the
 		 * left and on the right.
-		 * Merge all three into a single extent list entry.
+		 * Merge all three into a single extent record.
 		 */
 		xfs_bmap_trace_pre_update(fname, "LC|RC", ip, idx - 1,
 			whichfork);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			left.br_blockcount + new->br_blockcount +
 			right.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LC|RC", ip, idx - 1,
 			whichfork);
 		xfs_bmap_trace_delete(fname, "LC|RC", ip,
 			idx, 1, whichfork);
-		xfs_bmap_delete_exlist(ip, idx, 1, whichfork);
+		xfs_iext_remove(ifp, idx, 1);
 		ifp->if_lastex = idx - 1;
 		XFS_IFORK_NEXT_SET(ip, whichfork,
 			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
@@ -2062,7 +2045,7 @@ xfs_bmap_add_extent_hole_real(
 		 * Merge the new allocation with the left neighbor.
 		 */
 		xfs_bmap_trace_pre_update(fname, "LC", ip, idx - 1, whichfork);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			left.br_blockcount + new->br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LC", ip, idx - 1, whichfork);
 		ifp->if_lastex = idx - 1;
@@ -2116,7 +2099,7 @@ xfs_bmap_add_extent_hole_real(
 		 */
 		xfs_bmap_trace_insert(fname, "0", ip, idx, 1, new, NULL,
 			whichfork);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, whichfork);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ifp->if_lastex = idx;
 		XFS_IFORK_NEXT_SET(ip, whichfork,
 			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
@@ -2311,25 +2294,15 @@ xfs_bmap_extsize_align(
 
 #define XFS_ALLOC_GAP_UNITS	4
 
-/*
- * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
- * It figures out where to ask the underlying allocator to put the new extent.
- */
 STATIC int
-xfs_bmap_alloc(
+xfs_bmap_adjacent(
 	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
 {
 	xfs_fsblock_t	adjust;		/* adjustment to block numbers */
-	xfs_alloctype_t	atype=0;	/* type for allocation routines */
-	int		error;		/* error return value */
 	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
 	xfs_mount_t	*mp;		/* mount point structure */
 	int		nullfb;		/* true if ap->firstblock isn't set */
 	int		rt;		/* true if inode is realtime */
-	xfs_extlen_t	prod = 0;	/* product factor for allocators */
-	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
-	xfs_extlen_t	align;		/* minimum allocation alignment */
-	xfs_rtblock_t	rtx;
 
 #define	ISVALID(x,y)	\
 	(rt ? \
@@ -2338,75 +2311,10 @@ xfs_bmap_alloc(
 		XFS_FSB_TO_AGNO(mp, x) < mp->m_sb.sb_agcount && \
 		XFS_FSB_TO_AGBNO(mp, x) < mp->m_sb.sb_agblocks)
 
-	/*
-	 * Set up variables.
-	 */
 	mp = ap->ip->i_mount;
 	nullfb = ap->firstblock == NULLFSBLOCK;
 	rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata;
 	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
-	if (rt) {
-		align = ap->ip->i_d.di_extsize ?
-			ap->ip->i_d.di_extsize : mp->m_sb.sb_rextsize;
-		/* Set prod to match the extent size */
-		prod = align / mp->m_sb.sb_rextsize;
-
-		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
-						align, rt, ap->eof, 0,
-						ap->conv, &ap->off, &ap->alen);
-		if (error)
-			return error;
-		ASSERT(ap->alen);
-		ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);
-
-		/*
-		 * If the offset & length are not perfectly aligned
-		 * then kill prod, it will just get us in trouble.
-		 */
-		if (do_mod(ap->off, align) || ap->alen % align)
-			prod = 1;
-		/*
-		 * Set ralen to be the actual requested length in rtextents.
-		 */
-		ralen = ap->alen / mp->m_sb.sb_rextsize;
-		/*
-		 * If the old value was close enough to MAXEXTLEN that
-		 * we rounded up to it, cut it back so it's valid again.
-		 * Note that if it's a really large request (bigger than
-		 * MAXEXTLEN), we don't hear about that number, and can't
-		 * adjust the starting point to match it.
-		 */
-		if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
-			ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
-		/*
-		 * If it's an allocation to an empty file at offset 0,
-		 * pick an extent that will space things out in the rt area.
-		 */
-		if (ap->eof && ap->off == 0) {
-			error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
-			if (error)
-				return error;
-			ap->rval = rtx * mp->m_sb.sb_rextsize;
-		} else
-			ap->rval = 0;
-	} else {
-		align = (ap->userdata && ap->ip->i_d.di_extsize &&
-			(ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)) ?
-			ap->ip->i_d.di_extsize : 0;
-		if (unlikely(align)) {
-			error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
-							align, rt,
-							ap->eof, 0, ap->conv,
-							&ap->off, &ap->alen);
-			ASSERT(!error);
-			ASSERT(ap->alen);
-		}
-		if (nullfb)
-			ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
-		else
-			ap->rval = ap->firstblock;
-	}
-
 	/*
 	 * If allocating at eof, and there's a previous real block,
 	 * try to use it's last block as our starting point.
@@ -2531,287 +2439,384 @@ xfs_bmap_alloc(
 		else if (gotbno != NULLFSBLOCK)
 			ap->rval = gotbno;
 	}
+#undef ISVALID
+	return 0;
+}
+
+STATIC int
+xfs_bmap_rtalloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
+	int		error;		/* error return value */
+	xfs_mount_t	*mp;		/* mount point structure */
+	xfs_extlen_t	prod = 0;	/* product factor for allocators */
+	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
+	xfs_extlen_t	align;		/* minimum allocation alignment */
+	xfs_rtblock_t	rtx;		/* realtime extent number */
+	xfs_rtblock_t	rtb;
+
+	mp = ap->ip->i_mount;
+	align = ap->ip->i_d.di_extsize ?
+		ap->ip->i_d.di_extsize : mp->m_sb.sb_rextsize;
+	prod = align / mp->m_sb.sb_rextsize;
+	error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
+					align, 1, ap->eof, 0,
+					ap->conv, &ap->off, &ap->alen);
+	if (error)
+		return error;
+	ASSERT(ap->alen);
+	ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);
+
+	/*
+	 * If the offset & length are not perfectly aligned
+	 * then kill prod, it will just get us in trouble.
+	 */
+	if (do_mod(ap->off, align) || ap->alen % align)
+		prod = 1;
+	/*
+	 * Set ralen to be the actual requested length in rtextents.
+	 */
+	ralen = ap->alen / mp->m_sb.sb_rextsize;
+	/*
+	 * If the old value was close enough to MAXEXTLEN that
+	 * we rounded up to it, cut it back so it's valid again.
+	 * Note that if it's a really large request (bigger than
+	 * MAXEXTLEN), we don't hear about that number, and can't
+	 * adjust the starting point to match it.
+	 */
+	if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
+		ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
+	/*
+	 * If it's an allocation to an empty file at offset 0,
+	 * pick an extent that will space things out in the rt area.
+	 */
+	if (ap->eof && ap->off == 0) {
+		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
+		if (error)
+			return error;
+		ap->rval = rtx * mp->m_sb.sb_rextsize;
+	} else {
+		ap->rval = 0;
+	}
+
+	xfs_bmap_adjacent(ap);
+
+	/*
+	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 */
+	atype = ap->rval == 0 ?  XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
+	do_div(ap->rval, mp->m_sb.sb_rextsize);
+	rtb = ap->rval;
+	ap->alen = ralen;
+	if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
+				&ralen, atype, ap->wasdel, prod, &rtb)))
+		return error;
+	if (rtb == NULLFSBLOCK && prod > 1 &&
+	    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
+					   ap->alen, &ralen, atype,
+					   ap->wasdel, 1, &rtb)))
+		return error;
+	ap->rval = rtb;
+	if (ap->rval != NULLFSBLOCK) {
+		ap->rval *= mp->m_sb.sb_rextsize;
+		ralen *= mp->m_sb.sb_rextsize;
+		ap->alen = ralen;
+		ap->ip->i_d.di_nblocks += ralen;
+		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
+		if (ap->wasdel)
+			ap->ip->i_delayed_blks -= ralen;
+		/*
+		 * Adjust the disk quota also. This was reserved
+		 * earlier.
+		 */
+		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
+			ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
+					XFS_TRANS_DQ_RTBCOUNT, (long) ralen);
+	} else {
+		ap->alen = 0;
+	}
+	return 0;
+}
+
+STATIC int
+xfs_bmap_btalloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	xfs_mount_t	*mp;		/* mount point structure */
+	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
+	xfs_extlen_t	align;		/* minimum allocation alignment */
+	xfs_agnumber_t	ag;
+	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
+	xfs_agnumber_t	startag;
+	xfs_alloc_arg_t	args;
+	xfs_extlen_t	blen;
+	xfs_extlen_t	delta;
+	xfs_extlen_t	longest;
+	xfs_extlen_t	need;
+	xfs_extlen_t	nextminlen = 0;
+	xfs_perag_t	*pag;
+	int		nullfb;		/* true if ap->firstblock isn't set */
+	int		isaligned;
+	int		notinit;
+	int		tryagain;
+	int		error;
+
+	mp = ap->ip->i_mount;
+	align = (ap->userdata && ap->ip->i_d.di_extsize &&
+		(ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)) ?
+		ap->ip->i_d.di_extsize : 0;
+	if (unlikely(align)) {
+		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
+						align, 0, ap->eof, 0, ap->conv,
+						&ap->off, &ap->alen);
+		ASSERT(!error);
+		ASSERT(ap->alen);
+	}
+	nullfb = ap->firstblock == NULLFSBLOCK;
+	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
+	if (nullfb)
+		ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
+	else
+		ap->rval = ap->firstblock;
+
+	xfs_bmap_adjacent(ap);
+
 	/*
 	 * If allowed, use ap->rval; otherwise must use firstblock since
 	 * it's in the right allocation group.
 	 */
-	if (nullfb || rt || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
+	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
 		;
 	else
 		ap->rval = ap->firstblock;
 	/*
-	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 * Normal allocation, done through xfs_alloc_vextent.
 	 */
-	if (rt) {
-#ifndef __KERNEL__
-		ASSERT(0);
-#else
-		xfs_rtblock_t	rtb;
-
-		atype = ap->rval == 0 ?
-			XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
-		do_div(ap->rval, mp->m_sb.sb_rextsize);
-		rtb = ap->rval;
-		ap->alen = ralen;
-		if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
-				&ralen, atype, ap->wasdel, prod, &rtb)))
-			return error;
-		if (rtb == NULLFSBLOCK && prod > 1 &&
-		    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
-						   ap->alen, &ralen, atype,
-						   ap->wasdel, 1, &rtb)))
-			return error;
-		ap->rval = rtb;
-		if (ap->rval != NULLFSBLOCK) {
-			ap->rval *= mp->m_sb.sb_rextsize;
-			ralen *= mp->m_sb.sb_rextsize;
-			ap->alen = ralen;
-			ap->ip->i_d.di_nblocks += ralen;
-			xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
-			if (ap->wasdel)
-				ap->ip->i_delayed_blks -= ralen;
+	tryagain = isaligned = 0;
+	args.tp = ap->tp;
+	args.mp = mp;
+	args.fsbno = ap->rval;
+	args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
+	blen = 0;
+	if (nullfb) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.total = ap->total;
+		/*
+		 * Find the longest available space.
+		 * We're going to try for the whole allocation at once.
+		 */
+		startag = ag = XFS_FSB_TO_AGNO(mp, args.fsbno);
+		notinit = 0;
+		down_read(&mp->m_peraglock);
+		while (blen < ap->alen) {
+			pag = &mp->m_perag[ag];
+			if (!pag->pagf_init &&
+			    (error = xfs_alloc_pagf_init(mp, args.tp,
+				    ag, XFS_ALLOC_FLAG_TRYLOCK))) {
+				up_read(&mp->m_peraglock);
+				return error;
+			}
 			/*
-			 * Adjust the disk quota also. This was reserved
-			 * earlier.
+			 * See xfs_alloc_fix_freelist...
 			 */
-			XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
-				ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
-						XFS_TRANS_DQ_RTBCOUNT,
-				(long) ralen);
-		} else
-			ap->alen = 0;
-#endif	/* __KERNEL__ */
+			if (pag->pagf_init) {
+				need = XFS_MIN_FREELIST_PAG(pag, mp);
+				delta = need > pag->pagf_flcount ?
+					need - pag->pagf_flcount : 0;
+				longest = (pag->pagf_longest > delta) ?
+					(pag->pagf_longest - delta) :
+					(pag->pagf_flcount > 0 ||
+					 pag->pagf_longest > 0);
+				if (blen < longest)
+					blen = longest;
+			} else
+				notinit = 1;
+			if (++ag == mp->m_sb.sb_agcount)
+				ag = 0;
+			if (ag == startag)
+				break;
+		}
+		up_read(&mp->m_peraglock);
+		/*
+		 * Since the above loop did a BUF_TRYLOCK, it is
+		 * possible that there is space for this request.
+		 */
+		if (notinit || blen < ap->minlen)
+			args.minlen = ap->minlen;
+		/*
+		 * If the best seen length is less than the request
+		 * length, use the best as the minimum.
+		 */
+		else if (blen < ap->alen)
+			args.minlen = blen;
+		/*
+		 * Otherwise we've seen an extent as big as alen,
+		 * use that as the minimum.
+		 */
+		else
+			args.minlen = ap->alen;
+	} else if (ap->low) {
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.total = args.minlen = ap->minlen;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+		args.total = ap->total;
+		args.minlen = ap->minlen;
+	}
+	if (unlikely(ap->userdata && ap->ip->i_d.di_extsize &&
+		    (ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE))) {
+		args.prod = ap->ip->i_d.di_extsize;
+		if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
+			args.mod = (xfs_extlen_t)(args.prod - args.mod);
+	} else if (unlikely(mp->m_sb.sb_blocksize >= NBPP)) {
+		args.prod = 1;
+		args.mod = 0;
+	} else {
+		args.prod = NBPP >> mp->m_sb.sb_blocklog;
+		if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
+			args.mod = (xfs_extlen_t)(args.prod - args.mod);
 	}
 	/*
-	 * Normal allocation, done through xfs_alloc_vextent.
+	 * If we are not low on available data blocks, and the
+	 * underlying logical volume manager is a stripe, and
+	 * the file offset is zero then try to allocate data
+	 * blocks on stripe unit boundary.
+	 * NOTE: ap->aeof is only set if the allocation length
+	 * is >= the stripe unit and the allocation offset is
+	 * at the end of file.
 	 */
-	else {
-		xfs_agnumber_t	ag;
-		xfs_alloc_arg_t	args;
-		xfs_extlen_t	blen;
-		xfs_extlen_t	delta;
-		int		isaligned;
-		xfs_extlen_t	longest;
-		xfs_extlen_t	need;
-		xfs_extlen_t	nextminlen=0;
-		int		notinit;
-		xfs_perag_t	*pag;
-		xfs_agnumber_t	startag;
-		int		tryagain;
-
-		tryagain = isaligned = 0;
-		args.tp = ap->tp;
-		args.mp = mp;
-		args.fsbno = ap->rval;
-		args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
-		blen = 0;
-		if (nullfb) {
-			args.type = XFS_ALLOCTYPE_START_BNO;
-			args.total = ap->total;
-			/*
-			 * Find the longest available space.
-			 * We're going to try for the whole allocation at once.
-			 */
-			startag = ag = XFS_FSB_TO_AGNO(mp, args.fsbno);
-			notinit = 0;
-			down_read(&mp->m_peraglock);
-			while (blen < ap->alen) {
-				pag = &mp->m_perag[ag];
-				if (!pag->pagf_init &&
-				    (error = xfs_alloc_pagf_init(mp, args.tp,
-					    ag, XFS_ALLOC_FLAG_TRYLOCK))) {
-					up_read(&mp->m_peraglock);
-					return error;
-				}
-				/*
-				 * See xfs_alloc_fix_freelist...
-				 */
-				if (pag->pagf_init) {
-					need = XFS_MIN_FREELIST_PAG(pag, mp);
-					delta = need > pag->pagf_flcount ?
-						need - pag->pagf_flcount : 0;
-					longest = (pag->pagf_longest > delta) ?
-						(pag->pagf_longest - delta) :
-						(pag->pagf_flcount > 0 ||
-						 pag->pagf_longest > 0);
-					if (blen < longest)
-						blen = longest;
-				} else
-					notinit = 1;
-				if (++ag == mp->m_sb.sb_agcount)
-					ag = 0;
-				if (ag == startag)
-					break;
-			}
-			up_read(&mp->m_peraglock);
+	if (!ap->low && ap->aeof) {
+		if (!ap->off) {
+			args.alignment = mp->m_dalign;
+			atype = args.type;
+			isaligned = 1;
 			/*
-			 * Since the above loop did a BUF_TRYLOCK, it is
-			 * possible that there is space for this request.
+			 * Adjust for alignment
 			 */
-			if (notinit || blen < ap->minlen)
-				args.minlen = ap->minlen;
+			if (blen > args.alignment && blen <= ap->alen)
+				args.minlen = blen - args.alignment;
+			args.minalignslop = 0;
+		} else {
 			/*
-			 * If the best seen length is less than the request
-			 * length, use the best as the minimum.
+			 * First try an exact bno allocation.
+			 * If it fails then do a near or start bno
+			 * allocation with alignment turned on.
 			 */
-			else if (blen < ap->alen)
-				args.minlen = blen;
+			atype = args.type;
+			tryagain = 1;
+			args.type = XFS_ALLOCTYPE_THIS_BNO;
+			args.alignment = 1;
 			/*
-			 * Otherwise we've seen an extent as big as alen,
-			 * use that as the minimum.
+			 * Compute the minlen+alignment for the
+			 * next case.  Set slop so that the value
+			 * of minlen+alignment+slop doesn't go up
+			 * between the calls.
 			 */
+			if (blen > mp->m_dalign && blen <= ap->alen)
+				nextminlen = blen - mp->m_dalign;
 			else
-				args.minlen = ap->alen;
-		} else if (ap->low) {
-			args.type = XFS_ALLOCTYPE_FIRST_AG;
-			args.total = args.minlen = ap->minlen;
-		} else {
-			args.type = XFS_ALLOCTYPE_NEAR_BNO;
-			args.total = ap->total;
-			args.minlen = ap->minlen;
-		}
-		if (unlikely(ap->userdata && ap->ip->i_d.di_extsize &&
-			    (ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE))) {
-			args.prod = ap->ip->i_d.di_extsize;
-			if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
-				args.mod = (xfs_extlen_t)(args.prod - args.mod);
-		} else if (unlikely(mp->m_sb.sb_blocksize >= NBPP)) {
-			args.prod = 1;
-			args.mod = 0;
-		} else {
-			args.prod = NBPP >> mp->m_sb.sb_blocklog;
-			if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
-				args.mod = (xfs_extlen_t)(args.prod - args.mod);
+				nextminlen = args.minlen;
+			if (nextminlen + mp->m_dalign > args.minlen + 1)
+				args.minalignslop =
+					nextminlen + mp->m_dalign -
+					args.minlen - 1;
+			else
+				args.minalignslop = 0;
 		}
+	} else {
+		args.alignment = 1;
+		args.minalignslop = 0;
+	}
+	args.minleft = ap->minleft;
+	args.wasdel = ap->wasdel;
+	args.isfl = 0;
+	args.userdata = ap->userdata;
+	if ((error = xfs_alloc_vextent(&args)))
+		return error;
+	if (tryagain && args.fsbno == NULLFSBLOCK) {
 		/*
-		 * If we are not low on available data blocks, and the
-		 * underlying logical volume manager is a stripe, and
-		 * the file offset is zero then try to allocate data
-		 * blocks on stripe unit boundary.
-		 * NOTE: ap->aeof is only set if the allocation length
-		 * is >= the stripe unit and the allocation offset is
-		 * at the end of file.
+		 * Exact allocation failed. Now try with alignment
+		 * turned on.
 		 */
-		if (!ap->low && ap->aeof) {
-			if (!ap->off) {
-				args.alignment = mp->m_dalign;
-				atype = args.type;
-				isaligned = 1;
-				/*
-				 * Adjust for alignment
-				 */
-				if (blen > args.alignment && blen <= ap->alen)
-					args.minlen = blen - args.alignment;
-				args.minalignslop = 0;
-			} else {
-				/*
-				 * First try an exact bno allocation.
-				 * If it fails then do a near or start bno
-				 * allocation with alignment turned on.
-				 */
-				atype = args.type;
-				tryagain = 1;
-				args.type = XFS_ALLOCTYPE_THIS_BNO;
-				args.alignment = 1;
-				/*
-				 * Compute the minlen+alignment for the
-				 * next case.  Set slop so that the value
-				 * of minlen+alignment+slop doesn't go up
-				 * between the calls.
-				 */
-				if (blen > mp->m_dalign && blen <= ap->alen)
-					nextminlen = blen - mp->m_dalign;
-				else
-					nextminlen = args.minlen;
-				if (nextminlen + mp->m_dalign > args.minlen + 1)
-					args.minalignslop =
-						nextminlen + mp->m_dalign -
-						args.minlen - 1;
-				else
-					args.minalignslop = 0;
-			}
-		} else {
-			args.alignment = 1;
-			args.minalignslop = 0;
-		}
-		args.minleft = ap->minleft;
-		args.wasdel = ap->wasdel;
-		args.isfl = 0;
-		args.userdata = ap->userdata;
+		args.type = atype;
+		args.fsbno = ap->rval;
+		args.alignment = mp->m_dalign;
+		args.minlen = nextminlen;
+		args.minalignslop = 0;
+		isaligned = 1;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	}
+	if (isaligned && args.fsbno == NULLFSBLOCK) {
+		/*
+		 * allocation failed, so turn off alignment and
+		 * try again.
+		 */
+		args.type = atype;
+		args.fsbno = ap->rval;
+		args.alignment = 0;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	}
+	if (args.fsbno == NULLFSBLOCK && nullfb &&
+	    args.minlen > ap->minlen) {
+		args.minlen = ap->minlen;
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.fsbno = ap->rval;
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
-		if (tryagain && args.fsbno == NULLFSBLOCK) {
-			/*
-			 * Exact allocation failed. Now try with alignment
-			 * turned on.
-			 */
-			args.type = atype;
-			args.fsbno = ap->rval;
-			args.alignment = mp->m_dalign;
-			args.minlen = nextminlen;
-			args.minalignslop = 0;
-			isaligned = 1;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (isaligned && args.fsbno == NULLFSBLOCK) {
-			/*
-			 * allocation failed, so turn off alignment and
-			 * try again.
-			 */
-			args.type = atype;
-			args.fsbno = ap->rval;
-			args.alignment = 0;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (args.fsbno == NULLFSBLOCK && nullfb &&
-		    args.minlen > ap->minlen) {
-			args.minlen = ap->minlen;
-			args.type = XFS_ALLOCTYPE_START_BNO;
-			args.fsbno = ap->rval;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (args.fsbno == NULLFSBLOCK && nullfb) {
-			args.fsbno = 0;
-			args.type = XFS_ALLOCTYPE_FIRST_AG;
-			args.total = ap->minlen;
-			args.minleft = 0;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-			ap->low = 1;
-		}
-		if (args.fsbno != NULLFSBLOCK) {
-			ap->firstblock = ap->rval = args.fsbno;
-			ASSERT(nullfb || fb_agno == args.agno ||
-			       (ap->low && fb_agno < args.agno));
-			ap->alen = args.len;
-			ap->ip->i_d.di_nblocks += args.len;
-			xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
-			if (ap->wasdel)
-				ap->ip->i_delayed_blks -= args.len;
-			/*
-			 * Adjust the disk quota also. This was reserved
-			 * earlier.
-			 */
-			XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
-				ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
-						XFS_TRANS_DQ_BCOUNT,
-				(long) args.len);
-		} else {
-			ap->rval = NULLFSBLOCK;
-			ap->alen = 0;
-		}
+	}
+	if (args.fsbno == NULLFSBLOCK && nullfb) {
+		args.fsbno = 0;
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.total = ap->minlen;
+		args.minleft = 0;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+		ap->low = 1;
+	}
+	if (args.fsbno != NULLFSBLOCK) {
+		ap->firstblock = ap->rval = args.fsbno;
+		ASSERT(nullfb || fb_agno == args.agno ||
+		       (ap->low && fb_agno < args.agno));
+		ap->alen = args.len;
+		ap->ip->i_d.di_nblocks += args.len;
+		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
+		if (ap->wasdel)
+			ap->ip->i_delayed_blks -= args.len;
+		/*
+		 * Adjust the disk quota also. This was reserved
+		 * earlier.
+		 */
+		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
+			ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
+					XFS_TRANS_DQ_BCOUNT,
+			(long) args.len);
+	} else {
+		ap->rval = NULLFSBLOCK;
+		ap->alen = 0;
 	}
 	return 0;
-#undef	ISVALID
+}
+
+/*
+ * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
+ * It figures out where to ask the underlying allocator to put the new extent.
+ */
+STATIC int
+xfs_bmap_alloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	if ((ap->ip->i_d.di_flags & XFS_DIFLAG_REALTIME) && ap->userdata)
+		return xfs_bmap_rtalloc(ap);
+	return xfs_bmap_btalloc(ap);
 }
 
 /*
  * Transform a btree format file with only one leaf node, where the
  * extents list will fit in the inode, into an extents format file.
- * Since the extent list is already in-core, all we have to do is
+ * Since the file extents are already in-core, all we have to do is
  * give up the space for the btree root and pitch the leaf block.
  */
 STATIC int				/* error */
@@ -2868,7 +2873,7 @@ xfs_bmap_btree_to_extents(
 }
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after removing space (or undoing a delayed allocation).
  */
 STATIC int				/* error */
@@ -2878,7 +2883,7 @@ xfs_bmap_del_extent(
 	xfs_extnum_t		idx,	/* extent number to update/delete */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*del,	/* data to remove from extent list */
+	xfs_bmbt_irec_t		*del,	/* data to remove from extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork, /* data or attr fork */
 	int			rsvd)	/* OK to allocate reserved blocks */
@@ -2903,7 +2908,6 @@ xfs_bmap_del_extent(
 	xfs_filblks_t		nblks;	/* quota/sb block count */
 	xfs_bmbt_irec_t		new;	/* new record to be inserted */
 	/* REFERENCED */
-	xfs_extnum_t		nextents;	/* number of extents in list */
 	uint			qfield;	/* quota field to update */
 	xfs_filblks_t		temp;	/* for indirect length calculations */
 	xfs_filblks_t		temp2;	/* for indirect length calculations */
@@ -2911,10 +2915,10 @@ xfs_bmap_del_extent(
 	XFS_STATS_INC(xs_del_exlist);
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	ASSERT(idx >= 0 && idx < nextents);
+	ASSERT((idx >= 0) && (idx < ifp->if_bytes /
+		(uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(del->br_blockcount > 0);
-	ep = &ifp->if_u1.if_extents[idx];
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &got);
 	ASSERT(got.br_startoff <= del->br_startoff);
 	del_endoff = del->br_startoff + del->br_blockcount;
@@ -2990,7 +2994,7 @@ xfs_bmap_del_extent(
 		 * Matches the whole extent.  Delete the entry.
 		 */
 		xfs_bmap_trace_delete(fname, "3", ip, idx, 1, whichfork);
-		xfs_bmap_delete_exlist(ip, idx, 1, whichfork);
+		xfs_iext_remove(ifp, idx, 1);
 		ifp->if_lastex = idx;
 		if (delay)
 			break;
@@ -3160,7 +3164,7 @@ xfs_bmap_del_extent(
 		xfs_bmap_trace_post_update(fname, "0", ip, idx, whichfork);
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 1, &new, NULL,
 			whichfork);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, &new, whichfork);
+		xfs_iext_insert(ifp, idx + 1, 1, &new);
 		ifp->if_lastex = idx + 1;
 		break;
 	}
@@ -3213,31 +3217,6 @@ xfs_bmap_del_free(
 }
 
 /*
- * Remove count entries from the extents array for inode "ip", starting
- * at index "idx".  Copies the remaining items down over the deleted ones,
- * and gives back the excess memory.
- */
-STATIC void
-xfs_bmap_delete_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting delete index */
-	xfs_extnum_t	count,		/* count of items to delete */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_t	*base;		/* base of extent list */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extnum_t	nextents;	/* number of extents in list after */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	base = ifp->if_u1.if_extents;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - count;
-	memmove(&base[idx], &base[idx + count],
-		(nextents - idx) * sizeof(*base));
-	xfs_iext_realloc(ip, -count, whichfork);
-}
-
-/*
  * Convert an extents-format file into a btree-format file.
  * The new file will have a root block (in the inode) and a single child block.
  */
@@ -3258,13 +3237,13 @@ xfs_bmap_extents_to_btree(
 	xfs_bmbt_rec_t		*arp;		/* child record pointer */
 	xfs_bmbt_block_t	*block;		/* btree root block */
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
-	xfs_bmbt_rec_t		*ep;		/* extent list pointer */
+	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
 	int			error;		/* error return value */
-	xfs_extnum_t		i, cnt;		/* extent list index */
+	xfs_extnum_t		i, cnt;		/* extent record index */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	xfs_bmbt_key_t		*kp;		/* root block key pointer */
 	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* extent list size */
+	xfs_extnum_t		nextents;	/* number of file extents */
 	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
@@ -3343,7 +3322,8 @@ xfs_bmap_extents_to_btree(
 	ablock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
 	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (ep = ifp->if_u1.if_extents, cnt = i = 0; i < nextents; i++, ep++) {
+	for (cnt = i = 0; i < nextents; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		if (!ISNULLSTARTBLOCK(xfs_bmbt_get_startblock(ep))) {
 			arp->l0 = INT_GET(ep->l0, ARCH_CONVERT);
 			arp->l1 = INT_GET(ep->l1, ARCH_CONVERT);
@@ -3373,34 +3353,6 @@ xfs_bmap_extents_to_btree(
 }
 
 /*
- * Insert new item(s) in the extent list for inode "ip".
- * Count new items are inserted at offset idx.
- */
-STATIC void
-xfs_bmap_insert_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting index of new items */
-	xfs_extnum_t	count,		/* number of inserted items */
-	xfs_bmbt_irec_t	*new,		/* items to insert */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_t	*base;		/* extent list base */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extnum_t	nextents;	/* extent list size */
-	xfs_extnum_t	to;		/* extent list index */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	xfs_iext_realloc(ip, count, whichfork);
-	base = ifp->if_u1.if_extents;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	memmove(&base[idx + count], &base[idx],
-		(nextents - (idx + count)) * sizeof(*base));
-	for (to = idx; to < idx + count; to++, new++)
-		xfs_bmbt_set_all(&base[to], new);
-}
-
-/*
  * Helper routine to reset inode di_forkoff field when switching
  * attribute fork from local to extent format - we reset it where
  * possible to make space available for inline data fork extents.
@@ -3457,12 +3409,13 @@ xfs_bmap_local_to_extents(
 	error = 0;
 	if (ifp->if_bytes) {
 		xfs_alloc_arg_t	args;	/* allocation arguments */
-		xfs_buf_t	*bp;	/* buffer for extent list block */
-		xfs_bmbt_rec_t	*ep;	/* extent list pointer */
+		xfs_buf_t	*bp;	/* buffer for extent block */
+		xfs_bmbt_rec_t	*ep;	/* extent record pointer */
 
 		args.tp = tp;
 		args.mp = ip->i_mount;
-		ASSERT(ifp->if_flags & XFS_IFINLINE);
+		ASSERT((ifp->if_flags &
+			(XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) == XFS_IFINLINE);
 		/*
 		 * Allocate a block.  We know we need only one, since the
 		 * file currently fits in an inode.
@@ -3492,8 +3445,8 @@ xfs_bmap_local_to_extents(
 		xfs_trans_log_buf(tp, bp, 0, ifp->if_bytes - 1);
 		xfs_bmap_forkoff_reset(args.mp, ip, whichfork);
 		xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
-		xfs_iext_realloc(ip, 1, whichfork);
-		ep = ifp->if_u1.if_extents;
+		xfs_iext_add(ifp, 0, 1);
+		ep = xfs_iext_get_ext(ifp, 0);
 		xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
 		xfs_bmap_trace_post_update(fname, "new", ip, 0, whichfork);
 		XFS_IFORK_NEXT_SET(ip, whichfork, 1);
@@ -3518,7 +3471,7 @@ xfs_bmbt_rec_t *			/* pointer to found extent entry */
 xfs_bmap_do_search_extents(
 	xfs_bmbt_rec_t	*base,		/* base of extent list */
 	xfs_extnum_t	lastx,		/* last extent index used */
-	xfs_extnum_t	nextents,	/* extent list size */
+	xfs_extnum_t	nextents,	/* number of file extents */
 	xfs_fileoff_t	bno,		/* block number searched for */
 	int		*eofp,		/* out: end of file found */
 	xfs_extnum_t	*lastxp,	/* out: last extent index */
@@ -3569,9 +3522,9 @@ xfs_bmap_do_search_extents(
 		got.br_blockcount = xfs_bmbt_get_blockcount(ep);
 		*eofp = 0;
 	} else {
-		/* binary search the extents array */
 		low = 0;
 		high = nextents - 1;
+		/* binary search the extents array */
 		while (low <= high) {
 			XFS_STATS_INC(xs_cmp_exlist);
 			lastx = (low + high) >> 1;
@@ -3622,6 +3575,57 @@ xfs_bmap_do_search_extents(
 }
 
 /*
+ * Search the extent records for the entry containing block bno.
+ * If bno lies in a hole, point to the next entry.  If bno lies
+ * past eof, *eofp will be set, and *prevp will contain the last
+ * entry (null if none).  Else, *lastxp will be set to the index
+ * of the found entry; *gotp will contain the entry.
+ */
+xfs_bmbt_rec_t *			/* pointer to found extent entry */
+xfs_bmap_search_multi_extents(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number searched for */
+	int		*eofp,		/* out: end of file found */
+	xfs_extnum_t	*lastxp,	/* out: last extent index */
+	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
+	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
+{
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	lastx;		/* last extent index */
+
+	/*
+	 * Initialize the extent entry structure to catch access to
+	 * uninitialized br_startblock field.
+	 */
+	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
+	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
+	gotp->br_state = XFS_EXT_INVALID;
+#if XFS_BIG_BLKNOS
+	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
+#else
+	gotp->br_startblock = 0xffffa5a5;
+#endif
+	prevp->br_startoff = NULLFILEOFF;
+
+	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
+	if (lastx > 0) {
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
+	}
+	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
+		xfs_bmbt_get_all(ep, gotp);
+		*eofp = 0;
+	} else {
+		if (lastx > 0) {
+			*gotp = *prevp;
+		}
+		*eofp = 1;
+		ep = NULL;
+	}
+	*lastxp = lastx;
+	return ep;
+}
+
+/*
  * Search the extents list for the inode, for the extent containing bno.
  * If bno lies in a hole, point to the next entry.  If bno lies past eof,
  * *eofp will be set, and *prevp will contain the last entry (null if none).
@@ -3639,20 +3643,14 @@ xfs_bmap_search_extents(
 	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
 {
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t  *base;          /* base of extent list */
-	xfs_extnum_t    lastx;          /* last extent index used */
-	xfs_extnum_t    nextents;       /* extent list size */
-	xfs_bmbt_rec_t  *ep;            /* extent list entry pointer */
+	xfs_bmbt_rec_t  *ep;            /* extent record pointer */
 	int		rt;		/* realtime flag    */
 
 	XFS_STATS_INC(xs_look_exlist);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	lastx = ifp->if_lastex;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
 
-	ep = xfs_bmap_do_search_extents(base, lastx, nextents, bno, eofp,
-					  lastxp, gotp, prevp);
+	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
+
 	rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
 	if (unlikely(!rt && !gotp->br_startblock && (*lastxp != NULLEXTNUM))) {
                 cmn_err(CE_PANIC,"Access to block zero: fs: <%s> inode: %lld "
@@ -3732,7 +3730,7 @@ xfs_bmap_trace_addentry(
 }
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_delete_exlist.
+ * Add bmap trace entry prior to a call to xfs_iext_remove.
  */
 STATIC void
 xfs_bmap_trace_delete(
@@ -3747,13 +3745,13 @@ xfs_bmap_trace_delete(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_DELETE, fname, desc, ip, idx,
-		cnt, &ifp->if_u1.if_extents[idx],
-		cnt == 2 ? &ifp->if_u1.if_extents[idx + 1] : NULL,
+		cnt, xfs_iext_get_ext(ifp, idx),
+		cnt == 2 ? xfs_iext_get_ext(ifp, idx + 1) : NULL,
 		whichfork);
 }
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_insert_exlist, or
+ * Add bmap trace entry prior to a call to xfs_iext_insert, or
  * reading in the extents list from the disk (in the btree).
  */
 STATIC void
@@ -3783,7 +3781,7 @@ xfs_bmap_trace_insert(
 }
 
 /*
- * Add bmap trace entry after updating an extent list entry in place.
+ * Add bmap trace entry after updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_post_update(
@@ -3797,11 +3795,11 @@ xfs_bmap_trace_post_update(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_POST_UP, fname, desc, ip, idx,
-		1, &ifp->if_u1.if_extents[idx], NULL, whichfork);
+		1, xfs_iext_get_ext(ifp, idx), NULL, whichfork);
 }
 
 /*
- * Add bmap trace entry prior to updating an extent list entry in place.
+ * Add bmap trace entry prior to updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_pre_update(
@@ -3815,7 +3813,7 @@ xfs_bmap_trace_pre_update(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_PRE_UP, fname, desc, ip, idx, 1,
-		&ifp->if_u1.if_extents[idx], NULL, whichfork);
+		xfs_iext_get_ext(ifp, idx), NULL, whichfork);
 }
 #endif	/* XFS_BMAP_TRACE */
 
@@ -3892,7 +3890,7 @@ xfs_bmap_add_attrfork(
 	int			rsvd)		/* xact may use reserved blks */
 {
 	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
-	xfs_bmap_free_t		flist;		/* freed extent list */
+	xfs_bmap_free_t		flist;		/* freed extent records */
 	xfs_mount_t		*mp;		/* mount structure */
 	xfs_trans_t		*tp;		/* transaction pointer */
 	unsigned long		s;		/* spinlock spl value */
@@ -4146,7 +4144,7 @@ xfs_bmap_finish(
 	xfs_efd_log_item_t	*efd;		/* extent free data */
 	xfs_efi_log_item_t	*efi;		/* extent free intention */
 	int			error;		/* error return value */
-	xfs_bmap_free_item_t	*free;		/* free extent list item */
+	xfs_bmap_free_item_t	*free;		/* free extent item */
 	unsigned int		logres;		/* new log reservation */
 	unsigned int		logcount;	/* new log count */
 	xfs_mount_t		*mp;		/* filesystem mount structure */
@@ -4242,9 +4240,9 @@ xfs_bmap_first_unused(
 	xfs_fileoff_t	*first_unused,		/* unused block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;			/* error return value */
+	int		idx;			/* extent record index */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
 	xfs_fileoff_t	lastaddr;		/* last block number seen */
 	xfs_fileoff_t	lowest;			/* lowest useful block */
@@ -4265,10 +4263,8 @@ xfs_bmap_first_unused(
 		return error;
 	lowest = *first_unused;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
-	for (lastaddr = 0, max = lowest, ep = base;
-	     ep < &base[nextents];
-	     ep++) {
+	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		off = xfs_bmbt_get_startoff(ep);
 		/*
 		 * See if the hole before this extent will work.
@@ -4287,8 +4283,8 @@ xfs_bmap_first_unused(
 /*
  * Returns the file-relative block number of the last block + 1 before
  * last_block (input value) in the file.
- * This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
+ * This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
  */
 int						/* error */
 xfs_bmap_last_before(
@@ -4335,8 +4331,8 @@ xfs_bmap_last_before(
 
 /*
  * Returns the file-relative block number of the first block past eof in
- * the file.  This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
+ * the file.  This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
  */
 int						/* error */
 xfs_bmap_last_offset(
@@ -4345,7 +4341,6 @@ xfs_bmap_last_offset(
 	xfs_fileoff_t	*last_block,		/* last block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to last extent */
 	int		error;			/* error return value */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
@@ -4368,9 +4363,7 @@ xfs_bmap_last_offset(
 		*last_block = 0;
 		return 0;
 	}
-	base = &ifp->if_u1.if_extents[0];
-	ASSERT(base != NULL);
-	ep = &base[nextents - 1];
+	ep = xfs_iext_get_ext(ifp, nextents - 1);
 	*last_block = xfs_bmbt_get_startoff(ep) + xfs_bmbt_get_blockcount(ep);
 	return 0;
 }
@@ -4400,7 +4393,7 @@ xfs_bmap_one_block(
 		return 0;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	ep = ifp->if_u1.if_extents;
+	ep = xfs_iext_get_ext(ifp, 0);
 	xfs_bmbt_get_all(ep, &s);
 	rval = s.br_startoff == 0 && s.br_blockcount == 1;
 	if (rval && whichfork == XFS_DATA_FORK)
@@ -4435,7 +4428,6 @@ xfs_bmap_read_extents(
 	xfs_bmbt_ptr_t		*pp;	/* pointer to block address */
 	/* REFERENCED */
 	xfs_extnum_t		room;	/* number of entries there's room for */
-	xfs_bmbt_rec_t		*trp;	/* target record pointer */
 
 	bno = NULLFSBLOCK;
 	mp = ip->i_mount;
@@ -4478,16 +4470,16 @@ xfs_bmap_read_extents(
 	/*
 	 * Here with bp and block set to the leftmost leaf node in the tree.
 	 */
-	room = ifp->if_bytes / (uint)sizeof(*trp);
-	trp = ifp->if_u1.if_extents;
+	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 	i = 0;
 	/*
-	 * Loop over all leaf nodes.  Copy information to the extent list.
+	 * Loop over all leaf nodes.  Copy information to the extent records.
 	 */
 	for (;;) {
-		xfs_bmbt_rec_t	*frp, *temp;
+		xfs_bmbt_rec_t	*frp, *trp;
 		xfs_fsblock_t	nextbno;
 		xfs_extnum_t	num_recs;
+		xfs_extnum_t	start;
 
 
 		num_recs = be16_to_cpu(block->bb_numrecs);
@@ -4511,12 +4503,13 @@ xfs_bmap_read_extents(
 		if (nextbno != NULLFSBLOCK)
 			xfs_btree_reada_bufl(mp, nextbno, 1);
 		/*
-		 * Copy records into the extent list.
+		 * Copy records into the extent records.
 		 */
 		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
 			block, 1, mp->m_bmap_dmxr[0]);
-		temp = trp;
-		for (j = 0; j < num_recs; j++, frp++, trp++) {
+		start = i;
+		for (j = 0; j < num_recs; j++, i++, frp++) {
+			trp = xfs_iext_get_ext(ifp, i);
 			trp->l0 = INT_GET(frp->l0, ARCH_CONVERT);
 			trp->l1 = INT_GET(frp->l1, ARCH_CONVERT);
 		}
@@ -4526,14 +4519,14 @@ xfs_bmap_read_extents(
 			 * any "older" data bmap btree records for a
 			 * set bit in the "extent flag" position.
 			 */
-			if (unlikely(xfs_check_nostate_extents(temp, num_recs))) {
+			if (unlikely(xfs_check_nostate_extents(ifp,
+					start, num_recs))) {
 				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
 						 XFS_ERRLEVEL_LOW,
 						 ip->i_mount);
 				goto error0;
 			}
 		}
-		i += num_recs;
 		xfs_trans_brelse(tp, bp);
 		bno = nextbno;
 		/*
@@ -4546,7 +4539,7 @@ xfs_bmap_read_extents(
 			return error;
 		block = XFS_BUF_TO_BMBT_BLOCK(bp);
 	}
-	ASSERT(i == ifp->if_bytes / (uint)sizeof(*trp));
+	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
 	xfs_bmap_trace_exlist(fname, ip, i, whichfork);
 	return 0;
@@ -4557,7 +4550,7 @@ error0:
 
 #ifdef XFS_BMAP_TRACE
 /*
- * Add bmap trace insert entries for all the contents of the extent list.
+ * Add bmap trace insert entries for all the contents of the extent records.
  */
 void
 xfs_bmap_trace_exlist(
@@ -4566,16 +4559,15 @@ xfs_bmap_trace_exlist(
 	xfs_extnum_t	cnt,		/* count of entries in the list */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;		/* base of extent list */
-	xfs_bmbt_rec_t	*ep;		/* current entry in extent list */
-	xfs_extnum_t	idx;		/* extent list entry number */
+	xfs_bmbt_rec_t	*ep;		/* current extent record */
+	xfs_extnum_t	idx;		/* extent record index */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_irec_t	s;		/* extent list record */
+	xfs_bmbt_irec_t	s;		/* file extent record */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(cnt == ifp->if_bytes / (uint)sizeof(*base));
-	base = ifp->if_u1.if_extents;
-	for (idx = 0, ep = base; idx < cnt; idx++, ep++) {
+	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
+	for (idx = 0; idx < cnt; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_get_all(ep, &s);
 		xfs_bmap_trace_insert(fname, "exlist", ip, idx, 1, &s, NULL,
 			whichfork);
@@ -4661,14 +4653,10 @@ xfs_bmapi(
 	xfs_bmalloca_t	bma;		/* args for xfs_bmap_alloc */
 	xfs_btree_cur_t	*cur;		/* bmap btree cursor */
 	xfs_fileoff_t	end;		/* end of mapped file region */
-	int		eof;		/* we've hit the end of extent list */
-	char		contig;		/* allocation must be one extent */
-	char		delay;		/* this request is for delayed alloc */
-	char		exact;		/* don't do all of wasdelayed extent */
-	char		convert;	/* unwritten extent I/O completion */
-	xfs_bmbt_rec_t	*ep;		/* extent list entry pointer */
+	int		eof;		/* we've hit the end of extents */
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
 	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current extent list record */
+	xfs_bmbt_irec_t	got;		/* current file extent record */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	xfs_extlen_t	indlen;		/* indirect blocks length */
 	xfs_extnum_t	lastx;		/* last useful extent number */
@@ -4680,17 +4668,13 @@ xfs_bmapi(
 	int		nallocs;	/* number of extents alloc\'d */
 	xfs_extnum_t	nextents;	/* number of extents in file */
 	xfs_fileoff_t	obno;		/* old block number (offset) */
-	xfs_bmbt_irec_t	prev;		/* previous extent list record */
+	xfs_bmbt_irec_t	prev;		/* previous file extent record */
 	int		tmp_logflags;	/* temp flags holder */
 	int		whichfork;	/* data or attr fork */
 	char		inhole;		/* current location is hole in file */
-	char		stateless;	/* ignore state flag set */
-	char		trim;		/* output trimmed to match range */
-	char		userdata;	/* allocating non-metadata */
 	char		wasdelay;	/* old extent was delayed */
 	char		wr;		/* this is a write request */
 	char		rt;		/* this is a realtime file */
-	char		rsvd;		/* OK to allocate reserved blocks */
 #ifdef DEBUG
 	xfs_fileoff_t	orig_bno;	/* original block number value */
 	int		orig_flags;	/* original flags arg value */
@@ -4727,15 +4711,8 @@ xfs_bmapi(
 		XFS_STATS_INC(xs_blk_mapw);
 	else
 		XFS_STATS_INC(xs_blk_mapr);
-	delay = (flags & XFS_BMAPI_DELAY) != 0;
-	trim = (flags & XFS_BMAPI_ENTIRE) == 0;
-	userdata = (flags & XFS_BMAPI_METADATA) == 0;
-	convert = (flags & XFS_BMAPI_CONVERT) != 0;
-	exact = (flags & XFS_BMAPI_EXACT) != 0;
-	rsvd = (flags & XFS_BMAPI_RSVBLOCKS) != 0;
-	contig = (flags & XFS_BMAPI_CONTIG) != 0;
 	/*
-	 * stateless is used to combine extents which
+	 * IGSTATE flag is used to combine extents which
 	 * differ only due to the state of the extents.
 	 * This technique is used from xfs_getbmap()
 	 * when the caller does not wish to see the
@@ -4751,10 +4728,9 @@ xfs_bmapi(
 	 * xfs_strat_comp(), where the xfs_bmapi() call
 	 * is transactioned, and the extents combined.
 	 */
-	stateless = (flags & XFS_BMAPI_IGSTATE) != 0;
-	if (stateless && wr)	/* if writing unwritten space, no */
-		wr = 0;		/* allocations are allowed */
-	ASSERT(wr || !delay);
+	if ((flags & XFS_BMAPI_IGSTATE) && wr)	/* if writing unwritten space */
+		wr = 0;				/* no allocations are allowed */
+	ASSERT(wr || !(flags & XFS_BMAPI_DELAY));
 	logflags = 0;
 	nallocs = 0;
 	cur = NULL;
@@ -4789,7 +4765,7 @@ xfs_bmapi(
 		if (eof && !wr)
 			got.br_startoff = end;
 		inhole = eof || got.br_startoff > bno;
-		wasdelay = wr && !inhole && !delay &&
+		wasdelay = wr && !inhole && !(flags & XFS_BMAPI_DELAY) &&
 			ISNULLSTARTBLOCK(got.br_startblock);
 		/*
 		 * First, deal with the hole before the allocated space
@@ -4801,11 +4777,11 @@ xfs_bmapi(
 			 * allocate the stuff asked for in this bmap call
 			 * but that wouldn't be as good.
 			 */
-			if (wasdelay && !exact) {
+			if (wasdelay && !(flags & XFS_BMAPI_EXACT)) {
 				alen = (xfs_extlen_t)got.br_blockcount;
 				aoff = got.br_startoff;
 				if (lastx != NULLEXTNUM && lastx) {
-					ep = &ifp->if_u1.if_extents[lastx - 1];
+					ep = xfs_iext_get_ext(ifp, lastx - 1);
 					xfs_bmbt_get_all(ep, &prev);
 				}
 			} else if (wasdelay) {
@@ -4823,8 +4799,8 @@ xfs_bmapi(
 							got.br_startoff - bno);
 				aoff = bno;
 			}
-			minlen = contig ? alen : 1;
-			if (delay) {
+			minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1;
+			if (flags & XFS_BMAPI_DELAY) {
 				xfs_extlen_t	extsz;
 
 				/* Figure out the extent size, adjust alen */
@@ -4837,7 +4813,9 @@ xfs_bmapi(
 				if (extsz) {
 					error = xfs_bmap_extsize_align(mp,
 							&got, &prev, extsz,
-							rt, eof, delay, convert,
+							rt, eof,
+							flags&XFS_BMAPI_DELAY,
+							flags&XFS_BMAPI_CONVERT,
 							&aoff, &alen);
 					ASSERT(!error);
 				}
@@ -4875,24 +4853,29 @@ xfs_bmapi(
 				if (rt) {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FREXTENTS,
-							-(extsz), rsvd);
+							-(extsz), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				} else {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							-(alen), rsvd);
+							-(alen), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				}
 				if (!error) {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							-(indlen), rsvd);
+							-(indlen), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 					if (error && rt)
 						xfs_mod_incore_sb(mp,
 							XFS_SBS_FREXTENTS,
-							extsz, rsvd);
+							extsz, (flags &
+							XFS_BMAPI_RSVBLOCKS));
 					else if (error)
 						xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							alen, rsvd);
+							alen, (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				}
 
 				if (error) {
@@ -4925,7 +4908,7 @@ xfs_bmapi(
 				/* Indicate if this is the first user data
 				 * in the file, or just any user data.
 				 */
-				if (userdata) {
+				if (!(flags & XFS_BMAPI_METADATA)) {
 					bma.userdata = (aoff == 0) ?
 						XFS_ALLOC_INITIAL_USER_DATA :
 						XFS_ALLOC_USERDATA;
@@ -4937,7 +4920,7 @@ xfs_bmapi(
 				bma.firstblock = *firstblock;
 				bma.alen = alen;
 				bma.off = aoff;
-				bma.conv = convert;
+				bma.conv = (flags & XFS_BMAPI_CONVERT);
 				bma.wasdel = wasdelay;
 				bma.minlen = minlen;
 				bma.low = flist->xbf_low;
@@ -4948,7 +4931,8 @@ xfs_bmapi(
 				 * is larger than a stripe unit.
 				 */
 				if (mp->m_dalign && alen >= mp->m_dalign &&
-				    userdata && whichfork == XFS_DATA_FORK) {
+				    (!(flags & XFS_BMAPI_METADATA)) &&
+				    (whichfork == XFS_DATA_FORK)) {
 					if ((error = xfs_bmap_isaeof(ip, aoff,
 							whichfork, &bma.aeof)))
 						goto error0;
@@ -5011,19 +4995,19 @@ xfs_bmapi(
 			}
 			error = xfs_bmap_add_extent(ip, lastx, &cur, &got,
 				firstblock, flist, &tmp_logflags, whichfork,
-				rsvd);
+				(flags & XFS_BMAPI_RSVBLOCKS));
 			logflags |= tmp_logflags;
 			if (error)
 				goto error0;
 			lastx = ifp->if_lastex;
-			ep = &ifp->if_u1.if_extents[lastx];
+			ep = xfs_iext_get_ext(ifp, lastx);
 			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 			xfs_bmbt_get_all(ep, &got);
 			ASSERT(got.br_startoff <= aoff);
 			ASSERT(got.br_startoff + got.br_blockcount >=
 				aoff + alen);
 #ifdef DEBUG
-			if (delay) {
+			if (flags & XFS_BMAPI_DELAY) {
 				ASSERT(ISNULLSTARTBLOCK(got.br_startblock));
 				ASSERT(STARTBLOCKVAL(got.br_startblock) > 0);
 			}
@@ -5052,14 +5036,15 @@ xfs_bmapi(
 		 * Then deal with the allocated space we found.
 		 */
 		ASSERT(ep != NULL);
-		if (trim && (got.br_startoff + got.br_blockcount > obno)) {
+		if (!(flags & XFS_BMAPI_ENTIRE) &&
+		    (got.br_startoff + got.br_blockcount > obno)) {
 			if (obno > bno)
 				bno = obno;
 			ASSERT((bno >= obno) || (n == 0));
 			ASSERT(bno < end);
 			mval->br_startoff = bno;
 			if (ISNULLSTARTBLOCK(got.br_startblock)) {
-				ASSERT(!wr || delay);
+				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
 				mval->br_startblock = DELAYSTARTBLOCK;
 			} else
 				mval->br_startblock =
@@ -5081,7 +5066,7 @@ xfs_bmapi(
 		} else {
 			*mval = got;
 			if (ISNULLSTARTBLOCK(mval->br_startblock)) {
-				ASSERT(!wr || delay);
+				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
 				mval->br_startblock = DELAYSTARTBLOCK;
 			}
 		}
@@ -5107,12 +5092,12 @@ xfs_bmapi(
 			mval->br_state = XFS_EXT_NORM;
 			error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
 				firstblock, flist, &tmp_logflags, whichfork,
-				rsvd);
+				(flags & XFS_BMAPI_RSVBLOCKS));
 			logflags |= tmp_logflags;
 			if (error)
 				goto error0;
 			lastx = ifp->if_lastex;
-			ep = &ifp->if_u1.if_extents[lastx];
+			ep = xfs_iext_get_ext(ifp, lastx);
 			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 			xfs_bmbt_get_all(ep, &got);
 			/*
@@ -5124,9 +5109,10 @@ xfs_bmapi(
 				continue;
 		}
 
-		ASSERT(!trim ||
+		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
 		       ((mval->br_startoff + mval->br_blockcount) <= end));
-		ASSERT(!trim || (mval->br_blockcount <= len) ||
+		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
+		       (mval->br_blockcount <= len) ||
 		       (mval->br_startoff < obno));
 		bno = mval->br_startoff + mval->br_blockcount;
 		len = end - bno;
@@ -5141,7 +5127,8 @@ xfs_bmapi(
 			   mval[-1].br_startblock != HOLESTARTBLOCK &&
 			   mval->br_startblock ==
 			   mval[-1].br_startblock + mval[-1].br_blockcount &&
-			   (stateless || mval[-1].br_state == mval->br_state)) {
+			   ((flags & XFS_BMAPI_IGSTATE) ||
+				mval[-1].br_state == mval->br_state)) {
 			ASSERT(mval->br_startoff ==
 			       mval[-1].br_startoff + mval[-1].br_blockcount);
 			mval[-1].br_blockcount += mval->br_blockcount;
@@ -5168,8 +5155,7 @@ xfs_bmapi(
 		/*
 		 * Else go on to the next record.
 		 */
-		ep++;
-		lastx++;
+		ep = xfs_iext_get_ext(ifp, ++lastx);
 		if (lastx >= nextents) {
 			eof = 1;
 			prev = got;
@@ -5199,7 +5185,7 @@ xfs_bmapi(
 error0:
 	/*
 	 * Log everything.  Do this after conversion, there's no point in
-	 * logging the extent list if we've converted to btree format.
+	 * logging the extent records if we've converted to btree format.
 	 */
 	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
 	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
@@ -5252,12 +5238,12 @@ xfs_bmapi_single(
 	xfs_fsblock_t	*fsb,		/* output: mapped block */
 	xfs_fileoff_t	bno)		/* starting file offs. mapped */
 {
-	int		eof;		/* we've hit the end of extent list */
+	int		eof;		/* we've hit the end of extents */
 	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current extent list record */
+	xfs_bmbt_irec_t	got;		/* current file extent record */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	xfs_extnum_t	lastx;		/* last useful extent number */
-	xfs_bmbt_irec_t	prev;		/* previous extent list record */
+	xfs_bmbt_irec_t	prev;		/* previous file extent record */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if (unlikely(
@@ -5312,18 +5298,18 @@ xfs_bunmapi(
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	xfs_bmbt_irec_t		del;		/* extent being deleted */
 	int			eof;		/* is deleting at eof */
-	xfs_bmbt_rec_t		*ep;		/* extent list entry pointer */
+	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
 	int			error;		/* error return value */
 	xfs_extnum_t		extno;		/* extent number in list */
-	xfs_bmbt_irec_t		got;		/* current extent list entry */
+	xfs_bmbt_irec_t		got;		/* current extent record */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	int			isrt;		/* freeing in rt area */
 	xfs_extnum_t		lastx;		/* last extent index used */
 	int			logflags;	/* transaction logging flags */
 	xfs_extlen_t		mod;		/* rt extent offset */
 	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* size of extent list */
-	xfs_bmbt_irec_t		prev;		/* previous extent list entry */
+	xfs_extnum_t		nextents;	/* number of file extents */
+	xfs_bmbt_irec_t		prev;		/* previous extent record */
 	xfs_fileoff_t		start;		/* first file offset deleted */
 	int			tmp_logflags;	/* partial logging flags */
 	int			wasdel;		/* was a delayed alloc extent */
@@ -5369,7 +5355,7 @@ xfs_bunmapi(
 	 * file, back up to the last block if so...
 	 */
 	if (eof) {
-		ep = &ifp->if_u1.if_extents[--lastx];
+		ep = xfs_iext_get_ext(ifp, --lastx);
 		xfs_bmbt_get_all(ep, &got);
 		bno = got.br_startoff + got.br_blockcount - 1;
 	}
@@ -5393,7 +5379,7 @@ xfs_bunmapi(
 		if (got.br_startoff > bno) {
 			if (--lastx < 0)
 				break;
-			ep--;
+			ep = xfs_iext_get_ext(ifp, lastx);
 			xfs_bmbt_get_all(ep, &got);
 		}
 		/*
@@ -5440,7 +5426,8 @@ xfs_bunmapi(
 					del.br_blockcount : mod;
 				if (bno < got.br_startoff) {
 					if (--lastx >= 0)
-						xfs_bmbt_get_all(--ep, &got);
+						xfs_bmbt_get_all(xfs_iext_get_ext(
+							ifp, lastx), &got);
 				}
 				continue;
 			}
@@ -5500,7 +5487,8 @@ xfs_bunmapi(
 				 * try again.
 				 */
 				ASSERT(lastx > 0);
-				xfs_bmbt_get_all(ep - 1, &prev);
+				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
+						lastx - 1), &prev);
 				ASSERT(prev.br_state == XFS_EXT_NORM);
 				ASSERT(!ISNULLSTARTBLOCK(prev.br_startblock));
 				ASSERT(del.br_startblock ==
@@ -5587,12 +5575,12 @@ nodelete:
 		 * If not done go on to the next (previous) record.
 		 * Reset ep in case the extents array was re-alloced.
 		 */
-		ep = &ifp->if_u1.if_extents[lastx];
+		ep = xfs_iext_get_ext(ifp, lastx);
 		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
 			if (lastx >= XFS_IFORK_NEXTENTS(ip, whichfork) ||
 			    xfs_bmbt_get_startoff(ep) > bno) {
-				lastx--;
-				ep--;
+				if (--lastx >= 0)
+					ep = xfs_iext_get_ext(ifp, lastx);
 			}
 			if (lastx >= 0)
 				xfs_bmbt_get_all(ep, &got);
@@ -5636,7 +5624,7 @@ nodelete:
 error0:
 	/*
 	 * Log everything.  Do this after conversion, there's no point in
-	 * logging the extent list if we've converted to btree format.
+	 * logging the extent records if we've converted to btree format.
 	 */
 	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
 	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
@@ -5892,9 +5880,9 @@ xfs_bmap_isaeof(
 {
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent list entry pointer */
-	xfs_extnum_t	nextents;	/* size of extent list */
-	xfs_bmbt_irec_t	s;		/* expanded extent list entry */
+	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of file extents */
+	xfs_bmbt_irec_t	s;		/* expanded extent record */
 
 	ASSERT(whichfork == XFS_DATA_FORK);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
@@ -5909,7 +5897,7 @@ xfs_bmap_isaeof(
 	/*
 	 * Go to the last extent
 	 */
-	lastrec = &ifp->if_u1.if_extents[nextents - 1];
+	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
 	xfs_bmbt_get_all(lastrec, &s);
 	/*
 	 * Check we are allocating in the last extent (for delayed allocations)
@@ -5936,8 +5924,8 @@ xfs_bmap_eof(
 	xfs_fsblock_t	blockcount;	/* extent block count */
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent list entry pointer */
-	xfs_extnum_t	nextents;	/* size of extent list */
+	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of file extents */
 	xfs_fileoff_t	startoff;	/* extent starting file offset */
 
 	ASSERT(whichfork == XFS_DATA_FORK);
@@ -5953,7 +5941,7 @@ xfs_bmap_eof(
 	/*
 	 * Go to the last extent
 	 */
-	lastrec = &ifp->if_u1.if_extents[nextents - 1];
+	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
 	startoff = xfs_bmbt_get_startoff(lastrec);
 	blockcount = xfs_bmbt_get_blockcount(lastrec);
 	*eof = endoff >= startoff + blockcount;
@@ -5969,18 +5957,21 @@ xfs_bmap_check_extents(
 	xfs_inode_t		*ip,		/* incore inode pointer */
 	int			whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t		*base;		/* base of extents list */
 	xfs_bmbt_rec_t		*ep;		/* current extent entry */
+	xfs_extnum_t		idx;		/* extent record index */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	xfs_extnum_t		nextents;	/* number of extents in list */
+	xfs_bmbt_rec_t		*nextp;		/* next extent entry */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	base = ifp->if_u1.if_extents;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (ep = base; ep < &base[nextents - 1]; ep++) {
+	ep = xfs_iext_get_ext(ifp, 0);
+	for (idx = 0; idx < nextents - 1; idx++) {
+		nextp = xfs_iext_get_ext(ifp, idx + 1);
 		xfs_btree_check_rec(XFS_BTNUM_BMAP, (void *)ep,
-			(void *)(ep + 1));
+			(void *)(nextp));
+		ep = nextp;
 	}
 }
 
@@ -6119,12 +6110,14 @@ xfs_bmap_check_leaf_extents(
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
-	xfs_extnum_t		i=0;	/* index into the extents list */
+	xfs_extnum_t		i=0, j;	/* index into the extents list */
 	xfs_ifork_t		*ifp;	/* fork structure */
 	int			level;	/* btree level, for checking */
 	xfs_mount_t		*mp;	/* file system mount structure */
 	xfs_bmbt_ptr_t		*pp;	/* pointer to block address */
-	xfs_bmbt_rec_t		*ep, *lastp;	/* extent pointers in block entry */
+	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
+	xfs_bmbt_rec_t		*lastp; /* pointer to previous extent */
+	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
 	int			bp_release = 0;
 
 	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
@@ -6194,7 +6187,6 @@ xfs_bmap_check_leaf_extents(
 	 */
 	lastp = NULL;
 	for (;;) {
-		xfs_bmbt_rec_t	*frp;
 		xfs_fsblock_t	nextbno;
 		xfs_extnum_t	num_recs;
 
@@ -6213,18 +6205,20 @@ xfs_bmap_check_leaf_extents(
 		 * conform with the first entry in this one.
 		 */
 
-		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
+		ep = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
 			block, 1, mp->m_bmap_dmxr[0]);
-
-		for (ep = frp;ep < frp + (num_recs - 1); ep++) {
+		for (j = 1; j < num_recs; j++) {
+			nextp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
+				block, j + 1, mp->m_bmap_dmxr[0]);
 			if (lastp) {
 				xfs_btree_check_rec(XFS_BTNUM_BMAP,
 					(void *)lastp, (void *)ep);
 			}
 			xfs_btree_check_rec(XFS_BTNUM_BMAP, (void *)ep,
-				(void *)(ep + 1));
+				(void *)(nextp));
+			lastp = ep;
+			ep = nextp;
 		}
-		lastp = frp + num_recs - 1; /* For the next iteration */
 
 		i += num_recs;
 		if (bp_release) {
@@ -6288,7 +6282,7 @@ xfs_bmap_count_blocks(
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
-		if (unlikely(xfs_bmap_count_leaves(ifp->if_u1.if_extents,
+		if (unlikely(xfs_bmap_count_leaves(ifp, 0,
 			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
 			count) < 0)) {
 			XFS_ERROR_REPORT("xfs_bmap_count_blocks(1)",
@@ -6310,7 +6304,7 @@ xfs_bmap_count_blocks(
 	ASSERT(XFS_FSB_TO_AGBNO(mp, INT_GET(*pp, ARCH_CONVERT)) < mp->m_sb.sb_agblocks);
 	bno = INT_GET(*pp, ARCH_CONVERT);
 
-	if (unlikely(xfs_bmap_count_tree(mp, tp, bno, level, count) < 0)) {
+	if (unlikely(xfs_bmap_count_tree(mp, tp, ifp, bno, level, count) < 0)) {
 		XFS_ERROR_REPORT("xfs_bmap_count_blocks(2)", XFS_ERRLEVEL_LOW,
 				 mp);
 		return XFS_ERROR(EFSCORRUPTED);
@@ -6327,6 +6321,7 @@ int                                     /* error */
 xfs_bmap_count_tree(
 	xfs_mount_t     *mp,            /* file system mount point */
 	xfs_trans_t     *tp,            /* transaction pointer */
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fsblock_t   blockno,	/* file system block number */
 	int             levelin,	/* level in btree */
 	int		*count)		/* Count of blocks */
@@ -6339,7 +6334,6 @@ xfs_bmap_count_tree(
 	xfs_fsblock_t		nextbno;
 	xfs_bmbt_block_t        *block, *nextblock;
 	int			numrecs;
-	xfs_bmbt_rec_t		*frp;
 
 	if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF)))
 		return error;
@@ -6364,7 +6358,7 @@ xfs_bmap_count_tree(
 			xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
 		bno = INT_GET(*pp, ARCH_CONVERT);
 		if (unlikely((error =
-		     xfs_bmap_count_tree(mp, tp, bno, level, count)) < 0)) {
+		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
 			xfs_trans_brelse(tp, bp);
 			XFS_ERROR_REPORT("xfs_bmap_count_tree(1)",
 					 XFS_ERRLEVEL_LOW, mp);
@@ -6376,9 +6370,8 @@ xfs_bmap_count_tree(
 		for (;;) {
 			nextbno = be64_to_cpu(block->bb_rightsib);
 			numrecs = be16_to_cpu(block->bb_numrecs);
-			frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize,
-				xfs_bmbt, block, 1, mp->m_bmap_dmxr[0]);
-			if (unlikely(xfs_bmap_disk_count_leaves(frp, numrecs, count) < 0)) {
+			if (unlikely(xfs_bmap_disk_count_leaves(ifp, mp,
+					0, block, numrecs, count) < 0)) {
 				xfs_trans_brelse(tp, bp);
 				XFS_ERROR_REPORT("xfs_bmap_count_tree(2)",
 						 XFS_ERRLEVEL_LOW, mp);
@@ -6399,33 +6392,45 @@ xfs_bmap_count_tree(
 }
 
 /*
- * Count leaf blocks given a pointer to an extent list.
+ * Count leaf blocks given a range of extent records.
  */
 int
 xfs_bmap_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	int			numrecs,
 	int			*count)
 {
 	int		b;
+	xfs_bmbt_rec_t	*frp;
 
-	for ( b = 1; b <= numrecs; b++, frp++)
+	for (b = 0; b < numrecs; b++) {
+		frp = xfs_iext_get_ext(ifp, idx + b);
 		*count += xfs_bmbt_get_blockcount(frp);
+	}
 	return 0;
 }
 
 /*
- * Count leaf blocks given a pointer to an extent list originally in btree format.
+ * Count leaf blocks given a range of extent records originally
+ * in btree format.
  */
 int
 xfs_bmap_disk_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_mount_t		*mp,
+	xfs_extnum_t		idx,
+	xfs_bmbt_block_t	*block,
 	int			numrecs,
 	int			*count)
 {
 	int		b;
+	xfs_bmbt_rec_t	*frp;
 
-	for ( b = 1; b <= numrecs; b++, frp++)
+	for (b = 1; b <= numrecs; b++) {
+		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize,
+			xfs_bmbt, block, idx + b, mp->m_bmap_dmxr[0]);
 		*count += xfs_bmbt_disk_get_blockcount(frp);
+	}
 	return 0;
 }
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 12cc63dfc2c4..011ccaa9a1c0 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -20,6 +20,7 @@
 
 struct getbmap;
 struct xfs_bmbt_irec;
+struct xfs_ifork;
 struct xfs_inode;
 struct xfs_mount;
 struct xfs_trans;
@@ -347,9 +348,28 @@ xfs_bmap_count_blocks(
  */
 int
 xfs_check_nostate_extents(
-	xfs_bmbt_rec_t		*ep,
+	struct xfs_ifork	*ifp,
+	xfs_extnum_t		idx,
 	xfs_extnum_t		num);
 
+/*
+ * Call xfs_bmap_do_search_extents() to search for the extent
+ * record containing block bno. If in multi-level in-core extent
+ * allocation mode, find and extract the target extent buffer,
+ * otherwise just use the direct extent list.
+ */
+xfs_bmbt_rec_t *
+xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
+			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
+
+/*
+ * Search an extent list for the extent which includes block
+ * bno.
+ */
+xfs_bmbt_rec_t *xfs_bmap_do_search_extents(xfs_bmbt_rec_t *,
+			xfs_extnum_t, xfs_extnum_t, xfs_fileoff_t, int *,
+			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
+
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 3f1383d160e8..bea44709afbe 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -2754,7 +2754,7 @@ xfs_bmbt_update(
 }
 
 /*
- * Check an extent list, which has just been read, for
+ * Check extent records, which have just been read, for
  * any bit in the extent flag field. ASSERT on debug
  * kernels, as this condition should not occur.
  * Return an error condition (1) if any flags found,
@@ -2763,10 +2763,14 @@ xfs_bmbt_update(
 
 int
 xfs_check_nostate_extents(
-	xfs_bmbt_rec_t		*ep,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	xfs_extnum_t		num)
 {
-	for (; num > 0; num--, ep++) {
+	xfs_bmbt_rec_t		*ep;
+
+	for (; num > 0; num--, idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		if ((ep->l0 >>
 		     (64 - BMBT_EXNTFLAG_BITLEN)) != 0) {
 			ASSERT(0);
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index e095a2d344ae..6478cfa0e539 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -372,14 +372,6 @@ extern int xfs_bmbt_get_rec(struct xfs_btree_cur *, xfs_fileoff_t *,
 				xfs_exntst_t *, int *);
 #endif
 
-/*
- * Search an extent list for the extent which includes block
- * bno.
- */
-xfs_bmbt_rec_t *xfs_bmap_do_search_extents(xfs_bmbt_rec_t *,
-			xfs_extnum_t, xfs_extnum_t, xfs_fileoff_t, int *,
-			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
-
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_clnt.h b/fs/xfs/xfs_clnt.h
index f57cc9ac875e..022fff62085b 100644
--- a/fs/xfs/xfs_clnt.h
+++ b/fs/xfs/xfs_clnt.h
@@ -68,8 +68,6 @@ struct xfs_mount_args {
 						 * enforcement */
 #define XFSMNT_PQUOTAENF	0x00000040	/* IRIX project quota limit
 						 * enforcement */
-#define XFSMNT_NOATIME		0x00000100	/* don't modify access
-						 * times on reads */
 #define XFSMNT_NOALIGN		0x00000200	/* don't allocate at
 						 * stripe boundaries*/
 #define XFSMNT_RETERR		0x00000400	/* return error to user */
diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c
index 473671fa5c13..4bae3a76c678 100644
--- a/fs/xfs/xfs_da_btree.c
+++ b/fs/xfs/xfs_da_btree.c
@@ -126,10 +126,10 @@ xfs_da_node_create(xfs_da_args_t *args, xfs_dablk_t blkno, int level,
 	node = bp->data;
 	node->hdr.info.forw = 0;
 	node->hdr.info.back = 0;
-	INT_SET(node->hdr.info.magic, ARCH_CONVERT, XFS_DA_NODE_MAGIC);
+	node->hdr.info.magic = cpu_to_be16(XFS_DA_NODE_MAGIC);
 	node->hdr.info.pad = 0;
 	node->hdr.count = 0;
-	INT_SET(node->hdr.level, ARCH_CONVERT, level);
+	node->hdr.level = cpu_to_be16(level);
 
 	xfs_da_log_buf(tp, bp,
 		XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
@@ -290,28 +290,28 @@ xfs_da_split(xfs_da_state_t *state)
 
 	node = oldblk->bp->data;
 	if (node->hdr.info.forw) {
-		if (INT_GET(node->hdr.info.forw, ARCH_CONVERT) == addblk->blkno) {
+		if (be32_to_cpu(node->hdr.info.forw) == addblk->blkno) {
 			bp = addblk->bp;
 		} else {
 			ASSERT(state->extravalid);
 			bp = state->extrablk.bp;
 		}
 		node = bp->data;
-		INT_SET(node->hdr.info.back, ARCH_CONVERT, oldblk->blkno);
+		node->hdr.info.back = cpu_to_be32(oldblk->blkno);
 		xfs_da_log_buf(state->args->trans, bp,
 		    XFS_DA_LOGRANGE(node, &node->hdr.info,
 		    sizeof(node->hdr.info)));
 	}
 	node = oldblk->bp->data;
-	if (INT_GET(node->hdr.info.back, ARCH_CONVERT)) {
-		if (INT_GET(node->hdr.info.back, ARCH_CONVERT) == addblk->blkno) {
+	if (node->hdr.info.back) {
+		if (be32_to_cpu(node->hdr.info.back) == addblk->blkno) {
 			bp = addblk->bp;
 		} else {
 			ASSERT(state->extravalid);
 			bp = state->extrablk.bp;
 		}
 		node = bp->data;
-		INT_SET(node->hdr.info.forw, ARCH_CONVERT, oldblk->blkno);
+		node->hdr.info.forw = cpu_to_be32(oldblk->blkno);
 		xfs_da_log_buf(state->args->trans, bp,
 		    XFS_DA_LOGRANGE(node, &node->hdr.info,
 		    sizeof(node->hdr.info)));
@@ -359,14 +359,14 @@ xfs_da_root_split(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(bp != NULL);
 	node = bp->data;
 	oldroot = blk1->bp->data;
-	if (INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
-		size = (int)((char *)&oldroot->btree[INT_GET(oldroot->hdr.count, ARCH_CONVERT)] -
+	if (be16_to_cpu(oldroot->hdr.info.magic) == XFS_DA_NODE_MAGIC) {
+		size = (int)((char *)&oldroot->btree[be16_to_cpu(oldroot->hdr.count)] -
 			     (char *)oldroot);
 	} else {
 		ASSERT(XFS_DIR_IS_V2(mp));
-		ASSERT(INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+		ASSERT(be16_to_cpu(oldroot->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 		leaf = (xfs_dir2_leaf_t *)oldroot;
-		size = (int)((char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] -
+		size = (int)((char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] -
 			     (char *)leaf);
 	}
 	memcpy(node, oldroot, size);
@@ -381,18 +381,18 @@ xfs_da_root_split(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	error = xfs_da_node_create(args,
 		args->whichfork == XFS_DATA_FORK &&
 		XFS_DIR_IS_V2(mp) ? mp->m_dirleafblk : 0,
-		INT_GET(node->hdr.level, ARCH_CONVERT) + 1, &bp, args->whichfork);
+		be16_to_cpu(node->hdr.level) + 1, &bp, args->whichfork);
 	if (error)
 		return(error);
 	node = bp->data;
-	INT_SET(node->btree[0].hashval, ARCH_CONVERT, blk1->hashval);
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blk1->blkno);
-	INT_SET(node->btree[1].hashval, ARCH_CONVERT, blk2->hashval);
-	INT_SET(node->btree[1].before, ARCH_CONVERT, blk2->blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 2);
+	node->btree[0].hashval = cpu_to_be32(blk1->hashval);
+	node->btree[0].before = cpu_to_be32(blk1->blkno);
+	node->btree[1].hashval = cpu_to_be32(blk2->hashval);
+	node->btree[1].before = cpu_to_be32(blk2->blkno);
+	node->hdr.count = cpu_to_be16(2);
 
 #ifdef DEBUG
-	if (INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+	if (be16_to_cpu(oldroot->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC) {
 		ASSERT(blk1->blkno >= mp->m_dirleafblk &&
 		       blk1->blkno < mp->m_dirfreeblk);
 		ASSERT(blk2->blkno >= mp->m_dirleafblk &&
@@ -424,7 +424,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	int useextra;
 
 	node = oldblk->bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 
 	/*
 	 * With V2 the extra block is data or freespace.
@@ -435,7 +435,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	/*
 	 * Do we have to split the node?
 	 */
-	if ((INT_GET(node->hdr.count, ARCH_CONVERT) + newcount) > state->node_ents) {
+	if ((be16_to_cpu(node->hdr.count) + newcount) > state->node_ents) {
 		/*
 		 * Allocate a new node, add to the doubly linked chain of
 		 * nodes, then move some of our excess entries into it.
@@ -472,7 +472,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 * If we had double-split op below us, then add the extra block too.
 	 */
 	node = oldblk->bp->data;
-	if (oldblk->index <= INT_GET(node->hdr.count, ARCH_CONVERT)) {
+	if (oldblk->index <= be16_to_cpu(node->hdr.count)) {
 		oldblk->index++;
 		xfs_da_node_add(state, oldblk, addblk);
 		if (useextra) {
@@ -516,17 +516,17 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * Figure out how many entries need to move, and in which direction.
 	 * Swap the nodes around if that makes it simpler.
 	 */
-	if ((INT_GET(node1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(node2->hdr.count, ARCH_CONVERT) > 0) &&
-	    ((INT_GET(node2->btree[ 0 ].hashval, ARCH_CONVERT) < INT_GET(node1->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	     (INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))) {
+	if ((be16_to_cpu(node1->hdr.count) > 0) && (be16_to_cpu(node2->hdr.count) > 0) &&
+	    ((be32_to_cpu(node2->btree[0].hashval) < be32_to_cpu(node1->btree[0].hashval)) ||
+	     (be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval) <
+	      be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval)))) {
 		tmpnode = node1;
 		node1 = node2;
 		node2 = tmpnode;
 	}
-	ASSERT(INT_GET(node1->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT(INT_GET(node2->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	count = (INT_GET(node1->hdr.count, ARCH_CONVERT) - INT_GET(node2->hdr.count, ARCH_CONVERT)) / 2;
+	ASSERT(be16_to_cpu(node1->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node2->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	count = (be16_to_cpu(node1->hdr.count) - be16_to_cpu(node2->hdr.count)) / 2;
 	if (count == 0)
 		return;
 	tp = state->args->trans;
@@ -537,7 +537,7 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		/*
 		 * Move elements in node2 up to make a hole.
 		 */
-		if ((tmp = INT_GET(node2->hdr.count, ARCH_CONVERT)) > 0) {
+		if ((tmp = be16_to_cpu(node2->hdr.count)) > 0) {
 			tmp *= (uint)sizeof(xfs_da_node_entry_t);
 			btree_s = &node2->btree[0];
 			btree_d = &node2->btree[count];
@@ -548,13 +548,12 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Move the req'd B-tree elements from high in node1 to
 		 * low in node2.
 		 */
-		INT_MOD(node2->hdr.count, ARCH_CONVERT, count);
+		be16_add(&node2->hdr.count, count);
 		tmp = count * (uint)sizeof(xfs_da_node_entry_t);
-		btree_s = &node1->btree[INT_GET(node1->hdr.count, ARCH_CONVERT) - count];
+		btree_s = &node1->btree[be16_to_cpu(node1->hdr.count) - count];
 		btree_d = &node2->btree[0];
 		memcpy(btree_d, btree_s, tmp);
-		INT_MOD(node1->hdr.count, ARCH_CONVERT, -(count));
-
+		be16_add(&node1->hdr.count, -count);
 	} else {
 		/*
 		 * Move the req'd B-tree elements from low in node2 to
@@ -563,21 +562,21 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		count = -count;
 		tmp = count * (uint)sizeof(xfs_da_node_entry_t);
 		btree_s = &node2->btree[0];
-		btree_d = &node1->btree[INT_GET(node1->hdr.count, ARCH_CONVERT)];
+		btree_d = &node1->btree[be16_to_cpu(node1->hdr.count)];
 		memcpy(btree_d, btree_s, tmp);
-		INT_MOD(node1->hdr.count, ARCH_CONVERT, count);
+		be16_add(&node1->hdr.count, count);
 		xfs_da_log_buf(tp, blk1->bp,
 			XFS_DA_LOGRANGE(node1, btree_d, tmp));
 
 		/*
 		 * Move elements in node2 down to fill the hole.
 		 */
-		tmp  = INT_GET(node2->hdr.count, ARCH_CONVERT) - count;
+		tmp  = be16_to_cpu(node2->hdr.count) - count;
 		tmp *= (uint)sizeof(xfs_da_node_entry_t);
 		btree_s = &node2->btree[count];
 		btree_d = &node2->btree[0];
 		memmove(btree_d, btree_s, tmp);
-		INT_MOD(node2->hdr.count, ARCH_CONVERT, -(count));
+		be16_add(&node2->hdr.count, -count);
 	}
 
 	/*
@@ -588,7 +587,7 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	xfs_da_log_buf(tp, blk2->bp,
 		XFS_DA_LOGRANGE(node2, &node2->hdr,
 			sizeof(node2->hdr) +
-			sizeof(node2->btree[0]) * INT_GET(node2->hdr.count, ARCH_CONVERT)));
+			sizeof(node2->btree[0]) * be16_to_cpu(node2->hdr.count)));
 
 	/*
 	 * Record the last hashval from each block for upward propagation.
@@ -596,15 +595,15 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 */
 	node1 = blk1->bp->data;
 	node2 = blk2->bp->data;
-	blk1->hashval = INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
-	blk2->hashval = INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	blk1->hashval = be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval);
+	blk2->hashval = be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval);
 
 	/*
 	 * Adjust the expected index for insertion.
 	 */
-	if (blk1->index >= INT_GET(node1->hdr.count, ARCH_CONVERT)) {
-		blk2->index = blk1->index - INT_GET(node1->hdr.count, ARCH_CONVERT);
-		blk1->index = INT_GET(node1->hdr.count, ARCH_CONVERT) + 1;	/* make it invalid */
+	if (blk1->index >= be16_to_cpu(node1->hdr.count)) {
+		blk2->index = blk1->index - be16_to_cpu(node1->hdr.count);
+		blk1->index = be16_to_cpu(node1->hdr.count) + 1;	/* make it invalid */
 	}
 }
 
@@ -622,8 +621,8 @@ xfs_da_node_add(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 
 	node = oldblk->bp->data;
 	mp = state->mp;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT((oldblk->index >= 0) && (oldblk->index <= INT_GET(node->hdr.count, ARCH_CONVERT)));
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT((oldblk->index >= 0) && (oldblk->index <= be16_to_cpu(node->hdr.count)));
 	ASSERT(newblk->blkno != 0);
 	if (state->args->whichfork == XFS_DATA_FORK && XFS_DIR_IS_V2(mp))
 		ASSERT(newblk->blkno >= mp->m_dirleafblk &&
@@ -634,22 +633,22 @@ xfs_da_node_add(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 */
 	tmp = 0;
 	btree = &node->btree[ oldblk->index ];
-	if (oldblk->index < INT_GET(node->hdr.count, ARCH_CONVERT)) {
-		tmp = (INT_GET(node->hdr.count, ARCH_CONVERT) - oldblk->index) * (uint)sizeof(*btree);
+	if (oldblk->index < be16_to_cpu(node->hdr.count)) {
+		tmp = (be16_to_cpu(node->hdr.count) - oldblk->index) * (uint)sizeof(*btree);
 		memmove(btree + 1, btree, tmp);
 	}
-	INT_SET(btree->hashval, ARCH_CONVERT, newblk->hashval);
-	INT_SET(btree->before, ARCH_CONVERT, newblk->blkno);
+	btree->hashval = cpu_to_be32(newblk->hashval);
+	btree->before = cpu_to_be32(newblk->blkno);
 	xfs_da_log_buf(state->args->trans, oldblk->bp,
 		XFS_DA_LOGRANGE(node, btree, tmp + sizeof(*btree)));
-	INT_MOD(node->hdr.count, ARCH_CONVERT, +1);
+	be16_add(&node->hdr.count, 1);
 	xfs_da_log_buf(state->args->trans, oldblk->bp,
 		XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
 
 	/*
 	 * Copy the last hash value from the oldblk to propagate upwards.
 	 */
-	oldblk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	oldblk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1 ].hashval);
 }
 
 /*========================================================================
@@ -768,21 +767,21 @@ xfs_da_root_join(xfs_da_state_t *state, xfs_da_state_blk_t *root_blk)
 	ASSERT(args != NULL);
 	ASSERT(root_blk->magic == XFS_DA_NODE_MAGIC);
 	oldroot = root_blk->bp->data;
-	ASSERT(INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(oldroot->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	ASSERT(!oldroot->hdr.info.forw);
 	ASSERT(!oldroot->hdr.info.back);
 
 	/*
 	 * If the root has more than one child, then don't do anything.
 	 */
-	if (INT_GET(oldroot->hdr.count, ARCH_CONVERT) > 1)
+	if (be16_to_cpu(oldroot->hdr.count) > 1)
 		return(0);
 
 	/*
 	 * Read in the (only) child block, then copy those bytes into
 	 * the root block's buffer and free the original child block.
 	 */
-	child = INT_GET(oldroot->btree[ 0 ].before, ARCH_CONVERT);
+	child = be32_to_cpu(oldroot->btree[0].before);
 	ASSERT(child != 0);
 	error = xfs_da_read_buf(args->trans, args->dp, child, -1, &bp,
 					     args->whichfork);
@@ -790,11 +789,11 @@ xfs_da_root_join(xfs_da_state_t *state, xfs_da_state_blk_t *root_blk)
 		return(error);
 	ASSERT(bp != NULL);
 	blkinfo = bp->data;
-	if (INT_GET(oldroot->hdr.level, ARCH_CONVERT) == 1) {
-		ASSERT(INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+	if (be16_to_cpu(oldroot->hdr.level) == 1) {
+		ASSERT(be16_to_cpu(blkinfo->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(blkinfo->magic) == XFS_ATTR_LEAF_MAGIC);
 	} else {
-		ASSERT(INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(blkinfo->magic) == XFS_DA_NODE_MAGIC);
 	}
 	ASSERT(!blkinfo->forw);
 	ASSERT(!blkinfo->back);
@@ -830,9 +829,9 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC);
 	node = (xfs_da_intnode_t *)info;
-	count = INT_GET(node->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(node->hdr.count);
 	if (count > (state->node_ents >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);	/* blk over 50%, don't try to join */
@@ -849,7 +848,7 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -871,13 +870,12 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 	 * to shrink a directory over time.
 	 */
 	/* start with smaller blk num */
-	forward = (INT_GET(info->forw, ARCH_CONVERT)
-				< INT_GET(info->back, ARCH_CONVERT));
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -889,10 +887,10 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 		node = (xfs_da_intnode_t *)info;
 		count  = state->node_ents;
 		count -= state->node_ents >> 2;
-		count -= INT_GET(node->hdr.count, ARCH_CONVERT);
+		count -= be16_to_cpu(node->hdr.count);
 		node = bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-		count -= INT_GET(node->hdr.count, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+		count -= be16_to_cpu(node->hdr.count);
 		xfs_da_brelse(state->args->trans, bp);
 		if (count >= 0)
 			break;	/* fits with at least 25% to spare */
@@ -973,16 +971,16 @@ xfs_da_fixhashpath(xfs_da_state_t *state, xfs_da_state_path_t *path)
 	}
 	for (blk--, level--; level >= 0; blk--, level--) {
 		node = blk->bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 		btree = &node->btree[ blk->index ];
-		if (INT_GET(btree->hashval, ARCH_CONVERT) == lasthash)
+		if (be32_to_cpu(btree->hashval) == lasthash)
 			break;
 		blk->hashval = lasthash;
-		INT_SET(btree->hashval, ARCH_CONVERT, lasthash);
+		btree->hashval = cpu_to_be32(lasthash);
 		xfs_da_log_buf(state->args->trans, blk->bp,
 				  XFS_DA_LOGRANGE(node, btree, sizeof(*btree)));
 
-		lasthash = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+		lasthash = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 	}
 }
 
@@ -997,25 +995,25 @@ xfs_da_node_remove(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk)
 	int tmp;
 
 	node = drop_blk->bp->data;
-	ASSERT(drop_blk->index < INT_GET(node->hdr.count, ARCH_CONVERT));
+	ASSERT(drop_blk->index < be16_to_cpu(node->hdr.count));
 	ASSERT(drop_blk->index >= 0);
 
 	/*
 	 * Copy over the offending entry, or just zero it out.
 	 */
 	btree = &node->btree[drop_blk->index];
-	if (drop_blk->index < (INT_GET(node->hdr.count, ARCH_CONVERT)-1)) {
-		tmp  = INT_GET(node->hdr.count, ARCH_CONVERT) - drop_blk->index - 1;
+	if (drop_blk->index < (be16_to_cpu(node->hdr.count)-1)) {
+		tmp  = be16_to_cpu(node->hdr.count) - drop_blk->index - 1;
 		tmp *= (uint)sizeof(xfs_da_node_entry_t);
 		memmove(btree, btree + 1, tmp);
 		xfs_da_log_buf(state->args->trans, drop_blk->bp,
 		    XFS_DA_LOGRANGE(node, btree, tmp));
-		btree = &node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ];
+		btree = &node->btree[be16_to_cpu(node->hdr.count)-1];
 	}
 	memset((char *)btree, 0, sizeof(xfs_da_node_entry_t));
 	xfs_da_log_buf(state->args->trans, drop_blk->bp,
 	    XFS_DA_LOGRANGE(node, btree, sizeof(*btree)));
-	INT_MOD(node->hdr.count, ARCH_CONVERT, -1);
+	be16_add(&node->hdr.count, -1);
 	xfs_da_log_buf(state->args->trans, drop_blk->bp,
 	    XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
 
@@ -1023,7 +1021,7 @@ xfs_da_node_remove(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk)
 	 * Copy the last hash value from the block to propagate upwards.
 	 */
 	btree--;
-	drop_blk->hashval = INT_GET(btree->hashval, ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(btree->hashval);
 }
 
 /*
@@ -1041,40 +1039,40 @@ xfs_da_node_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 
 	drop_node = drop_blk->bp->data;
 	save_node = save_blk->bp->data;
-	ASSERT(INT_GET(drop_node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT(INT_GET(save_node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(drop_node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(save_node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	tp = state->args->trans;
 
 	/*
 	 * If the dying block has lower hashvals, then move all the
 	 * elements in the remaining block up to make a hole.
 	 */
-	if ((INT_GET(drop_node->btree[ 0 ].hashval, ARCH_CONVERT) < INT_GET(save_node->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	    (INT_GET(drop_node->btree[ INT_GET(drop_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	     INT_GET(save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))
+	if ((be32_to_cpu(drop_node->btree[0].hashval) < be32_to_cpu(save_node->btree[ 0 ].hashval)) ||
+	    (be32_to_cpu(drop_node->btree[be16_to_cpu(drop_node->hdr.count)-1].hashval) <
+	     be32_to_cpu(save_node->btree[be16_to_cpu(save_node->hdr.count)-1].hashval)))
 	{
-		btree = &save_node->btree[ INT_GET(drop_node->hdr.count, ARCH_CONVERT) ];
-		tmp = INT_GET(save_node->hdr.count, ARCH_CONVERT) * (uint)sizeof(xfs_da_node_entry_t);
+		btree = &save_node->btree[be16_to_cpu(drop_node->hdr.count)];
+		tmp = be16_to_cpu(save_node->hdr.count) * (uint)sizeof(xfs_da_node_entry_t);
 		memmove(btree, &save_node->btree[0], tmp);
 		btree = &save_node->btree[0];
 		xfs_da_log_buf(tp, save_blk->bp,
 			XFS_DA_LOGRANGE(save_node, btree,
-				(INT_GET(save_node->hdr.count, ARCH_CONVERT) + INT_GET(drop_node->hdr.count, ARCH_CONVERT)) *
+				(be16_to_cpu(save_node->hdr.count) + be16_to_cpu(drop_node->hdr.count)) *
 				sizeof(xfs_da_node_entry_t)));
 	} else {
-		btree = &save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT) ];
+		btree = &save_node->btree[be16_to_cpu(save_node->hdr.count)];
 		xfs_da_log_buf(tp, save_blk->bp,
 			XFS_DA_LOGRANGE(save_node, btree,
-				INT_GET(drop_node->hdr.count, ARCH_CONVERT) *
+				be16_to_cpu(drop_node->hdr.count) *
 				sizeof(xfs_da_node_entry_t)));
 	}
 
 	/*
 	 * Move all the B-tree elements from drop_blk to save_blk.
 	 */
-	tmp = INT_GET(drop_node->hdr.count, ARCH_CONVERT) * (uint)sizeof(xfs_da_node_entry_t);
+	tmp = be16_to_cpu(drop_node->hdr.count) * (uint)sizeof(xfs_da_node_entry_t);
 	memcpy(btree, &drop_node->btree[0], tmp);
-	INT_MOD(save_node->hdr.count, ARCH_CONVERT, INT_GET(drop_node->hdr.count, ARCH_CONVERT));
+	be16_add(&save_node->hdr.count, be16_to_cpu(drop_node->hdr.count));
 
 	xfs_da_log_buf(tp, save_blk->bp,
 		XFS_DA_LOGRANGE(save_node, &save_node->hdr,
@@ -1083,7 +1081,7 @@ xfs_da_node_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	/*
 	 * Save the last hashval in the remaining block for upward propagation.
 	 */
-	save_blk->hashval = INT_GET(save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	save_blk->hashval = be32_to_cpu(save_node->btree[be16_to_cpu(save_node->hdr.count)-1].hashval);
 }
 
 /*========================================================================
@@ -1138,46 +1136,46 @@ xfs_da_node_lookup_int(xfs_da_state_t *state, int *result)
 			return(error);
 		}
 		curr = blk->bp->data;
-		ASSERT(INT_GET(curr->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC ||
-		       INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(curr->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+		ASSERT(be16_to_cpu(curr->magic) == XFS_DA_NODE_MAGIC ||
+		       be16_to_cpu(curr->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(curr->magic) == XFS_ATTR_LEAF_MAGIC);
 
 		/*
 		 * Search an intermediate node for a match.
 		 */
-		blk->magic = INT_GET(curr->magic, ARCH_CONVERT);
-		if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+		blk->magic = be16_to_cpu(curr->magic);
+		if (blk->magic == XFS_DA_NODE_MAGIC) {
 			node = blk->bp->data;
-			blk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+			blk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 
 			/*
 			 * Binary search.  (note: small blocks will skip loop)
 			 */
-			max = INT_GET(node->hdr.count, ARCH_CONVERT);
+			max = be16_to_cpu(node->hdr.count);
 			probe = span = max / 2;
 			hashval = args->hashval;
 			for (btree = &node->btree[probe]; span > 4;
 				   btree = &node->btree[probe]) {
 				span /= 2;
-				if (INT_GET(btree->hashval, ARCH_CONVERT) < hashval)
+				if (be32_to_cpu(btree->hashval) < hashval)
 					probe += span;
-				else if (INT_GET(btree->hashval, ARCH_CONVERT) > hashval)
+				else if (be32_to_cpu(btree->hashval) > hashval)
 					probe -= span;
 				else
 					break;
 			}
 			ASSERT((probe >= 0) && (probe < max));
-			ASSERT((span <= 4) || (INT_GET(btree->hashval, ARCH_CONVERT) == hashval));
+			ASSERT((span <= 4) || (be32_to_cpu(btree->hashval) == hashval));
 
 			/*
 			 * Since we may have duplicate hashval's, find the first
 			 * matching hashval in the node.
 			 */
-			while ((probe > 0) && (INT_GET(btree->hashval, ARCH_CONVERT) >= hashval)) {
+			while ((probe > 0) && (be32_to_cpu(btree->hashval) >= hashval)) {
 				btree--;
 				probe--;
 			}
-			while ((probe < max) && (INT_GET(btree->hashval, ARCH_CONVERT) < hashval)) {
+			while ((probe < max) && (be32_to_cpu(btree->hashval) < hashval)) {
 				btree++;
 				probe++;
 			}
@@ -1187,21 +1185,21 @@ xfs_da_node_lookup_int(xfs_da_state_t *state, int *result)
 			 */
 			if (probe == max) {
 				blk->index = max-1;
-				blkno = INT_GET(node->btree[ max-1 ].before, ARCH_CONVERT);
+				blkno = be32_to_cpu(node->btree[max-1].before);
 			} else {
 				blk->index = probe;
-				blkno = INT_GET(btree->before, ARCH_CONVERT);
+				blkno = be32_to_cpu(btree->before);
 			}
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_ATTR_LEAF_MAGIC) {
 			blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL);
 			break;
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_DIR_LEAF_MAGIC) {
 			blk->hashval = xfs_dir_leaf_lasthash(blk->bp, NULL);
 			break;
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_DIR2_LEAFN_MAGIC) {
 			blk->hashval = xfs_dir2_leafn_lasthash(blk->bp, NULL);
 			break;
 		}
@@ -1274,8 +1272,8 @@ xfs_da_blk_link(xfs_da_state_t *state, xfs_da_state_blk_t *old_blk,
 	ASSERT(old_blk->magic == XFS_DA_NODE_MAGIC ||
 	       old_blk->magic == XFS_DIRX_LEAF_MAGIC(state->mp) ||
 	       old_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(old_blk->magic == INT_GET(old_info->magic, ARCH_CONVERT));
-	ASSERT(new_blk->magic == INT_GET(new_info->magic, ARCH_CONVERT));
+	ASSERT(old_blk->magic == be16_to_cpu(old_info->magic));
+	ASSERT(new_blk->magic == be16_to_cpu(new_info->magic));
 	ASSERT(old_blk->magic == new_blk->magic);
 
 	switch (old_blk->magic) {
@@ -1302,47 +1300,44 @@ xfs_da_blk_link(xfs_da_state_t *state, xfs_da_state_blk_t *old_blk,
 		/*
 		 * Link new block in before existing block.
 		 */
-		INT_SET(new_info->forw, ARCH_CONVERT, old_blk->blkno);
-		new_info->back = old_info->back; /* INT_: direct copy */
-		if (INT_GET(old_info->back, ARCH_CONVERT)) {
+		new_info->forw = cpu_to_be32(old_blk->blkno);
+		new_info->back = old_info->back;
+		if (old_info->back) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(old_info->back,
-							ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(old_info->back),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT) == INT_GET(old_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->forw, ARCH_CONVERT) == old_blk->blkno);
-			INT_SET(tmp_info->forw, ARCH_CONVERT, new_blk->blkno);
+			ASSERT(be16_to_cpu(tmp_info->magic) == be16_to_cpu(old_info->magic));
+			ASSERT(be32_to_cpu(tmp_info->forw) == old_blk->blkno);
+			tmp_info->forw = cpu_to_be32(new_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1);
 			xfs_da_buf_done(bp);
 		}
-		INT_SET(old_info->back, ARCH_CONVERT, new_blk->blkno);
+		old_info->back = cpu_to_be32(new_blk->blkno);
 	} else {
 		/*
 		 * Link new block in after existing block.
 		 */
-		new_info->forw = old_info->forw; /* INT_: direct copy */
-		INT_SET(new_info->back, ARCH_CONVERT, old_blk->blkno);
-		if (INT_GET(old_info->forw, ARCH_CONVERT)) {
+		new_info->forw = old_info->forw;
+		new_info->back = cpu_to_be32(old_blk->blkno);
+		if (old_info->forw) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(old_info->forw, ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(old_info->forw),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT)
-				    == INT_GET(old_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->back, ARCH_CONVERT)
-				    == old_blk->blkno);
-			INT_SET(tmp_info->back, ARCH_CONVERT, new_blk->blkno);
+			ASSERT(tmp_info->magic == old_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->back) == old_blk->blkno);
+			tmp_info->back = cpu_to_be32(new_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1);
 			xfs_da_buf_done(bp);
 		}
-		INT_SET(old_info->forw, ARCH_CONVERT, new_blk->blkno);
+		old_info->forw = cpu_to_be32(new_blk->blkno);
 	}
 
 	xfs_da_log_buf(args->trans, old_blk->bp, 0, sizeof(*tmp_info) - 1);
@@ -1360,13 +1355,13 @@ xfs_da_node_order(xfs_dabuf_t *node1_bp, xfs_dabuf_t *node2_bp)
 
 	node1 = node1_bp->data;
 	node2 = node2_bp->data;
-	ASSERT((INT_GET(node1->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) &&
-	       (INT_GET(node2->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC));
-	if ((INT_GET(node1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(node2->hdr.count, ARCH_CONVERT) > 0) &&
-	    ((INT_GET(node2->btree[ 0 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	     (INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))) {
+	ASSERT((be16_to_cpu(node1->hdr.info.magic) == XFS_DA_NODE_MAGIC) &&
+	       (be16_to_cpu(node2->hdr.info.magic) == XFS_DA_NODE_MAGIC));
+	if ((be16_to_cpu(node1->hdr.count) > 0) && (be16_to_cpu(node2->hdr.count) > 0) &&
+	    ((be32_to_cpu(node2->btree[0].hashval) <
+	      be32_to_cpu(node1->btree[0].hashval)) ||
+	     (be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval) <
+	      be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval)))) {
 		return(1);
 	}
 	return(0);
@@ -1381,12 +1376,12 @@ xfs_da_node_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_da_intnode_t *node;
 
 	node = bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	if (count)
-		*count = INT_GET(node->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(node->hdr.count);
 	if (!node->hdr.count)
 		return(0);
-	return(INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT));
+	return be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 }
 
 /*
@@ -1411,50 +1406,47 @@ xfs_da_blk_unlink(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_DA_NODE_MAGIC ||
 	       save_blk->magic == XFS_DIRX_LEAF_MAGIC(state->mp) ||
 	       save_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(save_blk->magic == INT_GET(save_info->magic, ARCH_CONVERT));
-	ASSERT(drop_blk->magic == INT_GET(drop_info->magic, ARCH_CONVERT));
+	ASSERT(save_blk->magic == be16_to_cpu(save_info->magic));
+	ASSERT(drop_blk->magic == be16_to_cpu(drop_info->magic));
 	ASSERT(save_blk->magic == drop_blk->magic);
-	ASSERT((INT_GET(save_info->forw, ARCH_CONVERT) == drop_blk->blkno) ||
-	       (INT_GET(save_info->back, ARCH_CONVERT) == drop_blk->blkno));
-	ASSERT((INT_GET(drop_info->forw, ARCH_CONVERT) == save_blk->blkno) ||
-	       (INT_GET(drop_info->back, ARCH_CONVERT) == save_blk->blkno));
+	ASSERT((be32_to_cpu(save_info->forw) == drop_blk->blkno) ||
+	       (be32_to_cpu(save_info->back) == drop_blk->blkno));
+	ASSERT((be32_to_cpu(drop_info->forw) == save_blk->blkno) ||
+	       (be32_to_cpu(drop_info->back) == save_blk->blkno));
 
 	/*
 	 * Unlink the leaf block from the doubly linked chain of leaves.
 	 */
-	if (INT_GET(save_info->back, ARCH_CONVERT) == drop_blk->blkno) {
-		save_info->back = drop_info->back; /* INT_: direct copy */
-		if (INT_GET(drop_info->back, ARCH_CONVERT)) {
+	if (be32_to_cpu(save_info->back) == drop_blk->blkno) {
+		save_info->back = drop_info->back;
+		if (drop_info->back) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(drop_info->back,
-							ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(drop_info->back),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT) == INT_GET(save_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->forw, ARCH_CONVERT) == drop_blk->blkno);
-			INT_SET(tmp_info->forw, ARCH_CONVERT, save_blk->blkno);
+			ASSERT(tmp_info->magic == save_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->forw) == drop_blk->blkno);
+			tmp_info->forw = cpu_to_be32(save_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0,
 						    sizeof(*tmp_info) - 1);
 			xfs_da_buf_done(bp);
 		}
 	} else {
-		save_info->forw = drop_info->forw; /* INT_: direct copy */
-		if (INT_GET(drop_info->forw, ARCH_CONVERT)) {
+		save_info->forw = drop_info->forw;
+		if (drop_info->forw) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(drop_info->forw, ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(drop_info->forw),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT)
-				    == INT_GET(save_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->back, ARCH_CONVERT)
-				    == drop_blk->blkno);
-			INT_SET(tmp_info->back, ARCH_CONVERT, save_blk->blkno);
+			ASSERT(tmp_info->magic == save_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->back) == drop_blk->blkno);
+			tmp_info->back = cpu_to_be32(save_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0,
 						    sizeof(*tmp_info) - 1);
 			xfs_da_buf_done(bp);
@@ -1497,14 +1489,14 @@ xfs_da_path_shift(xfs_da_state_t *state, xfs_da_state_path_t *path,
 	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
 		ASSERT(blk->bp != NULL);
 		node = blk->bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-		if (forward && (blk->index < INT_GET(node->hdr.count, ARCH_CONVERT)-1)) {
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+		if (forward && (blk->index < be16_to_cpu(node->hdr.count)-1)) {
 			blk->index++;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 			break;
 		} else if (!forward && (blk->index > 0)) {
 			blk->index--;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 			break;
 		}
 	}
@@ -1536,18 +1528,18 @@ xfs_da_path_shift(xfs_da_state_t *state, xfs_da_state_path_t *path,
 			return(error);
 		ASSERT(blk->bp != NULL);
 		info = blk->bp->data;
-		ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC ||
-		       INT_GET(info->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
-		blk->magic = INT_GET(info->magic, ARCH_CONVERT);
-		if (INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+		ASSERT(be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC ||
+		       be16_to_cpu(info->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC);
+		blk->magic = be16_to_cpu(info->magic);
+		if (blk->magic == XFS_DA_NODE_MAGIC) {
 			node = (xfs_da_intnode_t *)info;
-			blk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+			blk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 			if (forward)
 				blk->index = 0;
 			else
-				blk->index = INT_GET(node->hdr.count, ARCH_CONVERT)-1;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+				blk->index = be16_to_cpu(node->hdr.count)-1;
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 		} else {
 			ASSERT(level == path->active-1);
 			blk->index = 0;
@@ -1788,40 +1780,40 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * Get values from the moved block.
 	 */
-	if (INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) {
+	if (be16_to_cpu(dead_info->magic) == XFS_DIR_LEAF_MAGIC) {
 		ASSERT(XFS_DIR_IS_V1(mp));
 		dead_leaf = (xfs_dir_leafblock_t *)dead_info;
 		dead_level = 0;
 		dead_hash =
 			INT_GET(dead_leaf->entries[INT_GET(dead_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
-	} else if (INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+	} else if (be16_to_cpu(dead_info->magic) == XFS_DIR2_LEAFN_MAGIC) {
 		ASSERT(XFS_DIR_IS_V2(mp));
 		dead_leaf2 = (xfs_dir2_leaf_t *)dead_info;
 		dead_level = 0;
-		dead_hash = INT_GET(dead_leaf2->ents[INT_GET(dead_leaf2->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+		dead_hash = be32_to_cpu(dead_leaf2->ents[be16_to_cpu(dead_leaf2->hdr.count) - 1].hashval);
 	} else {
-		ASSERT(INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(dead_info->magic) == XFS_DA_NODE_MAGIC);
 		dead_node = (xfs_da_intnode_t *)dead_info;
-		dead_level = INT_GET(dead_node->hdr.level, ARCH_CONVERT);
-		dead_hash = INT_GET(dead_node->btree[INT_GET(dead_node->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+		dead_level = be16_to_cpu(dead_node->hdr.level);
+		dead_hash = be32_to_cpu(dead_node->btree[be16_to_cpu(dead_node->hdr.count) - 1].hashval);
 	}
 	sib_buf = par_buf = NULL;
 	/*
 	 * If the moved block has a left sibling, fix up the pointers.
 	 */
-	if ((sib_blkno = INT_GET(dead_info->back, ARCH_CONVERT))) {
+	if ((sib_blkno = be32_to_cpu(dead_info->back))) {
 		if ((error = xfs_da_read_buf(tp, ip, sib_blkno, -1, &sib_buf, w)))
 			goto done;
 		sib_info = sib_buf->data;
 		if (unlikely(
-		    INT_GET(sib_info->forw, ARCH_CONVERT) != last_blkno ||
-		    INT_GET(sib_info->magic, ARCH_CONVERT) != INT_GET(dead_info->magic, ARCH_CONVERT))) {
+		    be32_to_cpu(sib_info->forw) != last_blkno ||
+		    sib_info->magic != dead_info->magic)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(2)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		INT_SET(sib_info->forw, ARCH_CONVERT, dead_blkno);
+		sib_info->forw = cpu_to_be32(dead_blkno);
 		xfs_da_log_buf(tp, sib_buf,
 			XFS_DA_LOGRANGE(sib_info, &sib_info->forw,
 					sizeof(sib_info->forw)));
@@ -1831,20 +1823,19 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * If the moved block has a right sibling, fix up the pointers.
 	 */
-	if ((sib_blkno = INT_GET(dead_info->forw, ARCH_CONVERT))) {
+	if ((sib_blkno = be32_to_cpu(dead_info->forw))) {
 		if ((error = xfs_da_read_buf(tp, ip, sib_blkno, -1, &sib_buf, w)))
 			goto done;
 		sib_info = sib_buf->data;
 		if (unlikely(
-		       INT_GET(sib_info->back, ARCH_CONVERT) != last_blkno
-		    || INT_GET(sib_info->magic, ARCH_CONVERT)
-				!= INT_GET(dead_info->magic, ARCH_CONVERT))) {
+		       be32_to_cpu(sib_info->back) != last_blkno ||
+		       sib_info->magic != dead_info->magic)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(3)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		INT_SET(sib_info->back, ARCH_CONVERT, dead_blkno);
+		sib_info->back = cpu_to_be32(dead_blkno);
 		xfs_da_log_buf(tp, sib_buf,
 			XFS_DA_LOGRANGE(sib_info, &sib_info->back,
 					sizeof(sib_info->back)));
@@ -1861,26 +1852,26 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 			goto done;
 		par_node = par_buf->data;
 		if (unlikely(
-		    INT_GET(par_node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC ||
-		    (level >= 0 && level != INT_GET(par_node->hdr.level, ARCH_CONVERT) + 1))) {
+		    be16_to_cpu(par_node->hdr.info.magic) != XFS_DA_NODE_MAGIC ||
+		    (level >= 0 && level != be16_to_cpu(par_node->hdr.level) + 1))) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(4)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		level = INT_GET(par_node->hdr.level, ARCH_CONVERT);
+		level = be16_to_cpu(par_node->hdr.level);
 		for (entno = 0;
-		     entno < INT_GET(par_node->hdr.count, ARCH_CONVERT) &&
-		     INT_GET(par_node->btree[entno].hashval, ARCH_CONVERT) < dead_hash;
+		     entno < be16_to_cpu(par_node->hdr.count) &&
+		     be32_to_cpu(par_node->btree[entno].hashval) < dead_hash;
 		     entno++)
 			continue;
-		if (unlikely(entno == INT_GET(par_node->hdr.count, ARCH_CONVERT))) {
+		if (unlikely(entno == be16_to_cpu(par_node->hdr.count))) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(5)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		par_blkno = INT_GET(par_node->btree[entno].before, ARCH_CONVERT);
+		par_blkno = be32_to_cpu(par_node->btree[entno].before);
 		if (level == dead_level + 1)
 			break;
 		xfs_da_brelse(tp, par_buf);
@@ -1892,13 +1883,13 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	 */
 	for (;;) {
 		for (;
-		     entno < INT_GET(par_node->hdr.count, ARCH_CONVERT) &&
-		     INT_GET(par_node->btree[entno].before, ARCH_CONVERT) != last_blkno;
+		     entno < be16_to_cpu(par_node->hdr.count) &&
+		     be32_to_cpu(par_node->btree[entno].before) != last_blkno;
 		     entno++)
 			continue;
-		if (entno < INT_GET(par_node->hdr.count, ARCH_CONVERT))
+		if (entno < be16_to_cpu(par_node->hdr.count))
 			break;
-		par_blkno = INT_GET(par_node->hdr.info.forw, ARCH_CONVERT);
+		par_blkno = be32_to_cpu(par_node->hdr.info.forw);
 		xfs_da_brelse(tp, par_buf);
 		par_buf = NULL;
 		if (unlikely(par_blkno == 0)) {
@@ -1911,8 +1902,8 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 			goto done;
 		par_node = par_buf->data;
 		if (unlikely(
-		    INT_GET(par_node->hdr.level, ARCH_CONVERT) != level ||
-		    INT_GET(par_node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC)) {
+		    be16_to_cpu(par_node->hdr.level) != level ||
+		    be16_to_cpu(par_node->hdr.info.magic) != XFS_DA_NODE_MAGIC)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(7)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
@@ -1923,7 +1914,7 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * Update the parent entry pointing to the moved block.
 	 */
-	INT_SET(par_node->btree[entno].before, ARCH_CONVERT, dead_blkno);
+	par_node->btree[entno].before = cpu_to_be32(dead_blkno);
 	xfs_da_log_buf(tp, par_buf,
 		XFS_DA_LOGRANGE(par_node, &par_node->btree[entno].before,
 				sizeof(par_node->btree[entno].before)));
@@ -2203,8 +2194,8 @@ xfs_da_do_buf(
 		info = rbp->data;
 		data = rbp->data;
 		free = rbp->data;
-		magic = INT_GET(info->magic, ARCH_CONVERT);
-		magic1 = INT_GET(data->hdr.magic, ARCH_CONVERT);
+		magic = be16_to_cpu(info->magic);
+		magic1 = be32_to_cpu(data->hdr.magic);
 		if (unlikely(
 		    XFS_TEST_ERROR((magic != XFS_DA_NODE_MAGIC) &&
 				   (magic != XFS_DIR_LEAF_MAGIC) &&
@@ -2213,7 +2204,7 @@ xfs_da_do_buf(
 				   (magic != XFS_DIR2_LEAFN_MAGIC) &&
 				   (magic1 != XFS_DIR2_BLOCK_MAGIC) &&
 				   (magic1 != XFS_DIR2_DATA_MAGIC) &&
-				   (INT_GET(free->hdr.magic, ARCH_CONVERT) != XFS_DIR2_FREE_MAGIC),
+				   (be32_to_cpu(free->hdr.magic) != XFS_DIR2_FREE_MAGIC),
 				mp, XFS_ERRTAG_DA_READ_BUF,
 				XFS_RANDOM_DA_READ_BUF))) {
 			xfs_buftrace("DA READ ERROR", rbp->bps[0]);
diff --git a/fs/xfs/xfs_da_btree.h b/fs/xfs/xfs_da_btree.h
index 41352113721a..243a730d5ec8 100644
--- a/fs/xfs/xfs_da_btree.h
+++ b/fs/xfs/xfs_da_btree.h
@@ -45,10 +45,10 @@ struct zone;
 	(XFS_DIR_IS_V1(mp) ? XFS_DIR_LEAF_MAGIC : XFS_DIR2_LEAFN_MAGIC)
 
 typedef struct xfs_da_blkinfo {
-	xfs_dablk_t forw;			/* previous block in list */
-	xfs_dablk_t back;			/* following block in list */
-	__uint16_t magic;			/* validity check on block */
-	__uint16_t pad;				/* unused */
+	__be32		forw;			/* previous block in list */
+	__be32		back;			/* following block in list */
+	__be16		magic;			/* validity check on block */
+	__be16		pad;			/* unused */
 } xfs_da_blkinfo_t;
 
 /*
@@ -65,12 +65,12 @@ typedef struct xfs_da_blkinfo {
 typedef struct xfs_da_intnode {
 	struct xfs_da_node_hdr {	/* constant-structure header block */
 		xfs_da_blkinfo_t info;	/* block type, links, etc. */
-		__uint16_t count;	/* count of active entries */
-		__uint16_t level;	/* level above leaves (leaf == 0) */
+		__be16	count;		/* count of active entries */
+		__be16	level;		/* level above leaves (leaf == 0) */
 	} hdr;
 	struct xfs_da_node_entry {
-		xfs_dahash_t hashval;	/* hash value for this descendant */
-		xfs_dablk_t before;	/* Btree block before this key */
+		__be32	hashval;	/* hash value for this descendant */
+		__be32	before;		/* Btree block before this key */
 	} btree[1];			/* variable sized array of keys */
 } xfs_da_intnode_t;
 typedef struct xfs_da_node_hdr xfs_da_node_hdr_t;
diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c
index c6191d00ad27..4968a6358e61 100644
--- a/fs/xfs/xfs_dfrag.c
+++ b/fs/xfs/xfs_dfrag.c
@@ -83,7 +83,7 @@ xfs_swapext(
 
 	/* Pull information for the target fd */
 	if (((fp = fget((int)sxp->sx_fdtarget)) == NULL) ||
-	    ((vp = LINVFS_GET_VP(fp->f_dentry->d_inode)) == NULL))  {
+	    ((vp = vn_from_inode(fp->f_dentry->d_inode)) == NULL))  {
 		error = XFS_ERROR(EINVAL);
 		goto error0;
 	}
@@ -95,7 +95,7 @@ xfs_swapext(
 	}
 
 	if (((tfp = fget((int)sxp->sx_fdtmp)) == NULL) ||
-	    ((tvp = LINVFS_GET_VP(tfp->f_dentry->d_inode)) == NULL)) {
+	    ((tvp = vn_from_inode(tfp->f_dentry->d_inode)) == NULL)) {
 		error = XFS_ERROR(EINVAL);
 		goto error0;
 	}
diff --git a/fs/xfs/xfs_dir.c b/fs/xfs/xfs_dir.c
index bb87d2a700a9..9cc702a839a3 100644
--- a/fs/xfs/xfs_dir.c
+++ b/fs/xfs/xfs_dir.c
@@ -634,7 +634,7 @@ xfs_dir_leaf_removename(xfs_da_args_t *args, int *count, int *totallen)
 		return(retval);
 	ASSERT(bp != NULL);
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	retval = xfs_dir_leaf_lookup_int(bp, args, &index);
 	if (retval == EEXIST) {
 		(void)xfs_dir_leaf_remove(args->trans, bp, index);
@@ -912,7 +912,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			return(error);
 		if (bp)
 			leaf = bp->data;
-		if (bp && INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC) {
+		if (bp && be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC) {
 			xfs_dir_trace_g_dub("node: block not a leaf",
 						   dp, uio, bno);
 			xfs_da_brelse(trans, bp);
@@ -949,17 +949,17 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			if (bp == NULL)
 				return(XFS_ERROR(EFSCORRUPTED));
 			node = bp->data;
-			if (INT_GET(node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC)
+			if (be16_to_cpu(node->hdr.info.magic) != XFS_DA_NODE_MAGIC)
 				break;
 			btree = &node->btree[0];
 			xfs_dir_trace_g_dun("node: node detail", dp, uio, node);
-			for (i = 0; i < INT_GET(node->hdr.count, ARCH_CONVERT); btree++, i++) {
-				if (INT_GET(btree->hashval, ARCH_CONVERT) >= cookhash) {
-					bno = INT_GET(btree->before, ARCH_CONVERT);
+			for (i = 0; i < be16_to_cpu(node->hdr.count); btree++, i++) {
+				if (be32_to_cpu(btree->hashval) >= cookhash) {
+					bno = be32_to_cpu(btree->before);
 					break;
 				}
 			}
-			if (i == INT_GET(node->hdr.count, ARCH_CONVERT)) {
+			if (i == be16_to_cpu(node->hdr.count)) {
 				xfs_da_brelse(trans, bp);
 				xfs_dir_trace_g_du("node: hash beyond EOF",
 							  dp, uio);
@@ -982,7 +982,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 	 */
 	for (;;) {
 		leaf = bp->data;
-		if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC)) {
+		if (unlikely(be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC)) {
 			xfs_dir_trace_g_dul("node: not a leaf", dp, uio, leaf);
 			xfs_da_brelse(trans, bp);
 			XFS_CORRUPTION_ERROR("xfs_dir_node_getdents(1)",
@@ -990,7 +990,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			return XFS_ERROR(EFSCORRUPTED);
 		}
 		xfs_dir_trace_g_dul("node: leaf detail", dp, uio, leaf);
-		if ((nextbno = INT_GET(leaf->hdr.info.forw, ARCH_CONVERT))) {
+		if ((nextbno = be32_to_cpu(leaf->hdr.info.forw))) {
 			nextda = xfs_da_reada_buf(trans, dp, nextbno,
 						  XFS_DATA_FORK);
 		} else
@@ -1118,21 +1118,20 @@ void
 xfs_dir_trace_g_dun(char *where, xfs_inode_t *dp, uio_t *uio,
 			xfs_da_intnode_t *node)
 {
-	int	last = INT_GET(node->hdr.count, ARCH_CONVERT) - 1;
+	int	last = be16_to_cpu(node->hdr.count) - 1;
 
 	xfs_dir_trace_enter(XFS_DIR_KTRACE_G_DUN, where,
 		     (void *)dp, (void *)dp->i_mount,
 		     (void *)((unsigned long)(uio->uio_offset >> 32)),
 		     (void *)((unsigned long)(uio->uio_offset & 0xFFFFFFFF)),
 		     (void *)(unsigned long)uio->uio_resid,
+		     (void *)(unsigned long)be32_to_cpu(node->hdr.info.forw),
 		     (void *)(unsigned long)
-			INT_GET(node->hdr.info.forw, ARCH_CONVERT),
+			be16_to_cpu(node->hdr.count),
 		     (void *)(unsigned long)
-			INT_GET(node->hdr.count, ARCH_CONVERT),
+			be32_to_cpu(node->btree[0].hashval),
 		     (void *)(unsigned long)
-			INT_GET(node->btree[0].hashval, ARCH_CONVERT),
-		     (void *)(unsigned long)
-			INT_GET(node->btree[last].hashval, ARCH_CONVERT),
+			be32_to_cpu(node->btree[last].hashval),
 		     NULL, NULL, NULL);
 }
 
@@ -1150,8 +1149,7 @@ xfs_dir_trace_g_dul(char *where, xfs_inode_t *dp, uio_t *uio,
 		     (void *)((unsigned long)(uio->uio_offset >> 32)),
 		     (void *)((unsigned long)(uio->uio_offset & 0xFFFFFFFF)),
 		     (void *)(unsigned long)uio->uio_resid,
-		     (void *)(unsigned long)
-			INT_GET(leaf->hdr.info.forw, ARCH_CONVERT),
+		     (void *)(unsigned long)be32_to_cpu(leaf->hdr.info.forw),
 		     (void *)(unsigned long)
 			INT_GET(leaf->hdr.count, ARCH_CONVERT),
 		     (void *)(unsigned long)
diff --git a/fs/xfs/xfs_dir2.h b/fs/xfs/xfs_dir2.h
index 3158f5dc431f..7dd364b1e038 100644
--- a/fs/xfs/xfs_dir2.h
+++ b/fs/xfs/xfs_dir2.h
@@ -55,16 +55,16 @@ typedef	__uint32_t	xfs_dir2_db_t;
 /*
  * Byte offset in a directory.
  */
-typedef	xfs_off_t		xfs_dir2_off_t;
+typedef	xfs_off_t	xfs_dir2_off_t;
 
 /*
  * For getdents, argument struct for put routines.
  */
 typedef int (*xfs_dir2_put_t)(struct xfs_dir2_put_args *pa);
 typedef struct xfs_dir2_put_args {
-	xfs_off_t		cook;		/* cookie of (next) entry */
+	xfs_off_t	cook;		/* cookie of (next) entry */
 	xfs_intino_t	ino;		/* inode number */
-	struct xfs_dirent	*dbp;		/* buffer pointer */
+	xfs_dirent_t	*dbp;		/* buffer pointer */
 	char		*name;		/* directory entry name */
 	int		namelen;	/* length of name */
 	int		done;		/* output: set if value was stored */
@@ -75,18 +75,13 @@ typedef struct xfs_dir2_put_args {
 /*
  * Other interfaces used by the rest of the dir v2 code.
  */
-extern int
-	xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
-			    xfs_dir2_db_t *dbp);
-
-extern int
-	xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *vp);
-
-extern int
-	xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *vp);
-
-extern int
-	xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
-			      struct xfs_dabuf *bp);
+extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
+				xfs_dir2_db_t *dbp);
+extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp,
+				int *vp);
+extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp,
+				int *vp);
+extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
+				struct xfs_dabuf *bp);
 
 #endif	/* __XFS_DIR2_H__ */
diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c
index 31bc99faa704..bd5cee6aa51a 100644
--- a/fs/xfs/xfs_dir2_block.c
+++ b/fs/xfs/xfs_dir2_block.c
@@ -81,7 +81,7 @@ xfs_dir2_block_addname(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			needlog;	/* need to log header */
 	int			needscan;	/* need to rescan freespace */
-	xfs_dir2_data_off_t	*tagp;		/* pointer to tag value */
+	__be16			*tagp;		/* pointer to tag value */
 	xfs_trans_t		*tp;		/* transaction structure */
 
 	xfs_dir2_trace_args("block_addname", args);
@@ -100,8 +100,7 @@ xfs_dir2_block_addname(
 	/*
 	 * Check the magic number, corrupted if wrong.
 	 */
-	if (unlikely(INT_GET(block->hdr.magic, ARCH_CONVERT)
-						!= XFS_DIR2_BLOCK_MAGIC)) {
+	if (unlikely(be32_to_cpu(block->hdr.magic) != XFS_DIR2_BLOCK_MAGIC)) {
 		XFS_CORRUPTION_ERROR("xfs_dir2_block_addname",
 				     XFS_ERRLEVEL_LOW, mp, block);
 		xfs_da_brelse(tp, bp);
@@ -121,38 +120,38 @@ xfs_dir2_block_addname(
 		/*
 		 * Tag just before the first leaf entry.
 		 */
-		tagp = (xfs_dir2_data_off_t *)blp - 1;
+		tagp = (__be16 *)blp - 1;
 		/*
 		 * Data object just before the first leaf entry.
 		 */
-		enddup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+		enddup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 		/*
 		 * If it's not free then can't do this add without cleaning up:
 		 * the space before the first leaf entry needs to be free so it
 		 * can be expanded to hold the pointer to the new entry.
 		 */
-		if (INT_GET(enddup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		if (be16_to_cpu(enddup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			dup = enddup = NULL;
 		/*
 		 * Check out the biggest freespace and see if it's the same one.
 		 */
 		else {
 			dup = (xfs_dir2_data_unused_t *)
-			      ((char *)block + INT_GET(bf[0].offset, ARCH_CONVERT));
+			      ((char *)block + be16_to_cpu(bf[0].offset));
 			if (dup == enddup) {
 				/*
 				 * It is the biggest freespace, is it too small
 				 * to hold the new leaf too?
 				 */
-				if (INT_GET(dup->length, ARCH_CONVERT) < len + (uint)sizeof(*blp)) {
+				if (be16_to_cpu(dup->length) < len + (uint)sizeof(*blp)) {
 					/*
 					 * Yes, we use the second-largest
 					 * entry instead if it works.
 					 */
-					if (INT_GET(bf[1].length, ARCH_CONVERT) >= len)
+					if (be16_to_cpu(bf[1].length) >= len)
 						dup = (xfs_dir2_data_unused_t *)
 						      ((char *)block +
-						       INT_GET(bf[1].offset, ARCH_CONVERT));
+						       be16_to_cpu(bf[1].offset));
 					else
 						dup = NULL;
 				}
@@ -161,7 +160,7 @@ xfs_dir2_block_addname(
 				 * Not the same free entry,
 				 * just check its length.
 				 */
-				if (INT_GET(dup->length, ARCH_CONVERT) < len) {
+				if (be16_to_cpu(dup->length) < len) {
 					dup = NULL;
 				}
 			}
@@ -172,9 +171,9 @@ xfs_dir2_block_addname(
 	 * If there are stale entries we'll use one for the leaf.
 	 * Is the biggest entry enough to avoid compaction?
 	 */
-	else if (INT_GET(bf[0].length, ARCH_CONVERT) >= len) {
+	else if (be16_to_cpu(bf[0].length) >= len) {
 		dup = (xfs_dir2_data_unused_t *)
-		      ((char *)block + INT_GET(bf[0].offset, ARCH_CONVERT));
+		      ((char *)block + be16_to_cpu(bf[0].offset));
 		compact = 0;
 	}
 	/*
@@ -184,20 +183,20 @@ xfs_dir2_block_addname(
 		/*
 		 * Tag just before the first leaf entry.
 		 */
-		tagp = (xfs_dir2_data_off_t *)blp - 1;
+		tagp = (__be16 *)blp - 1;
 		/*
 		 * Data object just before the first leaf entry.
 		 */
-		dup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+		dup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 		/*
 		 * If it's not free then the data will go where the
 		 * leaf data starts now, if it works at all.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			if (INT_GET(dup->length, ARCH_CONVERT) + (INT_GET(btp->stale, ARCH_CONVERT) - 1) *
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			if (be16_to_cpu(dup->length) + (be32_to_cpu(btp->stale) - 1) *
 			    (uint)sizeof(*blp) < len)
 				dup = NULL;
-		} else if ((INT_GET(btp->stale, ARCH_CONVERT) - 1) * (uint)sizeof(*blp) < len)
+		} else if ((be32_to_cpu(btp->stale) - 1) * (uint)sizeof(*blp) < len)
 			dup = NULL;
 		else
 			dup = (xfs_dir2_data_unused_t *)blp;
@@ -243,11 +242,11 @@ xfs_dir2_block_addname(
 		int	fromidx;		/* source leaf index */
 		int	toidx;			/* target leaf index */
 
-		for (fromidx = toidx = INT_GET(btp->count, ARCH_CONVERT) - 1,
+		for (fromidx = toidx = be32_to_cpu(btp->count) - 1,
 			highstale = lfloghigh = -1;
 		     fromidx >= 0;
 		     fromidx--) {
-			if (INT_GET(blp[fromidx].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR) {
+			if (be32_to_cpu(blp[fromidx].address) == XFS_DIR2_NULL_DATAPTR) {
 				if (highstale == -1)
 					highstale = toidx;
 				else {
@@ -260,15 +259,15 @@ xfs_dir2_block_addname(
 				blp[toidx] = blp[fromidx];
 			toidx--;
 		}
-		lfloglow = toidx + 1 - (INT_GET(btp->stale, ARCH_CONVERT) - 1);
-		lfloghigh -= INT_GET(btp->stale, ARCH_CONVERT) - 1;
-		INT_MOD(btp->count, ARCH_CONVERT, -(INT_GET(btp->stale, ARCH_CONVERT) - 1));
+		lfloglow = toidx + 1 - (be32_to_cpu(btp->stale) - 1);
+		lfloghigh -= be32_to_cpu(btp->stale) - 1;
+		be32_add(&btp->count, -(be32_to_cpu(btp->stale) - 1));
 		xfs_dir2_data_make_free(tp, bp,
 			(xfs_dir2_data_aoff_t)((char *)blp - (char *)block),
-			(xfs_dir2_data_aoff_t)((INT_GET(btp->stale, ARCH_CONVERT) - 1) * sizeof(*blp)),
+			(xfs_dir2_data_aoff_t)((be32_to_cpu(btp->stale) - 1) * sizeof(*blp)),
 			&needlog, &needscan);
-		blp += INT_GET(btp->stale, ARCH_CONVERT) - 1;
-		INT_SET(btp->stale, ARCH_CONVERT, 1);
+		blp += be32_to_cpu(btp->stale) - 1;
+		btp->stale = cpu_to_be32(1);
 		/*
 		 * If we now need to rebuild the bestfree map, do so.
 		 * This needs to happen before the next call to use_free.
@@ -283,23 +282,23 @@ xfs_dir2_block_addname(
 	 * Set leaf logging boundaries to impossible state.
 	 * For the no-stale case they're set explicitly.
 	 */
-	else if (INT_GET(btp->stale, ARCH_CONVERT)) {
-		lfloglow = INT_GET(btp->count, ARCH_CONVERT);
+	else if (btp->stale) {
+		lfloglow = be32_to_cpu(btp->count);
 		lfloghigh = -1;
 	}
 	/*
 	 * Find the slot that's first lower than our hash value, -1 if none.
 	 */
-	for (low = 0, high = INT_GET(btp->count, ARCH_CONVERT) - 1; low <= high; ) {
+	for (low = 0, high = be32_to_cpu(btp->count) - 1; low <= high; ) {
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(blp[mid].hashval, ARCH_CONVERT)) == args->hashval)
+		if ((hash = be32_to_cpu(blp[mid].hashval)) == args->hashval)
 			break;
 		if (hash < args->hashval)
 			low = mid + 1;
 		else
 			high = mid - 1;
 	}
-	while (mid >= 0 && INT_GET(blp[mid].hashval, ARCH_CONVERT) >= args->hashval) {
+	while (mid >= 0 && be32_to_cpu(blp[mid].hashval) >= args->hashval) {
 		mid--;
 	}
 	/*
@@ -311,14 +310,14 @@ xfs_dir2_block_addname(
 		 */
 		xfs_dir2_data_use_free(tp, bp, enddup,
 			(xfs_dir2_data_aoff_t)
-			((char *)enddup - (char *)block + INT_GET(enddup->length, ARCH_CONVERT) -
+			((char *)enddup - (char *)block + be16_to_cpu(enddup->length) -
 			 sizeof(*blp)),
 			(xfs_dir2_data_aoff_t)sizeof(*blp),
 			&needlog, &needscan);
 		/*
 		 * Update the tail (entry count).
 		 */
-		INT_MOD(btp->count, ARCH_CONVERT, +1);
+		be32_add(&btp->count, 1);
 		/*
 		 * If we now need to rebuild the bestfree map, do so.
 		 * This needs to happen before the next call to use_free.
@@ -346,12 +345,12 @@ xfs_dir2_block_addname(
 	else {
 		for (lowstale = mid;
 		     lowstale >= 0 &&
-			INT_GET(blp[lowstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR;
+			be32_to_cpu(blp[lowstale].address) != XFS_DIR2_NULL_DATAPTR;
 		     lowstale--)
 			continue;
 		for (highstale = mid + 1;
-		     highstale < INT_GET(btp->count, ARCH_CONVERT) &&
-			INT_GET(blp[highstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR &&
+		     highstale < be32_to_cpu(btp->count) &&
+			be32_to_cpu(blp[highstale].address) != XFS_DIR2_NULL_DATAPTR &&
 			(lowstale < 0 || mid - lowstale > highstale - mid);
 		     highstale++)
 			continue;
@@ -359,7 +358,7 @@ xfs_dir2_block_addname(
 		 * Move entries toward the low-numbered stale entry.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(btp->count, ARCH_CONVERT) ||
+		    (highstale == be32_to_cpu(btp->count) ||
 		     mid - lowstale <= highstale - mid)) {
 			if (mid - lowstale)
 				memmove(&blp[lowstale], &blp[lowstale + 1],
@@ -371,7 +370,7 @@ xfs_dir2_block_addname(
 		 * Move entries toward the high-numbered stale entry.
 		 */
 		else {
-			ASSERT(highstale < INT_GET(btp->count, ARCH_CONVERT));
+			ASSERT(highstale < be32_to_cpu(btp->count));
 			mid++;
 			if (highstale - mid)
 				memmove(&blp[mid + 1], &blp[mid],
@@ -379,7 +378,7 @@ xfs_dir2_block_addname(
 			lfloglow = MIN(mid, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(btp->stale, ARCH_CONVERT, -1);
+		be32_add(&btp->stale, -1);
 	}
 	/*
 	 * Point to the new data entry.
@@ -388,8 +387,9 @@ xfs_dir2_block_addname(
 	/*
 	 * Fill in the leaf entry.
 	 */
-	INT_SET(blp[mid].hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(blp[mid].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[mid].hashval = cpu_to_be32(args->hashval);
+	blp[mid].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	xfs_dir2_block_log_leaf(tp, bp, lfloglow, lfloghigh);
 	/*
 	 * Mark space for the data entry used.
@@ -404,7 +404,7 @@ xfs_dir2_block_addname(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, args->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	/*
 	 * Clean up the bestfree array and log the header, tail, and entry.
 	 */
@@ -485,8 +485,8 @@ xfs_dir2_block_getdents(
 		/*
 		 * Unused, skip it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			ptr += INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			ptr += be16_to_cpu(dup->length);
 			continue;
 		}
 
@@ -622,7 +622,7 @@ xfs_dir2_block_lookup(
 	 * Get the offset from the leaf entry, to point to the data.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	/*
 	 * Fill in inode number, release the block.
 	 */
@@ -674,10 +674,10 @@ xfs_dir2_block_lookup_int(
 	 * Loop doing a binary search for our hash value.
 	 * Find our entry, ENOENT if it's not there.
 	 */
-	for (low = 0, high = INT_GET(btp->count, ARCH_CONVERT) - 1; ; ) {
+	for (low = 0, high = be32_to_cpu(btp->count) - 1; ; ) {
 		ASSERT(low <= high);
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(blp[mid].hashval, ARCH_CONVERT)) == args->hashval)
+		if ((hash = be32_to_cpu(blp[mid].hashval)) == args->hashval)
 			break;
 		if (hash < args->hashval)
 			low = mid + 1;
@@ -692,7 +692,7 @@ xfs_dir2_block_lookup_int(
 	/*
 	 * Back up to the first one with the right hash value.
 	 */
-	while (mid > 0 && INT_GET(blp[mid - 1].hashval, ARCH_CONVERT) == args->hashval) {
+	while (mid > 0 && be32_to_cpu(blp[mid - 1].hashval) == args->hashval) {
 		mid--;
 	}
 	/*
@@ -700,7 +700,7 @@ xfs_dir2_block_lookup_int(
 	 * right hash value looking for our name.
 	 */
 	do {
-		if ((addr = INT_GET(blp[mid].address, ARCH_CONVERT)) == XFS_DIR2_NULL_DATAPTR)
+		if ((addr = be32_to_cpu(blp[mid].address)) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Get pointer to the entry from the leaf.
@@ -717,7 +717,7 @@ xfs_dir2_block_lookup_int(
 			*entno = mid;
 			return 0;
 		}
-	} while (++mid < INT_GET(btp->count, ARCH_CONVERT) && INT_GET(blp[mid].hashval, ARCH_CONVERT) == hash);
+	} while (++mid < be32_to_cpu(btp->count) && be32_to_cpu(blp[mid].hashval) == hash);
 	/*
 	 * No match, release the buffer and return ENOENT.
 	 */
@@ -767,7 +767,7 @@ xfs_dir2_block_removename(
 	 * Point to the data entry using the leaf entry.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	/*
 	 * Mark the data entry's space free.
 	 */
@@ -778,12 +778,12 @@ xfs_dir2_block_removename(
 	/*
 	 * Fix up the block tail.
 	 */
-	INT_MOD(btp->stale, ARCH_CONVERT, +1);
+	be32_add(&btp->stale, 1);
 	xfs_dir2_block_log_tail(tp, bp);
 	/*
 	 * Remove the leaf entry by marking it stale.
 	 */
-	INT_SET(blp[ent].address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	blp[ent].address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_block_log_leaf(tp, bp, ent, ent);
 	/*
 	 * Fix up bestfree, log the header if necessary.
@@ -843,7 +843,7 @@ xfs_dir2_block_replace(
 	 * Point to the data entry we need to change.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	ASSERT(INT_GET(dep->inumber, ARCH_CONVERT) != args->inumber);
 	/*
 	 * Change the inode number to the new value.
@@ -868,8 +868,8 @@ xfs_dir2_block_sort(
 
 	la = a;
 	lb = b;
-	return INT_GET(la->hashval, ARCH_CONVERT) < INT_GET(lb->hashval, ARCH_CONVERT) ? -1 :
-		(INT_GET(la->hashval, ARCH_CONVERT) > INT_GET(lb->hashval, ARCH_CONVERT) ? 1 : 0);
+	return be32_to_cpu(la->hashval) < be32_to_cpu(lb->hashval) ? -1 :
+		(be32_to_cpu(la->hashval) > be32_to_cpu(lb->hashval) ? 1 : 0);
 }
 
 /*
@@ -881,7 +881,7 @@ xfs_dir2_leaf_to_block(
 	xfs_dabuf_t		*lbp,		/* leaf buffer */
 	xfs_dabuf_t		*dbp)		/* data buffer */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf bests table */
+	__be16			*bestsp;	/* leaf bests table */
 	xfs_dir2_block_t	*block;		/* block structure */
 	xfs_dir2_block_tail_t	*btp;		/* block tail */
 	xfs_inode_t		*dp;		/* incore directory inode */
@@ -896,7 +896,7 @@ xfs_dir2_leaf_to_block(
 	int			needscan;	/* need to scan for bestfree */
 	xfs_dir2_sf_hdr_t	sfh;		/* shortform header */
 	int			size;		/* bytes used */
-	xfs_dir2_data_off_t	*tagp;		/* end of entry (tag) */
+	__be16			*tagp;		/* end of entry (tag) */
 	int			to;		/* block/leaf to index */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
@@ -905,7 +905,7 @@ xfs_dir2_leaf_to_block(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = lbp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	/*
 	 * If there are data blocks other than the first one, take this
@@ -915,11 +915,11 @@ xfs_dir2_leaf_to_block(
 	 */
 	while (dp->i_d.di_size > mp->m_dirblksize) {
 		bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-		if (INT_GET(bestsp[INT_GET(ltp->bestcount, ARCH_CONVERT) - 1], ARCH_CONVERT) ==
+		if (be16_to_cpu(bestsp[be32_to_cpu(ltp->bestcount) - 1]) ==
 		    mp->m_dirblksize - (uint)sizeof(block->hdr)) {
 			if ((error =
 			    xfs_dir2_leaf_trim_data(args, lbp,
-				    (xfs_dir2_db_t)(INT_GET(ltp->bestcount, ARCH_CONVERT) - 1))))
+				    (xfs_dir2_db_t)(be32_to_cpu(ltp->bestcount) - 1))))
 				goto out;
 		} else {
 			error = 0;
@@ -935,28 +935,29 @@ xfs_dir2_leaf_to_block(
 		goto out;
 	}
 	block = dbp->data;
-	ASSERT(INT_GET(block->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+	ASSERT(be32_to_cpu(block->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 	/*
 	 * Size of the "leaf" area in the block.
 	 */
 	size = (uint)sizeof(block->tail) +
-	       (uint)sizeof(*lep) * (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT));
+	       (uint)sizeof(*lep) * (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale));
 	/*
 	 * Look at the last data entry.
 	 */
-	tagp = (xfs_dir2_data_off_t *)((char *)block + mp->m_dirblksize) - 1;
-	dup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+	tagp = (__be16 *)((char *)block + mp->m_dirblksize) - 1;
+	dup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 	/*
 	 * If it's not free or is too short we can't do it.
 	 */
-	if (INT_GET(dup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG || INT_GET(dup->length, ARCH_CONVERT) < size) {
+	if (be16_to_cpu(dup->freetag) != XFS_DIR2_DATA_FREE_TAG ||
+	    be16_to_cpu(dup->length) < size) {
 		error = 0;
 		goto out;
 	}
 	/*
 	 * Start converting it to block form.
 	 */
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_BLOCK_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
 	needlog = 1;
 	needscan = 0;
 	/*
@@ -968,20 +969,20 @@ xfs_dir2_leaf_to_block(
 	 * Initialize the block tail.
 	 */
 	btp = XFS_DIR2_BLOCK_TAIL_P(mp, block);
-	INT_SET(btp->count, ARCH_CONVERT, INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT));
+	btp->count = cpu_to_be32(be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale));
 	btp->stale = 0;
 	xfs_dir2_block_log_tail(tp, dbp);
 	/*
 	 * Initialize the block leaf area.  We compact out stale entries.
 	 */
 	lep = XFS_DIR2_BLOCK_LEAF_P(btp);
-	for (from = to = 0; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
-		if (INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (from = to = 0; from < be16_to_cpu(leaf->hdr.count); from++) {
+		if (be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		lep[to++] = leaf->ents[from];
 	}
-	ASSERT(to == INT_GET(btp->count, ARCH_CONVERT));
-	xfs_dir2_block_log_leaf(tp, dbp, 0, INT_GET(btp->count, ARCH_CONVERT) - 1);
+	ASSERT(to == be32_to_cpu(btp->count));
+	xfs_dir2_block_log_leaf(tp, dbp, 0, be32_to_cpu(btp->count) - 1);
 	/*
 	 * Scan the bestfree if we need it and log the data block header.
 	 */
@@ -1043,7 +1044,7 @@ xfs_dir2_sf_to_block(
 	int			offset;		/* target block offset */
 	xfs_dir2_sf_entry_t	*sfep;		/* sf entry pointer */
 	xfs_dir2_sf_t		*sfp;		/* shortform structure */
-	xfs_dir2_data_off_t	*tagp;		/* end of data entry */
+	__be16			*tagp;		/* end of data entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	xfs_dir2_trace_args("sf_to_block", args);
@@ -1095,12 +1096,12 @@ xfs_dir2_sf_to_block(
 		return error;
 	}
 	block = bp->data;
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_BLOCK_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
 	/*
 	 * Compute size of block "tail" area.
 	 */
 	i = (uint)sizeof(*btp) +
-	    (INT_GET(sfp->hdr.count, ARCH_CONVERT) + 2) * (uint)sizeof(xfs_dir2_leaf_entry_t);
+	    (sfp->hdr.count + 2) * (uint)sizeof(xfs_dir2_leaf_entry_t);
 	/*
 	 * The whole thing is initialized to free by the init routine.
 	 * Say we're using the leaf and tail area.
@@ -1114,7 +1115,7 @@ xfs_dir2_sf_to_block(
 	 * Fill in the tail.
 	 */
 	btp = XFS_DIR2_BLOCK_TAIL_P(mp, block);
-	INT_SET(btp->count, ARCH_CONVERT, INT_GET(sfp->hdr.count, ARCH_CONVERT) + 2);	/* ., .. */
+	btp->count = cpu_to_be32(sfp->hdr.count + 2);	/* ., .. */
 	btp->stale = 0;
 	blp = XFS_DIR2_BLOCK_LEAF_P(btp);
 	endoffset = (uint)((char *)blp - (char *)block);
@@ -1123,7 +1124,7 @@ xfs_dir2_sf_to_block(
 	 */
 	xfs_dir2_data_use_free(tp, bp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)block),
-		INT_GET(dup->length, ARCH_CONVERT), &needlog, &needscan);
+		be16_to_cpu(dup->length), &needlog, &needscan);
 	/*
 	 * Create entry for .
 	 */
@@ -1133,10 +1134,11 @@ xfs_dir2_sf_to_block(
 	dep->namelen = 1;
 	dep->name[0] = '.';
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	xfs_dir2_data_log_entry(tp, bp, dep);
-	INT_SET(blp[0].hashval, ARCH_CONVERT, xfs_dir_hash_dot);
-	INT_SET(blp[0].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[0].hashval = cpu_to_be32(xfs_dir_hash_dot);
+	blp[0].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	/*
 	 * Create entry for ..
 	 */
@@ -1146,15 +1148,16 @@ xfs_dir2_sf_to_block(
 	dep->namelen = 2;
 	dep->name[0] = dep->name[1] = '.';
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	xfs_dir2_data_log_entry(tp, bp, dep);
-	INT_SET(blp[1].hashval, ARCH_CONVERT, xfs_dir_hash_dotdot);
-	INT_SET(blp[1].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[1].hashval = cpu_to_be32(xfs_dir_hash_dotdot);
+	blp[1].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	offset = XFS_DIR2_DATA_FIRST_OFFSET;
 	/*
 	 * Loop over existing entries, stuff them in.
 	 */
-	if ((i = 0) == INT_GET(sfp->hdr.count, ARCH_CONVERT))
+	if ((i = 0) == sfp->hdr.count)
 		sfep = NULL;
 	else
 		sfep = XFS_DIR2_SF_FIRSTENTRY(sfp);
@@ -1176,15 +1179,14 @@ xfs_dir2_sf_to_block(
 		if (offset < newoffset) {
 			dup = (xfs_dir2_data_unused_t *)
 			      ((char *)block + offset);
-			INT_SET(dup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-			INT_SET(dup->length, ARCH_CONVERT, newoffset - offset);
-			INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT,
-				(xfs_dir2_data_off_t)
+			dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+			dup->length = cpu_to_be16(newoffset - offset);
+			*XFS_DIR2_DATA_UNUSED_TAG_P(dup) = cpu_to_be16(
 				((char *)dup - (char *)block));
 			xfs_dir2_data_log_unused(tp, bp, dup);
 			(void)xfs_dir2_data_freeinsert((xfs_dir2_data_t *)block,
 				dup, &dummy);
-			offset += INT_GET(dup->length, ARCH_CONVERT);
+			offset += be16_to_cpu(dup->length);
 			continue;
 		}
 		/*
@@ -1196,13 +1198,14 @@ xfs_dir2_sf_to_block(
 		dep->namelen = sfep->namelen;
 		memcpy(dep->name, sfep->name, dep->namelen);
 		tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-		INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+		*tagp = cpu_to_be16((char *)dep - (char *)block);
 		xfs_dir2_data_log_entry(tp, bp, dep);
-		INT_SET(blp[2 + i].hashval, ARCH_CONVERT, xfs_da_hashname((char *)sfep->name, sfep->namelen));
-		INT_SET(blp[2 + i].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp,
+		blp[2 + i].hashval = cpu_to_be32(xfs_da_hashname(
+					(char *)sfep->name, sfep->namelen));
+		blp[2 + i].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
 						 (char *)dep - (char *)block));
 		offset = (int)((char *)(tagp + 1) - (char *)block);
-		if (++i == INT_GET(sfp->hdr.count, ARCH_CONVERT))
+		if (++i == sfp->hdr.count)
 			sfep = NULL;
 		else
 			sfep = XFS_DIR2_SF_NEXTENTRY(sfp, sfep);
@@ -1212,13 +1215,13 @@ xfs_dir2_sf_to_block(
 	/*
 	 * Sort the leaf entries by hash value.
 	 */
-	xfs_sort(blp, INT_GET(btp->count, ARCH_CONVERT), sizeof(*blp), xfs_dir2_block_sort);
+	xfs_sort(blp, be32_to_cpu(btp->count), sizeof(*blp), xfs_dir2_block_sort);
 	/*
 	 * Log the leaf entry area and tail.
 	 * Already logged the header in data_init, ignore needlog.
 	 */
 	ASSERT(needscan == 0);
-	xfs_dir2_block_log_leaf(tp, bp, 0, INT_GET(btp->count, ARCH_CONVERT) - 1);
+	xfs_dir2_block_log_leaf(tp, bp, 0, be32_to_cpu(btp->count) - 1);
 	xfs_dir2_block_log_tail(tp, bp);
 	xfs_dir2_data_check(dp, bp);
 	xfs_da_buf_done(bp);
diff --git a/fs/xfs/xfs_dir2_block.h b/fs/xfs/xfs_dir2_block.h
index a2e5cb98a838..6722effd0b20 100644
--- a/fs/xfs/xfs_dir2_block.h
+++ b/fs/xfs/xfs_dir2_block.h
@@ -43,8 +43,8 @@ struct xfs_trans;
 #define	XFS_DIR2_BLOCK_MAGIC	0x58443242	/* XD2B: for one block dirs */
 
 typedef struct xfs_dir2_block_tail {
-	__uint32_t	count;			/* count of leaf entries */
-	__uint32_t	stale;			/* count of stale lf entries */
+	__be32		count;			/* count of leaf entries */
+	__be32		stale;			/* count of stale lf entries */
 } xfs_dir2_block_tail_t;
 
 /*
@@ -75,8 +75,7 @@ xfs_dir2_block_tail_p(struct xfs_mount *mp, xfs_dir2_block_t *block)
 static inline struct xfs_dir2_leaf_entry *
 xfs_dir2_block_leaf_p(xfs_dir2_block_tail_t *btp)
 {
-	return (((struct xfs_dir2_leaf_entry *)
-		(btp)) - INT_GET((btp)->count, ARCH_CONVERT));
+	return ((struct xfs_dir2_leaf_entry *)btp) - be32_to_cpu(btp->count);
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c
index 5b7c47e2f14a..bb3d03ff002b 100644
--- a/fs/xfs/xfs_dir2_data.c
+++ b/fs/xfs/xfs_dir2_data.c
@@ -70,11 +70,11 @@ xfs_dir2_data_check(
 
 	mp = dp->i_mount;
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	bf = d->hdr.bestfree;
 	p = (char *)d->u;
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		lep = XFS_DIR2_BLOCK_LEAF_P(btp);
 		endp = (char *)lep;
@@ -96,8 +96,8 @@ xfs_dir2_data_check(
 		ASSERT(!bf[2].offset);
 		freeseen |= 1 << 2;
 	}
-	ASSERT(INT_GET(bf[0].length, ARCH_CONVERT) >= INT_GET(bf[1].length, ARCH_CONVERT));
-	ASSERT(INT_GET(bf[1].length, ARCH_CONVERT) >= INT_GET(bf[2].length, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(bf[0].length) >= be16_to_cpu(bf[1].length));
+	ASSERT(be16_to_cpu(bf[1].length) >= be16_to_cpu(bf[2].length));
 	/*
 	 * Loop over the data/unused entries.
 	 */
@@ -108,18 +108,20 @@ xfs_dir2_data_check(
 		 * If we find it, account for that, else make sure it
 		 * doesn't need to be there.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
 			ASSERT(lastfree == 0);
-			ASSERT(INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT) ==
+			ASSERT(be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)) ==
 			       (char *)dup - (char *)d);
 			dfp = xfs_dir2_data_freefind(d, dup);
 			if (dfp) {
 				i = (int)(dfp - bf);
 				ASSERT((freeseen & (1 << i)) == 0);
 				freeseen |= 1 << i;
-			} else
-				ASSERT(INT_GET(dup->length, ARCH_CONVERT) <= INT_GET(bf[2].length, ARCH_CONVERT));
-			p += INT_GET(dup->length, ARCH_CONVERT);
+			} else {
+				ASSERT(be16_to_cpu(dup->length) <=
+				       be16_to_cpu(bf[2].length));
+			}
+			p += be16_to_cpu(dup->length);
 			lastfree = 1;
 			continue;
 		}
@@ -132,21 +134,21 @@ xfs_dir2_data_check(
 		dep = (xfs_dir2_data_entry_t *)p;
 		ASSERT(dep->namelen != 0);
 		ASSERT(xfs_dir_ino_validate(mp, INT_GET(dep->inumber, ARCH_CONVERT)) == 0);
-		ASSERT(INT_GET(*XFS_DIR2_DATA_ENTRY_TAG_P(dep), ARCH_CONVERT) ==
+		ASSERT(be16_to_cpu(*XFS_DIR2_DATA_ENTRY_TAG_P(dep)) ==
 		       (char *)dep - (char *)d);
 		count++;
 		lastfree = 0;
-		if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+		if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 			addr = XFS_DIR2_DB_OFF_TO_DATAPTR(mp, mp->m_dirdatablk,
 				(xfs_dir2_data_aoff_t)
 				((char *)dep - (char *)d));
 			hash = xfs_da_hashname((char *)dep->name, dep->namelen);
-			for (i = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-				if (INT_GET(lep[i].address, ARCH_CONVERT) == addr &&
-				    INT_GET(lep[i].hashval, ARCH_CONVERT) == hash)
+			for (i = 0; i < be32_to_cpu(btp->count); i++) {
+				if (be32_to_cpu(lep[i].address) == addr &&
+				    be32_to_cpu(lep[i].hashval) == hash)
 					break;
 			}
-			ASSERT(i < INT_GET(btp->count, ARCH_CONVERT));
+			ASSERT(i < be32_to_cpu(btp->count));
 		}
 		p += XFS_DIR2_DATA_ENTSIZE(dep->namelen);
 	}
@@ -154,15 +156,15 @@ xfs_dir2_data_check(
 	 * Need to have seen all the entries and all the bestfree slots.
 	 */
 	ASSERT(freeseen == 7);
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
-		for (i = stale = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-			if (INT_GET(lep[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
+		for (i = stale = 0; i < be32_to_cpu(btp->count); i++) {
+			if (be32_to_cpu(lep[i].address) == XFS_DIR2_NULL_DATAPTR)
 				stale++;
 			if (i > 0)
-				ASSERT(INT_GET(lep[i].hashval, ARCH_CONVERT) >= INT_GET(lep[i - 1].hashval, ARCH_CONVERT));
+				ASSERT(be32_to_cpu(lep[i].hashval) >= be32_to_cpu(lep[i - 1].hashval));
 		}
-		ASSERT(count == INT_GET(btp->count, ARCH_CONVERT) - INT_GET(btp->stale, ARCH_CONVERT));
-		ASSERT(stale == INT_GET(btp->stale, ARCH_CONVERT));
+		ASSERT(count == be32_to_cpu(btp->count) - be32_to_cpu(btp->stale));
+		ASSERT(stale == be32_to_cpu(btp->stale));
 	}
 }
 #endif
@@ -190,8 +192,8 @@ xfs_dir2_data_freefind(
 	 * Check order, non-overlapping entries, and if we find the
 	 * one we're looking for it has to be exact.
 	 */
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	for (dfp = &d->hdr.bestfree[0], seenzero = matched = 0;
 	     dfp < &d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT];
 	     dfp++) {
@@ -201,23 +203,24 @@ xfs_dir2_data_freefind(
 			continue;
 		}
 		ASSERT(seenzero == 0);
-		if (INT_GET(dfp->offset, ARCH_CONVERT) == off) {
+		if (be16_to_cpu(dfp->offset) == off) {
 			matched = 1;
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(dup->length, ARCH_CONVERT));
-		} else if (off < INT_GET(dfp->offset, ARCH_CONVERT))
-			ASSERT(off + INT_GET(dup->length, ARCH_CONVERT) <= INT_GET(dfp->offset, ARCH_CONVERT));
+			ASSERT(dfp->length == dup->length);
+		} else if (off < be16_to_cpu(dfp->offset))
+			ASSERT(off + be16_to_cpu(dup->length) <= be16_to_cpu(dfp->offset));
 		else
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) + INT_GET(dfp->length, ARCH_CONVERT) <= off);
-		ASSERT(matched || INT_GET(dfp->length, ARCH_CONVERT) >= INT_GET(dup->length, ARCH_CONVERT));
+			ASSERT(be16_to_cpu(dfp->offset) + be16_to_cpu(dfp->length) <= off);
+		ASSERT(matched || be16_to_cpu(dfp->length) >= be16_to_cpu(dup->length));
 		if (dfp > &d->hdr.bestfree[0])
-			ASSERT(INT_GET(dfp[-1].length, ARCH_CONVERT) >= INT_GET(dfp[0].length, ARCH_CONVERT));
+			ASSERT(be16_to_cpu(dfp[-1].length) >= be16_to_cpu(dfp[0].length));
 	}
 #endif
 	/*
 	 * If this is smaller than the smallest bestfree entry,
 	 * it can't be there since they're sorted.
 	 */
-	if (INT_GET(dup->length, ARCH_CONVERT) < INT_GET(d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT - 1].length, ARCH_CONVERT))
+	if (be16_to_cpu(dup->length) <
+	    be16_to_cpu(d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT - 1].length))
 		return NULL;
 	/*
 	 * Look at the three bestfree entries for our guy.
@@ -227,7 +230,7 @@ xfs_dir2_data_freefind(
 	     dfp++) {
 		if (!dfp->offset)
 			return NULL;
-		if (INT_GET(dfp->offset, ARCH_CONVERT) == off)
+		if (be16_to_cpu(dfp->offset) == off)
 			return dfp;
 	}
 	/*
@@ -249,29 +252,29 @@ xfs_dir2_data_freeinsert(
 	xfs_dir2_data_free_t	new;		/* new bestfree entry */
 
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	dfp = d->hdr.bestfree;
-	INT_COPY(new.length, dup->length, ARCH_CONVERT);
-	INT_SET(new.offset, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dup - (char *)d));
+	new.length = dup->length;
+	new.offset = cpu_to_be16((char *)dup - (char *)d);
 	/*
 	 * Insert at position 0, 1, or 2; or not at all.
 	 */
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[0].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[0].length)) {
 		dfp[2] = dfp[1];
 		dfp[1] = dfp[0];
 		dfp[0] = new;
 		*loghead = 1;
 		return &dfp[0];
 	}
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[1].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[1].length)) {
 		dfp[2] = dfp[1];
 		dfp[1] = new;
 		*loghead = 1;
 		return &dfp[1];
 	}
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[2].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[2].length)) {
 		dfp[2] = new;
 		*loghead = 1;
 		return &dfp[2];
@@ -289,8 +292,8 @@ xfs_dir2_data_freeremove(
 	int			*loghead)	/* out: log data header */
 {
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	/*
 	 * It's the first entry, slide the next 2 up.
@@ -334,8 +337,8 @@ xfs_dir2_data_freescan(
 	char			*p;		/* current entry pointer */
 
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	/*
 	 * Start by clearing the table.
@@ -348,7 +351,7 @@ xfs_dir2_data_freescan(
 	p = (char *)d->u;
 	if (aendp)
 		endp = aendp;
-	else if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+	else if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		endp = (char *)XFS_DIR2_BLOCK_LEAF_P(btp);
 	} else
@@ -361,11 +364,11 @@ xfs_dir2_data_freescan(
 		/*
 		 * If it's a free entry, insert it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
 			ASSERT((char *)dup - (char *)d ==
-			       INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT));
+			       be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)));
 			xfs_dir2_data_freeinsert(d, dup, loghead);
-			p += INT_GET(dup->length, ARCH_CONVERT);
+			p += be16_to_cpu(dup->length);
 		}
 		/*
 		 * For active entries, check their tags and skip them.
@@ -373,7 +376,7 @@ xfs_dir2_data_freescan(
 		else {
 			dep = (xfs_dir2_data_entry_t *)p;
 			ASSERT((char *)dep - (char *)d ==
-			       INT_GET(*XFS_DIR2_DATA_ENTRY_TAG_P(dep), ARCH_CONVERT));
+			       be16_to_cpu(*XFS_DIR2_DATA_ENTRY_TAG_P(dep)));
 			p += XFS_DIR2_DATA_ENTSIZE(dep->namelen);
 		}
 	}
@@ -415,8 +418,8 @@ xfs_dir2_data_init(
 	 * Initialize the header.
 	 */
 	d = bp->data;
-	INT_SET(d->hdr.magic, ARCH_CONVERT, XFS_DIR2_DATA_MAGIC);
-	INT_SET(d->hdr.bestfree[0].offset, ARCH_CONVERT, (xfs_dir2_data_off_t)sizeof(d->hdr));
+	d->hdr.magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
+	d->hdr.bestfree[0].offset = cpu_to_be16(sizeof(d->hdr));
 	for (i = 1; i < XFS_DIR2_DATA_FD_COUNT; i++) {
 		d->hdr.bestfree[i].length = 0;
 		d->hdr.bestfree[i].offset = 0;
@@ -425,13 +428,12 @@ xfs_dir2_data_init(
 	 * Set up an unused entry for the block's body.
 	 */
 	dup = &d->u[0].unused;
-	INT_SET(dup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
+	dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
 
 	t=mp->m_dirblksize - (uint)sizeof(d->hdr);
-	INT_SET(d->hdr.bestfree[0].length, ARCH_CONVERT, t);
-	INT_SET(dup->length, ARCH_CONVERT, t);
-	INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT,
-		(xfs_dir2_data_off_t)((char *)dup - (char *)d));
+	d->hdr.bestfree[0].length = cpu_to_be16(t);
+	dup->length = cpu_to_be16(t);
+	*XFS_DIR2_DATA_UNUSED_TAG_P(dup) = cpu_to_be16((char *)dup - (char *)d);
 	/*
 	 * Log it and return it.
 	 */
@@ -453,8 +455,8 @@ xfs_dir2_data_log_entry(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)dep - (char *)d),
 		(uint)((char *)(XFS_DIR2_DATA_ENTRY_TAG_P(dep) + 1) -
 		       (char *)d - 1));
@@ -471,8 +473,8 @@ xfs_dir2_data_log_header(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&d->hdr - (char *)d),
 		(uint)(sizeof(d->hdr) - 1));
 }
@@ -489,8 +491,8 @@ xfs_dir2_data_log_unused(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	/*
 	 * Log the first part of the unused entry.
 	 */
@@ -533,12 +535,12 @@ xfs_dir2_data_make_free(
 	/*
 	 * Figure out where the end of the data area is.
 	 */
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC)
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC)
 		endptr = (char *)d + mp->m_dirblksize;
 	else {
 		xfs_dir2_block_tail_t	*btp;	/* block tail */
 
-		ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+		ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		endptr = (char *)XFS_DIR2_BLOCK_LEAF_P(btp);
 	}
@@ -547,11 +549,11 @@ xfs_dir2_data_make_free(
 	 * the previous entry and see if it's free.
 	 */
 	if (offset > sizeof(d->hdr)) {
-		xfs_dir2_data_off_t	*tagp;	/* tag just before us */
+		__be16			*tagp;	/* tag just before us */
 
-		tagp = (xfs_dir2_data_off_t *)((char *)d + offset) - 1;
-		prevdup = (xfs_dir2_data_unused_t *)((char *)d + INT_GET(*tagp, ARCH_CONVERT));
-		if (INT_GET(prevdup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		tagp = (__be16 *)((char *)d + offset) - 1;
+		prevdup = (xfs_dir2_data_unused_t *)((char *)d + be16_to_cpu(*tagp));
+		if (be16_to_cpu(prevdup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			prevdup = NULL;
 	} else
 		prevdup = NULL;
@@ -562,7 +564,7 @@ xfs_dir2_data_make_free(
 	if ((char *)d + offset + len < endptr) {
 		postdup =
 			(xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		if (INT_GET(postdup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		if (be16_to_cpu(postdup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			postdup = NULL;
 	} else
 		postdup = NULL;
@@ -586,13 +588,13 @@ xfs_dir2_data_make_free(
 		 * since the third bestfree is there, there might be more
 		 * entries.
 		 */
-		needscan = d->hdr.bestfree[2].length;
+		needscan = (d->hdr.bestfree[2].length != 0);
 		/*
 		 * Fix up the new big freespace.
 		 */
-		INT_MOD(prevdup->length, ARCH_CONVERT, len + INT_GET(postdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)prevdup - (char *)d));
+		be16_add(&prevdup->length, len + be16_to_cpu(postdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup) =
+			cpu_to_be16((char *)prevdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, prevdup);
 		if (!needscan) {
 			/*
@@ -614,7 +616,7 @@ xfs_dir2_data_make_free(
 			 */
 			dfp = xfs_dir2_data_freeinsert(d, prevdup, needlogp);
 			ASSERT(dfp == &d->hdr.bestfree[0]);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(prevdup->length, ARCH_CONVERT));
+			ASSERT(dfp->length == prevdup->length);
 			ASSERT(!dfp[1].length);
 			ASSERT(!dfp[2].length);
 		}
@@ -624,9 +626,9 @@ xfs_dir2_data_make_free(
 	 */
 	else if (prevdup) {
 		dfp = xfs_dir2_data_freefind(d, prevdup);
-		INT_MOD(prevdup->length, ARCH_CONVERT, len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)prevdup - (char *)d));
+		be16_add(&prevdup->length, len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup) =
+			cpu_to_be16((char *)prevdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, prevdup);
 		/*
 		 * If the previous entry was in the table, the new entry
@@ -640,8 +642,10 @@ xfs_dir2_data_make_free(
 		/*
 		 * Otherwise we need a scan if the new entry is big enough.
 		 */
-		else
-			needscan = INT_GET(prevdup->length, ARCH_CONVERT) > INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT);
+		else {
+			needscan = be16_to_cpu(prevdup->length) >
+				   be16_to_cpu(d->hdr.bestfree[2].length);
+		}
 	}
 	/*
 	 * The following entry is free, merge with it.
@@ -649,10 +653,10 @@ xfs_dir2_data_make_free(
 	else if (postdup) {
 		dfp = xfs_dir2_data_freefind(d, postdup);
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, len + INT_GET(postdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(len + be16_to_cpu(postdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If the following entry was in the table, the new entry
@@ -666,18 +670,20 @@ xfs_dir2_data_make_free(
 		/*
 		 * Otherwise we need a scan if the new entry is big enough.
 		 */
-		else
-			needscan = INT_GET(newdup->length, ARCH_CONVERT) > INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT);
+		else {
+			needscan = be16_to_cpu(newdup->length) >
+				   be16_to_cpu(d->hdr.bestfree[2].length);
+		}
 	}
 	/*
 	 * Neither neighbor is free.  Make a new entry.
 	 */
 	else {
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		(void)xfs_dir2_data_freeinsert(d, newdup, needlogp);
 	}
@@ -707,18 +713,18 @@ xfs_dir2_data_use_free(
 	int			oldlen;		/* old unused entry's length */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
-	ASSERT(INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG);
 	ASSERT(offset >= (char *)dup - (char *)d);
-	ASSERT(offset + len <= (char *)dup + INT_GET(dup->length, ARCH_CONVERT) - (char *)d);
-	ASSERT((char *)dup - (char *)d == INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT));
+	ASSERT(offset + len <= (char *)dup + be16_to_cpu(dup->length) - (char *)d);
+	ASSERT((char *)dup - (char *)d == be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)));
 	/*
 	 * Look up the entry in the bestfree table.
 	 */
 	dfp = xfs_dir2_data_freefind(d, dup);
-	oldlen = INT_GET(dup->length, ARCH_CONVERT);
-	ASSERT(dfp || oldlen <= INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT));
+	oldlen = be16_to_cpu(dup->length);
+	ASSERT(dfp || oldlen <= be16_to_cpu(d->hdr.bestfree[2].length));
 	/*
 	 * Check for alignment with front and back of the entry.
 	 */
@@ -732,7 +738,7 @@ xfs_dir2_data_use_free(
 	 */
 	if (matchfront && matchback) {
 		if (dfp) {
-			needscan = d->hdr.bestfree[2].offset;
+			needscan = (d->hdr.bestfree[2].offset != 0);
 			if (!needscan)
 				xfs_dir2_data_freeremove(d, dfp, needlogp);
 		}
@@ -743,10 +749,10 @@ xfs_dir2_data_use_free(
 	 */
 	else if (matchfront) {
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, oldlen - len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(oldlen - len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
@@ -755,8 +761,8 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(d, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(d, newdup, needlogp);
 			ASSERT(dfp != NULL);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(newdup->length, ARCH_CONVERT));
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) == (char *)newdup - (char *)d);
+			ASSERT(dfp->length == newdup->length);
+			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)d);
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -771,10 +777,9 @@ xfs_dir2_data_use_free(
 	 */
 	else if (matchback) {
 		newdup = dup;
-		INT_SET(newdup->length, ARCH_CONVERT, (xfs_dir2_data_off_t)
-			(((char *)d + offset) - (char *)newdup));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->length = cpu_to_be16(((char *)d + offset) - (char *)newdup);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
@@ -783,8 +788,8 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(d, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(d, newdup, needlogp);
 			ASSERT(dfp != NULL);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(newdup->length, ARCH_CONVERT));
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) == (char *)newdup - (char *)d);
+			ASSERT(dfp->length == newdup->length);
+			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)d);
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -799,16 +804,15 @@ xfs_dir2_data_use_free(
 	 */
 	else {
 		newdup = dup;
-		INT_SET(newdup->length, ARCH_CONVERT, (xfs_dir2_data_off_t)
-			(((char *)d + offset) - (char *)newdup));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->length = cpu_to_be16(((char *)d + offset) - (char *)newdup);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		newdup2 = (xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		INT_SET(newdup2->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup2->length, ARCH_CONVERT, oldlen - len - INT_GET(newdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup2), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup2 - (char *)d));
+		newdup2->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup2->length = cpu_to_be16(oldlen - len - be16_to_cpu(newdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup2) =
+			cpu_to_be16((char *)newdup2 - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup2);
 		/*
 		 * If the old entry was in the table, we need to scan
@@ -819,7 +823,7 @@ xfs_dir2_data_use_free(
 		 * the 2 new will work.
 		 */
 		if (dfp) {
-			needscan = d->hdr.bestfree[2].length;
+			needscan = (d->hdr.bestfree[2].length != 0);
 			if (!needscan) {
 				xfs_dir2_data_freeremove(d, dfp, needlogp);
 				(void)xfs_dir2_data_freeinsert(d, newdup,
diff --git a/fs/xfs/xfs_dir2_data.h b/fs/xfs/xfs_dir2_data.h
index 5e3a7f9ec735..0847cbb53e17 100644
--- a/fs/xfs/xfs_dir2_data.h
+++ b/fs/xfs/xfs_dir2_data.h
@@ -65,8 +65,8 @@ struct xfs_trans;
  * The freespace will be formatted as a xfs_dir2_data_unused_t.
  */
 typedef struct xfs_dir2_data_free {
-	xfs_dir2_data_off_t	offset;		/* start of freespace */
-	xfs_dir2_data_off_t	length;		/* length of freespace */
+	__be16			offset;		/* start of freespace */
+	__be16			length;		/* length of freespace */
 } xfs_dir2_data_free_t;
 
 /*
@@ -75,7 +75,7 @@ typedef struct xfs_dir2_data_free {
  * The code knows that XFS_DIR2_DATA_FD_COUNT is 3.
  */
 typedef struct xfs_dir2_data_hdr {
-	__uint32_t		magic;		/* XFS_DIR2_DATA_MAGIC */
+	__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;
@@ -97,10 +97,10 @@ typedef struct xfs_dir2_data_entry {
  * Tag appears as the last 2 bytes.
  */
 typedef struct xfs_dir2_data_unused {
-	__uint16_t		freetag;	/* XFS_DIR2_DATA_FREE_TAG */
-	xfs_dir2_data_off_t	length;		/* total free length */
+	__be16			freetag;	/* XFS_DIR2_DATA_FREE_TAG */
+	__be16			length;		/* total free length */
 						/* variable offset */
-	xfs_dir2_data_off_t	tag;		/* starting offset of us */
+	__be16			tag;		/* starting offset of us */
 } xfs_dir2_data_unused_t;
 
 typedef union {
@@ -134,12 +134,11 @@ static inline int xfs_dir2_data_entsize(int n)
  * Pointer to an entry's tag word.
  */
 #define	XFS_DIR2_DATA_ENTRY_TAG_P(dep)	xfs_dir2_data_entry_tag_p(dep)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_data_entry_tag_p(xfs_dir2_data_entry_t *dep)
 {
-	return (xfs_dir2_data_off_t *) \
-		 ((char *)(dep) + XFS_DIR2_DATA_ENTSIZE((dep)->namelen) - \
-		  (uint)sizeof(xfs_dir2_data_off_t));
+	return (__be16 *)((char *)dep +
+		XFS_DIR2_DATA_ENTSIZE(dep->namelen) - sizeof(__be16));
 }
 
 /*
@@ -147,12 +146,11 @@ xfs_dir2_data_entry_tag_p(xfs_dir2_data_entry_t *dep)
  */
 #define	XFS_DIR2_DATA_UNUSED_TAG_P(dup) \
 	xfs_dir2_data_unused_tag_p(dup)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_data_unused_tag_p(xfs_dir2_data_unused_t *dup)
 {
-	return (xfs_dir2_data_off_t *) \
-		 ((char *)(dup) + INT_GET((dup)->length, ARCH_CONVERT) \
-				- (uint)sizeof(xfs_dir2_data_off_t));
+	return (__be16 *)((char *)dup +
+			be16_to_cpu(dup->length) - sizeof(__be16));
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c
index d342b6b55239..08648b18265c 100644
--- a/fs/xfs/xfs_dir2_leaf.c
+++ b/fs/xfs/xfs_dir2_leaf.c
@@ -66,7 +66,7 @@ xfs_dir2_block_to_leaf(
 	xfs_da_args_t		*args,		/* operation arguments */
 	xfs_dabuf_t		*dbp)		/* input block's buffer */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf's bestsp entries */
+	__be16			*bestsp;	/* leaf's bestsp entries */
 	xfs_dablk_t		blkno;		/* leaf block's bno */
 	xfs_dir2_block_t	*block;		/* block structure */
 	xfs_dir2_leaf_entry_t	*blp;		/* block's leaf entries */
@@ -111,14 +111,14 @@ xfs_dir2_block_to_leaf(
 	/*
 	 * Set the counts in the leaf header.
 	 */
-	INT_COPY(leaf->hdr.count, btp->count, ARCH_CONVERT); /* INT_: type change */
-	INT_COPY(leaf->hdr.stale, btp->stale, ARCH_CONVERT); /* INT_: type change */
+	leaf->hdr.count = cpu_to_be16(be32_to_cpu(btp->count));
+	leaf->hdr.stale = cpu_to_be16(be32_to_cpu(btp->stale));
 	/*
 	 * Could compact these but I think we always do the conversion
 	 * after squeezing out stale entries.
 	 */
-	memcpy(leaf->ents, blp, INT_GET(btp->count, ARCH_CONVERT) * sizeof(xfs_dir2_leaf_entry_t));
-	xfs_dir2_leaf_log_ents(tp, lbp, 0, INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1);
+	memcpy(leaf->ents, blp, be32_to_cpu(btp->count) * sizeof(xfs_dir2_leaf_entry_t));
+	xfs_dir2_leaf_log_ents(tp, lbp, 0, be16_to_cpu(leaf->hdr.count) - 1);
 	needscan = 0;
 	needlog = 1;
 	/*
@@ -133,7 +133,7 @@ xfs_dir2_block_to_leaf(
 	/*
 	 * Fix up the block header, make it a data block.
 	 */
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_DATA_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
 	if (needscan)
 		xfs_dir2_data_freescan(mp, (xfs_dir2_data_t *)block, &needlog,
 			NULL);
@@ -141,9 +141,9 @@ xfs_dir2_block_to_leaf(
 	 * Set up leaf tail and bests table.
 	 */
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	INT_SET(ltp->bestcount, ARCH_CONVERT, 1);
+	ltp->bestcount = cpu_to_be32(1);
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	INT_COPY(bestsp[0], block->hdr.bestfree[0].length, ARCH_CONVERT);
+	bestsp[0] =  block->hdr.bestfree[0].length;
 	/*
 	 * Log the data header and leaf bests table.
 	 */
@@ -163,7 +163,7 @@ int						/* error */
 xfs_dir2_leaf_addname(
 	xfs_da_args_t		*args)		/* operation arguments */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* freespace table in leaf */
+	__be16			*bestsp;	/* freespace table in leaf */
 	int			compact;	/* need to compact leaves */
 	xfs_dir2_data_t		*data;		/* data block structure */
 	xfs_dabuf_t		*dbp;		/* data block buffer */
@@ -187,7 +187,7 @@ xfs_dir2_leaf_addname(
 	int			needbytes;	/* leaf block bytes needed */
 	int			needlog;	/* need to log data header */
 	int			needscan;	/* need to rescan data free */
-	xfs_dir2_data_off_t	*tagp;		/* end of data entry */
+	__be16			*tagp;		/* end of data entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 	xfs_dir2_db_t		use_block;	/* data block number */
 
@@ -222,14 +222,14 @@ xfs_dir2_leaf_addname(
 	 * in a data block, improving the lookup of those entries.
 	 */
 	for (use_block = -1, lep = &leaf->ents[index];
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     index++, lep++) {
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
-		i = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
-		ASSERT(i < INT_GET(ltp->bestcount, ARCH_CONVERT));
-		ASSERT(INT_GET(bestsp[i], ARCH_CONVERT) != NULLDATAOFF);
-		if (INT_GET(bestsp[i], ARCH_CONVERT) >= length) {
+		i = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
+		ASSERT(i < be32_to_cpu(ltp->bestcount));
+		ASSERT(be16_to_cpu(bestsp[i]) != NULLDATAOFF);
+		if (be16_to_cpu(bestsp[i]) >= length) {
 			use_block = i;
 			break;
 		}
@@ -238,13 +238,13 @@ xfs_dir2_leaf_addname(
 	 * Didn't find a block yet, linear search all the data blocks.
 	 */
 	if (use_block == -1) {
-		for (i = 0; i < INT_GET(ltp->bestcount, ARCH_CONVERT); i++) {
+		for (i = 0; i < be32_to_cpu(ltp->bestcount); i++) {
 			/*
 			 * Remember a block we see that's missing.
 			 */
-			if (INT_GET(bestsp[i], ARCH_CONVERT) == NULLDATAOFF && use_block == -1)
+			if (be16_to_cpu(bestsp[i]) == NULLDATAOFF && use_block == -1)
 				use_block = i;
-			else if (INT_GET(bestsp[i], ARCH_CONVERT) >= length) {
+			else if (be16_to_cpu(bestsp[i]) >= length) {
 				use_block = i;
 				break;
 			}
@@ -260,21 +260,21 @@ xfs_dir2_leaf_addname(
 	 * Now kill use_block if it refers to a missing block, so we
 	 * can use it as an indication of allocation needed.
 	 */
-	if (use_block != -1 && INT_GET(bestsp[use_block], ARCH_CONVERT) == NULLDATAOFF)
+	if (use_block != -1 && be16_to_cpu(bestsp[use_block]) == NULLDATAOFF)
 		use_block = -1;
 	/*
 	 * If we don't have enough free bytes but we can make enough
 	 * by compacting out stale entries, we'll do that.
 	 */
-	if ((char *)bestsp - (char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] < needbytes &&
-	    INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1) {
+	if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] < needbytes &&
+	    be16_to_cpu(leaf->hdr.stale) > 1) {
 		compact = 1;
 	}
 	/*
 	 * Otherwise if we don't have enough free bytes we need to
 	 * convert to node form.
 	 */
-	else if ((char *)bestsp - (char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] <
+	else if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <
 		 needbytes) {
 		/*
 		 * Just checking or no space reservation, give up.
@@ -330,8 +330,8 @@ xfs_dir2_leaf_addname(
 	 * There are stale entries, so we'll need log-low and log-high
 	 * impossibly bad values later.
 	 */
-	else if (INT_GET(leaf->hdr.stale, ARCH_CONVERT)) {
-		lfloglow = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+	else if (be16_to_cpu(leaf->hdr.stale)) {
+		lfloglow = be16_to_cpu(leaf->hdr.count);
 		lfloghigh = -1;
 	}
 	/*
@@ -358,13 +358,13 @@ xfs_dir2_leaf_addname(
 		 * If we're adding a new data block on the end we need to
 		 * extend the bests table.  Copy it up one entry.
 		 */
-		if (use_block >= INT_GET(ltp->bestcount, ARCH_CONVERT)) {
+		if (use_block >= be32_to_cpu(ltp->bestcount)) {
 			bestsp--;
 			memmove(&bestsp[0], &bestsp[1],
-				INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(bestsp[0]));
-			INT_MOD(ltp->bestcount, ARCH_CONVERT, +1);
+				be32_to_cpu(ltp->bestcount) * sizeof(bestsp[0]));
+			be32_add(&ltp->bestcount, 1);
 			xfs_dir2_leaf_log_tail(tp, lbp);
-			xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 		}
 		/*
 		 * If we're filling in a previously empty block just log it.
@@ -372,7 +372,7 @@ xfs_dir2_leaf_addname(
 		else
 			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
 		data = dbp->data;
-		INT_COPY(bestsp[use_block], data->hdr.bestfree[0].length, ARCH_CONVERT);
+		bestsp[use_block] = data->hdr.bestfree[0].length;
 		grown = 1;
 	}
 	/*
@@ -394,8 +394,8 @@ xfs_dir2_leaf_addname(
 	 * Point to the biggest freespace in our data block.
 	 */
 	dup = (xfs_dir2_data_unused_t *)
-	      ((char *)data + INT_GET(data->hdr.bestfree[0].offset, ARCH_CONVERT));
-	ASSERT(INT_GET(dup->length, ARCH_CONVERT) >= length);
+	      ((char *)data + be16_to_cpu(data->hdr.bestfree[0].offset));
+	ASSERT(be16_to_cpu(dup->length) >= length);
 	needscan = needlog = 0;
 	/*
 	 * Mark the initial part of our freespace in use for the new entry.
@@ -411,7 +411,7 @@ xfs_dir2_leaf_addname(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, dep->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)data));
+	*tagp = cpu_to_be16((char *)dep - (char *)data);
 	/*
 	 * Need to scan fix up the bestfree table.
 	 */
@@ -427,8 +427,8 @@ xfs_dir2_leaf_addname(
 	 * If the bests table needs to be changed, do it.
 	 * Log the change unless we've already done that.
 	 */
-	if (INT_GET(bestsp[use_block], ARCH_CONVERT) != INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
-		INT_COPY(bestsp[use_block], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(bestsp[use_block]) != be16_to_cpu(data->hdr.bestfree[0].length)) {
+		bestsp[use_block] = data->hdr.bestfree[0].length;
 		if (!grown)
 			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
 	}
@@ -440,15 +440,15 @@ xfs_dir2_leaf_addname(
 		/*
 		 * lep is still good as the index leaf entry.
 		 */
-		if (index < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+		if (index < be16_to_cpu(leaf->hdr.count))
 			memmove(lep + 1, lep,
-				(INT_GET(leaf->hdr.count, ARCH_CONVERT) - index) * sizeof(*lep));
+				(be16_to_cpu(leaf->hdr.count) - index) * sizeof(*lep));
 		/*
 		 * Record low and high logging indices for the leaf.
 		 */
 		lfloglow = index;
-		lfloghigh = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		INT_MOD(leaf->hdr.count, ARCH_CONVERT, +1);
+		lfloghigh = be16_to_cpu(leaf->hdr.count);
+		be16_add(&leaf->hdr.count, 1);
 	}
 	/*
 	 * There are stale entries.
@@ -468,7 +468,7 @@ xfs_dir2_leaf_addname(
 			 */
 			for (lowstale = index - 1;
 			     lowstale >= 0 &&
-				INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) !=
+				be32_to_cpu(leaf->ents[lowstale].address) !=
 				XFS_DIR2_NULL_DATAPTR;
 			     lowstale--)
 				continue;
@@ -478,8 +478,8 @@ xfs_dir2_leaf_addname(
 			 * lowstale entry would be better.
 			 */
 			for (highstale = index;
-			     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-				INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) !=
+			     highstale < be16_to_cpu(leaf->hdr.count) &&
+				be32_to_cpu(leaf->ents[highstale].address) !=
 				XFS_DIR2_NULL_DATAPTR &&
 				(lowstale < 0 ||
 				 index - lowstale - 1 >= highstale - index);
@@ -490,10 +490,10 @@ xfs_dir2_leaf_addname(
 		 * If the low one is better, use it.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+		    (highstale == be16_to_cpu(leaf->hdr.count) ||
 		     index - lowstale - 1 < highstale - index)) {
 			ASSERT(index - lowstale - 1 >= 0);
-			ASSERT(INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[lowstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			/*
 			 * Copy entries up to cover the stale entry
@@ -512,7 +512,7 @@ xfs_dir2_leaf_addname(
 		 */
 		else {
 			ASSERT(highstale - index >= 0);
-			ASSERT(INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[highstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			/*
 			 * Copy entries down to copver the stale entry
@@ -526,13 +526,14 @@ xfs_dir2_leaf_addname(
 			lfloglow = MIN(index, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(leaf->hdr.stale, ARCH_CONVERT, -1);
+		be16_add(&leaf->hdr.stale, -1);
 	}
 	/*
 	 * Fill in the new leaf entry.
 	 */
-	INT_SET(lep->hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_DB_OFF_TO_DATAPTR(mp, use_block, INT_GET(*tagp, ARCH_CONVERT)));
+	lep->hashval = cpu_to_be32(args->hashval);
+	lep->address = cpu_to_be32(XFS_DIR2_DB_OFF_TO_DATAPTR(mp, use_block,
+				be16_to_cpu(*tagp)));
 	/*
 	 * Log the leaf fields and give up the buffers.
 	 */
@@ -563,30 +564,30 @@ xfs_dir2_leaf_check(
 
 	leaf = bp->data;
 	mp = dp->i_mount;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	/*
 	 * This value is not restrictive enough.
 	 * Should factor in the size of the bests table as well.
 	 * We can deduce a value for that from di_size.
 	 */
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
+	ASSERT(be16_to_cpu(leaf->hdr.count) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	/*
 	 * Leaves and bests don't overlap.
 	 */
-	ASSERT((char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] <=
+	ASSERT((char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <=
 	       (char *)XFS_DIR2_LEAF_BESTS_P(ltp));
 	/*
 	 * Check hash value order, count stale entries.
 	 */
-	for (i = stale = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); i++) {
-		if (i + 1 < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-			ASSERT(INT_GET(leaf->ents[i].hashval, ARCH_CONVERT) <=
-			       INT_GET(leaf->ents[i + 1].hashval, ARCH_CONVERT));
-		if (INT_GET(leaf->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (i = stale = 0; i < be16_to_cpu(leaf->hdr.count); i++) {
+		if (i + 1 < be16_to_cpu(leaf->hdr.count))
+			ASSERT(be32_to_cpu(leaf->ents[i].hashval) <=
+			       be32_to_cpu(leaf->ents[i + 1].hashval));
+		if (be32_to_cpu(leaf->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 			stale++;
 	}
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == stale);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == stale);
 }
 #endif	/* DEBUG */
 
@@ -611,8 +612,8 @@ xfs_dir2_leaf_compact(
 	/*
 	 * Compress out the stale entries in place.
 	 */
-	for (from = to = 0, loglow = -1; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
-		if (INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (from = to = 0, loglow = -1; from < be16_to_cpu(leaf->hdr.count); from++) {
+		if (be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Only actually copy the entries that are different.
@@ -627,8 +628,8 @@ xfs_dir2_leaf_compact(
 	/*
 	 * Update and log the header, log the leaf entries.
 	 */
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == from - to);
-	INT_MOD(leaf->hdr.count, ARCH_CONVERT, -(INT_GET(leaf->hdr.stale, ARCH_CONVERT)));
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == from - to);
+	be16_add(&leaf->hdr.count, -(be16_to_cpu(leaf->hdr.stale)));
 	leaf->hdr.stale = 0;
 	xfs_dir2_leaf_log_header(args->trans, bp);
 	if (loglow != -1)
@@ -662,14 +663,14 @@ xfs_dir2_leaf_compact_x1(
 	int		to;		/* destination copy index */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) > 1);
 	index = *indexp;
 	/*
 	 * Find the first stale entry before our index, if any.
 	 */
 	for (lowstale = index - 1;
 	     lowstale >= 0 &&
-		INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR;
+		be32_to_cpu(leaf->ents[lowstale].address) != XFS_DIR2_NULL_DATAPTR;
 	     lowstale--)
 		continue;
 	/*
@@ -677,8 +678,8 @@ xfs_dir2_leaf_compact_x1(
 	 * Stop if the answer would be worse than lowstale.
 	 */
 	for (highstale = index;
-	     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-		INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR &&
+	     highstale < be16_to_cpu(leaf->hdr.count) &&
+		be32_to_cpu(leaf->ents[highstale].address) != XFS_DIR2_NULL_DATAPTR &&
 		(lowstale < 0 || index - lowstale > highstale - index);
 	     highstale++)
 		continue;
@@ -686,7 +687,7 @@ xfs_dir2_leaf_compact_x1(
 	 * Pick the better of lowstale and highstale.
 	 */
 	if (lowstale >= 0 &&
-	    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+	    (highstale == be16_to_cpu(leaf->hdr.count) ||
 	     index - lowstale <= highstale - index))
 		keepstale = lowstale;
 	else
@@ -695,14 +696,14 @@ xfs_dir2_leaf_compact_x1(
 	 * Copy the entries in place, removing all the stale entries
 	 * except keepstale.
 	 */
-	for (from = to = 0; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
+	for (from = to = 0; from < be16_to_cpu(leaf->hdr.count); from++) {
 		/*
 		 * Notice the new value of index.
 		 */
 		if (index == from)
 			newindex = to;
 		if (from != keepstale &&
-		    INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR) {
+		    be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR) {
 			if (from == to)
 				*lowlogp = to;
 			continue;
@@ -730,8 +731,8 @@ xfs_dir2_leaf_compact_x1(
 	/*
 	 * Adjust the leaf header values.
 	 */
-	INT_MOD(leaf->hdr.count, ARCH_CONVERT, -(from - to));
-	INT_SET(leaf->hdr.stale, ARCH_CONVERT, 1);
+	be16_add(&leaf->hdr.count, -(from - to));
+	leaf->hdr.stale = cpu_to_be16(1);
 	/*
 	 * Remember the low/high stale value only in the "right"
 	 * direction.
@@ -739,8 +740,8 @@ xfs_dir2_leaf_compact_x1(
 	if (lowstale >= newindex)
 		lowstale = -1;
 	else
-		highstale = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-	*highlogp = INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1;
+		highstale = be16_to_cpu(leaf->hdr.count);
+	*highlogp = be16_to_cpu(leaf->hdr.count) - 1;
 	*lowstalep = lowstale;
 	*highstalep = highstale;
 }
@@ -766,7 +767,7 @@ xfs_dir2_leaf_getdents(
 	xfs_dir2_data_entry_t	*dep;		/* data entry */
 	xfs_dir2_data_unused_t	*dup;		/* unused entry */
 	int			eof;		/* reached end of directory */
-	int			error=0;		/* error return value */
+	int			error = 0;	/* error return value */
 	int			i;		/* temporary loop index */
 	int			j;		/* temporary loop index */
 	int			length;		/* temporary length value */
@@ -778,8 +779,8 @@ xfs_dir2_leaf_getdents(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	xfs_dir2_off_t		newoff;		/* new curoff after new blk */
 	int			nmap;		/* mappings to ask xfs_bmapi */
-	xfs_dir2_put_args_t	p;		/* formatting arg bundle */
-	char			*ptr=NULL;		/* pointer to current data */
+	xfs_dir2_put_args_t	*p;		/* formatting arg bundle */
+	char			*ptr = NULL;	/* pointer to current data */
 	int			ra_current;	/* number of read-ahead blks */
 	int			ra_index;	/* *map index for read-ahead */
 	int			ra_offset;	/* map entry offset for ra */
@@ -797,9 +798,10 @@ xfs_dir2_leaf_getdents(
 	/*
 	 * Setup formatting arguments.
 	 */
-	p.dbp = dbp;
-	p.put = put;
-	p.uio = uio;
+	p = kmem_alloc(sizeof(*p), KM_SLEEP);
+	p->dbp = dbp;
+	p->put = put;
+	p->uio = uio;
 	/*
 	 * Set up to bmap a number of blocks based on the caller's
 	 * buffer size, the directory block size, and the filesystem
@@ -1046,11 +1048,10 @@ xfs_dir2_leaf_getdents(
 				while ((char *)ptr - (char *)data < byteoff) {
 					dup = (xfs_dir2_data_unused_t *)ptr;
 
-					if (INT_GET(dup->freetag, ARCH_CONVERT)
+					if (be16_to_cpu(dup->freetag)
 						  == XFS_DIR2_DATA_FREE_TAG) {
 
-						length = INT_GET(dup->length,
-								 ARCH_CONVERT);
+						length = be16_to_cpu(dup->length);
 						ptr += length;
 						continue;
 					}
@@ -1079,9 +1080,8 @@ xfs_dir2_leaf_getdents(
 		/*
 		 * No, it's unused, skip over it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT)
-						== XFS_DIR2_DATA_FREE_TAG) {
-			length = INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			length = be16_to_cpu(dup->length);
 			ptr += length;
 			curoff += length;
 			continue;
@@ -1092,24 +1092,24 @@ xfs_dir2_leaf_getdents(
 		 */
 		dep = (xfs_dir2_data_entry_t *)ptr;
 
-		p.namelen = dep->namelen;
+		p->namelen = dep->namelen;
 
-		length = XFS_DIR2_DATA_ENTSIZE(p.namelen);
+		length = XFS_DIR2_DATA_ENTSIZE(p->namelen);
 
-		p.cook = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff + length);
+		p->cook = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff + length);
 
-		p.ino = INT_GET(dep->inumber, ARCH_CONVERT);
+		p->ino = INT_GET(dep->inumber, ARCH_CONVERT);
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		p->ino += mp->m_inoadd;
 #endif
-		p.name = (char *)dep->name;
+		p->name = (char *)dep->name;
 
-		error = p.put(&p);
+		error = p->put(p);
 
 		/*
 		 * Won't fit.  Return to caller.
 		 */
-		if (!p.done) {
+		if (!p->done) {
 			eof = 0;
 			break;
 		}
@@ -1129,6 +1129,7 @@ xfs_dir2_leaf_getdents(
 	else
 		uio->uio_offset = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff);
 	kmem_free(map, map_size * sizeof(*map));
+	kmem_free(p, sizeof(*p));
 	if (bp)
 		xfs_da_brelse(tp, bp);
 	return error;
@@ -1171,7 +1172,7 @@ xfs_dir2_leaf_init(
 	/*
 	 * Initialize the header.
 	 */
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, magic);
+	leaf->hdr.info.magic = cpu_to_be16(magic);
 	leaf->hdr.info.forw = 0;
 	leaf->hdr.info.back = 0;
 	leaf->hdr.count = 0;
@@ -1201,13 +1202,13 @@ xfs_dir2_leaf_log_bests(
 	int			first,		/* first entry to log */
 	int			last)		/* last entry to log */
 {
-	xfs_dir2_data_off_t	*firstb;	/* pointer to first entry */
-	xfs_dir2_data_off_t	*lastb;		/* pointer to last entry */
+	__be16			*firstb;	/* pointer to first entry */
+	__be16			*lastb;		/* pointer to last entry */
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(tp->t_mountp, leaf);
 	firstb = XFS_DIR2_LEAF_BESTS_P(ltp) + first;
 	lastb = XFS_DIR2_LEAF_BESTS_P(ltp) + last;
@@ -1230,8 +1231,8 @@ xfs_dir2_leaf_log_ents(
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC ||
-	       INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
+	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	firstlep = &leaf->ents[first];
 	lastlep = &leaf->ents[last];
 	xfs_da_log_buf(tp, bp, (uint)((char *)firstlep - (char *)leaf),
@@ -1249,8 +1250,8 @@ xfs_dir2_leaf_log_header(
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC ||
-	       INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
+	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&leaf->hdr - (char *)leaf),
 		(uint)(sizeof(leaf->hdr) - 1));
 }
@@ -1269,7 +1270,7 @@ xfs_dir2_leaf_log_tail(
 
 	mp = tp->t_mountp;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	xfs_da_log_buf(tp, bp, (uint)((char *)ltp - (char *)leaf),
 		(uint)(mp->m_dirblksize - 1));
@@ -1314,7 +1315,7 @@ xfs_dir2_leaf_lookup(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->data +
-	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, INT_GET(lep->address, ARCH_CONVERT)));
+	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
 	/*
 	 * Return the found inode number.
 	 */
@@ -1373,17 +1374,17 @@ xfs_dir2_leaf_lookup_int(
 	 * looking to match the name.
 	 */
 	for (lep = &leaf->ents[index], dbp = NULL, curdb = -1;
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     lep++, index++) {
 		/*
 		 * Skip over stale leaf entries.
 		 */
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Get the new data block number.
 		 */
-		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 		/*
 		 * If it's not the same as the old data block number,
 		 * need to pitch the old one and read the new one.
@@ -1406,7 +1407,7 @@ xfs_dir2_leaf_lookup_int(
 		 */
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)dbp->data +
-		       XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+		       XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 		/*
 		 * If it matches then return it.
 		 */
@@ -1435,7 +1436,7 @@ int						/* error */
 xfs_dir2_leaf_removename(
 	xfs_da_args_t		*args)		/* operation arguments */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf block best freespace */
+	__be16			*bestsp;	/* leaf block best freespace */
 	xfs_dir2_data_t		*data;		/* data block structure */
 	xfs_dir2_db_t		db;		/* data block number */
 	xfs_dabuf_t		*dbp;		/* data block buffer */
@@ -1471,14 +1472,14 @@ xfs_dir2_leaf_removename(
 	 * Point to the leaf entry, use that to point to the data entry.
 	 */
 	lep = &leaf->ents[index];
-	db = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+	db = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)data + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+	      ((char *)data + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 	needscan = needlog = 0;
-	oldbest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+	oldbest = be16_to_cpu(data->hdr.bestfree[0].length);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	ASSERT(INT_GET(bestsp[db], ARCH_CONVERT) == oldbest);
+	ASSERT(be16_to_cpu(bestsp[db]) == oldbest);
 	/*
 	 * Mark the former data entry unused.
 	 */
@@ -1488,9 +1489,9 @@ xfs_dir2_leaf_removename(
 	/*
 	 * We just mark the leaf entry stale by putting a null in it.
 	 */
-	INT_MOD(leaf->hdr.stale, ARCH_CONVERT, +1);
+	be16_add(&leaf->hdr.stale, 1);
 	xfs_dir2_leaf_log_header(tp, lbp);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_leaf_log_ents(tp, lbp, index, index);
 	/*
 	 * Scan the freespace in the data block again if necessary,
@@ -1504,15 +1505,15 @@ xfs_dir2_leaf_removename(
 	 * If the longest freespace in the data block has changed,
 	 * put the new value in the bests table and log that.
 	 */
-	if (INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) != oldbest) {
-		INT_COPY(bestsp[db], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(data->hdr.bestfree[0].length) != oldbest) {
+		bestsp[db] = data->hdr.bestfree[0].length;
 		xfs_dir2_leaf_log_bests(tp, lbp, db, db);
 	}
 	xfs_dir2_data_check(dp, dbp);
 	/*
 	 * If the data block is now empty then get rid of the data block.
 	 */
-	if (INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) ==
+	if (be16_to_cpu(data->hdr.bestfree[0].length) ==
 	    mp->m_dirblksize - (uint)sizeof(data->hdr)) {
 		ASSERT(db != mp->m_dirdatablk);
 		if ((error = xfs_dir2_shrink_inode(args, db, dbp))) {
@@ -1535,12 +1536,12 @@ xfs_dir2_leaf_removename(
 		 * If this is the last data block then compact the
 		 * bests table by getting rid of entries.
 		 */
-		if (db == INT_GET(ltp->bestcount, ARCH_CONVERT) - 1) {
+		if (db == be32_to_cpu(ltp->bestcount) - 1) {
 			/*
 			 * Look for the last active entry (i).
 			 */
 			for (i = db - 1; i > 0; i--) {
-				if (INT_GET(bestsp[i], ARCH_CONVERT) != NULLDATAOFF)
+				if (be16_to_cpu(bestsp[i]) != NULLDATAOFF)
 					break;
 			}
 			/*
@@ -1548,12 +1549,12 @@ xfs_dir2_leaf_removename(
 			 * end are removed.
 			 */
 			memmove(&bestsp[db - i], bestsp,
-				(INT_GET(ltp->bestcount, ARCH_CONVERT) - (db - i)) * sizeof(*bestsp));
-			INT_MOD(ltp->bestcount, ARCH_CONVERT, -(db - i));
+				(be32_to_cpu(ltp->bestcount) - (db - i)) * sizeof(*bestsp));
+			be32_add(&ltp->bestcount, -(db - i));
 			xfs_dir2_leaf_log_tail(tp, lbp);
-			xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 		} else
-			INT_SET(bestsp[db], ARCH_CONVERT, NULLDATAOFF);
+			bestsp[db] = cpu_to_be16(NULLDATAOFF);
 	}
 	/*
 	 * If the data block was not the first one, drop it.
@@ -1604,7 +1605,7 @@ xfs_dir2_leaf_replace(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->data +
-	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, INT_GET(lep->address, ARCH_CONVERT)));
+	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
 	ASSERT(args->inumber != INT_GET(dep->inumber, ARCH_CONVERT));
 	/*
 	 * Put the new inode number in, log it.
@@ -1645,11 +1646,11 @@ xfs_dir2_leaf_search_hash(
 	 * Note, the table cannot be empty, so we have to go through the loop.
 	 * Binary search the leaf entries looking for our hash value.
 	 */
-	for (lep = leaf->ents, low = 0, high = INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1,
+	for (lep = leaf->ents, low = 0, high = be16_to_cpu(leaf->hdr.count) - 1,
 		hashwant = args->hashval;
 	     low <= high; ) {
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(lep[mid].hashval, ARCH_CONVERT)) == hashwant)
+		if ((hash = be32_to_cpu(lep[mid].hashval)) == hashwant)
 			break;
 		if (hash < hashwant)
 			low = mid + 1;
@@ -1660,7 +1661,7 @@ xfs_dir2_leaf_search_hash(
 	 * Found one, back up through all the equal hash values.
 	 */
 	if (hash == hashwant) {
-		while (mid > 0 && INT_GET(lep[mid - 1].hashval, ARCH_CONVERT) == hashwant) {
+		while (mid > 0 && be32_to_cpu(lep[mid - 1].hashval) == hashwant) {
 			mid--;
 		}
 	}
@@ -1682,7 +1683,7 @@ xfs_dir2_leaf_trim_data(
 	xfs_dabuf_t		*lbp,		/* leaf buffer */
 	xfs_dir2_db_t		db)		/* data block number */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf bests table */
+	__be16			*bestsp;	/* leaf bests table */
 #ifdef DEBUG
 	xfs_dir2_data_t		*data;		/* data block structure */
 #endif
@@ -1706,7 +1707,7 @@ xfs_dir2_leaf_trim_data(
 	}
 #ifdef DEBUG
 	data = dbp->data;
-	ASSERT(INT_GET(data->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+	ASSERT(be32_to_cpu(data->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 #endif
 	/* this seems to be an error
 	 * data is only valid if DEBUG is defined?
@@ -1715,9 +1716,9 @@ xfs_dir2_leaf_trim_data(
 
 	leaf = lbp->data;
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	ASSERT(INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) ==
+	ASSERT(be16_to_cpu(data->hdr.bestfree[0].length) ==
 	       mp->m_dirblksize - (uint)sizeof(data->hdr));
-	ASSERT(db == INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+	ASSERT(db == be32_to_cpu(ltp->bestcount) - 1);
 	/*
 	 * Get rid of the data block.
 	 */
@@ -1730,10 +1731,10 @@ xfs_dir2_leaf_trim_data(
 	 * Eliminate the last bests entry from the table.
 	 */
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	INT_MOD(ltp->bestcount, ARCH_CONVERT, -1);
-	memmove(&bestsp[1], &bestsp[0], INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(*bestsp));
+	be32_add(&ltp->bestcount, -1);
+	memmove(&bestsp[1], &bestsp[0], be32_to_cpu(ltp->bestcount) * sizeof(*bestsp));
 	xfs_dir2_leaf_log_tail(tp, lbp);
-	xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 	return 0;
 }
 
@@ -1805,7 +1806,7 @@ xfs_dir2_node_to_leaf(
 		return 0;
 	lbp = state->path.blk[0].bp;
 	leaf = lbp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Read the freespace block.
 	 */
@@ -1814,15 +1815,15 @@ xfs_dir2_node_to_leaf(
 		return error;
 	}
 	free = fbp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	ASSERT(!free->hdr.firstdb);
 	/*
 	 * Now see if the leafn and free data will fit in a leaf1.
 	 * If not, release the buffer and give up.
 	 */
 	if ((uint)sizeof(leaf->hdr) +
-	    (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT)) * (uint)sizeof(leaf->ents[0]) +
-	    INT_GET(free->hdr.nvalid, ARCH_CONVERT) * (uint)sizeof(leaf->bests[0]) +
+	    (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale)) * (uint)sizeof(leaf->ents[0]) +
+	    be32_to_cpu(free->hdr.nvalid) * (uint)sizeof(leaf->bests[0]) +
 	    (uint)sizeof(leaf->tail) >
 	    mp->m_dirblksize) {
 		xfs_da_brelse(tp, fbp);
@@ -1832,22 +1833,22 @@ xfs_dir2_node_to_leaf(
 	 * If the leaf has any stale entries in it, compress them out.
 	 * The compact routine will log the header.
 	 */
-	if (INT_GET(leaf->hdr.stale, ARCH_CONVERT))
+	if (be16_to_cpu(leaf->hdr.stale))
 		xfs_dir2_leaf_compact(args, lbp);
 	else
 		xfs_dir2_leaf_log_header(tp, lbp);
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, XFS_DIR2_LEAF1_MAGIC);
+	leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAF1_MAGIC);
 	/*
 	 * Set up the leaf tail from the freespace block.
 	 */
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	INT_COPY(ltp->bestcount, free->hdr.nvalid, ARCH_CONVERT);
+	ltp->bestcount = free->hdr.nvalid;
 	/*
 	 * Set up the leaf bests table.
 	 */
 	memcpy(XFS_DIR2_LEAF_BESTS_P(ltp), free->bests,
-		INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(leaf->bests[0]));
-	xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+		be32_to_cpu(ltp->bestcount) * sizeof(leaf->bests[0]));
+	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 	xfs_dir2_leaf_log_tail(tp, lbp);
 	xfs_dir2_leaf_check(dp, lbp);
 	/*
diff --git a/fs/xfs/xfs_dir2_leaf.h b/fs/xfs/xfs_dir2_leaf.h
index 1393993d61e9..f57ca1162412 100644
--- a/fs/xfs/xfs_dir2_leaf.h
+++ b/fs/xfs/xfs_dir2_leaf.h
@@ -46,23 +46,23 @@ typedef	__uint32_t	xfs_dir2_dataptr_t;
  */
 typedef struct xfs_dir2_leaf_hdr {
 	xfs_da_blkinfo_t	info;		/* header for da routines */
-	__uint16_t		count;		/* count of entries */
-	__uint16_t		stale;		/* count of stale entries */
+	__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 {
-	xfs_dahash_t		hashval;	/* hash value of name */
-	xfs_dir2_dataptr_t	address;	/* address of data 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 {
-	__uint32_t		bestcount;
+	__be32			bestcount;
 } xfs_dir2_leaf_tail_t;
 
 /*
@@ -105,11 +105,10 @@ xfs_dir2_leaf_tail_p(struct xfs_mount *mp, xfs_dir2_leaf_t *lp)
  * Get address of the bests array in the single-leaf block.
  */
 #define	XFS_DIR2_LEAF_BESTS_P(ltp)	xfs_dir2_leaf_bests_p(ltp)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_leaf_bests_p(xfs_dir2_leaf_tail_t *ltp)
 {
-	return (xfs_dir2_data_off_t *)
-		(ltp) - INT_GET((ltp)->bestcount, ARCH_CONVERT);
+	return (__be16 *)ltp - be32_to_cpu(ltp->bestcount);
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c
index 641f8633d254..af556f16a0c7 100644
--- a/fs/xfs/xfs_dir2_node.c
+++ b/fs/xfs/xfs_dir2_node.c
@@ -76,7 +76,7 @@ xfs_dir2_free_log_bests(
 	xfs_dir2_free_t		*free;		/* freespace structure */
 
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	xfs_da_log_buf(tp, bp,
 		(uint)((char *)&free->bests[first] - (char *)free),
 		(uint)((char *)&free->bests[last] - (char *)free +
@@ -94,7 +94,7 @@ xfs_dir2_free_log_header(
 	xfs_dir2_free_t		*free;		/* freespace structure */
 
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&free->hdr - (char *)free),
 		(uint)(sizeof(xfs_dir2_free_hdr_t) - 1));
 }
@@ -114,14 +114,14 @@ xfs_dir2_leaf_to_node(
 	xfs_dabuf_t		*fbp;		/* freespace buffer */
 	xfs_dir2_db_t		fdb;		/* freespace block number */
 	xfs_dir2_free_t		*free;		/* freespace structure */
-	xfs_dir2_data_off_t	*from;		/* pointer to freespace entry */
+	__be16			*from;		/* pointer to freespace entry */
 	int			i;		/* leaf freespace index */
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			n;		/* count of live freespc ents */
 	xfs_dir2_data_off_t	off;		/* freespace entry value */
-	xfs_dir2_data_off_t	*to;		/* pointer to freespace entry */
+	__be16			*to;		/* pointer to freespace entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	xfs_dir2_trace_args_b("leaf_to_node", args, lbp);
@@ -149,28 +149,28 @@ xfs_dir2_leaf_to_node(
 	/*
 	 * Initialize the freespace block header.
 	 */
-	INT_SET(free->hdr.magic, ARCH_CONVERT, XFS_DIR2_FREE_MAGIC);
+	free->hdr.magic = cpu_to_be32(XFS_DIR2_FREE_MAGIC);
 	free->hdr.firstdb = 0;
-	ASSERT(INT_GET(ltp->bestcount, ARCH_CONVERT) <= (uint)dp->i_d.di_size / mp->m_dirblksize);
-	INT_COPY(free->hdr.nvalid, ltp->bestcount, ARCH_CONVERT);
+	ASSERT(be32_to_cpu(ltp->bestcount) <= (uint)dp->i_d.di_size / mp->m_dirblksize);
+	free->hdr.nvalid = ltp->bestcount;
 	/*
 	 * Copy freespace entries from the leaf block to the new block.
 	 * Count active entries.
 	 */
 	for (i = n = 0, from = XFS_DIR2_LEAF_BESTS_P(ltp), to = free->bests;
-	     i < INT_GET(ltp->bestcount, ARCH_CONVERT); i++, from++, to++) {
-		if ((off = INT_GET(*from, ARCH_CONVERT)) != NULLDATAOFF)
+	     i < be32_to_cpu(ltp->bestcount); i++, from++, to++) {
+		if ((off = be16_to_cpu(*from)) != NULLDATAOFF)
 			n++;
-		INT_SET(*to, ARCH_CONVERT, off);
+		*to = cpu_to_be16(off);
 	}
-	INT_SET(free->hdr.nused, ARCH_CONVERT, n);
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, XFS_DIR2_LEAFN_MAGIC);
+	free->hdr.nused = cpu_to_be32(n);
+	leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Log everything.
 	 */
 	xfs_dir2_leaf_log_header(tp, lbp);
 	xfs_dir2_free_log_header(tp, fbp);
-	xfs_dir2_free_log_bests(tp, fbp, 0, INT_GET(free->hdr.nvalid, ARCH_CONVERT) - 1);
+	xfs_dir2_free_log_bests(tp, fbp, 0, be32_to_cpu(free->hdr.nvalid) - 1);
 	xfs_da_buf_done(fbp);
 	xfs_dir2_leafn_check(dp, lbp);
 	return 0;
@@ -217,15 +217,15 @@ xfs_dir2_leafn_add(
 	 * a compact.
 	 */
 
-	if (INT_GET(leaf->hdr.count, ARCH_CONVERT) == XFS_DIR2_MAX_LEAF_ENTS(mp)) {
+	if (be16_to_cpu(leaf->hdr.count) == XFS_DIR2_MAX_LEAF_ENTS(mp)) {
 		if (!leaf->hdr.stale)
 			return XFS_ERROR(ENOSPC);
-		compact = INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1;
+		compact = be16_to_cpu(leaf->hdr.stale) > 1;
 	} else
 		compact = 0;
-	ASSERT(index == 0 || INT_GET(leaf->ents[index - 1].hashval, ARCH_CONVERT) <= args->hashval);
-	ASSERT(index == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
-	       INT_GET(leaf->ents[index].hashval, ARCH_CONVERT) >= args->hashval);
+	ASSERT(index == 0 || be32_to_cpu(leaf->ents[index - 1].hashval) <= args->hashval);
+	ASSERT(index == be16_to_cpu(leaf->hdr.count) ||
+	       be32_to_cpu(leaf->ents[index].hashval) >= args->hashval);
 
 	if (args->justcheck)
 		return 0;
@@ -242,7 +242,7 @@ xfs_dir2_leafn_add(
 	 * Set impossible logging indices for this case.
 	 */
 	else if (leaf->hdr.stale) {
-		lfloglow = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		lfloglow = be16_to_cpu(leaf->hdr.count);
 		lfloghigh = -1;
 	}
 	/*
@@ -250,12 +250,12 @@ xfs_dir2_leafn_add(
 	 */
 	if (!leaf->hdr.stale) {
 		lep = &leaf->ents[index];
-		if (index < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+		if (index < be16_to_cpu(leaf->hdr.count))
 			memmove(lep + 1, lep,
-				(INT_GET(leaf->hdr.count, ARCH_CONVERT) - index) * sizeof(*lep));
+				(be16_to_cpu(leaf->hdr.count) - index) * sizeof(*lep));
 		lfloglow = index;
-		lfloghigh = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		INT_MOD(leaf->hdr.count, ARCH_CONVERT, +1);
+		lfloghigh = be16_to_cpu(leaf->hdr.count);
+		be16_add(&leaf->hdr.count, 1);
 	}
 	/*
 	 * There are stale entries.  We'll use one for the new entry.
@@ -271,7 +271,7 @@ xfs_dir2_leafn_add(
 			 */
 			for (lowstale = index - 1;
 			     lowstale >= 0 &&
-				INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) !=
+				be32_to_cpu(leaf->ents[lowstale].address) !=
 				XFS_DIR2_NULL_DATAPTR;
 			     lowstale--)
 				continue;
@@ -281,8 +281,8 @@ xfs_dir2_leafn_add(
 			 * lowstale already found.
 			 */
 			for (highstale = index;
-			     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-				INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) !=
+			     highstale < be16_to_cpu(leaf->hdr.count) &&
+				be32_to_cpu(leaf->ents[highstale].address) !=
 				XFS_DIR2_NULL_DATAPTR &&
 				(lowstale < 0 ||
 				 index - lowstale - 1 >= highstale - index);
@@ -294,9 +294,9 @@ xfs_dir2_leafn_add(
 		 * Shift entries up toward the stale slot.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+		    (highstale == be16_to_cpu(leaf->hdr.count) ||
 		     index - lowstale - 1 < highstale - index)) {
-			ASSERT(INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[lowstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			ASSERT(index - lowstale - 1 >= 0);
 			if (index - lowstale - 1 > 0)
@@ -312,7 +312,7 @@ xfs_dir2_leafn_add(
 		 * Shift entries down toward the stale slot.
 		 */
 		else {
-			ASSERT(INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[highstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			ASSERT(highstale - index >= 0);
 			if (highstale - index > 0)
@@ -323,13 +323,14 @@ xfs_dir2_leafn_add(
 			lfloglow = MIN(index, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(leaf->hdr.stale, ARCH_CONVERT, -1);
+		be16_add(&leaf->hdr.stale, -1);
 	}
 	/*
 	 * Insert the new entry, log everything.
 	 */
-	INT_SET(lep->hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_DB_OFF_TO_DATAPTR(mp, args->blkno, args->index));
+	lep->hashval = cpu_to_be32(args->hashval);
+	lep->address = cpu_to_be32(XFS_DIR2_DB_OFF_TO_DATAPTR(mp,
+				args->blkno, args->index));
 	xfs_dir2_leaf_log_header(tp, bp);
 	xfs_dir2_leaf_log_ents(tp, bp, lfloglow, lfloghigh);
 	xfs_dir2_leafn_check(dp, bp);
@@ -352,17 +353,17 @@ xfs_dir2_leafn_check(
 
 	leaf = bp->data;
 	mp = dp->i_mount;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
-	for (i = stale = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); i++) {
-		if (i + 1 < INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
-			ASSERT(INT_GET(leaf->ents[i].hashval, ARCH_CONVERT) <=
-			       INT_GET(leaf->ents[i + 1].hashval, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
+	for (i = stale = 0; i < be16_to_cpu(leaf->hdr.count); i++) {
+		if (i + 1 < be16_to_cpu(leaf->hdr.count)) {
+			ASSERT(be32_to_cpu(leaf->ents[i].hashval) <=
+			       be32_to_cpu(leaf->ents[i + 1].hashval));
 		}
-		if (INT_GET(leaf->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(leaf->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 			stale++;
 	}
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == stale);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == stale);
 }
 #endif	/* DEBUG */
 
@@ -378,12 +379,12 @@ xfs_dir2_leafn_lasthash(
 	xfs_dir2_leaf_t	*leaf;			/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	if (count)
-		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(leaf->hdr.count);
 	if (!leaf->hdr.count)
 		return 0;
-	return INT_GET(leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+	return be32_to_cpu(leaf->ents[be16_to_cpu(leaf->hdr.count) - 1].hashval);
 }
 
 /*
@@ -419,9 +420,9 @@ xfs_dir2_leafn_lookup_int(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 #ifdef __KERNEL__
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) > 0);
+	ASSERT(be16_to_cpu(leaf->hdr.count) > 0);
 #endif
 	xfs_dir2_leafn_check(dp, bp);
 	/*
@@ -443,7 +444,7 @@ xfs_dir2_leafn_lookup_int(
 		curdb = -1;
 		length = XFS_DIR2_DATA_ENTSIZE(args->namelen);
 		if ((free = (curbp ? curbp->data : NULL)))
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	}
 	/*
 	 * For others, it's a data block buffer, get the block number.
@@ -456,17 +457,17 @@ xfs_dir2_leafn_lookup_int(
 	 * Loop over leaf entries with the right hash value.
 	 */
 	for (lep = &leaf->ents[index];
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     lep++, index++) {
 		/*
 		 * Skip stale leaf entries.
 		 */
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Pull the data block number from the entry.
 		 */
-		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 		/*
 		 * For addname, we're looking for a place to put the new entry.
 		 * We want to use a data block with an entry of equal
@@ -506,15 +507,15 @@ xfs_dir2_leafn_lookup_int(
 					}
 					curfdb = newfdb;
 					free = curbp->data;
-					ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) ==
+					ASSERT(be32_to_cpu(free->hdr.magic) ==
 					       XFS_DIR2_FREE_MAGIC);
-					ASSERT((INT_GET(free->hdr.firstdb, ARCH_CONVERT) %
+					ASSERT((be32_to_cpu(free->hdr.firstdb) %
 						XFS_DIR2_MAX_FREE_BESTS(mp)) ==
 					       0);
-					ASSERT(INT_GET(free->hdr.firstdb, ARCH_CONVERT) <= curdb);
+					ASSERT(be32_to_cpu(free->hdr.firstdb) <= curdb);
 					ASSERT(curdb <
-					       INT_GET(free->hdr.firstdb, ARCH_CONVERT) +
-					       INT_GET(free->hdr.nvalid, ARCH_CONVERT));
+					       be32_to_cpu(free->hdr.firstdb) +
+					       be32_to_cpu(free->hdr.nvalid));
 				}
 				/*
 				 * Get the index for our entry.
@@ -523,12 +524,12 @@ xfs_dir2_leafn_lookup_int(
 				/*
 				 * If it has room, return it.
 				 */
-				if (unlikely(INT_GET(free->bests[fi], ARCH_CONVERT) == NULLDATAOFF)) {
+				if (unlikely(be16_to_cpu(free->bests[fi]) == NULLDATAOFF)) {
 					XFS_ERROR_REPORT("xfs_dir2_leafn_lookup_int",
 							 XFS_ERRLEVEL_LOW, mp);
 					return XFS_ERROR(EFSCORRUPTED);
 				}
-				if (INT_GET(free->bests[fi], ARCH_CONVERT) >= length) {
+				if (be16_to_cpu(free->bests[fi]) >= length) {
 					*indexp = index;
 					state->extravalid = 1;
 					state->extrablk.bp = curbp;
@@ -572,7 +573,7 @@ xfs_dir2_leafn_lookup_int(
 			 */
 			dep = (xfs_dir2_data_entry_t *)
 			      ((char *)curbp->data +
-			       XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+			       XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 			/*
 			 * Compare the entry, return it if it matches.
 			 */
@@ -619,7 +620,7 @@ xfs_dir2_leafn_lookup_int(
 	 * Return the final index, that will be the insertion point.
 	 */
 	*indexp = index;
-	ASSERT(index == INT_GET(leaf->hdr.count, ARCH_CONVERT) || args->oknoent);
+	ASSERT(index == be16_to_cpu(leaf->hdr.count) || args->oknoent);
 	return XFS_ERROR(ENOENT);
 }
 
@@ -657,12 +658,12 @@ xfs_dir2_leafn_moveents(
 	 * destination leaf entries, open up a hole in the destination
 	 * to hold the new entries.
 	 */
-	if (start_d < INT_GET(leaf_d->hdr.count, ARCH_CONVERT)) {
+	if (start_d < be16_to_cpu(leaf_d->hdr.count)) {
 		memmove(&leaf_d->ents[start_d + count], &leaf_d->ents[start_d],
-			(INT_GET(leaf_d->hdr.count, ARCH_CONVERT) - start_d) *
+			(be16_to_cpu(leaf_d->hdr.count) - start_d) *
 			sizeof(xfs_dir2_leaf_entry_t));
 		xfs_dir2_leaf_log_ents(tp, bp_d, start_d + count,
-			count + INT_GET(leaf_d->hdr.count, ARCH_CONVERT) - 1);
+			count + be16_to_cpu(leaf_d->hdr.count) - 1);
 	}
 	/*
 	 * If the source has stale leaves, count the ones in the copy range
@@ -672,7 +673,7 @@ xfs_dir2_leafn_moveents(
 		int	i;			/* temp leaf index */
 
 		for (i = start_s, stale = 0; i < start_s + count; i++) {
-			if (INT_GET(leaf_s->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+			if (be32_to_cpu(leaf_s->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 				stale++;
 		}
 	} else
@@ -687,7 +688,7 @@ xfs_dir2_leafn_moveents(
 	 * If there are source entries after the ones we copied,
 	 * delete the ones we copied by sliding the next ones down.
 	 */
-	if (start_s + count < INT_GET(leaf_s->hdr.count, ARCH_CONVERT)) {
+	if (start_s + count < be16_to_cpu(leaf_s->hdr.count)) {
 		memmove(&leaf_s->ents[start_s], &leaf_s->ents[start_s + count],
 			count * sizeof(xfs_dir2_leaf_entry_t));
 		xfs_dir2_leaf_log_ents(tp, bp_s, start_s, start_s + count - 1);
@@ -695,10 +696,10 @@ xfs_dir2_leafn_moveents(
 	/*
 	 * Update the headers and log them.
 	 */
-	INT_MOD(leaf_s->hdr.count, ARCH_CONVERT, -(count));
-	INT_MOD(leaf_s->hdr.stale, ARCH_CONVERT, -(stale));
-	INT_MOD(leaf_d->hdr.count, ARCH_CONVERT, count);
-	INT_MOD(leaf_d->hdr.stale, ARCH_CONVERT, stale);
+	be16_add(&leaf_s->hdr.count, -(count));
+	be16_add(&leaf_s->hdr.stale, -(stale));
+	be16_add(&leaf_d->hdr.count, count);
+	be16_add(&leaf_d->hdr.stale, stale);
 	xfs_dir2_leaf_log_header(tp, bp_s);
 	xfs_dir2_leaf_log_header(tp, bp_d);
 	xfs_dir2_leafn_check(args->dp, bp_s);
@@ -719,13 +720,13 @@ xfs_dir2_leafn_order(
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0 &&
-	    INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0 &&
-	    (INT_GET(leaf2->ents[0].hashval, ARCH_CONVERT) < INT_GET(leaf1->ents[0].hashval, ARCH_CONVERT) ||
-	     INT_GET(leaf2->ents[INT_GET(leaf2->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT) <
-	     INT_GET(leaf1->ents[INT_GET(leaf1->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT)))
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	if (be16_to_cpu(leaf1->hdr.count) > 0 &&
+	    be16_to_cpu(leaf2->hdr.count) > 0 &&
+	    (be32_to_cpu(leaf2->ents[0].hashval) < be32_to_cpu(leaf1->ents[0].hashval) ||
+	     be32_to_cpu(leaf2->ents[be16_to_cpu(leaf2->hdr.count) - 1].hashval) <
+	     be32_to_cpu(leaf1->ents[be16_to_cpu(leaf1->hdr.count) - 1].hashval)))
 		return 1;
 	return 0;
 }
@@ -768,9 +769,9 @@ xfs_dir2_leafn_rebalance(
 	}
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	oldsum = INT_GET(leaf1->hdr.count, ARCH_CONVERT) + INT_GET(leaf2->hdr.count, ARCH_CONVERT);
+	oldsum = be16_to_cpu(leaf1->hdr.count) + be16_to_cpu(leaf2->hdr.count);
 #ifdef DEBUG
-	oldstale = INT_GET(leaf1->hdr.stale, ARCH_CONVERT) + INT_GET(leaf2->hdr.stale, ARCH_CONVERT);
+	oldstale = be16_to_cpu(leaf1->hdr.stale) + be16_to_cpu(leaf2->hdr.stale);
 #endif
 	mid = oldsum >> 1;
 	/*
@@ -780,10 +781,10 @@ xfs_dir2_leafn_rebalance(
 	if (oldsum & 1) {
 		xfs_dahash_t	midhash;	/* middle entry hash value */
 
-		if (mid >= INT_GET(leaf1->hdr.count, ARCH_CONVERT))
-			midhash = INT_GET(leaf2->ents[mid - INT_GET(leaf1->hdr.count, ARCH_CONVERT)].hashval, ARCH_CONVERT);
+		if (mid >= be16_to_cpu(leaf1->hdr.count))
+			midhash = be32_to_cpu(leaf2->ents[mid - be16_to_cpu(leaf1->hdr.count)].hashval);
 		else
-			midhash = INT_GET(leaf1->ents[mid].hashval, ARCH_CONVERT);
+			midhash = be32_to_cpu(leaf1->ents[mid].hashval);
 		isleft = args->hashval <= midhash;
 	}
 	/*
@@ -797,30 +798,30 @@ xfs_dir2_leafn_rebalance(
 	 * Calculate moved entry count.  Positive means left-to-right,
 	 * negative means right-to-left.  Then move the entries.
 	 */
-	count = INT_GET(leaf1->hdr.count, ARCH_CONVERT) - mid + (isleft == 0);
+	count = be16_to_cpu(leaf1->hdr.count) - mid + (isleft == 0);
 	if (count > 0)
 		xfs_dir2_leafn_moveents(args, blk1->bp,
-			INT_GET(leaf1->hdr.count, ARCH_CONVERT) - count, blk2->bp, 0, count);
+			be16_to_cpu(leaf1->hdr.count) - count, blk2->bp, 0, count);
 	else if (count < 0)
 		xfs_dir2_leafn_moveents(args, blk2->bp, 0, blk1->bp,
-			INT_GET(leaf1->hdr.count, ARCH_CONVERT), count);
-	ASSERT(INT_GET(leaf1->hdr.count, ARCH_CONVERT) + INT_GET(leaf2->hdr.count, ARCH_CONVERT) == oldsum);
-	ASSERT(INT_GET(leaf1->hdr.stale, ARCH_CONVERT) + INT_GET(leaf2->hdr.stale, ARCH_CONVERT) == oldstale);
+			be16_to_cpu(leaf1->hdr.count), count);
+	ASSERT(be16_to_cpu(leaf1->hdr.count) + be16_to_cpu(leaf2->hdr.count) == oldsum);
+	ASSERT(be16_to_cpu(leaf1->hdr.stale) + be16_to_cpu(leaf2->hdr.stale) == oldstale);
 	/*
 	 * Mark whether we're inserting into the old or new leaf.
 	 */
-	if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) < INT_GET(leaf2->hdr.count, ARCH_CONVERT))
+	if (be16_to_cpu(leaf1->hdr.count) < be16_to_cpu(leaf2->hdr.count))
 		state->inleaf = swap;
-	else if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > INT_GET(leaf2->hdr.count, ARCH_CONVERT))
+	else if (be16_to_cpu(leaf1->hdr.count) > be16_to_cpu(leaf2->hdr.count))
 		state->inleaf = !swap;
 	else
 		state->inleaf =
-			swap ^ (blk1->index <= INT_GET(leaf1->hdr.count, ARCH_CONVERT));
+			swap ^ (blk1->index <= be16_to_cpu(leaf1->hdr.count));
 	/*
 	 * Adjust the expected index for insertion.
 	 */
 	if (!state->inleaf)
-		blk2->index = blk1->index - INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+		blk2->index = blk1->index - be16_to_cpu(leaf1->hdr.count);
 	
 	/* 
 	 * Finally sanity check just to make sure we are not returning a negative index 
@@ -867,7 +868,7 @@ xfs_dir2_leafn_remove(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Point to the entry we're removing.
 	 */
@@ -875,17 +876,17 @@ xfs_dir2_leafn_remove(
 	/*
 	 * Extract the data block and offset from the entry.
 	 */
-	db = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+	db = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 	ASSERT(dblk->blkno == db);
-	off = XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT));
+	off = XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address));
 	ASSERT(dblk->index == off);
 	/*
 	 * Kill the leaf entry by marking it stale.
 	 * Log the leaf block changes.
 	 */
-	INT_MOD(leaf->hdr.stale, ARCH_CONVERT, +1);
+	be16_add(&leaf->hdr.stale, 1);
 	xfs_dir2_leaf_log_header(tp, bp);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_leaf_log_ents(tp, bp, index, index);
 	/*
 	 * Make the data entry free.  Keep track of the longest freespace
@@ -894,7 +895,7 @@ xfs_dir2_leafn_remove(
 	dbp = dblk->bp;
 	data = dbp->data;
 	dep = (xfs_dir2_data_entry_t *)((char *)data + off);
-	longest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+	longest = be16_to_cpu(data->hdr.bestfree[0].length);
 	needlog = needscan = 0;
 	xfs_dir2_data_make_free(tp, dbp, off,
 		XFS_DIR2_DATA_ENTSIZE(dep->namelen), &needlog, &needscan);
@@ -911,7 +912,7 @@ xfs_dir2_leafn_remove(
 	 * If the longest data block freespace changes, need to update
 	 * the corresponding freeblock entry.
 	 */
-	if (longest < INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
+	if (longest < be16_to_cpu(data->hdr.bestfree[0].length)) {
 		int		error;		/* error return value */
 		xfs_dabuf_t	*fbp;		/* freeblock buffer */
 		xfs_dir2_db_t	fdb;		/* freeblock block number */
@@ -929,15 +930,15 @@ xfs_dir2_leafn_remove(
 			return error;
 		}
 		free = fbp->data;
-		ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
-		ASSERT(INT_GET(free->hdr.firstdb, ARCH_CONVERT) ==
+		ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
+		ASSERT(be32_to_cpu(free->hdr.firstdb) ==
 		       XFS_DIR2_MAX_FREE_BESTS(mp) *
 		       (fdb - XFS_DIR2_FREE_FIRSTDB(mp)));
 		/*
 		 * Calculate which entry we need to fix.
 		 */
 		findex = XFS_DIR2_DB_TO_FDINDEX(mp, db);
-		longest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+		longest = be16_to_cpu(data->hdr.bestfree[0].length);
 		/*
 		 * If the data block is now empty we can get rid of it
 		 * (usually).
@@ -969,7 +970,7 @@ xfs_dir2_leafn_remove(
 			/*
 			 * One less used entry in the free table.
 			 */
-			INT_MOD(free->hdr.nused, ARCH_CONVERT, -1);
+			free->hdr.nused = cpu_to_be32(-1);
 			xfs_dir2_free_log_header(tp, fbp);
 			/*
 			 * If this was the last entry in the table, we can
@@ -977,21 +978,21 @@ xfs_dir2_leafn_remove(
 			 * entries at the end referring to non-existent
 			 * data blocks, get those too.
 			 */
-			if (findex == INT_GET(free->hdr.nvalid, ARCH_CONVERT) - 1) {
+			if (findex == be32_to_cpu(free->hdr.nvalid) - 1) {
 				int	i;		/* free entry index */
 
 				for (i = findex - 1;
-				     i >= 0 && INT_GET(free->bests[i], ARCH_CONVERT) == NULLDATAOFF;
+				     i >= 0 && be16_to_cpu(free->bests[i]) == NULLDATAOFF;
 				     i--)
 					continue;
-				INT_SET(free->hdr.nvalid, ARCH_CONVERT, i + 1);
+				free->hdr.nvalid = cpu_to_be32(i + 1);
 				logfree = 0;
 			}
 			/*
 			 * Not the last entry, just punch it out.
 			 */
 			else {
-				INT_SET(free->bests[findex], ARCH_CONVERT, NULLDATAOFF);
+				free->bests[findex] = cpu_to_be16(NULLDATAOFF);
 				logfree = 1;
 			}
 			/*
@@ -1017,7 +1018,7 @@ xfs_dir2_leafn_remove(
 		 * the new value.
 		 */
 		else {
-			INT_SET(free->bests[findex], ARCH_CONVERT, longest);
+			free->bests[findex] = cpu_to_be16(longest);
 			logfree = 1;
 		}
 		/*
@@ -1039,7 +1040,7 @@ xfs_dir2_leafn_remove(
 	*rval =
 		((uint)sizeof(leaf->hdr) +
 		 (uint)sizeof(leaf->ents[0]) *
-		 (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT))) <
+		 (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale))) <
 		mp->m_dir_magicpct;
 	return 0;
 }
@@ -1138,9 +1139,9 @@ xfs_dir2_leafn_toosmall(
 	 */
 	blk = &state->path.blk[state->path.active - 1];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DIR2_LEAFN_MAGIC);
 	leaf = (xfs_dir2_leaf_t *)info;
-	count = INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+	count = be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 	bytes = (uint)sizeof(leaf->hdr) + count * (uint)sizeof(leaf->ents[0]);
 	if (bytes > (state->blocksize >> 1)) {
 		/*
@@ -1160,7 +1161,7 @@ xfs_dir2_leafn_toosmall(
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward, 0,
 			&rval);
@@ -1176,9 +1177,9 @@ xfs_dir2_leafn_toosmall(
 	 * We prefer coalescing with the lower numbered sibling so as
 	 * to shrink a directory over time.
 	 */
-	forward = INT_GET(info->forw, ARCH_CONVERT) < INT_GET(info->back, ARCH_CONVERT);
+	forward = be32_to_cpu(info->forw) < be32_to_cpu(info->back);
 	for (i = 0, bp = NULL; i < 2; forward = !forward, i++) {
-		blkno = forward ?INT_GET( info->forw, ARCH_CONVERT) : INT_GET(info->back, ARCH_CONVERT);
+		blkno = forward ? be32_to_cpu(info->forw) : be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		/*
@@ -1194,11 +1195,11 @@ xfs_dir2_leafn_toosmall(
 		 * Count bytes in the two blocks combined.
 		 */
 		leaf = (xfs_dir2_leaf_t *)info;
-		count = INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+		count = be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 		bytes = state->blocksize - (state->blocksize >> 2);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-		count += INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+		count += be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 		bytes -= count * (uint)sizeof(leaf->ents[0]);
 		/*
 		 * Fits with at least 25% to spare.
@@ -1256,27 +1257,27 @@ xfs_dir2_leafn_unbalance(
 	ASSERT(save_blk->magic == XFS_DIR2_LEAFN_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * If there are any stale leaf entries, take this opportunity
 	 * to purge them.
 	 */
-	if (INT_GET(drop_leaf->hdr.stale, ARCH_CONVERT))
+	if (drop_leaf->hdr.stale)
 		xfs_dir2_leaf_compact(args, drop_blk->bp);
-	if (INT_GET(save_leaf->hdr.stale, ARCH_CONVERT))
+	if (save_leaf->hdr.stale)
 		xfs_dir2_leaf_compact(args, save_blk->bp);
 	/*
 	 * Move the entries from drop to the appropriate end of save.
 	 */
-	drop_blk->hashval = INT_GET(drop_leaf->ents[INT_GET(drop_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(drop_leaf->ents[be16_to_cpu(drop_leaf->hdr.count) - 1].hashval);
 	if (xfs_dir2_leafn_order(save_blk->bp, drop_blk->bp))
 		xfs_dir2_leafn_moveents(args, drop_blk->bp, 0, save_blk->bp, 0,
-			INT_GET(drop_leaf->hdr.count, ARCH_CONVERT));
+			be16_to_cpu(drop_leaf->hdr.count));
 	else
 		xfs_dir2_leafn_moveents(args, drop_blk->bp, 0, save_blk->bp,
-			INT_GET(save_leaf->hdr.count, ARCH_CONVERT), INT_GET(drop_leaf->hdr.count, ARCH_CONVERT));
-	save_blk->hashval = INT_GET(save_leaf->ents[INT_GET(save_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+			be16_to_cpu(save_leaf->hdr.count), be16_to_cpu(drop_leaf->hdr.count));
+	save_blk->hashval = be32_to_cpu(save_leaf->ents[be16_to_cpu(save_leaf->hdr.count) - 1].hashval);
 	xfs_dir2_leafn_check(args->dp, save_blk->bp);
 }
 
@@ -1378,7 +1379,7 @@ xfs_dir2_node_addname_int(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			needlog;	/* need to log data header */
 	int			needscan;	/* need to rescan data frees */
-	xfs_dir2_data_off_t	*tagp;		/* data entry tag pointer */
+	__be16			*tagp;		/* data entry tag pointer */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	dp = args->dp;
@@ -1397,7 +1398,7 @@ xfs_dir2_node_addname_int(
 		 */
 		ifbno = fblk->blkno;
 		free = fbp->data;
-		ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+		ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 		findex = fblk->index;
 		/*
 		 * This means the free entry showed that the data block had
@@ -1405,10 +1406,10 @@ xfs_dir2_node_addname_int(
 		 * Use that data block.
 		 */
 		if (findex >= 0) {
-			ASSERT(findex < INT_GET(free->hdr.nvalid, ARCH_CONVERT));
-			ASSERT(INT_GET(free->bests[findex], ARCH_CONVERT) != NULLDATAOFF);
-			ASSERT(INT_GET(free->bests[findex], ARCH_CONVERT) >= length);
-			dbno = INT_GET(free->hdr.firstdb, ARCH_CONVERT) + findex;
+			ASSERT(findex < be32_to_cpu(free->hdr.nvalid));
+			ASSERT(be16_to_cpu(free->bests[findex]) != NULLDATAOFF);
+			ASSERT(be16_to_cpu(free->bests[findex]) >= length);
+			dbno = be32_to_cpu(free->hdr.firstdb) + findex;
 		}
 		/*
 		 * The data block looked at didn't have enough room.
@@ -1481,20 +1482,20 @@ xfs_dir2_node_addname_int(
 				continue;
 			}
 			free = fbp->data;
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 			findex = 0;
 		}
 		/*
 		 * Look at the current free entry.  Is it good enough?
 		 */
-		if (INT_GET(free->bests[findex], ARCH_CONVERT) != NULLDATAOFF &&
-		    INT_GET(free->bests[findex], ARCH_CONVERT) >= length)
-			dbno = INT_GET(free->hdr.firstdb, ARCH_CONVERT) + findex;
+		if (be16_to_cpu(free->bests[findex]) != NULLDATAOFF &&
+		    be16_to_cpu(free->bests[findex]) >= length)
+			dbno = be32_to_cpu(free->hdr.firstdb) + findex;
 		else {
 			/*
 			 * Are we done with the freeblock?
 			 */
-			if (++findex == INT_GET(free->hdr.nvalid, ARCH_CONVERT)) {
+			if (++findex == be32_to_cpu(free->hdr.nvalid)) {
 				/*
 				 * Drop the block.
 				 */
@@ -1608,15 +1609,15 @@ xfs_dir2_node_addname_int(
 			 * its first slot as our empty slot.
 			 */
 			free = fbp->data;
-			INT_SET(free->hdr.magic, ARCH_CONVERT, XFS_DIR2_FREE_MAGIC);
-			INT_SET(free->hdr.firstdb, ARCH_CONVERT,
+			free->hdr.magic = cpu_to_be32(XFS_DIR2_FREE_MAGIC);
+			free->hdr.firstdb = cpu_to_be32(
 				(fbno - XFS_DIR2_FREE_FIRSTDB(mp)) *
 				XFS_DIR2_MAX_FREE_BESTS(mp));
 			free->hdr.nvalid = 0;
 			free->hdr.nused = 0;
 		} else {
 			free = fbp->data;
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 		}
 
 		/*
@@ -1627,20 +1628,20 @@ xfs_dir2_node_addname_int(
 		 * If it's after the end of the current entries in the
 		 * freespace block, extend that table.
 		 */
-		if (findex >= INT_GET(free->hdr.nvalid, ARCH_CONVERT)) {
+		if (findex >= be32_to_cpu(free->hdr.nvalid)) {
 			ASSERT(findex < XFS_DIR2_MAX_FREE_BESTS(mp));
-			INT_SET(free->hdr.nvalid, ARCH_CONVERT, findex + 1);
+			free->hdr.nvalid = cpu_to_be32(findex + 1);
 			/*
 			 * Tag new entry so nused will go up.
 			 */
-			INT_SET(free->bests[findex], ARCH_CONVERT, NULLDATAOFF);
+			free->bests[findex] = cpu_to_be16(NULLDATAOFF);
 		}
 		/*
 		 * If this entry was for an empty data block
 		 * (this should always be true) then update the header.
 		 */
-		if (INT_GET(free->bests[findex], ARCH_CONVERT) == NULLDATAOFF) {
-			INT_MOD(free->hdr.nused, ARCH_CONVERT, +1);
+		if (be16_to_cpu(free->bests[findex]) == NULLDATAOFF) {
+			be32_add(&free->hdr.nused, 1);
 			xfs_dir2_free_log_header(tp, fbp);
 		}
 		/*
@@ -1649,7 +1650,7 @@ xfs_dir2_node_addname_int(
 		 * change again.
 		 */
 		data = dbp->data;
-		INT_COPY(free->bests[findex], data->hdr.bestfree[0].length, ARCH_CONVERT);
+		free->bests[findex] = data->hdr.bestfree[0].length;
 		logfree = 1;
 	}
 	/*
@@ -1677,12 +1678,12 @@ xfs_dir2_node_addname_int(
 		data = dbp->data;
 		logfree = 0;
 	}
-	ASSERT(INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) >= length);
+	ASSERT(be16_to_cpu(data->hdr.bestfree[0].length) >= length);
 	/*
 	 * Point to the existing unused space.
 	 */
 	dup = (xfs_dir2_data_unused_t *)
-	      ((char *)data + INT_GET(data->hdr.bestfree[0].offset, ARCH_CONVERT));
+	      ((char *)data + be16_to_cpu(data->hdr.bestfree[0].offset));
 	needscan = needlog = 0;
 	/*
 	 * Mark the first part of the unused space, inuse for us.
@@ -1698,7 +1699,7 @@ xfs_dir2_node_addname_int(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, dep->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)data));
+	*tagp = cpu_to_be16((char *)dep - (char *)data);
 	xfs_dir2_data_log_entry(tp, dbp, dep);
 	/*
 	 * Rescan the block for bestfree if needed.
@@ -1713,8 +1714,8 @@ xfs_dir2_node_addname_int(
 	/*
 	 * If the freespace entry is now wrong, update it.
 	 */
-	if (INT_GET(free->bests[findex], ARCH_CONVERT) != INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
-		INT_COPY(free->bests[findex], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(free->bests[findex]) != be16_to_cpu(data->hdr.bestfree[0].length)) {
+		free->bests[findex] = data->hdr.bestfree[0].length;
 		logfree = 1;
 	}
 	/*
@@ -1731,7 +1732,7 @@ xfs_dir2_node_addname_int(
 	 * Return the data block and offset in args, then drop the data block.
 	 */
 	args->blkno = (xfs_dablk_t)dbno;
-	args->index = INT_GET(*tagp, ARCH_CONVERT);
+	args->index = be16_to_cpu(*tagp);
 	xfs_da_buf_done(dbp);
 	return 0;
 }
@@ -1900,10 +1901,10 @@ xfs_dir2_node_replace(
 		 * Point to the data entry.
 		 */
 		data = state->extrablk.bp->data;
-		ASSERT(INT_GET(data->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+		ASSERT(be32_to_cpu(data->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)data +
-		       XFS_DIR2_DATAPTR_TO_OFF(state->mp, INT_GET(lep->address, ARCH_CONVERT)));
+		       XFS_DIR2_DATAPTR_TO_OFF(state->mp, be32_to_cpu(lep->address)));
 		ASSERT(inum != INT_GET(dep->inumber, ARCH_CONVERT));
 		/*
 		 * Fill in the new inode number and log the entry.
@@ -1966,11 +1967,11 @@ xfs_dir2_node_trim_free(
 		return 0;
 	}
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	/*
 	 * If there are used entries, there's nothing to do.
 	 */
-	if (INT_GET(free->hdr.nused, ARCH_CONVERT) > 0) {
+	if (be32_to_cpu(free->hdr.nused) > 0) {
 		xfs_da_brelse(tp, bp);
 		*rvalp = 0;
 		return 0;
diff --git a/fs/xfs/xfs_dir2_node.h b/fs/xfs/xfs_dir2_node.h
index 0ab8fbd59512..c7c870ee7857 100644
--- a/fs/xfs/xfs_dir2_node.h
+++ b/fs/xfs/xfs_dir2_node.h
@@ -41,15 +41,15 @@ struct xfs_trans;
 #define	XFS_DIR2_FREE_MAGIC	0x58443246	/* XD2F */
 
 typedef	struct xfs_dir2_free_hdr {
-	__uint32_t		magic;		/* XFS_DIR2_FREE_MAGIC */
-	__int32_t		firstdb;	/* db of first entry */
-	__int32_t		nvalid;		/* count of valid entries */
-	__int32_t		nused;		/* count of used entries */
+	__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 */
-	xfs_dir2_data_off_t	bests[1];	/* best free counts */
+	__be16			bests[1];	/* best free counts */
 						/* unused entries are -1 */
 } xfs_dir2_free_t;
 
diff --git a/fs/xfs/xfs_dir2_sf.c b/fs/xfs/xfs_dir2_sf.c
index ec8e7476c8b7..d98a41d1fe63 100644
--- a/fs/xfs/xfs_dir2_sf.c
+++ b/fs/xfs/xfs_dir2_sf.c
@@ -98,8 +98,8 @@ xfs_dir2_block_sfsize(
 	/*
 	 * Iterate over the block's data entries by using the leaf pointers.
 	 */
-	for (i = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-		if ((addr = INT_GET(blp[i].address, ARCH_CONVERT)) == XFS_DIR2_NULL_DATAPTR)
+	for (i = 0; i < be32_to_cpu(btp->count); i++) {
+		if ((addr = be32_to_cpu(blp[i].address)) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Calculate the pointer to the entry at hand.
@@ -220,8 +220,8 @@ xfs_dir2_block_to_sf(
 		 * If it's unused, just skip over it.
 		 */
 		dup = (xfs_dir2_data_unused_t *)ptr;
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			ptr += INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			ptr += be16_to_cpu(dup->length);
 			continue;
 		}
 		dep = (xfs_dir2_data_entry_t *)ptr;
diff --git a/fs/xfs/xfs_dir_leaf.c b/fs/xfs/xfs_dir_leaf.c
index e83074016abb..ee88751c3be6 100644
--- a/fs/xfs/xfs_dir_leaf.c
+++ b/fs/xfs/xfs_dir_leaf.c
@@ -176,7 +176,7 @@ xfs_dir_shortform_addname(xfs_da_args_t *args)
 	ASSERT(dp->i_df.if_u1.if_data != NULL);
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    args->name[0] == sfe->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0)
@@ -193,7 +193,7 @@ xfs_dir_shortform_addname(xfs_da_args_t *args)
 	XFS_DIR_SF_PUT_DIRINO(&args->inumber, &sfe->inumber);
 	sfe->namelen = args->namelen;
 	memcpy(sfe->name, args->name, sfe->namelen);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, +1);
+	sf->hdr.count++;
 
 	dp->i_d.di_size += size;
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA);
@@ -227,7 +227,7 @@ xfs_dir_shortform_removename(xfs_da_args_t *args)
 	base = sizeof(xfs_dir_sf_hdr_t);
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		size = XFS_DIR_SF_ENTSIZE_BYENTRY(sfe);
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
@@ -245,7 +245,7 @@ xfs_dir_shortform_removename(xfs_da_args_t *args)
 		memmove(&((char *)sf)[base], &((char *)sf)[base+size],
 					      dp->i_d.di_size - (base+size));
 	}
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, -1);
+	sf->hdr.count--;
 
 	xfs_idata_realloc(dp, -size, XFS_DATA_FORK);
 	dp->i_d.di_size -= size;
@@ -288,7 +288,7 @@ xfs_dir_shortform_lookup(xfs_da_args_t *args)
 		return(XFS_ERROR(EEXIST));
 	}
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0) {
@@ -375,7 +375,7 @@ xfs_dir_shortform_to_leaf(xfs_da_args_t *iargs)
 		goto out;
 
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0; i < sf->hdr.count; i++) {
 		args.name = (char *)(sfe->name);
 		args.namelen = sfe->namelen;
 		args.hashval = xfs_da_hashname((char *)(sfe->name),
@@ -428,7 +428,7 @@ xfs_dir_shortform_getdents(xfs_inode_t *dp, uio_t *uio, int *eofp,
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	cookhash = XFS_DA_COOKIE_HASH(mp, uio->uio_offset);
 	want_entno = XFS_DA_COOKIE_ENTRY(mp, uio->uio_offset);
-	nsbuf = INT_GET(sf->hdr.count, ARCH_CONVERT) + 2;
+	nsbuf = sf->hdr.count + 2;
 	sbsize = (nsbuf + 1) * sizeof(*sbuf);
 	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP);
 
@@ -460,8 +460,7 @@ xfs_dir_shortform_getdents(xfs_inode_t *dp, uio_t *uio, int *eofp,
 	/*
 	 * Scan the directory data for the rest of the entries.
 	 */
-	for (i = 0, sfe = &sf->list[0];
-			i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 
 		if (unlikely(
 		    ((char *)sfe < (char *)sf) ||
@@ -600,7 +599,7 @@ xfs_dir_shortform_replace(xfs_da_args_t *args)
 	}
 	ASSERT(args->namelen != 1 || args->name[0] != '.');
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0) {
@@ -644,7 +643,7 @@ xfs_dir_leaf_to_shortform(xfs_da_args_t *iargs)
 	ASSERT(bp != NULL);
 	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(dp->i_mount));
 	leaf = (xfs_dir_leafblock_t *)tmpbuffer;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	memset(bp->data, 0, XFS_LBSIZE(dp->i_mount));
 
 	/*
@@ -742,11 +741,13 @@ xfs_dir_leaf_to_node(xfs_da_args_t *args)
 	}
 	node = bp1->data;
 	leaf = bp2->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	INT_SET(node->btree[0].hashval, ARCH_CONVERT, INT_GET(leaf->entries[ INT_GET(leaf->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	node->btree[0].hashval = cpu_to_be32(
+		INT_GET(leaf->entries[
+			INT_GET(leaf->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
 	xfs_da_buf_done(bp2);
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 1);
+	node->btree[0].before = cpu_to_be32(blkno);
+	node->hdr.count = cpu_to_be16(1);
 	xfs_da_log_buf(args->trans, bp1,
 		XFS_DA_LOGRANGE(node, &node->btree[0], sizeof(node->btree[0])));
 	xfs_da_buf_done(bp1);
@@ -781,7 +782,7 @@ xfs_dir_leaf_create(xfs_da_args_t *args, xfs_dablk_t blkno, xfs_dabuf_t **bpp)
 	leaf = bp->data;
 	memset((char *)leaf, 0, XFS_LBSIZE(dp->i_mount));
 	hdr = &leaf->hdr;
-	INT_SET(hdr->info.magic, ARCH_CONVERT, XFS_DIR_LEAF_MAGIC);
+	hdr->info.magic = cpu_to_be16(XFS_DIR_LEAF_MAGIC);
 	INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount));
 	if (!hdr->firstused)
 		INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount) - 1);
@@ -860,7 +861,7 @@ xfs_dir_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args, int index)
 	int tablesize, entsize, sum, i, tmp, error;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	ASSERT((index >= 0) && (index <= INT_GET(leaf->hdr.count, ARCH_CONVERT)));
 	hdr = &leaf->hdr;
 	entsize = XFS_DIR_LEAF_ENTSIZE_BYNAME(args->namelen);
@@ -940,7 +941,7 @@ xfs_dir_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int index,
 	int tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	ASSERT((mapindex >= 0) && (mapindex < XFS_DIR_LEAF_MAPSIZE));
 	ASSERT((index >= 0) && (index <= INT_GET(hdr->count, ARCH_CONVERT)));
@@ -1097,8 +1098,8 @@ xfs_dir_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(blk2->magic == XFS_DIR_LEAF_MAGIC);
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 
 	/*
 	 * Check ordering of blocks, reverse if it makes things simpler.
@@ -1325,7 +1326,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DIR_LEAF_MAGIC);
 	leaf = (xfs_dir_leafblock_t *)info;
 	count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
 	bytes = (uint)sizeof(xfs_dir_leaf_hdr_t) +
@@ -1348,7 +1349,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -1369,12 +1370,12 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * We prefer coalescing with the lower numbered sibling so as
 	 * to shrink a directory over time.
 	 */
-	forward = (INT_GET(info->forw, ARCH_CONVERT) < INT_GET(info->back, ARCH_CONVERT));	/* start with smaller blk num */
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));	/* start with smaller blk num */
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -1389,7 +1390,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 		bytes  = state->blocksize - (state->blocksize>>2);
 		bytes -= INT_GET(leaf->hdr.namebytes, ARCH_CONVERT);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 		count += INT_GET(leaf->hdr.count, ARCH_CONVERT);
 		bytes -= INT_GET(leaf->hdr.namebytes, ARCH_CONVERT);
 		bytes -= count * ((uint)sizeof(xfs_dir_leaf_name_t) - 1);
@@ -1447,7 +1448,7 @@ xfs_dir_leaf_remove(xfs_trans_t *trans, xfs_dabuf_t *bp, int index)
 	xfs_mount_t *mp;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	mp = trans->t_mountp;
 	ASSERT((INT_GET(hdr->count, ARCH_CONVERT) > 0) && (INT_GET(hdr->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
@@ -1599,8 +1600,8 @@ xfs_dir_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_DIR_LEAF_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	drop_hdr = &drop_leaf->hdr;
 	save_hdr = &save_leaf->hdr;
 
@@ -1695,7 +1696,7 @@ xfs_dir_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args, int *index)
 	xfs_dahash_t hashval;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) < (XFS_LBSIZE(args->dp->i_mount)/8));
 
 	/*
@@ -1782,8 +1783,8 @@ xfs_dir_leaf_moveents(xfs_dir_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Set up environment.
 	 */
-	ASSERT(INT_GET(leaf_s->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf_d->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_s->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_d->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
 	ASSERT((INT_GET(hdr_s->count, ARCH_CONVERT) > 0) && (INT_GET(hdr_s->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
@@ -1883,8 +1884,8 @@ xfs_dir_leaf_order(xfs_dabuf_t *leaf1_bp, xfs_dabuf_t *leaf2_bp)
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT((INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) &&
-	       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC));
+	ASSERT((be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR_LEAF_MAGIC) &&
+	       (be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR_LEAF_MAGIC));
 	if ((INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0) &&
 	    ((INT_GET(leaf2->entries[ 0 ].hashval, ARCH_CONVERT) <
 	      INT_GET(leaf1->entries[ 0 ].hashval, ARCH_CONVERT)) ||
@@ -1904,7 +1905,7 @@ xfs_dir_leaf_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_dir_leafblock_t *leaf;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	if (count)
 		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
 	if (!leaf->hdr.count)
@@ -1940,7 +1941,7 @@ xfs_dir_leaf_getdents_int(
 
 	mp = dp->i_mount;
 	leaf = bp->data;
-	if (INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC) {
+	if (be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC) {
 		*eobp = 1;
 		return XFS_ERROR(ENOENT);	/* XXX wrong code */
 	}
@@ -1992,7 +1993,7 @@ xfs_dir_leaf_getdents_int(
 
 	if (i == INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
 		xfs_dir_trace_g_du("leaf: hash not found", dp, uio);
-		if (!INT_GET(leaf->hdr.info.forw, ARCH_CONVERT))
+		if (!leaf->hdr.info.forw)
 			uio->uio_offset =
 				XFS_DA_MAKE_COOKIE(mp, 0, 0, XFS_DA_MAXHASH);
 		/*
@@ -2047,8 +2048,7 @@ xfs_dir_leaf_getdents_int(
 			xfs_dir_trace_g_duc("leaf: middle cookie  ",
 						   dp, uio, p.cook.o);
 
-		} else if ((thishash = INT_GET(leaf->hdr.info.forw,
-							ARCH_CONVERT))) {
+		} else if ((thishash = be32_to_cpu(leaf->hdr.info.forw))) {
 			xfs_dabuf_t *bp2;
 			xfs_dir_leafblock_t *leaf2;
 
@@ -2064,9 +2064,9 @@ xfs_dir_leaf_getdents_int(
 			leaf2 = bp2->data;
 
 			if (unlikely(
-			       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
+			       (be16_to_cpu(leaf2->hdr.info.magic)
 						!= XFS_DIR_LEAF_MAGIC)
-			    || (INT_GET(leaf2->hdr.info.back, ARCH_CONVERT)
+			    || (be32_to_cpu(leaf2->hdr.info.back)
 						!= bno))) {	/* GROT */
 				XFS_CORRUPTION_ERROR("xfs_dir_leaf_getdents_int(3)",
 						     XFS_ERRLEVEL_LOW, mp,
diff --git a/fs/xfs/xfs_dir_sf.h b/fs/xfs/xfs_dir_sf.h
index fe44c6f4d560..5b20b4d3f57d 100644
--- a/fs/xfs/xfs_dir_sf.h
+++ b/fs/xfs/xfs_dir_sf.h
@@ -35,19 +35,21 @@ typedef struct { __uint8_t i[sizeof(xfs_ino_t)]; } xfs_dir_ino_t;
  * and the elements much be memcpy'd out into a work area to get correct
  * alignment for the inode number fields.
  */
+typedef struct xfs_dir_sf_hdr {		/* constant-structure header block */
+	xfs_dir_ino_t	parent;		/* parent dir inode number */
+	__uint8_t	count;		/* count of active entries */
+} xfs_dir_sf_hdr_t;
+
+typedef struct xfs_dir_sf_entry {
+	xfs_dir_ino_t	inumber;	/* referenced inode number */
+	__uint8_t	namelen;	/* actual length of name (no NULL) */
+	__uint8_t	name[1];	/* name */
+} xfs_dir_sf_entry_t;
+
 typedef struct xfs_dir_shortform {
-	struct xfs_dir_sf_hdr {		/* constant-structure header block */
-		xfs_dir_ino_t parent;	/* parent dir inode number */
-		__uint8_t count;	/* count of active entries */
-	} hdr;
-	struct xfs_dir_sf_entry {
-		xfs_dir_ino_t inumber;	/* referenced inode number */
-		__uint8_t namelen;	/* actual length of name (no NULL) */
-		__uint8_t name[1];	/* name */
-	} list[1];			/* variable sized array */
+	xfs_dir_sf_hdr_t	hdr;
+	xfs_dir_sf_entry_t	list[1];	/* variable sized array */
 } xfs_dir_shortform_t;
-typedef struct xfs_dir_sf_hdr xfs_dir_sf_hdr_t;
-typedef struct xfs_dir_sf_entry xfs_dir_sf_entry_t;
 
 /*
  * We generate this then sort it, so that readdirs are returned in
diff --git a/fs/xfs/xfs_dmapi.h b/fs/xfs/xfs_dmapi.h
index b4c7f2bc55a0..00b1540f8108 100644
--- a/fs/xfs/xfs_dmapi.h
+++ b/fs/xfs/xfs_dmapi.h
@@ -191,14 +191,4 @@ typedef enum {
 
 extern struct bhv_vfsops xfs_dmops;
 
-#ifdef CONFIG_XFS_DMAPI
-void xfs_dm_init(struct file_system_type *);
-void xfs_dm_exit(struct file_system_type *);
-#define XFS_DM_INIT(fstype)	xfs_dm_init(fstype)
-#define XFS_DM_EXIT(fstype)	xfs_dm_exit(fstype)
-#else
-#define XFS_DM_INIT(fstype)
-#define XFS_DM_EXIT(fstype)
-#endif
-
 #endif  /* __XFS_DMAPI_H__ */
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index b4d971b01588..56caa88713ab 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -462,6 +462,7 @@ xfs_fs_counts(
 {
 	unsigned long	s;
 
+	xfs_icsb_sync_counters_lazy(mp);
 	s = XFS_SB_LOCK(mp);
 	cnt->freedata = mp->m_sb.sb_fdblocks;
 	cnt->freertx = mp->m_sb.sb_frextents;
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 8f3fae1aa98a..0024892841a3 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -138,8 +138,6 @@ xfs_ialloc_ag_alloc(
 	int		version;	/* inode version number to use */
 	int		isaligned;	/* inode allocation at stripe unit */
 					/* boundary */
-	xfs_dinode_core_t dic;          /* a dinode_core to copy to new */
-					/* inodes */
 
 	args.tp = tp;
 	args.mp = tp->t_mountp;
@@ -250,10 +248,6 @@ xfs_ialloc_ag_alloc(
 	else
 		version = XFS_DINODE_VERSION_1;
 
-	memset(&dic, 0, sizeof(xfs_dinode_core_t));
-	INT_SET(dic.di_magic, ARCH_CONVERT, XFS_DINODE_MAGIC);
-	INT_SET(dic.di_version, ARCH_CONVERT, version);
-
 	for (j = 0; j < nbufs; j++) {
 		/*
 		 * Get the block.
@@ -266,12 +260,13 @@ xfs_ialloc_ag_alloc(
 		ASSERT(fbuf);
 		ASSERT(!XFS_BUF_GETERROR(fbuf));
 		/*
-		 * Loop over the inodes in this buffer.
+		 * Set initial values for the inodes in this buffer.
 		 */
-
+		xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
 		for (i = 0; i < ninodes; i++) {
 			free = XFS_MAKE_IPTR(args.mp, fbuf, i);
-			memcpy(&(free->di_core), &dic, sizeof(xfs_dinode_core_t));
+			INT_SET(free->di_core.di_magic, ARCH_CONVERT, XFS_DINODE_MAGIC);
+			INT_SET(free->di_core.di_version, ARCH_CONVERT, version);
 			INT_SET(free->di_next_unlinked, ARCH_CONVERT, NULLAGINO);
 			xfs_ialloc_log_di(tp, fbuf, i,
 				XFS_DI_CORE_BITS | XFS_DI_NEXT_UNLINKED);
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index 8e380a1fb79b..3ce35a6f700b 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -258,7 +258,7 @@ again:
 				goto finish_inode;
 
 			} else if (vp != inode_vp) {
-				struct inode *inode = LINVFS_GET_IP(inode_vp);
+				struct inode *inode = vn_to_inode(inode_vp);
 
 				/* The inode is being torn down, pause and
 				 * try again.
@@ -495,7 +495,7 @@ retry:
 	if ((inode = iget_locked(XFS_MTOVFS(mp)->vfs_super, ino))) {
 		xfs_inode_t	*ip;
 
-		vp = LINVFS_GET_VP(inode);
+		vp = vn_from_inode(inode);
 		if (inode->i_state & I_NEW) {
 			vn_initialize(inode);
 			error = xfs_iget_core(vp, mp, tp, ino, flags,
@@ -617,7 +617,7 @@ xfs_iput_new(xfs_inode_t	*ip,
 	     uint		lock_flags)
 {
 	vnode_t		*vp = XFS_ITOV(ip);
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	vn_trace_entry(vp, "xfs_iput_new", (inst_t *)__return_address);
 
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 1d7f5a7e063e..88a517fad07b 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -76,16 +76,18 @@ STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int);
  */
 STATIC void
 xfs_validate_extents(
-	xfs_bmbt_rec_t		*ep,
+	xfs_ifork_t		*ifp,
 	int			nrecs,
 	int			disk,
 	xfs_exntfmt_t		fmt)
 {
+	xfs_bmbt_rec_t		*ep;
 	xfs_bmbt_irec_t		irec;
 	xfs_bmbt_rec_t		rec;
 	int			i;
 
 	for (i = 0; i < nrecs; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		rec.l0 = get_unaligned((__uint64_t*)&ep->l0);
 		rec.l1 = get_unaligned((__uint64_t*)&ep->l1);
 		if (disk)
@@ -94,11 +96,10 @@ xfs_validate_extents(
 			xfs_bmbt_get_all(&rec, &irec);
 		if (fmt == XFS_EXTFMT_NOSTATE)
 			ASSERT(irec.br_state == XFS_EXT_NORM);
-		ep++;
 	}
 }
 #else /* DEBUG */
-#define xfs_validate_extents(ep, nrecs, disk, fmt)
+#define xfs_validate_extents(ifp, nrecs, disk, fmt)
 #endif /* DEBUG */
 
 /*
@@ -252,7 +253,8 @@ xfs_itobp(
 	xfs_inode_t	*ip,
 	xfs_dinode_t	**dipp,
 	xfs_buf_t	**bpp,
-	xfs_daddr_t	bno)
+	xfs_daddr_t	bno,
+	uint		imap_flags)
 {
 	xfs_buf_t	*bp;
 	int		error;
@@ -268,10 +270,9 @@ xfs_itobp(
 		 * inode on disk.
 		 */
 		imap.im_blkno = bno;
-		error = xfs_imap(mp, tp, ip->i_ino, &imap, XFS_IMAP_LOOKUP);
-		if (error != 0) {
+		if ((error = xfs_imap(mp, tp, ip->i_ino, &imap,
+					XFS_IMAP_LOOKUP | imap_flags)))
 			return error;
-		}
 
 		/*
 		 * If the inode number maps to a block outside the bounds
@@ -335,9 +336,10 @@ xfs_itobp(
 	 * (if DEBUG kernel) or the first inode in the buffer, otherwise.
 	 */
 #ifdef DEBUG
-	ni = BBTOB(imap.im_len) >> mp->m_sb.sb_inodelog;
+	ni = (imap_flags & XFS_IMAP_BULKSTAT) ? 0 :
+		(BBTOB(imap.im_len) >> mp->m_sb.sb_inodelog);
 #else
-	ni = 1;
+	ni = (imap_flags & XFS_IMAP_BULKSTAT) ? 0 : 1;
 #endif
 	for (i = 0; i < ni; i++) {
 		int		di_ok;
@@ -504,7 +506,7 @@ xfs_iformat(
 	switch (INT_GET(dip->di_core.di_aformat, ARCH_CONVERT)) {
 	case XFS_DINODE_FMT_LOCAL:
 		atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip);
-		size = (int)INT_GET(atp->hdr.totsize, ARCH_CONVERT);
+		size = be16_to_cpu(atp->hdr.totsize);
 		error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK, size);
 		break;
 	case XFS_DINODE_FMT_EXTENTS:
@@ -597,7 +599,6 @@ xfs_iformat_extents(
 	xfs_bmbt_rec_t	*ep, *dp;
 	xfs_ifork_t	*ifp;
 	int		nex;
-	int		real_size;
 	int		size;
 	int		i;
 
@@ -619,23 +620,20 @@ xfs_iformat_extents(
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 
-	real_size = 0;
+	ifp->if_real_bytes = 0;
 	if (nex == 0)
 		ifp->if_u1.if_extents = NULL;
 	else if (nex <= XFS_INLINE_EXTS)
 		ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
-	else {
-		ifp->if_u1.if_extents = kmem_alloc(size, KM_SLEEP);
-		ASSERT(ifp->if_u1.if_extents != NULL);
-		real_size = size;
-	}
+	else
+		xfs_iext_add(ifp, 0, nex);
+
 	ifp->if_bytes = size;
-	ifp->if_real_bytes = real_size;
 	if (size) {
 		dp = (xfs_bmbt_rec_t *) XFS_DFORK_PTR(dip, whichfork);
-		xfs_validate_extents(dp, nex, 1, XFS_EXTFMT_INODE(ip));
-		ep = ifp->if_u1.if_extents;
-		for (i = 0; i < nex; i++, ep++, dp++) {
+		xfs_validate_extents(ifp, nex, 1, XFS_EXTFMT_INODE(ip));
+		for (i = 0; i < nex; i++, dp++) {
+			ep = xfs_iext_get_ext(ifp, i);
 			ep->l0 = INT_GET(get_unaligned((__uint64_t*)&dp->l0),
 								ARCH_CONVERT);
 			ep->l1 = INT_GET(get_unaligned((__uint64_t*)&dp->l1),
@@ -646,7 +644,7 @@ xfs_iformat_extents(
 		if (whichfork != XFS_DATA_FORK ||
 			XFS_EXTFMT_INODE(ip) == XFS_EXTFMT_NOSTATE)
 				if (unlikely(xfs_check_nostate_extents(
-				    ifp->if_u1.if_extents, nex))) {
+				    ifp, 0, nex))) {
 					XFS_ERROR_REPORT("xfs_iformat_extents(2)",
 							 XFS_ERRLEVEL_LOW,
 							 ip->i_mount);
@@ -871,9 +869,8 @@ xfs_iread(
 	 * return NULL as well.  Set i_blkno to 0 so that xfs_itobp() will
 	 * know that this is a new incore inode.
 	 */
-	error = xfs_itobp(mp, tp, ip, &dip, &bp, bno);
-
-	if (error != 0) {
+	error = xfs_itobp(mp, tp, ip, &dip, &bp, bno, 0);
+	if (error) {
 		kmem_zone_free(xfs_inode_zone, ip);
 		return error;
 	}
@@ -1015,6 +1012,7 @@ xfs_iread_extents(
 {
 	int		error;
 	xfs_ifork_t	*ifp;
+	xfs_extnum_t	nextents;
 	size_t		size;
 
 	if (unlikely(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) {
@@ -1022,26 +1020,24 @@ xfs_iread_extents(
 				 ip->i_mount);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
-	size = XFS_IFORK_NEXTENTS(ip, whichfork) * (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = XFS_IFORK_NEXTENTS(ip, whichfork);
+	size = nextents * sizeof(xfs_bmbt_rec_t);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
+
 	/*
 	 * We know that the size is valid (it's checked in iformat_btree)
 	 */
-	ifp->if_u1.if_extents = kmem_alloc(size, KM_SLEEP);
-	ASSERT(ifp->if_u1.if_extents != NULL);
 	ifp->if_lastex = NULLEXTNUM;
-	ifp->if_bytes = ifp->if_real_bytes = (int)size;
+	ifp->if_bytes = ifp->if_real_bytes = 0;
 	ifp->if_flags |= XFS_IFEXTENTS;
+	xfs_iext_add(ifp, 0, nextents);
 	error = xfs_bmap_read_extents(tp, ip, whichfork);
 	if (error) {
-		kmem_free(ifp->if_u1.if_extents, size);
-		ifp->if_u1.if_extents = NULL;
-		ifp->if_bytes = ifp->if_real_bytes = 0;
+		xfs_iext_destroy(ifp);
 		ifp->if_flags &= ~XFS_IFEXTENTS;
 		return error;
 	}
-	xfs_validate_extents((xfs_bmbt_rec_t *)ifp->if_u1.if_extents,
-		XFS_IFORK_NEXTENTS(ip, whichfork), 0, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, nextents, 0, XFS_EXTFMT_INODE(ip));
 	return 0;
 }
 
@@ -1376,10 +1372,10 @@ xfs_itrunc_trace(
 		     (void*)(unsigned long)((toss_finish >> 32) & 0xffffffff),
 		     (void*)(unsigned long)(toss_finish & 0xffffffff),
 		     (void*)(unsigned long)current_cpu(),
-		     (void*)0,
-		     (void*)0,
-		     (void*)0,
-		     (void*)0);
+		     (void*)(unsigned long)current_pid(),
+		     (void*)NULL,
+		     (void*)NULL,
+		     (void*)NULL);
 }
 #else
 #define	xfs_itrunc_trace(tag, ip, flag, new_size, toss_start, toss_finish)
@@ -1397,6 +1393,16 @@ xfs_itrunc_trace(
  * calling into the buffer/page cache code and we can't hold the
  * inode lock when we do so.
  *
+ * We need to wait for any direct I/Os in flight to complete before we
+ * proceed with the truncate. This is needed to prevent the extents
+ * being read or written by the direct I/Os from being removed while the
+ * I/O is in flight as there is no other method of synchronising
+ * direct I/O with the truncate operation.  Also, because we hold
+ * the IOLOCK in exclusive mode, we prevent new direct I/Os from being
+ * started until the truncate completes and drops the lock. Essentially,
+ * the vn_iowait() call forms an I/O barrier that provides strict ordering
+ * between direct I/Os and the truncate operation.
+ *
  * The flags parameter can have either the value XFS_ITRUNC_DEFINITE
  * or XFS_ITRUNC_MAYBE.  The XFS_ITRUNC_MAYBE value should be used
  * in the case that the caller is locking things out of order and
@@ -1424,6 +1430,9 @@ xfs_itruncate_start(
 
 	mp = ip->i_mount;
 	vp = XFS_ITOV(ip);
+
+	vn_iowait(vp);  /* wait for the completion of any pending DIOs */
+	
 	/*
 	 * Call VOP_TOSS_PAGES() or VOP_FLUSHINVAL_PAGES() to get rid of pages and buffers
 	 * overlapping the region being removed.  We have to use
@@ -1899,7 +1908,7 @@ xfs_iunlink(
 		 * Here we put the head pointer into our next pointer,
 		 * and then we fall through to point the head at us.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			return error;
 		}
@@ -2008,7 +2017,7 @@ xfs_iunlink_remove(
 		 * of dealing with the buffer when there is no need to
 		 * change it.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -2070,7 +2079,7 @@ xfs_iunlink_remove(
 		 * Now last_ibp points to the buffer previous to us on
 		 * the unlinked list.  Pull us from the list.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -2476,92 +2485,6 @@ xfs_iroot_realloc(
 
 
 /*
- * This is called when the amount of space needed for if_extents
- * is increased or decreased.  The change in size is indicated by
- * the number of extents that need to be added or deleted in the
- * ext_diff parameter.
- *
- * If the amount of space needed has decreased below the size of the
- * inline buffer, then switch to using the inline buffer.  Otherwise,
- * use kmem_realloc() or kmem_alloc() to adjust the size of the buffer
- * to what is needed.
- *
- * ip -- the inode whose if_extents area is changing
- * ext_diff -- the change in the number of extents, positive or negative,
- *	 requested for the if_extents array.
- */
-void
-xfs_iext_realloc(
-	xfs_inode_t	*ip,
-	int		ext_diff,
-	int		whichfork)
-{
-	int		byte_diff;
-	xfs_ifork_t	*ifp;
-	int		new_size;
-	uint		rnew_size;
-
-	if (ext_diff == 0) {
-		return;
-	}
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	byte_diff = ext_diff * (uint)sizeof(xfs_bmbt_rec_t);
-	new_size = (int)ifp->if_bytes + byte_diff;
-	ASSERT(new_size >= 0);
-
-	if (new_size == 0) {
-		if (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext) {
-			ASSERT(ifp->if_real_bytes != 0);
-			kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
-		}
-		ifp->if_u1.if_extents = NULL;
-		rnew_size = 0;
-	} else if (new_size <= sizeof(ifp->if_u2.if_inline_ext)) {
-		/*
-		 * If the valid extents can fit in if_inline_ext,
-		 * copy them from the malloc'd vector and free it.
-		 */
-		if (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext) {
-			/*
-			 * For now, empty files are format EXTENTS,
-			 * so the if_extents pointer is null.
-			 */
-			if (ifp->if_u1.if_extents) {
-				memcpy(ifp->if_u2.if_inline_ext,
-					ifp->if_u1.if_extents, new_size);
-				kmem_free(ifp->if_u1.if_extents,
-					  ifp->if_real_bytes);
-			}
-			ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
-		}
-		rnew_size = 0;
-	} else {
-		rnew_size = new_size;
-		if ((rnew_size & (rnew_size - 1)) != 0)
-			rnew_size = xfs_iroundup(rnew_size);
-		/*
-		 * Stuck with malloc/realloc.
-		 */
-		if (ifp->if_u1.if_extents == ifp->if_u2.if_inline_ext) {
-			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
-				kmem_alloc(rnew_size, KM_SLEEP);
-			memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext,
-			      sizeof(ifp->if_u2.if_inline_ext));
-		} else if (rnew_size != ifp->if_real_bytes) {
-			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
-			  kmem_realloc(ifp->if_u1.if_extents,
-					rnew_size,
-					ifp->if_real_bytes,
-					KM_NOFS);
-		}
-	}
-	ifp->if_real_bytes = rnew_size;
-	ifp->if_bytes = new_size;
-}
-
-
-/*
  * This is called when the amount of space needed for if_data
  * is increased or decreased.  The change in size is indicated by
  * the number of bytes that need to be added or deleted in the
@@ -2720,12 +2643,11 @@ xfs_idestroy_fork(
 			ifp->if_real_bytes = 0;
 		}
 	} else if ((ifp->if_flags & XFS_IFEXTENTS) &&
-		   (ifp->if_u1.if_extents != NULL) &&
-		   (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext)) {
+		   ((ifp->if_flags & XFS_IFEXTIREC) ||
+		    ((ifp->if_u1.if_extents != NULL) &&
+		     (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext)))) {
 		ASSERT(ifp->if_real_bytes != 0);
-		kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
-		ifp->if_u1.if_extents = NULL;
-		ifp->if_real_bytes = 0;
+		xfs_iext_destroy(ifp);
 	}
 	ASSERT(ifp->if_u1.if_extents == NULL ||
 	       ifp->if_u1.if_extents == ifp->if_u2.if_inline_ext);
@@ -2814,7 +2736,7 @@ xfs_iunpin(
 
 		/* make sync come back and flush this inode */
 		if (vp) {
-			struct inode	*inode = LINVFS_GET_IP(vp);
+			struct inode	*inode = vn_to_inode(vp);
 
 			if (!(inode->i_state & I_NEW))
 				mark_inode_dirty_sync(inode);
@@ -2902,16 +2824,15 @@ xfs_iextents_copy(
 	 * the delayed ones.  There must be at least one
 	 * non-delayed extent.
 	 */
-	ep = ifp->if_u1.if_extents;
 	dest_ep = buffer;
 	copied = 0;
 	for (i = 0; i < nrecs; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		start_block = xfs_bmbt_get_startblock(ep);
 		if (ISNULLSTARTBLOCK(start_block)) {
 			/*
 			 * It's a delayed allocation extent, so skip it.
 			 */
-			ep++;
 			continue;
 		}
 
@@ -2921,11 +2842,10 @@ xfs_iextents_copy(
 		put_unaligned(INT_GET(ep->l1, ARCH_CONVERT),
 			      (__uint64_t*)&dest_ep->l1);
 		dest_ep++;
-		ep++;
 		copied++;
 	}
 	ASSERT(copied != 0);
-	xfs_validate_extents(buffer, copied, 1, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, copied, 1, XFS_EXTFMT_INODE(ip));
 
 	return (copied * (uint)sizeof(xfs_bmbt_rec_t));
 }
@@ -2995,8 +2915,10 @@ xfs_iflush_fork(
 	case XFS_DINODE_FMT_EXTENTS:
 		ASSERT((ifp->if_flags & XFS_IFEXTENTS) ||
 		       !(iip->ili_format.ilf_fields & extflag[whichfork]));
-		ASSERT((ifp->if_u1.if_extents != NULL) || (ifp->if_bytes == 0));
-		ASSERT((ifp->if_u1.if_extents == NULL) || (ifp->if_bytes > 0));
+		ASSERT((xfs_iext_get_ext(ifp, 0) != NULL) ||
+			(ifp->if_bytes == 0));
+		ASSERT((xfs_iext_get_ext(ifp, 0) == NULL) ||
+			(ifp->if_bytes > 0));
 		if ((iip->ili_format.ilf_fields & extflag[whichfork]) &&
 		    (ifp->if_bytes > 0)) {
 			ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) > 0);
@@ -3114,8 +3036,8 @@ xfs_iflush(
 	/*
 	 * Get the buffer containing the on-disk inode.
 	 */
-	error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0);
-	if (error != 0) {
+	error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0, 0);
+	if (error) {
 		xfs_ifunlock(ip);
 		return error;
 	}
@@ -3610,7 +3532,7 @@ xfs_iaccess(
 {
 	int		error;
 	mode_t		orgmode = mode;
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 
 	if (mode & S_IWUSR) {
 		umode_t		imode = inode->i_mode;
@@ -3704,3 +3626,1100 @@ xfs_ilock_trace(xfs_inode_t *ip, int lock, unsigned int lockflags, inst_t *ra)
 		     NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL);
 }
 #endif
+
+/*
+ * Return a pointer to the extent record at file index idx.
+ */
+xfs_bmbt_rec_t *
+xfs_iext_get_ext(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx)		/* index of target extent */
+{
+	ASSERT(idx >= 0);
+	if ((ifp->if_flags & XFS_IFEXTIREC) && (idx == 0)) {
+		return ifp->if_u1.if_ext_irec->er_extbuf;
+	} else if (ifp->if_flags & XFS_IFEXTIREC) {
+		xfs_ext_irec_t	*erp;		/* irec pointer */
+		int		erp_idx = 0;	/* irec index */
+		xfs_extnum_t	page_idx = idx;	/* ext index in target list */
+
+		erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 0);
+		return &erp->er_extbuf[page_idx];
+	} else if (ifp->if_bytes) {
+		return &ifp->if_u1.if_extents[idx];
+	} else {
+		return NULL;
+	}
+}
+
+/*
+ * Insert new item(s) into the extent records for incore inode
+ * fork 'ifp'.  'count' new items are inserted at index 'idx'.
+ */
+void
+xfs_iext_insert(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* starting index of new items */
+	xfs_extnum_t	count,		/* number of inserted items */
+	xfs_bmbt_irec_t	*new)		/* items to insert */
+{
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	i;		/* extent record index */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	xfs_iext_add(ifp, idx, count);
+	for (i = idx; i < idx + count; i++, new++) {
+		ep = xfs_iext_get_ext(ifp, i);
+		xfs_bmbt_set_all(ep, new);
+	}
+}
+
+/*
+ * This is called when the amount of space required for incore file
+ * extents needs to be increased. The ext_diff parameter stores the
+ * number of new extents being added and the idx parameter contains
+ * the extent index where the new extents will be added. If the new
+ * extents are being appended, then we just need to (re)allocate and
+ * initialize the space. Otherwise, if the new extents are being
+ * inserted into the middle of the existing entries, a bit more work
+ * is required to make room for the new extents to be inserted. The
+ * caller is responsible for filling in the new extent entries upon
+ * return.
+ */
+void
+xfs_iext_add(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin adding exts */
+	int		ext_diff)	/* nubmer of extents to add */
+{
+	int		byte_diff;	/* new bytes being added */
+	int		new_size;	/* size of extents after adding */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT((idx >= 0) && (idx <= nextents));
+	byte_diff = ext_diff * sizeof(xfs_bmbt_rec_t);
+	new_size = ifp->if_bytes + byte_diff;
+	/*
+	 * If the new number of extents (nextents + ext_diff)
+	 * fits inside the inode, then continue to use the inline
+	 * extent buffer.
+	 */
+	if (nextents + ext_diff <= XFS_INLINE_EXTS) {
+		if (idx < nextents) {
+			memmove(&ifp->if_u2.if_inline_ext[idx + ext_diff],
+				&ifp->if_u2.if_inline_ext[idx],
+				(nextents - idx) * sizeof(xfs_bmbt_rec_t));
+			memset(&ifp->if_u2.if_inline_ext[idx], 0, byte_diff);
+		}
+		ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
+		ifp->if_real_bytes = 0;
+		ifp->if_lastex = nextents + ext_diff;
+	}
+	/*
+	 * Otherwise use a linear (direct) extent list.
+	 * If the extents are currently inside the inode,
+	 * xfs_iext_realloc_direct will switch us from
+	 * inline to direct extent allocation mode.
+	 */
+	else if (nextents + ext_diff <= XFS_LINEAR_EXTS) {
+		xfs_iext_realloc_direct(ifp, new_size);
+		if (idx < nextents) {
+			memmove(&ifp->if_u1.if_extents[idx + ext_diff],
+				&ifp->if_u1.if_extents[idx],
+				(nextents - idx) * sizeof(xfs_bmbt_rec_t));
+			memset(&ifp->if_u1.if_extents[idx], 0, byte_diff);
+		}
+	}
+	/* Indirection array */
+	else {
+		xfs_ext_irec_t	*erp;
+		int		erp_idx = 0;
+		int		page_idx = idx;
+
+		ASSERT(nextents + ext_diff > XFS_LINEAR_EXTS);
+		if (ifp->if_flags & XFS_IFEXTIREC) {
+			erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 1);
+		} else {
+			xfs_iext_irec_init(ifp);
+			ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+			erp = ifp->if_u1.if_ext_irec;
+		}
+		/* Extents fit in target extent page */
+		if (erp && erp->er_extcount + ext_diff <= XFS_LINEAR_EXTS) {
+			if (page_idx < erp->er_extcount) {
+				memmove(&erp->er_extbuf[page_idx + ext_diff],
+					&erp->er_extbuf[page_idx],
+					(erp->er_extcount - page_idx) *
+					sizeof(xfs_bmbt_rec_t));
+				memset(&erp->er_extbuf[page_idx], 0, byte_diff);
+			}
+			erp->er_extcount += ext_diff;
+			xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		}
+		/* Insert a new extent page */
+		else if (erp) {
+			xfs_iext_add_indirect_multi(ifp,
+				erp_idx, page_idx, ext_diff);
+		}
+		/*
+		 * If extent(s) are being appended to the last page in
+		 * the indirection array and the new extent(s) don't fit
+		 * in the page, then erp is NULL and erp_idx is set to
+		 * the next index needed in the indirection array.
+		 */
+		else {
+			int	count = ext_diff;
+
+			while (count) {
+				erp = xfs_iext_irec_new(ifp, erp_idx);
+				erp->er_extcount = count;
+				count -= MIN(count, (int)XFS_LINEAR_EXTS);
+				if (count) {
+					erp_idx++;
+				}
+			}
+		}
+	}
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This is called when incore extents are being added to the indirection
+ * array and the new extents do not fit in the target extent list. The
+ * erp_idx parameter contains the irec index for the target extent list
+ * in the indirection array, and the idx parameter contains the extent
+ * index within the list. The number of extents being added is stored
+ * in the count parameter.
+ *
+ *    |-------|   |-------|
+ *    |       |   |       |    idx - number of extents before idx
+ *    |  idx  |   | count |
+ *    |       |   |       |    count - number of extents being inserted at idx
+ *    |-------|   |-------|
+ *    | count |   | nex2  |    nex2 - number of extents after idx + count
+ *    |-------|   |-------|
+ */
+void
+xfs_iext_add_indirect_multi(
+	xfs_ifork_t	*ifp,			/* inode fork pointer */
+	int		erp_idx,		/* target extent irec index */
+	xfs_extnum_t	idx,			/* index within target list */
+	int		count)			/* new extents being added */
+{
+	int		byte_diff;		/* new bytes being added */
+	xfs_ext_irec_t	*erp;			/* pointer to irec entry */
+	xfs_extnum_t	ext_diff;		/* number of extents to add */
+	xfs_extnum_t	ext_cnt;		/* new extents still needed */
+	xfs_extnum_t	nex2;			/* extents after idx + count */
+	xfs_bmbt_rec_t	*nex2_ep = NULL;	/* temp list for nex2 extents */
+	int		nlists;			/* number of irec's (lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	erp = &ifp->if_u1.if_ext_irec[erp_idx];
+	nex2 = erp->er_extcount - idx;
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+
+	/*
+	 * Save second part of target extent list
+	 * (all extents past */
+	if (nex2) {
+		byte_diff = nex2 * sizeof(xfs_bmbt_rec_t);
+		nex2_ep = (xfs_bmbt_rec_t *) kmem_alloc(byte_diff, KM_SLEEP);
+		memmove(nex2_ep, &erp->er_extbuf[idx], byte_diff);
+		erp->er_extcount -= nex2;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -nex2);
+		memset(&erp->er_extbuf[idx], 0, byte_diff);
+	}
+
+	/*
+	 * Add the new extents to the end of the target
+	 * list, then allocate new irec record(s) and
+	 * extent buffer(s) as needed to store the rest
+	 * of the new extents.
+	 */
+	ext_cnt = count;
+	ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS - erp->er_extcount);
+	if (ext_diff) {
+		erp->er_extcount += ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		ext_cnt -= ext_diff;
+	}
+	while (ext_cnt) {
+		erp_idx++;
+		erp = xfs_iext_irec_new(ifp, erp_idx);
+		ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS);
+		erp->er_extcount = ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		ext_cnt -= ext_diff;
+	}
+
+	/* Add nex2 extents back to indirection array */
+	if (nex2) {
+		xfs_extnum_t	ext_avail;
+		int		i;
+
+		byte_diff = nex2 * sizeof(xfs_bmbt_rec_t);
+		ext_avail = XFS_LINEAR_EXTS - erp->er_extcount;
+		i = 0;
+		/*
+		 * If nex2 extents fit in the current page, append
+		 * nex2_ep after the new extents.
+		 */
+		if (nex2 <= ext_avail) {
+			i = erp->er_extcount;
+		}
+		/*
+		 * Otherwise, check if space is available in the
+		 * next page.
+		 */
+		else if ((erp_idx < nlists - 1) &&
+			 (nex2 <= (ext_avail = XFS_LINEAR_EXTS -
+			  ifp->if_u1.if_ext_irec[erp_idx+1].er_extcount))) {
+			erp_idx++;
+			erp++;
+			/* Create a hole for nex2 extents */
+			memmove(&erp->er_extbuf[nex2], erp->er_extbuf,
+				erp->er_extcount * sizeof(xfs_bmbt_rec_t));
+		}
+		/*
+		 * Final choice, create a new extent page for
+		 * nex2 extents.
+		 */
+		else {
+			erp_idx++;
+			erp = xfs_iext_irec_new(ifp, erp_idx);
+		}
+		memmove(&erp->er_extbuf[i], nex2_ep, byte_diff);
+		kmem_free(nex2_ep, byte_diff);
+		erp->er_extcount += nex2;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, nex2);
+	}
+}
+
+/*
+ * This is called when the amount of space required for incore file
+ * extents needs to be decreased. The ext_diff parameter stores the
+ * number of extents to be removed and the idx parameter contains
+ * the extent index where the extents will be removed from.
+ *
+ * If the amount of space needed has decreased below the linear
+ * limit, XFS_IEXT_BUFSZ, then switch to using the contiguous
+ * extent array.  Otherwise, use kmem_realloc() to adjust the
+ * size to what is needed.
+ */
+void
+xfs_iext_remove(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		new_size;	/* size of extents after removal */
+
+	ASSERT(ext_diff > 0);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	new_size = (nextents - ext_diff) * sizeof(xfs_bmbt_rec_t);
+
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	} else if (ifp->if_flags & XFS_IFEXTIREC) {
+		xfs_iext_remove_indirect(ifp, idx, ext_diff);
+	} else if (ifp->if_real_bytes) {
+		xfs_iext_remove_direct(ifp, idx, ext_diff);
+	} else {
+		xfs_iext_remove_inline(ifp, idx, ext_diff);
+	}
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This removes ext_diff extents from the inline buffer, beginning
+ * at extent index idx.
+ */
+void
+xfs_iext_remove_inline(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	int		nextents;	/* number of extents in file */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	ASSERT(idx < XFS_INLINE_EXTS);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(((nextents - ext_diff) > 0) &&
+		(nextents - ext_diff) < XFS_INLINE_EXTS);
+
+	if (idx + ext_diff < nextents) {
+		memmove(&ifp->if_u2.if_inline_ext[idx],
+			&ifp->if_u2.if_inline_ext[idx + ext_diff],
+			(nextents - (idx + ext_diff)) *
+			 sizeof(xfs_bmbt_rec_t));
+		memset(&ifp->if_u2.if_inline_ext[nextents - ext_diff],
+			0, ext_diff * sizeof(xfs_bmbt_rec_t));
+	} else {
+		memset(&ifp->if_u2.if_inline_ext[idx], 0,
+			ext_diff * sizeof(xfs_bmbt_rec_t));
+	}
+}
+
+/*
+ * This removes ext_diff extents from a linear (direct) extent list,
+ * beginning at extent index idx. If the extents are being removed
+ * from the end of the list (ie. truncate) then we just need to re-
+ * allocate the list to remove the extra space. Otherwise, if the
+ * extents are being removed from the middle of the existing extent
+ * entries, then we first need to move the extent records beginning
+ * at idx + ext_diff up in the list to overwrite the records being
+ * removed, then remove the extra space via kmem_realloc.
+ */
+void
+xfs_iext_remove_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		new_size;	/* size of extents after removal */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	new_size = ifp->if_bytes -
+		(ext_diff * sizeof(xfs_bmbt_rec_t));
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+		return;
+	}
+	/* Move extents up in the list (if needed) */
+	if (idx + ext_diff < nextents) {
+		memmove(&ifp->if_u1.if_extents[idx],
+			&ifp->if_u1.if_extents[idx + ext_diff],
+			(nextents - (idx + ext_diff)) *
+			 sizeof(xfs_bmbt_rec_t));
+	}
+	memset(&ifp->if_u1.if_extents[nextents - ext_diff],
+		0, ext_diff * sizeof(xfs_bmbt_rec_t));
+	/*
+	 * Reallocate the direct extent list. If the extents
+	 * will fit inside the inode then xfs_iext_realloc_direct
+	 * will switch from direct to inline extent allocation
+	 * mode for us.
+	 */
+	xfs_iext_realloc_direct(ifp, new_size);
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This is called when incore extents are being removed from the
+ * indirection array and the extents being removed span multiple extent
+ * buffers. The idx parameter contains the file extent index where we
+ * want to begin removing extents, and the count parameter contains
+ * how many extents need to be removed.
+ *
+ *    |-------|   |-------|
+ *    | nex1  |   |       |    nex1 - number of extents before idx
+ *    |-------|   | count |
+ *    |       |   |       |    count - number of extents being removed at idx
+ *    | count |   |-------|
+ *    |       |   | nex2  |    nex2 - number of extents after idx + count
+ *    |-------|   |-------|
+ */
+void
+xfs_iext_remove_indirect(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing extents */
+	int		count)		/* number of extents to remove */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		erp_idx = 0;	/* indirection array index */
+	xfs_extnum_t	ext_cnt;	/* extents left to remove */
+	xfs_extnum_t	ext_diff;	/* extents to remove in current list */
+	xfs_extnum_t	nex1;		/* number of extents before idx */
+	xfs_extnum_t	nex2;		/* extents after idx + count */
+	int		nlists;		/* entries in indirecton array */
+	int		page_idx = idx;	/* index in target extent list */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	erp = xfs_iext_idx_to_irec(ifp,  &page_idx, &erp_idx, 0);
+	ASSERT(erp != NULL);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	nex1 = page_idx;
+	ext_cnt = count;
+	while (ext_cnt) {
+		nex2 = MAX((erp->er_extcount - (nex1 + ext_cnt)), 0);
+		ext_diff = MIN(ext_cnt, (erp->er_extcount - nex1));
+		/*
+		 * Check for deletion of entire list;
+		 * xfs_iext_irec_remove() updates extent offsets.
+		 */
+		if (ext_diff == erp->er_extcount) {
+			xfs_iext_irec_remove(ifp, erp_idx);
+			ext_cnt -= ext_diff;
+			nex1 = 0;
+			if (ext_cnt) {
+				ASSERT(erp_idx < ifp->if_real_bytes /
+					XFS_IEXT_BUFSZ);
+				erp = &ifp->if_u1.if_ext_irec[erp_idx];
+				nex1 = 0;
+				continue;
+			} else {
+				break;
+			}
+		}
+		/* Move extents up (if needed) */
+		if (nex2) {
+			memmove(&erp->er_extbuf[nex1],
+				&erp->er_extbuf[nex1 + ext_diff],
+				nex2 * sizeof(xfs_bmbt_rec_t));
+		}
+		/* Zero out rest of page */
+		memset(&erp->er_extbuf[nex1 + nex2], 0, (XFS_IEXT_BUFSZ -
+			((nex1 + nex2) * sizeof(xfs_bmbt_rec_t))));
+		/* Update remaining counters */
+		erp->er_extcount -= ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -ext_diff);
+		ext_cnt -= ext_diff;
+		nex1 = 0;
+		erp_idx++;
+		erp++;
+	}
+	ifp->if_bytes -= count * sizeof(xfs_bmbt_rec_t);
+	xfs_iext_irec_compact(ifp);
+}
+
+/*
+ * Create, destroy, or resize a linear (direct) block of extents.
+ */
+void
+xfs_iext_realloc_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* new size of extents */
+{
+	int		rnew_size;	/* real new size of extents */
+
+	rnew_size = new_size;
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC) ||
+		((new_size >= 0) && (new_size <= XFS_IEXT_BUFSZ) &&
+		 (new_size != ifp->if_real_bytes)));
+
+	/* Free extent records */
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	}
+	/* Resize direct extent list and zero any new bytes */
+	else if (ifp->if_real_bytes) {
+		/* Check if extents will fit inside the inode */
+		if (new_size <= XFS_INLINE_EXTS * sizeof(xfs_bmbt_rec_t)) {
+			xfs_iext_direct_to_inline(ifp, new_size /
+				(uint)sizeof(xfs_bmbt_rec_t));
+			ifp->if_bytes = new_size;
+			return;
+		}
+		if ((new_size & (new_size - 1)) != 0) {
+			rnew_size = xfs_iroundup(new_size);
+		}
+		if (rnew_size != ifp->if_real_bytes) {
+			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+				kmem_realloc(ifp->if_u1.if_extents,
+						rnew_size,
+						ifp->if_real_bytes,
+						KM_SLEEP);
+		}
+		if (rnew_size > ifp->if_real_bytes) {
+			memset(&ifp->if_u1.if_extents[ifp->if_bytes /
+				(uint)sizeof(xfs_bmbt_rec_t)], 0,
+				rnew_size - ifp->if_real_bytes);
+		}
+	}
+	/*
+	 * Switch from the inline extent buffer to a direct
+	 * extent list. Be sure to include the inline extent
+	 * bytes in new_size.
+	 */
+	else {
+		new_size += ifp->if_bytes;
+		if ((new_size & (new_size - 1)) != 0) {
+			rnew_size = xfs_iroundup(new_size);
+		}
+		xfs_iext_inline_to_direct(ifp, rnew_size);
+	}
+	ifp->if_real_bytes = rnew_size;
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * Switch from linear (direct) extent records to inline buffer.
+ */
+void
+xfs_iext_direct_to_inline(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	nextents)	/* number of extents in file */
+{
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	ASSERT(nextents <= XFS_INLINE_EXTS);
+	/*
+	 * The inline buffer was zeroed when we switched
+	 * from inline to direct extent allocation mode,
+	 * so we don't need to clear it here.
+	 */
+	memcpy(ifp->if_u2.if_inline_ext, ifp->if_u1.if_extents,
+		nextents * sizeof(xfs_bmbt_rec_t));
+	kmem_free(ifp->if_u1.if_extents, KM_SLEEP);
+	ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
+	ifp->if_real_bytes = 0;
+}
+
+/*
+ * Switch from inline buffer to linear (direct) extent records.
+ * new_size should already be rounded up to the next power of 2
+ * by the caller (when appropriate), so use new_size as it is.
+ * However, since new_size may be rounded up, we can't update
+ * if_bytes here. It is the caller's responsibility to update
+ * if_bytes upon return.
+ */
+void
+xfs_iext_inline_to_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* number of extents in file */
+{
+	ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+		kmem_alloc(new_size, KM_SLEEP);
+	memset(ifp->if_u1.if_extents, 0, new_size);
+	if (ifp->if_bytes) {
+		memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext,
+			ifp->if_bytes);
+		memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS *
+			sizeof(xfs_bmbt_rec_t));
+	}
+	ifp->if_real_bytes = new_size;
+}
+
+/*
+ * Resize an extent indirection array to new_size bytes.
+ */
+void
+xfs_iext_realloc_indirect(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* new indirection array size */
+{
+	int		nlists;		/* number of irec's (ex lists) */
+	int		size;		/* current indirection array size */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	size = nlists * sizeof(xfs_ext_irec_t);
+	ASSERT(ifp->if_real_bytes);
+	ASSERT((new_size >= 0) && (new_size != size));
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	} else {
+		ifp->if_u1.if_ext_irec = (xfs_ext_irec_t *)
+			kmem_realloc(ifp->if_u1.if_ext_irec,
+				new_size, size, KM_SLEEP);
+	}
+}
+
+/*
+ * Switch from indirection array to linear (direct) extent allocations.
+ */
+void
+xfs_iext_indirect_to_direct(
+	 xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		size;		/* size of file extents */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(nextents <= XFS_LINEAR_EXTS);
+	size = nextents * sizeof(xfs_bmbt_rec_t);
+
+	xfs_iext_irec_compact_full(ifp);
+	ASSERT(ifp->if_real_bytes == XFS_IEXT_BUFSZ);
+
+	ep = ifp->if_u1.if_ext_irec->er_extbuf;
+	kmem_free(ifp->if_u1.if_ext_irec, sizeof(xfs_ext_irec_t));
+	ifp->if_flags &= ~XFS_IFEXTIREC;
+	ifp->if_u1.if_extents = ep;
+	ifp->if_bytes = size;
+	if (nextents < XFS_LINEAR_EXTS) {
+		xfs_iext_realloc_direct(ifp, size);
+	}
+}
+
+/*
+ * Free incore file extents.
+ */
+void
+xfs_iext_destroy(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		int	erp_idx;
+		int	nlists;
+
+		nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		for (erp_idx = nlists - 1; erp_idx >= 0 ; erp_idx--) {
+			xfs_iext_irec_remove(ifp, erp_idx);
+		}
+		ifp->if_flags &= ~XFS_IFEXTIREC;
+	} else if (ifp->if_real_bytes) {
+		kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
+	} else if (ifp->if_bytes) {
+		memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS *
+			sizeof(xfs_bmbt_rec_t));
+	}
+	ifp->if_u1.if_extents = NULL;
+	ifp->if_real_bytes = 0;
+	ifp->if_bytes = 0;
+}
+
+/*
+ * Return a pointer to the extent record for file system block bno.
+ */
+xfs_bmbt_rec_t *			/* pointer to found extent record */
+xfs_iext_bno_to_ext(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number to search for */
+	xfs_extnum_t	*idxp)		/* index of target extent */
+{
+	xfs_bmbt_rec_t	*base;		/* pointer to first extent */
+	xfs_filblks_t	blockcount = 0;	/* number of blocks in extent */
+	xfs_bmbt_rec_t	*ep = NULL;	/* pointer to target extent */
+	xfs_ext_irec_t	*erp = NULL;	/* indirection array pointer */
+	int		high;		/* upper boundry in search */
+	xfs_extnum_t	idx = 0;	/* index of target extent */
+	int		low;		/* lower boundry in search */
+	xfs_extnum_t	nextents;	/* number of file extents */
+	xfs_fileoff_t	startoff = 0;	/* start offset of extent */
+
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	if (nextents == 0) {
+		*idxp = 0;
+		return NULL;
+	}
+	low = 0;
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		/* Find target extent list */
+		int	erp_idx = 0;
+		erp = xfs_iext_bno_to_irec(ifp, bno, &erp_idx);
+		base = erp->er_extbuf;
+		high = erp->er_extcount - 1;
+	} else {
+		base = ifp->if_u1.if_extents;
+		high = nextents - 1;
+	}
+	/* Binary search extent records */
+	while (low <= high) {
+		idx = (low + high) >> 1;
+		ep = base + idx;
+		startoff = xfs_bmbt_get_startoff(ep);
+		blockcount = xfs_bmbt_get_blockcount(ep);
+		if (bno < startoff) {
+			high = idx - 1;
+		} else if (bno >= startoff + blockcount) {
+			low = idx + 1;
+		} else {
+			/* Convert back to file-based extent index */
+			if (ifp->if_flags & XFS_IFEXTIREC) {
+				idx += erp->er_extoff;
+			}
+			*idxp = idx;
+			return ep;
+		}
+	}
+	/* Convert back to file-based extent index */
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		idx += erp->er_extoff;
+	}
+	if (bno >= startoff + blockcount) {
+		if (++idx == nextents) {
+			ep = NULL;
+		} else {
+			ep = xfs_iext_get_ext(ifp, idx);
+		}
+	}
+	*idxp = idx;
+	return ep;
+}
+
+/*
+ * Return a pointer to the indirection array entry containing the
+ * extent record for filesystem block bno. Store the index of the
+ * target irec in *erp_idxp.
+ */
+xfs_ext_irec_t *			/* pointer to found extent record */
+xfs_iext_bno_to_irec(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number to search for */
+	int		*erp_idxp)	/* irec index of target ext list */
+{
+	xfs_ext_irec_t	*erp = NULL;	/* indirection array pointer */
+	xfs_ext_irec_t	*erp_next;	/* next indirection array entry */
+	int		erp_idx;	/* indirection array index */
+	int		nlists;		/* number of extent irec's (lists) */
+	int		high;		/* binary search upper limit */
+	int		low;		/* binary search lower limit */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp_idx = 0;
+	low = 0;
+	high = nlists - 1;
+	while (low <= high) {
+		erp_idx = (low + high) >> 1;
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		erp_next = erp_idx < nlists - 1 ? erp + 1 : NULL;
+		if (bno < xfs_bmbt_get_startoff(erp->er_extbuf)) {
+			high = erp_idx - 1;
+		} else if (erp_next && bno >=
+			   xfs_bmbt_get_startoff(erp_next->er_extbuf)) {
+			low = erp_idx + 1;
+		} else {
+			break;
+		}
+	}
+	*erp_idxp = erp_idx;
+	return erp;
+}
+
+/*
+ * Return a pointer to the indirection array entry containing the
+ * extent record at file extent index *idxp. Store the index of the
+ * target irec in *erp_idxp and store the page index of the target
+ * extent record in *idxp.
+ */
+xfs_ext_irec_t *
+xfs_iext_idx_to_irec(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	*idxp,		/* extent index (file -> page) */
+	int		*erp_idxp,	/* pointer to target irec */
+	int		realloc)	/* new bytes were just added */
+{
+	xfs_ext_irec_t	*prev;		/* pointer to previous irec */
+	xfs_ext_irec_t	*erp = NULL;	/* pointer to current irec */
+	int		erp_idx;	/* indirection array index */
+	int		nlists;		/* number of irec's (ex lists) */
+	int		high;		/* binary search upper limit */
+	int		low;		/* binary search lower limit */
+	xfs_extnum_t	page_idx = *idxp; /* extent index in target list */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	ASSERT(page_idx >= 0 && page_idx <=
+		ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp_idx = 0;
+	low = 0;
+	high = nlists - 1;
+
+	/* Binary search extent irec's */
+	while (low <= high) {
+		erp_idx = (low + high) >> 1;
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		prev = erp_idx > 0 ? erp - 1 : NULL;
+		if (page_idx < erp->er_extoff || (page_idx == erp->er_extoff &&
+		     realloc && prev && prev->er_extcount < XFS_LINEAR_EXTS)) {
+			high = erp_idx - 1;
+		} else if (page_idx > erp->er_extoff + erp->er_extcount ||
+			   (page_idx == erp->er_extoff + erp->er_extcount &&
+			    !realloc)) {
+			low = erp_idx + 1;
+		} else if (page_idx == erp->er_extoff + erp->er_extcount &&
+			   erp->er_extcount == XFS_LINEAR_EXTS) {
+			ASSERT(realloc);
+			page_idx = 0;
+			erp_idx++;
+			erp = erp_idx < nlists ? erp + 1 : NULL;
+			break;
+		} else {
+			page_idx -= erp->er_extoff;
+			break;
+		}
+	}
+	*idxp = page_idx;
+	*erp_idxp = erp_idx;
+	return(erp);
+}
+
+/*
+ * Allocate and initialize an indirection array once the space needed
+ * for incore extents increases above XFS_IEXT_BUFSZ.
+ */
+void
+xfs_iext_irec_init(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(nextents <= XFS_LINEAR_EXTS);
+
+	erp = (xfs_ext_irec_t *)
+		kmem_alloc(sizeof(xfs_ext_irec_t), KM_SLEEP);
+
+	if (nextents == 0) {
+		ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+			kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	} else if (!ifp->if_real_bytes) {
+		xfs_iext_inline_to_direct(ifp, XFS_IEXT_BUFSZ);
+	} else if (ifp->if_real_bytes < XFS_IEXT_BUFSZ) {
+		xfs_iext_realloc_direct(ifp, XFS_IEXT_BUFSZ);
+	}
+	erp->er_extbuf = ifp->if_u1.if_extents;
+	erp->er_extcount = nextents;
+	erp->er_extoff = 0;
+
+	ifp->if_flags |= XFS_IFEXTIREC;
+	ifp->if_real_bytes = XFS_IEXT_BUFSZ;
+	ifp->if_bytes = nextents * sizeof(xfs_bmbt_rec_t);
+	ifp->if_u1.if_ext_irec = erp;
+
+	return;
+}
+
+/*
+ * Allocate and initialize a new entry in the indirection array.
+ */
+xfs_ext_irec_t *
+xfs_iext_irec_new(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx)	/* index for new irec */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+
+	/* Resize indirection array */
+	xfs_iext_realloc_indirect(ifp, ++nlists *
+				  sizeof(xfs_ext_irec_t));
+	/*
+	 * Move records down in the array so the
+	 * new page can use erp_idx.
+	 */
+	erp = ifp->if_u1.if_ext_irec;
+	for (i = nlists - 1; i > erp_idx; i--) {
+		memmove(&erp[i], &erp[i-1], sizeof(xfs_ext_irec_t));
+	}
+	ASSERT(i == erp_idx);
+
+	/* Initialize new extent record */
+	erp = ifp->if_u1.if_ext_irec;
+	erp[erp_idx].er_extbuf = (xfs_bmbt_rec_t *)
+		kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
+	memset(erp[erp_idx].er_extbuf, 0, XFS_IEXT_BUFSZ);
+	erp[erp_idx].er_extcount = 0;
+	erp[erp_idx].er_extoff = erp_idx > 0 ?
+		erp[erp_idx-1].er_extoff + erp[erp_idx-1].er_extcount : 0;
+	return (&erp[erp_idx]);
+}
+
+/*
+ * Remove a record from the indirection array.
+ */
+void
+xfs_iext_irec_remove(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx)	/* irec index to remove */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp = &ifp->if_u1.if_ext_irec[erp_idx];
+	if (erp->er_extbuf) {
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1,
+			-erp->er_extcount);
+		kmem_free(erp->er_extbuf, XFS_IEXT_BUFSZ);
+	}
+	/* Compact extent records */
+	erp = ifp->if_u1.if_ext_irec;
+	for (i = erp_idx; i < nlists - 1; i++) {
+		memmove(&erp[i], &erp[i+1], sizeof(xfs_ext_irec_t));
+	}
+	/*
+	 * Manually free the last extent record from the indirection
+	 * array.  A call to xfs_iext_realloc_indirect() with a size
+	 * of zero would result in a call to xfs_iext_destroy() which
+	 * would in turn call this function again, creating a nasty
+	 * infinite loop.
+	 */
+	if (--nlists) {
+		xfs_iext_realloc_indirect(ifp,
+			nlists * sizeof(xfs_ext_irec_t));
+	} else {
+		kmem_free(ifp->if_u1.if_ext_irec,
+			sizeof(xfs_ext_irec_t));
+	}
+	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
+}
+
+/*
+ * This is called to clean up large amounts of unused memory allocated
+ * by the indirection array.  Before compacting anything though, verify
+ * that the indirection array is still needed and switch back to the
+ * linear extent list (or even the inline buffer) if possible.  The
+ * compaction policy is as follows:
+ *
+ *    Full Compaction: Extents fit into a single page (or inline buffer)
+ *    Full Compaction: Extents occupy less than 10% of allocated space
+ * Partial Compaction: Extents occupy > 10% and < 50% of allocated space
+ *      No Compaction: Extents occupy at least 50% of allocated space
+ */
+void
+xfs_iext_irec_compact(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+
+	if (nextents == 0) {
+		xfs_iext_destroy(ifp);
+	} else if (nextents <= XFS_INLINE_EXTS) {
+		xfs_iext_indirect_to_direct(ifp);
+		xfs_iext_direct_to_inline(ifp, nextents);
+	} else if (nextents <= XFS_LINEAR_EXTS) {
+		xfs_iext_indirect_to_direct(ifp);
+	} else if (nextents < (nlists * XFS_LINEAR_EXTS) >> 3) {
+		xfs_iext_irec_compact_full(ifp);
+	} else if (nextents < (nlists * XFS_LINEAR_EXTS) >> 1) {
+		xfs_iext_irec_compact_pages(ifp);
+	}
+}
+
+/*
+ * Combine extents from neighboring extent pages.
+ */
+void
+xfs_iext_irec_compact_pages(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_ext_irec_t	*erp, *erp_next;/* pointers to irec entries */
+	int		erp_idx = 0;	/* indirection array index */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	while (erp_idx < nlists - 1) {
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		erp_next = erp + 1;
+		if (erp_next->er_extcount <=
+		    (XFS_LINEAR_EXTS - erp->er_extcount)) {
+			memmove(&erp->er_extbuf[erp->er_extcount],
+				erp_next->er_extbuf, erp_next->er_extcount *
+				sizeof(xfs_bmbt_rec_t));
+			erp->er_extcount += erp_next->er_extcount;
+			/*
+			 * Free page before removing extent record
+			 * so er_extoffs don't get modified in
+			 * xfs_iext_irec_remove.
+			 */
+			kmem_free(erp_next->er_extbuf, XFS_IEXT_BUFSZ);
+			erp_next->er_extbuf = NULL;
+			xfs_iext_irec_remove(ifp, erp_idx + 1);
+			nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		} else {
+			erp_idx++;
+		}
+	}
+}
+
+/*
+ * Fully compact the extent records managed by the indirection array.
+ */
+void
+xfs_iext_irec_compact_full(
+	xfs_ifork_t	*ifp)			/* inode fork pointer */
+{
+	xfs_bmbt_rec_t	*ep, *ep_next;		/* extent record pointers */
+	xfs_ext_irec_t	*erp, *erp_next;	/* extent irec pointers */
+	int		erp_idx = 0;		/* extent irec index */
+	int		ext_avail;		/* empty entries in ex list */
+	int		ext_diff;		/* number of exts to add */
+	int		nlists;			/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp = ifp->if_u1.if_ext_irec;
+	ep = &erp->er_extbuf[erp->er_extcount];
+	erp_next = erp + 1;
+	ep_next = erp_next->er_extbuf;
+	while (erp_idx < nlists - 1) {
+		ext_avail = XFS_LINEAR_EXTS - erp->er_extcount;
+		ext_diff = MIN(ext_avail, erp_next->er_extcount);
+		memcpy(ep, ep_next, ext_diff * sizeof(xfs_bmbt_rec_t));
+		erp->er_extcount += ext_diff;
+		erp_next->er_extcount -= ext_diff;
+		/* Remove next page */
+		if (erp_next->er_extcount == 0) {
+			/*
+			 * Free page before removing extent record
+			 * so er_extoffs don't get modified in
+			 * xfs_iext_irec_remove.
+			 */
+			kmem_free(erp_next->er_extbuf,
+				erp_next->er_extcount * sizeof(xfs_bmbt_rec_t));
+			erp_next->er_extbuf = NULL;
+			xfs_iext_irec_remove(ifp, erp_idx + 1);
+			erp = &ifp->if_u1.if_ext_irec[erp_idx];
+			nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		/* Update next page */
+		} else {
+			/* Move rest of page up to become next new page */
+			memmove(erp_next->er_extbuf, ep_next,
+				erp_next->er_extcount * sizeof(xfs_bmbt_rec_t));
+			ep_next = erp_next->er_extbuf;
+			memset(&ep_next[erp_next->er_extcount], 0,
+				(XFS_LINEAR_EXTS - erp_next->er_extcount) *
+				sizeof(xfs_bmbt_rec_t));
+		}
+		if (erp->er_extcount == XFS_LINEAR_EXTS) {
+			erp_idx++;
+			if (erp_idx < nlists)
+				erp = &ifp->if_u1.if_ext_irec[erp_idx];
+			else
+				break;
+		}
+		ep = &erp->er_extbuf[erp->er_extcount];
+		erp_next = erp + 1;
+		ep_next = erp_next->er_extbuf;
+	}
+}
+
+/*
+ * This is called to update the er_extoff field in the indirection
+ * array when extents have been added or removed from one of the
+ * extent lists. erp_idx contains the irec index to begin updating
+ * at and ext_diff contains the number of extents that were added
+ * or removed.
+ */
+void
+xfs_iext_irec_update_extoffs(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx,	/* irec index to update */
+	int		ext_diff)	/* number of new extents */
+{
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	for (i = erp_idx; i < nlists; i++) {
+		ifp->if_u1.if_ext_irec[i].er_extoff += ext_diff;
+	}
+}
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1cfbcf18ce86..39ef9c36ea55 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -25,10 +25,37 @@
 #define	XFS_ATTR_FORK	1
 
 /*
+ * The following xfs_ext_irec_t struct introduces a second (top) level
+ * to the in-core extent allocation scheme. These structs are allocated
+ * in a contiguous block, creating an indirection array where each entry
+ * (irec) contains a pointer to a buffer of in-core extent records which
+ * it manages. Each extent buffer is 4k in size, since 4k is the system
+ * page size on Linux i386 and systems with larger page sizes don't seem
+ * to gain much, if anything, by using their native page size as the
+ * extent buffer size. Also, using 4k extent buffers everywhere provides
+ * a consistent interface for CXFS across different platforms.
+ *
+ * There is currently no limit on the number of irec's (extent lists)
+ * allowed, so heavily fragmented files may require an indirection array
+ * which spans multiple system pages of memory. The number of extents
+ * which would require this amount of contiguous memory is very large
+ * and should not cause problems in the foreseeable future. However,
+ * if the memory needed for the contiguous array ever becomes a problem,
+ * it is possible that a third level of indirection may be required.
+ */
+typedef struct xfs_ext_irec {
+	xfs_bmbt_rec_t	*er_extbuf;	/* block of extent records */
+	xfs_extnum_t	er_extoff;	/* extent offset in file */
+	xfs_extnum_t	er_extcount;	/* number of extents in page/block */
+} xfs_ext_irec_t;
+
+/*
  * File incore extent information, present for each of data & attr forks.
  */
-#define	XFS_INLINE_EXTS	2
-#define	XFS_INLINE_DATA	32
+#define	XFS_IEXT_BUFSZ		4096
+#define	XFS_LINEAR_EXTS		(XFS_IEXT_BUFSZ / (uint)sizeof(xfs_bmbt_rec_t))
+#define	XFS_INLINE_EXTS		2
+#define	XFS_INLINE_DATA		32
 typedef struct xfs_ifork {
 	int			if_bytes;	/* bytes in if_u1 */
 	int			if_real_bytes;	/* bytes allocated in if_u1 */
@@ -39,6 +66,7 @@ typedef struct xfs_ifork {
 	xfs_extnum_t		if_lastex;	/* last if_extents used */
 	union {
 		xfs_bmbt_rec_t	*if_extents;	/* linear map file exts */
+		xfs_ext_irec_t	*if_ext_irec;	/* irec map file exts */
 		char		*if_data;	/* inline file data */
 	} if_u1;
 	union {
@@ -61,20 +89,16 @@ typedef struct xfs_ifork {
 /*
  * Per-fork incore inode flags.
  */
-#define	XFS_IFINLINE	0x0001	/* Inline data is read in */
-#define	XFS_IFEXTENTS	0x0002	/* All extent pointers are read in */
-#define	XFS_IFBROOT	0x0004	/* i_broot points to the bmap b-tree root */
+#define	XFS_IFINLINE	0x01	/* Inline data is read in */
+#define	XFS_IFEXTENTS	0x02	/* All extent pointers are read in */
+#define	XFS_IFBROOT	0x04	/* i_broot points to the bmap b-tree root */
+#define	XFS_IFEXTIREC	0x08	/* Indirection array of extent blocks */
 
 /*
- * Flags for xfs_imap() and xfs_dilocate().
+ * Flags for xfs_itobp(), xfs_imap() and xfs_dilocate().
  */
-#define	XFS_IMAP_LOOKUP		0x1
-
-/*
- * Maximum number of extent pointers in if_u1.if_extents.
- */
-#define	XFS_MAX_INCORE_EXTENTS	32768
-
+#define XFS_IMAP_LOOKUP		0x1
+#define XFS_IMAP_BULKSTAT	0x2
 
 #ifdef __KERNEL__
 struct bhv_desc;
@@ -398,7 +422,7 @@ int		xfs_finish_reclaim_all(struct xfs_mount *, int);
  */
 int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
 			  xfs_inode_t *, xfs_dinode_t **, struct xfs_buf **,
-			  xfs_daddr_t);
+			  xfs_daddr_t, uint);
 int		xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			  xfs_inode_t **, xfs_daddr_t);
 int		xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int);
@@ -440,6 +464,32 @@ xfs_inode_t	*xfs_vtoi(struct vnode *vp);
 
 void		xfs_synchronize_atime(xfs_inode_t *);
 
+xfs_bmbt_rec_t	*xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
+void		xfs_iext_insert(xfs_ifork_t *, xfs_extnum_t, xfs_extnum_t,
+				xfs_bmbt_irec_t *);
+void		xfs_iext_add(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_add_indirect_multi(xfs_ifork_t *, int, xfs_extnum_t, int);
+void		xfs_iext_remove(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_realloc_direct(xfs_ifork_t *, int);
+void		xfs_iext_realloc_indirect(xfs_ifork_t *, int);
+void		xfs_iext_indirect_to_direct(xfs_ifork_t *);
+void		xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
+void		xfs_iext_inline_to_direct(xfs_ifork_t *, int);
+void		xfs_iext_destroy(xfs_ifork_t *);
+xfs_bmbt_rec_t	*xfs_iext_bno_to_ext(xfs_ifork_t *, xfs_fileoff_t, int *);
+xfs_ext_irec_t	*xfs_iext_bno_to_irec(xfs_ifork_t *, xfs_fileoff_t, int *);
+xfs_ext_irec_t	*xfs_iext_idx_to_irec(xfs_ifork_t *, xfs_extnum_t *, int *, int);
+void		xfs_iext_irec_init(xfs_ifork_t *);
+xfs_ext_irec_t *xfs_iext_irec_new(xfs_ifork_t *, int);
+void		xfs_iext_irec_remove(xfs_ifork_t *, int);
+void		xfs_iext_irec_compact(xfs_ifork_t *);
+void		xfs_iext_irec_compact_pages(xfs_ifork_t *);
+void		xfs_iext_irec_compact_full(xfs_ifork_t *);
+void		xfs_iext_irec_update_extoffs(xfs_ifork_t *, int, int);
+
 #define xfs_ipincount(ip)	((unsigned int) atomic_read(&ip->i_pincount))
 
 #ifdef DEBUG
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 788917f355c4..d5dfedcb8922 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -76,7 +76,7 @@ xfs_iomap_enter_trace(
 		(void *)((unsigned long)count),
 		(void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(io->io_new_size & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index c59450e1be40..32247b6bfee7 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -562,7 +562,8 @@ xfs_bulkstat(
 						if (bp)
 							xfs_buf_relse(bp);
 						error = xfs_itobp(mp, NULL, ip,
-								  &dip, &bp, bno);
+								&dip, &bp, bno,
+								XFS_IMAP_BULKSTAT);
 						if (!error)
 							clustidx = ip->i_boffset / mp->m_sb.sb_inodesize;
 						kmem_zone_free(xfs_inode_zone, ip);
@@ -570,6 +571,8 @@ xfs_bulkstat(
 								   mp, XFS_ERRTAG_BULKSTAT_READ_CHUNK,
 								   XFS_RANDOM_BULKSTAT_READ_CHUNK)) {
 							bp = NULL;
+							ubleft = 0;
+							rval = error;
 							break;
 						}
 					}
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 7d46cbd6a07a..add13f507ed2 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3249,7 +3249,7 @@ xlog_recover_process_iunlinks(
 					 * next inode in the bucket.
 					 */
 					error = xfs_itobp(mp, NULL, ip, &dip,
-							&ibp, 0);
+							&ibp, 0, 0);
 					ASSERT(error || (dip != NULL));
 				}
 
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 62188ea392c7..20e8abc16d18 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -51,11 +51,32 @@ STATIC int	xfs_uuid_mount(xfs_mount_t *);
 STATIC void	xfs_uuid_unmount(xfs_mount_t *mp);
 STATIC void	xfs_unmountfs_wait(xfs_mount_t *);
 
+
+#ifdef HAVE_PERCPU_SB
+STATIC void	xfs_icsb_destroy_counters(xfs_mount_t *);
+STATIC void	xfs_icsb_balance_counter(xfs_mount_t *, xfs_sb_field_t, int);
+STATIC void	xfs_icsb_sync_counters(xfs_mount_t *);
+STATIC int	xfs_icsb_modify_counters(xfs_mount_t *, xfs_sb_field_t,
+						int, int);
+STATIC int	xfs_icsb_modify_counters_locked(xfs_mount_t *, xfs_sb_field_t,
+						int, int);
+STATIC int	xfs_icsb_disable_counter(xfs_mount_t *, xfs_sb_field_t);
+
+#else
+
+#define xfs_icsb_destroy_counters(mp)			do { } while (0)
+#define xfs_icsb_balance_counter(mp, a, b)		do { } while (0)
+#define xfs_icsb_sync_counters(mp)			do { } while (0)
+#define xfs_icsb_modify_counters(mp, a, b, c)		do { } while (0)
+#define xfs_icsb_modify_counters_locked(mp, a, b, c)	do { } while (0)
+
+#endif
+
 static const struct {
-    short offset;
-    short type;     /* 0 = integer
-		* 1 = binary / string (no translation)
-		*/
+	short offset;
+	short type;	/* 0 = integer
+			 * 1 = binary / string (no translation)
+			 */
 } xfs_sb_info[] = {
     { offsetof(xfs_sb_t, sb_magicnum),   0 },
     { offsetof(xfs_sb_t, sb_blocksize),  0 },
@@ -113,7 +134,11 @@ xfs_mount_init(void)
 {
 	xfs_mount_t *mp;
 
-	mp = kmem_zalloc(sizeof(*mp), KM_SLEEP);
+	mp = kmem_zalloc(sizeof(xfs_mount_t), KM_SLEEP);
+
+	if (xfs_icsb_init_counters(mp)) {
+		mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;
+	}
 
 	AIL_LOCKINIT(&mp->m_ail_lock, "xfs_ail");
 	spinlock_init(&mp->m_sb_lock, "xfs_sb");
@@ -136,8 +161,8 @@ xfs_mount_init(void)
  */
 void
 xfs_mount_free(
-	xfs_mount_t *mp,
-	int	    remove_bhv)
+	xfs_mount_t	*mp,
+	int		remove_bhv)
 {
 	if (mp->m_ihash)
 		xfs_ihash_free(mp);
@@ -177,6 +202,7 @@ xfs_mount_free(
 		VFS_REMOVEBHV(vfsp, &mp->m_bhv);
 	}
 
+	xfs_icsb_destroy_counters(mp);
 	kmem_free(mp, sizeof(xfs_mount_t));
 }
 
@@ -242,9 +268,12 @@ xfs_mount_validate_sb(
 	    sbp->sb_blocklog > XFS_MAX_BLOCKSIZE_LOG			||
 	    sbp->sb_inodesize < XFS_DINODE_MIN_SIZE			||
 	    sbp->sb_inodesize > XFS_DINODE_MAX_SIZE			||
+	    sbp->sb_inodelog < XFS_DINODE_MIN_LOG			||
+	    sbp->sb_inodelog > XFS_DINODE_MAX_LOG			||
+	    (sbp->sb_blocklog - sbp->sb_inodelog != sbp->sb_inopblog)	||
 	    (sbp->sb_rextsize * sbp->sb_blocksize > XFS_MAX_RTEXTSIZE)	||
 	    (sbp->sb_rextsize * sbp->sb_blocksize < XFS_MIN_RTEXTSIZE)	||
-	    sbp->sb_imax_pct > 100)) {
+	    (sbp->sb_imax_pct > 100 || sbp->sb_imax_pct < 1))) {
 		cmn_err(CE_WARN, "XFS: SB sanity check 1 failed");
 		XFS_CORRUPTION_ERROR("xfs_mount_validate_sb(3)",
 				     XFS_ERRLEVEL_LOW, mp, sbp);
@@ -527,6 +556,10 @@ xfs_readsb(xfs_mount_t *mp)
 		ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
 	}
 
+	xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, 0);
+	xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, 0);
+	xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, 0);
+
 	mp->m_sb_bp = bp;
 	xfs_buf_relse(bp);
 	ASSERT(XFS_BUF_VALUSEMA(bp) > 0);
@@ -1154,6 +1187,9 @@ xfs_unmountfs_writesb(xfs_mount_t *mp)
 	sbp = xfs_getsb(mp, 0);
 	if (!(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY ||
 		XFS_FORCED_SHUTDOWN(mp))) {
+
+		xfs_icsb_sync_counters(mp);
+
 		/*
 		 * mark shared-readonly if desired
 		 */
@@ -1227,7 +1263,6 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
 
 	xfs_trans_log_buf(tp, bp, first, last);
 }
-
 /*
  * xfs_mod_incore_sb_unlocked() is a utility routine common used to apply
  * a delta to a specified field in the in-core superblock.  Simply
@@ -1237,7 +1272,7 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
  *
  * The SB_LOCK must be held when this routine is called.
  */
-STATIC int
+int
 xfs_mod_incore_sb_unlocked(xfs_mount_t *mp, xfs_sb_field_t field,
 			int delta, int rsvd)
 {
@@ -1406,9 +1441,26 @@ xfs_mod_incore_sb(xfs_mount_t *mp, xfs_sb_field_t field, int delta, int rsvd)
 	unsigned long	s;
 	int	status;
 
-	s = XFS_SB_LOCK(mp);
-	status = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
-	XFS_SB_UNLOCK(mp, s);
+	/* check for per-cpu counters */
+	switch (field) {
+#ifdef HAVE_PERCPU_SB
+	case XFS_SBS_ICOUNT:
+	case XFS_SBS_IFREE:
+	case XFS_SBS_FDBLOCKS:
+		if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+			status = xfs_icsb_modify_counters(mp, field,
+							delta, rsvd);
+			break;
+		}
+		/* FALLTHROUGH */
+#endif
+	default:
+		s = XFS_SB_LOCK(mp);
+		status = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
+		XFS_SB_UNLOCK(mp, s);
+		break;
+	}
+
 	return status;
 }
 
@@ -1445,8 +1497,26 @@ xfs_mod_incore_sb_batch(xfs_mount_t *mp, xfs_mod_sb_t *msb, uint nmsb, int rsvd)
 		 * from the loop so we'll fall into the undo loop
 		 * below.
 		 */
-		status = xfs_mod_incore_sb_unlocked(mp, msbp->msb_field,
-						    msbp->msb_delta, rsvd);
+		switch (msbp->msb_field) {
+#ifdef HAVE_PERCPU_SB
+		case XFS_SBS_ICOUNT:
+		case XFS_SBS_IFREE:
+		case XFS_SBS_FDBLOCKS:
+			if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+				status = xfs_icsb_modify_counters_locked(mp,
+							msbp->msb_field,
+							msbp->msb_delta, rsvd);
+				break;
+			}
+			/* FALLTHROUGH */
+#endif
+		default:
+			status = xfs_mod_incore_sb_unlocked(mp,
+						msbp->msb_field,
+						msbp->msb_delta, rsvd);
+			break;
+		}
+
 		if (status != 0) {
 			break;
 		}
@@ -1463,8 +1533,28 @@ xfs_mod_incore_sb_batch(xfs_mount_t *mp, xfs_mod_sb_t *msb, uint nmsb, int rsvd)
 	if (status != 0) {
 		msbp--;
 		while (msbp >= msb) {
-			status = xfs_mod_incore_sb_unlocked(mp,
-				    msbp->msb_field, -(msbp->msb_delta), rsvd);
+			switch (msbp->msb_field) {
+#ifdef HAVE_PERCPU_SB
+			case XFS_SBS_ICOUNT:
+			case XFS_SBS_IFREE:
+			case XFS_SBS_FDBLOCKS:
+				if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+					status =
+					    xfs_icsb_modify_counters_locked(mp,
+							msbp->msb_field,
+							-(msbp->msb_delta),
+							rsvd);
+					break;
+				}
+				/* FALLTHROUGH */
+#endif
+			default:
+				status = xfs_mod_incore_sb_unlocked(mp,
+							msbp->msb_field,
+							-(msbp->msb_delta),
+							rsvd);
+				break;
+			}
 			ASSERT(status == 0);
 			msbp--;
 		}
@@ -1577,3 +1667,525 @@ xfs_mount_log_sbunit(
 	xfs_mod_sb(tp, fields);
 	xfs_trans_commit(tp, 0, NULL);
 }
+
+
+#ifdef HAVE_PERCPU_SB
+/*
+ * Per-cpu incore superblock counters
+ *
+ * Simple concept, difficult implementation
+ *
+ * Basically, replace the incore superblock counters with a distributed per cpu
+ * counter for contended fields (e.g.  free block count).
+ *
+ * Difficulties arise in that the incore sb is used for ENOSPC checking, and
+ * hence needs to be accurately read when we are running low on space. Hence
+ * there is a method to enable and disable the per-cpu counters based on how
+ * much "stuff" is available in them.
+ *
+ * Basically, a counter is enabled if there is enough free resource to justify
+ * running a per-cpu fast-path. If the per-cpu counter runs out (i.e. a local
+ * ENOSPC), then we disable the counters to synchronise all callers and
+ * re-distribute the available resources.
+ *
+ * If, once we redistributed the available resources, we still get a failure,
+ * we disable the per-cpu counter and go through the slow path.
+ *
+ * The slow path is the current xfs_mod_incore_sb() function.  This means that
+ * when we disable a per-cpu counter, we need to drain it's resources back to
+ * the global superblock. We do this after disabling the counter to prevent
+ * more threads from queueing up on the counter.
+ *
+ * Essentially, this means that we still need a lock in the fast path to enable
+ * synchronisation between the global counters and the per-cpu counters. This
+ * is not a problem because the lock will be local to a CPU almost all the time
+ * and have little contention except when we get to ENOSPC conditions.
+ *
+ * Basically, this lock becomes a barrier that enables us to lock out the fast
+ * path while we do things like enabling and disabling counters and
+ * synchronising the counters.
+ *
+ * Locking rules:
+ *
+ * 	1. XFS_SB_LOCK() before picking up per-cpu locks
+ * 	2. per-cpu locks always picked up via for_each_online_cpu() order
+ * 	3. accurate counter sync requires XFS_SB_LOCK + per cpu locks
+ * 	4. modifying per-cpu counters requires holding per-cpu lock
+ * 	5. modifying global counters requires holding XFS_SB_LOCK
+ *	6. enabling or disabling a counter requires holding the XFS_SB_LOCK
+ *	   and _none_ of the per-cpu locks.
+ *
+ * Disabled counters are only ever re-enabled by a balance operation
+ * that results in more free resources per CPU than a given threshold.
+ * To ensure counters don't remain disabled, they are rebalanced when
+ * the global resource goes above a higher threshold (i.e. some hysteresis
+ * is present to prevent thrashing).
+ */
+
+/*
+ * hot-plug CPU notifier support.
+ *
+ * We cannot use the hotcpu_register() function because it does
+ * not allow notifier instances. We need a notifier per filesystem
+ * as we need to be able to identify the filesystem to balance
+ * the counters out. This is acheived by having a notifier block
+ * embedded in the xfs_mount_t and doing pointer magic to get the
+ * mount pointer from the notifier block address.
+ */
+STATIC int
+xfs_icsb_cpu_notify(
+	struct notifier_block *nfb,
+	unsigned long action,
+	void *hcpu)
+{
+	xfs_icsb_cnts_t *cntp;
+	xfs_mount_t	*mp;
+	int		s;
+
+	mp = (xfs_mount_t *)container_of(nfb, xfs_mount_t, m_icsb_notifier);
+	cntp = (xfs_icsb_cnts_t *)
+			per_cpu_ptr(mp->m_sb_cnts, (unsigned long)hcpu);
+	switch (action) {
+	case CPU_UP_PREPARE:
+		/* Easy Case - initialize the area and locks, and
+		 * then rebalance when online does everything else for us. */
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+		break;
+	case CPU_ONLINE:
+		xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, 0);
+		xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, 0);
+		xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, 0);
+		break;
+	case CPU_DEAD:
+		/* Disable all the counters, then fold the dead cpu's
+		 * count into the total on the global superblock and
+		 * re-enable the counters. */
+		s = XFS_SB_LOCK(mp);
+		xfs_icsb_disable_counter(mp, XFS_SBS_ICOUNT);
+		xfs_icsb_disable_counter(mp, XFS_SBS_IFREE);
+		xfs_icsb_disable_counter(mp, XFS_SBS_FDBLOCKS);
+
+		mp->m_sb.sb_icount += cntp->icsb_icount;
+		mp->m_sb.sb_ifree += cntp->icsb_ifree;
+		mp->m_sb.sb_fdblocks += cntp->icsb_fdblocks;
+
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+
+		xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, XFS_ICSB_SB_LOCKED);
+		xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, XFS_ICSB_SB_LOCKED);
+		xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, XFS_ICSB_SB_LOCKED);
+		XFS_SB_UNLOCK(mp, s);
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+int
+xfs_icsb_init_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	mp->m_sb_cnts = alloc_percpu(xfs_icsb_cnts_t);
+	if (mp->m_sb_cnts == NULL)
+		return -ENOMEM;
+
+	mp->m_icsb_notifier.notifier_call = xfs_icsb_cpu_notify;
+	mp->m_icsb_notifier.priority = 0;
+	register_cpu_notifier(&mp->m_icsb_notifier);
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+	}
+	/*
+	 * start with all counters disabled so that the
+	 * initial balance kicks us off correctly
+	 */
+	mp->m_icsb_counters = -1;
+	return 0;
+}
+
+STATIC void
+xfs_icsb_destroy_counters(
+	xfs_mount_t	*mp)
+{
+	if (mp->m_sb_cnts) {
+		unregister_cpu_notifier(&mp->m_icsb_notifier);
+		free_percpu(mp->m_sb_cnts);
+	}
+}
+
+STATIC inline void
+xfs_icsb_lock_cntr(
+	xfs_icsb_cnts_t	*icsbp)
+{
+	while (test_and_set_bit(XFS_ICSB_FLAG_LOCK, &icsbp->icsb_flags)) {
+		ndelay(1000);
+	}
+}
+
+STATIC inline void
+xfs_icsb_unlock_cntr(
+	xfs_icsb_cnts_t	*icsbp)
+{
+	clear_bit(XFS_ICSB_FLAG_LOCK, &icsbp->icsb_flags);
+}
+
+
+STATIC inline void
+xfs_icsb_lock_all_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		xfs_icsb_lock_cntr(cntp);
+	}
+}
+
+STATIC inline void
+xfs_icsb_unlock_all_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		xfs_icsb_unlock_cntr(cntp);
+	}
+}
+
+STATIC void
+xfs_icsb_count(
+	xfs_mount_t	*mp,
+	xfs_icsb_cnts_t	*cnt,
+	int		flags)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	memset(cnt, 0, sizeof(xfs_icsb_cnts_t));
+
+	if (!(flags & XFS_ICSB_LAZY_COUNT))
+		xfs_icsb_lock_all_counters(mp);
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		cnt->icsb_icount += cntp->icsb_icount;
+		cnt->icsb_ifree += cntp->icsb_ifree;
+		cnt->icsb_fdblocks += cntp->icsb_fdblocks;
+	}
+
+	if (!(flags & XFS_ICSB_LAZY_COUNT))
+		xfs_icsb_unlock_all_counters(mp);
+}
+
+STATIC int
+xfs_icsb_counter_disabled(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field)
+{
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+	return test_bit(field, &mp->m_icsb_counters);
+}
+
+STATIC int
+xfs_icsb_disable_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field)
+{
+	xfs_icsb_cnts_t	cnt;
+
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+
+	xfs_icsb_lock_all_counters(mp);
+	if (!test_and_set_bit(field, &mp->m_icsb_counters)) {
+		/* drain back to superblock */
+
+		xfs_icsb_count(mp, &cnt, XFS_ICSB_SB_LOCKED|XFS_ICSB_LAZY_COUNT);
+		switch(field) {
+		case XFS_SBS_ICOUNT:
+			mp->m_sb.sb_icount = cnt.icsb_icount;
+			break;
+		case XFS_SBS_IFREE:
+			mp->m_sb.sb_ifree = cnt.icsb_ifree;
+			break;
+		case XFS_SBS_FDBLOCKS:
+			mp->m_sb.sb_fdblocks = cnt.icsb_fdblocks;
+			break;
+		default:
+			BUG();
+		}
+	}
+
+	xfs_icsb_unlock_all_counters(mp);
+
+	return 0;
+}
+
+STATIC void
+xfs_icsb_enable_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	uint64_t	count,
+	uint64_t	resid)
+{
+	xfs_icsb_cnts_t	*cntp;
+	int		i;
+
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+
+	xfs_icsb_lock_all_counters(mp);
+	for_each_online_cpu(i) {
+		cntp = per_cpu_ptr(mp->m_sb_cnts, i);
+		switch (field) {
+		case XFS_SBS_ICOUNT:
+			cntp->icsb_icount = count + resid;
+			break;
+		case XFS_SBS_IFREE:
+			cntp->icsb_ifree = count + resid;
+			break;
+		case XFS_SBS_FDBLOCKS:
+			cntp->icsb_fdblocks = count + resid;
+			break;
+		default:
+			BUG();
+			break;
+		}
+		resid = 0;
+	}
+	clear_bit(field, &mp->m_icsb_counters);
+	xfs_icsb_unlock_all_counters(mp);
+}
+
+STATIC void
+xfs_icsb_sync_counters_int(
+	xfs_mount_t	*mp,
+	int		flags)
+{
+	xfs_icsb_cnts_t	cnt;
+	int		s;
+
+	/* Pass 1: lock all counters */
+	if ((flags & XFS_ICSB_SB_LOCKED) == 0)
+		s = XFS_SB_LOCK(mp);
+
+	xfs_icsb_count(mp, &cnt, flags);
+
+	/* Step 3: update mp->m_sb fields */
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_ICOUNT))
+		mp->m_sb.sb_icount = cnt.icsb_icount;
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_IFREE))
+		mp->m_sb.sb_ifree = cnt.icsb_ifree;
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_FDBLOCKS))
+		mp->m_sb.sb_fdblocks = cnt.icsb_fdblocks;
+
+	if ((flags & XFS_ICSB_SB_LOCKED) == 0)
+		XFS_SB_UNLOCK(mp, s);
+}
+
+/*
+ * Accurate update of per-cpu counters to incore superblock
+ */
+STATIC void
+xfs_icsb_sync_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_sync_counters_int(mp, 0);
+}
+
+/*
+ * lazy addition used for things like df, background sb syncs, etc
+ */
+void
+xfs_icsb_sync_counters_lazy(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_sync_counters_int(mp, XFS_ICSB_LAZY_COUNT);
+}
+
+/*
+ * Balance and enable/disable counters as necessary.
+ *
+ * Thresholds for re-enabling counters are somewhat magic.
+ * inode counts are chosen to be the same number as single
+ * on disk allocation chunk per CPU, and free blocks is
+ * something far enough zero that we aren't going thrash
+ * when we get near ENOSPC.
+ */
+#define XFS_ICSB_INO_CNTR_REENABLE	64
+#define XFS_ICSB_FDBLK_CNTR_REENABLE	512
+STATIC void
+xfs_icsb_balance_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t  field,
+	int		flags)
+{
+	uint64_t	count, resid = 0;
+	int		weight = num_online_cpus();
+	int		s;
+
+	if (!(flags & XFS_ICSB_SB_LOCKED))
+		s = XFS_SB_LOCK(mp);
+
+	/* disable counter and sync counter */
+	xfs_icsb_disable_counter(mp, field);
+
+	/* update counters  - first CPU gets residual*/
+	switch (field) {
+	case XFS_SBS_ICOUNT:
+		count = mp->m_sb.sb_icount;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_INO_CNTR_REENABLE)
+			goto out;
+		break;
+	case XFS_SBS_IFREE:
+		count = mp->m_sb.sb_ifree;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_INO_CNTR_REENABLE)
+			goto out;
+		break;
+	case XFS_SBS_FDBLOCKS:
+		count = mp->m_sb.sb_fdblocks;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_FDBLK_CNTR_REENABLE)
+			goto out;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	xfs_icsb_enable_counter(mp, field, count, resid);
+out:
+	if (!(flags & XFS_ICSB_SB_LOCKED))
+		XFS_SB_UNLOCK(mp, s);
+}
+
+STATIC int
+xfs_icsb_modify_counters_int(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd,
+	int		flags)
+{
+	xfs_icsb_cnts_t	*icsbp;
+	long long	lcounter;	/* long counter for 64 bit fields */
+	int		cpu, s, locked = 0;
+	int		ret = 0, balance_done = 0;
+
+again:
+	cpu = get_cpu();
+	icsbp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, cpu),
+	xfs_icsb_lock_cntr(icsbp);
+	if (unlikely(xfs_icsb_counter_disabled(mp, field)))
+		goto slow_path;
+
+	switch (field) {
+	case XFS_SBS_ICOUNT:
+		lcounter = icsbp->icsb_icount;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_icount = lcounter;
+		break;
+
+	case XFS_SBS_IFREE:
+		lcounter = icsbp->icsb_ifree;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_ifree = lcounter;
+		break;
+
+	case XFS_SBS_FDBLOCKS:
+		BUG_ON((mp->m_resblks - mp->m_resblks_avail) != 0);
+
+		lcounter = icsbp->icsb_fdblocks;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_fdblocks = lcounter;
+		break;
+	default:
+		BUG();
+		break;
+	}
+	xfs_icsb_unlock_cntr(icsbp);
+	put_cpu();
+	if (locked)
+		XFS_SB_UNLOCK(mp, s);
+	return 0;
+
+	/*
+	 * The slow path needs to be run with the SBLOCK
+	 * held so that we prevent other threads from
+	 * attempting to run this path at the same time.
+	 * this provides exclusion for the balancing code,
+	 * and exclusive fallback if the balance does not
+	 * provide enough resources to continue in an unlocked
+	 * manner.
+	 */
+slow_path:
+	xfs_icsb_unlock_cntr(icsbp);
+	put_cpu();
+
+	/* need to hold superblock incase we need
+	 * to disable a counter */
+	if (!(flags & XFS_ICSB_SB_LOCKED)) {
+		s = XFS_SB_LOCK(mp);
+		locked = 1;
+		flags |= XFS_ICSB_SB_LOCKED;
+	}
+	if (!balance_done) {
+		xfs_icsb_balance_counter(mp, field, flags);
+		balance_done = 1;
+		goto again;
+	} else {
+		/*
+		 * we might not have enough on this local
+		 * cpu to allocate for a bulk request.
+		 * We need to drain this field from all CPUs
+		 * and disable the counter fastpath
+		 */
+		xfs_icsb_disable_counter(mp, field);
+	}
+
+	ret = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
+
+	if (locked)
+		XFS_SB_UNLOCK(mp, s);
+	return ret;
+}
+
+STATIC int
+xfs_icsb_modify_counters(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd)
+{
+	return xfs_icsb_modify_counters_int(mp, field, delta, rsvd, 0);
+}
+
+/*
+ * Called when superblock is already locked
+ */
+STATIC int
+xfs_icsb_modify_counters_locked(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd)
+{
+	return xfs_icsb_modify_counters_int(mp, field, delta,
+						rsvd, XFS_ICSB_SB_LOCKED);
+}
+#endif
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index cd3cf9613a00..ebd73960e9db 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -267,6 +267,34 @@ typedef struct xfs_ioops {
 #define XFS_IODONE(vfsp) \
 	(*(mp)->m_io_ops.xfs_iodone)(vfsp)
 
+#ifdef HAVE_PERCPU_SB
+
+/*
+ * Valid per-cpu incore superblock counters. Note that if you add new counters,
+ * you may need to define new counter disabled bit field descriptors as there
+ * are more possible fields in the superblock that can fit in a bitfield on a
+ * 32 bit platform. The XFS_SBS_* values for the current current counters just
+ * fit.
+ */
+typedef struct xfs_icsb_cnts {
+	uint64_t	icsb_fdblocks;
+	uint64_t	icsb_ifree;
+	uint64_t	icsb_icount;
+	unsigned long	icsb_flags;
+} xfs_icsb_cnts_t;
+
+#define XFS_ICSB_FLAG_LOCK	(1 << 0)	/* counter lock bit */
+
+#define XFS_ICSB_SB_LOCKED	(1 << 0)	/* sb already locked */
+#define XFS_ICSB_LAZY_COUNT	(1 << 1)	/* accuracy not needed */
+
+extern int	xfs_icsb_init_counters(struct xfs_mount *);
+extern void	xfs_icsb_sync_counters_lazy(struct xfs_mount *);
+
+#else
+#define xfs_icsb_init_counters(mp)	(0)
+#define xfs_icsb_sync_counters_lazy(mp)	do { } while (0)
+#endif
 
 typedef struct xfs_mount {
 	bhv_desc_t		m_bhv;		/* vfs xfs behavior */
@@ -372,6 +400,11 @@ typedef struct xfs_mount {
 	struct xfs_qmops	m_qm_ops;	/* vector of XQM ops */
 	struct xfs_ioops	m_io_ops;	/* vector of I/O ops */
 	atomic_t		m_active_trans;	/* number trans frozen */
+#ifdef HAVE_PERCPU_SB
+	xfs_icsb_cnts_t		*m_sb_cnts;	/* per-cpu superblock counters */
+	unsigned long		m_icsb_counters; /* disabled per-cpu counters */
+	struct notifier_block	m_icsb_notifier; /* hotplug cpu notifier */
+#endif
 } xfs_mount_t;
 
 /*
@@ -386,8 +419,6 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_FS_SHUTDOWN	(1ULL << 4)	/* atomic stop of all filesystem
 						   operations, typically for
 						   disk errors in metadata */
-#define XFS_MOUNT_NOATIME	(1ULL << 5)	/* don't modify inode access
-						   times on reads */
 #define XFS_MOUNT_RETERR	(1ULL << 6)     /* return alignment errors to
 						   user */
 #define XFS_MOUNT_NOALIGN	(1ULL << 7)	/* turn off stripe alignment
@@ -411,6 +442,8 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_DIRSYNC	(1ULL << 21)	/* synchronous directory ops */
 #define XFS_MOUNT_COMPAT_IOSIZE	(1ULL << 22)	/* don't report large preferred
 						 * I/O size in stat() */
+#define XFS_MOUNT_NO_PERCPU_SB	(1ULL << 23)	/* don't use per-cpu superblock
+						   counters */
 
 
 /*
@@ -473,11 +506,6 @@ xfs_preferred_iosize(xfs_mount_t *mp)
 #define XFS_SHUTDOWN_REMOTE_REQ 0x10	/* Shutdown came from remote cell */
 
 /*
- * xflags for xfs_syncsub
- */
-#define XFS_XSYNC_RELOC		0x01
-
-/*
  * Flags for xfs_mountfs
  */
 #define XFS_MFSI_SECOND		0x01	/* Secondary mount -- skip stuff */
@@ -548,6 +576,8 @@ extern void	xfs_unmountfs_close(xfs_mount_t *, struct cred *);
 extern int	xfs_unmountfs_writesb(xfs_mount_t *);
 extern int	xfs_unmount_flush(xfs_mount_t *, int);
 extern int	xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int, int);
+extern int	xfs_mod_incore_sb_unlocked(xfs_mount_t *, xfs_sb_field_t,
+			int, int);
 extern int	xfs_mod_incore_sb_batch(xfs_mount_t *, xfs_mod_sb_t *,
 			uint, int);
 extern struct xfs_buf *xfs_getsb(xfs_mount_t *, int);
diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index de85eefb7966..e63795644478 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -89,6 +89,7 @@ extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
  */
 extern int xfs_rwlock(bhv_desc_t *bdp, vrwlock_t write_lock);
 extern void xfs_rwunlock(bhv_desc_t *bdp, vrwlock_t write_lock);
+extern int xfs_setattr(bhv_desc_t *bdp, vattr_t *vap, int flags, cred_t *credp);
 extern int xfs_change_file_space(bhv_desc_t *bdp, int cmd, xfs_flock64_t *bf,
 				 xfs_off_t offset, cred_t *credp, int flags);
 extern int xfs_set_dmattrs(bhv_desc_t *bdp, u_int evmask, u_int16_t state,
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index d3d714e6b32a..2918956553a5 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -55,10 +55,141 @@ STATIC void	xfs_trans_committed(xfs_trans_t *, int);
 STATIC void	xfs_trans_chunk_committed(xfs_log_item_chunk_t *, xfs_lsn_t, int);
 STATIC void	xfs_trans_free(xfs_trans_t *);
 
-kmem_zone_t		*xfs_trans_zone;
+kmem_zone_t	*xfs_trans_zone;
 
 
 /*
+ * Reservation functions here avoid a huge stack in xfs_trans_init
+ * due to register overflow from temporaries in the calculations.
+ */
+
+STATIC uint
+xfs_calc_write_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_WRITE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_itruncate_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ITRUNCATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_rename_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_RENAME_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_link_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_LINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_remove_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_REMOVE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_symlink_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_SYMLINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_create_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_CREATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_mkdir_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_MKDIR_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_ifree_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_IFREE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_ichange_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ICHANGE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_growdata_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWDATA_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtalloc_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTALLOC_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtzero_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTZERO_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtfree_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTFREE_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_swrite_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_SWRITE_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_writeid_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_WRITEID_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_addafork_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ADDAFORK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_attrinval_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRINVAL_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_attrset_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRSET_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_attrrm_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRRM_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_clear_agi_bucket_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_CLEAR_AGI_BUCKET_LOG_RES(mp);
+}
+
+/*
  * Initialize the precomputed transaction reservation values
  * in the mount structure.
  */
@@ -69,39 +200,27 @@ xfs_trans_init(
 	xfs_trans_reservations_t	*resp;
 
 	resp = &(mp->m_reservations);
-	resp->tr_write =
-		(uint)(XFS_CALC_WRITE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_itruncate =
-		(uint)(XFS_CALC_ITRUNCATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_rename =
-		(uint)(XFS_CALC_RENAME_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_link = (uint)XFS_CALC_LINK_LOG_RES(mp);
-	resp->tr_remove =
-		(uint)(XFS_CALC_REMOVE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_symlink =
-		(uint)(XFS_CALC_SYMLINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_create =
-		(uint)(XFS_CALC_CREATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_mkdir =
-		(uint)(XFS_CALC_MKDIR_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_ifree =
-		(uint)(XFS_CALC_IFREE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_ichange =
-		(uint)(XFS_CALC_ICHANGE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_growdata = (uint)XFS_CALC_GROWDATA_LOG_RES(mp);
-	resp->tr_swrite = (uint)XFS_CALC_SWRITE_LOG_RES(mp);
-	resp->tr_writeid = (uint)XFS_CALC_WRITEID_LOG_RES(mp);
-	resp->tr_addafork =
-		(uint)(XFS_CALC_ADDAFORK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_attrinval = (uint)XFS_CALC_ATTRINVAL_LOG_RES(mp);
-	resp->tr_attrset =
-		(uint)(XFS_CALC_ATTRSET_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_attrrm =
-		(uint)(XFS_CALC_ATTRRM_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_clearagi = (uint)XFS_CALC_CLEAR_AGI_BUCKET_LOG_RES(mp);
-	resp->tr_growrtalloc = (uint)XFS_CALC_GROWRTALLOC_LOG_RES(mp);
-	resp->tr_growrtzero = (uint)XFS_CALC_GROWRTZERO_LOG_RES(mp);
-	resp->tr_growrtfree = (uint)XFS_CALC_GROWRTFREE_LOG_RES(mp);
+	resp->tr_write = xfs_calc_write_reservation(mp);
+	resp->tr_itruncate = xfs_calc_itruncate_reservation(mp);
+	resp->tr_rename = xfs_calc_rename_reservation(mp);
+	resp->tr_link = xfs_calc_link_reservation(mp);
+	resp->tr_remove = xfs_calc_remove_reservation(mp);
+	resp->tr_symlink = xfs_calc_symlink_reservation(mp);
+	resp->tr_create = xfs_calc_create_reservation(mp);
+	resp->tr_mkdir = xfs_calc_mkdir_reservation(mp);
+	resp->tr_ifree = xfs_calc_ifree_reservation(mp);
+	resp->tr_ichange = xfs_calc_ichange_reservation(mp);
+	resp->tr_growdata = xfs_calc_growdata_reservation(mp);
+	resp->tr_swrite = xfs_calc_swrite_reservation(mp);
+	resp->tr_writeid = xfs_calc_writeid_reservation(mp);
+	resp->tr_addafork = xfs_calc_addafork_reservation(mp);
+	resp->tr_attrinval = xfs_calc_attrinval_reservation(mp);
+	resp->tr_attrset = xfs_calc_attrset_reservation(mp);
+	resp->tr_attrrm = xfs_calc_attrrm_reservation(mp);
+	resp->tr_clearagi = xfs_calc_clear_agi_bucket_reservation(mp);
+	resp->tr_growrtalloc = xfs_calc_growrtalloc_reservation(mp);
+	resp->tr_growrtzero = xfs_calc_growrtzero_reservation(mp);
+	resp->tr_growrtfree = xfs_calc_growrtfree_reservation(mp);
 }
 
 /*
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index d77901c07f63..e48befa4e337 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -380,7 +380,7 @@ typedef struct xfs_trans {
 	xfs_trans_header_t	t_header;	/* header for in-log trans */
 	unsigned int		t_busy_free;	/* busy descs free */
 	xfs_log_busy_chunk_t	t_busy;		/* busy/async free blocks */
-        xfs_pflags_t            t_pflags;       /* saved pflags state */
+	unsigned long		t_pflags;	/* saved process flags state */
 } xfs_trans_t;
 
 #endif	/* __KERNEL__ */
diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c
index b6ad370fab3d..d4ec4dfaf19c 100644
--- a/fs/xfs/xfs_vfsops.c
+++ b/fs/xfs/xfs_vfsops.c
@@ -55,7 +55,7 @@
 #include "xfs_clnt.h"
 #include "xfs_fsops.h"
 
-STATIC int xfs_sync(bhv_desc_t *, int, cred_t *);
+STATIC int	xfs_sync(bhv_desc_t *, int, cred_t *);
 
 int
 xfs_init(void)
@@ -77,11 +77,12 @@ xfs_init(void)
 						 "xfs_bmap_free_item");
 	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
 					    "xfs_btree_cur");
-	xfs_inode_zone = kmem_zone_init(sizeof(xfs_inode_t), "xfs_inode");
 	xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
 	xfs_da_state_zone =
 		kmem_zone_init(sizeof(xfs_da_state_t), "xfs_da_state");
 	xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
+	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
+	xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
 
 	/*
 	 * The size of the zone allocated buf log item is the maximum
@@ -93,17 +94,30 @@ xfs_init(void)
 				(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
 				  NBWORD) * sizeof(int))),
 			       "xfs_buf_item");
-	xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
-				       ((XFS_EFD_MAX_FAST_EXTENTS - 1) * sizeof(xfs_extent_t))),
+	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");
-	xfs_efi_zone = kmem_zone_init((sizeof(xfs_efi_log_item_t) +
-				       ((XFS_EFI_MAX_FAST_EXTENTS - 1) * sizeof(xfs_extent_t))),
+	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");
-	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
-	xfs_ili_zone = kmem_zone_init(sizeof(xfs_inode_log_item_t), "xfs_ili");
-	xfs_chashlist_zone = kmem_zone_init(sizeof(xfs_chashlist_t),
-					    "xfs_chashlist");
-	xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
+
+	/*
+	 * These zones warrant special memory allocator hints
+	 */
+	xfs_inode_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
+					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
+					KM_ZONE_SPREAD, NULL);
+	xfs_ili_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
+					KM_ZONE_SPREAD, NULL);
+	xfs_chashlist_zone =
+		kmem_zone_init_flags(sizeof(xfs_chashlist_t), "xfs_chashlist",
+					KM_ZONE_SPREAD, NULL);
 
 	/*
 	 * Allocate global trace buffers.
@@ -176,18 +190,18 @@ xfs_cleanup(void)
 	ktrace_free(xfs_alloc_trace_buf);
 #endif
 
-	kmem_cache_destroy(xfs_bmap_free_item_zone);
-	kmem_cache_destroy(xfs_btree_cur_zone);
-	kmem_cache_destroy(xfs_inode_zone);
-	kmem_cache_destroy(xfs_trans_zone);
-	kmem_cache_destroy(xfs_da_state_zone);
-	kmem_cache_destroy(xfs_dabuf_zone);
-	kmem_cache_destroy(xfs_buf_item_zone);
-	kmem_cache_destroy(xfs_efd_zone);
-	kmem_cache_destroy(xfs_efi_zone);
-	kmem_cache_destroy(xfs_ifork_zone);
-	kmem_cache_destroy(xfs_ili_zone);
-	kmem_cache_destroy(xfs_chashlist_zone);
+	kmem_zone_destroy(xfs_bmap_free_item_zone);
+	kmem_zone_destroy(xfs_btree_cur_zone);
+	kmem_zone_destroy(xfs_inode_zone);
+	kmem_zone_destroy(xfs_trans_zone);
+	kmem_zone_destroy(xfs_da_state_zone);
+	kmem_zone_destroy(xfs_dabuf_zone);
+	kmem_zone_destroy(xfs_buf_item_zone);
+	kmem_zone_destroy(xfs_efd_zone);
+	kmem_zone_destroy(xfs_efi_zone);
+	kmem_zone_destroy(xfs_ifork_zone);
+	kmem_zone_destroy(xfs_ili_zone);
+	kmem_zone_destroy(xfs_chashlist_zone);
 }
 
 /*
@@ -258,8 +272,6 @@ xfs_start_flags(
 		mp->m_inoadd = XFS_INO64_OFFSET;
 	}
 #endif
-	if (ap->flags & XFSMNT_NOATIME)
-		mp->m_flags |= XFS_MOUNT_NOATIME;
 	if (ap->flags & XFSMNT_RETERR)
 		mp->m_flags |= XFS_MOUNT_RETERR;
 	if (ap->flags & XFSMNT_NOALIGN)
@@ -620,7 +632,7 @@ xfs_quiesce_fs(
 	xfs_mount_t		*mp)
 {
 	int			count = 0, pincount;
-		
+
 	xfs_refcache_purge_mp(mp);
 	xfs_flush_buftarg(mp->m_ddev_targp, 0);
 	xfs_finish_reclaim_all(mp, 0);
@@ -631,7 +643,7 @@ xfs_quiesce_fs(
 	 * meta data (typically directory updates).
 	 * Which then must be flushed and logged before
 	 * we can write the unmount record.
-	 */ 
+	 */
 	do {
 		xfs_syncsub(mp, SYNC_REMOUNT|SYNC_ATTR|SYNC_WAIT, 0, NULL);
 		pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
@@ -654,11 +666,6 @@ xfs_mntupdate(
 	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
 	int		error;
 
-	if (args->flags & XFSMNT_NOATIME)
-		mp->m_flags |= XFS_MOUNT_NOATIME;
-	else
-		mp->m_flags &= ~XFS_MOUNT_NOATIME;
-
 	if (args->flags & XFSMNT_BARRIER)
 		mp->m_flags |= XFS_MOUNT_BARRIER;
 	else
@@ -814,6 +821,7 @@ xfs_statvfs(
 
 	statp->f_type = XFS_SB_MAGIC;
 
+	xfs_icsb_sync_counters_lazy(mp);
 	s = XFS_SB_LOCK(mp);
 	statp->f_bsize = sbp->sb_blocksize;
 	lsize = sbp->sb_logstart ? sbp->sb_logblocks : 0;
@@ -1221,7 +1229,7 @@ xfs_sync_inodes(
 					xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
 					error = xfs_itobp(mp, NULL, ip,
-							  &dip, &bp, 0);
+							  &dip, &bp, 0, 0);
 					if (!error) {
 						xfs_buf_relse(bp);
 					} else {
@@ -1690,10 +1698,7 @@ xfs_parseargs(
 	int			iosize;
 
 	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
-
-#if 0	/* XXX: off by default, until some remaining issues ironed out */
-	args->flags |= XFSMNT_IDELETE; /* default to on */
-#endif
+	args->flags |= XFSMNT_IDELETE;
 
 	if (!options)
 		goto done;
@@ -1903,7 +1908,6 @@ xfs_showargs(
 		{ XFS_MOUNT_NOUUID,		"," MNTOPT_NOUUID },
 		{ XFS_MOUNT_NORECOVERY,		"," MNTOPT_NORECOVERY },
 		{ XFS_MOUNT_OSYNCISOSYNC,	"," MNTOPT_OSYNCISOSYNC },
-		{ XFS_MOUNT_IDELETE,		"," MNTOPT_NOIKEEP },
 		{ 0, NULL }
 	};
 	struct proc_xfs_info	*xfs_infop;
@@ -1939,6 +1943,8 @@ xfs_showargs(
 		seq_printf(m, "," MNTOPT_SWIDTH "=%d",
 				(int)XFS_FSB_TO_BB(mp, mp->m_swidth));
 
+	if (!(mp->m_flags & XFS_MOUNT_IDELETE))
+		seq_printf(m, "," MNTOPT_IKEEP);
 	if (!(mp->m_flags & XFS_MOUNT_COMPAT_IOSIZE))
 		seq_printf(m, "," MNTOPT_LARGEIO);
 	if (mp->m_flags & XFS_MOUNT_BARRIER)
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index eaab355f5a89..a478f42e63ff 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -615,6 +615,7 @@ xfs_setattr(
 			code = xfs_igrow_start(ip, vap->va_size, credp);
 		}
 		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		vn_iowait(vp); /* wait for the completion of any pending DIOs */
 		if (!code)
 			code = xfs_itruncate_data(ip, vap->va_size);
 		if (code) {
@@ -1556,7 +1557,7 @@ xfs_release(
 			if ((error = xfs_inactive_free_eofblocks(mp, ip)))
 				return error;
 			/* Update linux inode block count after free above */
-			LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+			vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
 				ip->i_d.di_nblocks + ip->i_delayed_blks);
 		}
 	}
@@ -1637,7 +1638,7 @@ xfs_inactive(
 			if ((error = xfs_inactive_free_eofblocks(mp, ip)))
 				return VN_INACTIVE_CACHE;
 			/* Update linux inode block count after free above */
-			LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+			vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
 				ip->i_d.di_nblocks + ip->i_delayed_blks);
 		}
 		goto out;
@@ -3186,7 +3187,7 @@ xfs_rmdir(
 
 	/* Fall through to std_return with error = 0 or the errno
 	 * from xfs_trans_commit. */
-std_return:
+ std_return:
 	if (DM_EVENT_ENABLED(dir_vp->v_vfsp, dp, DM_EVENT_POSTREMOVE)) {
 		(void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE,
 					dir_vp, DM_RIGHT_NULL,
@@ -3196,12 +3197,12 @@ std_return:
 	}
 	return error;
 
-error1:
+ error1:
 	xfs_bmap_cancel(&free_list);
 	cancel_flags |= XFS_TRANS_ABORT;
 	/* FALLTHROUGH */
 
-error_return:
+ error_return:
 	xfs_trans_cancel(tp, cancel_flags);
 	goto std_return;
 }
@@ -4310,8 +4311,10 @@ xfs_free_file_space(
 	ASSERT(attr_flags & ATTR_NOLOCK ? attr_flags & ATTR_DMI : 1);
 	if (attr_flags & ATTR_NOLOCK)
 		need_iolock = 0;
-	if (need_iolock)
+	if (need_iolock) {
 		xfs_ilock(ip, XFS_IOLOCK_EXCL);
+		vn_iowait(vp);	/* wait for the completion of any pending DIOs */
+	}
 
 	rounding = MAX((__uint8_t)(1 << mp->m_sb.sb_blocklog),
 			(__uint8_t)NBPP);