/spare/repo/netdev-2.6 branch 'ieee80211'

69 files changed, 3051 insertions, 2152 deletions

diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index 9c09641..fca83e2 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -92,6 +92,7 @@ struct autofs_wait_queue {
 
 struct autofs_sb_info {
 	u32 magic;
+	struct dentry *root;
 	struct file *pipe;
 	pid_t oz_pgrp;
 	int catatonic;
diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c
index 4bb14cc..0a3c05d 100644
--- a/fs/autofs4/inode.c
+++ b/fs/autofs4/inode.c
@@ -16,6 +16,7 @@
 #include <linux/pagemap.h>
 #include <linux/parser.h>
 #include <linux/bitops.h>
+#include <linux/smp_lock.h>
 #include "autofs_i.h"
 #include <linux/module.h>
 
@@ -76,6 +77,66 @@ void autofs4_free_ino(struct autofs_info *ino)
 	kfree(ino);
 }
 
+/*
+ * Deal with the infamous "Busy inodes after umount ..." message.
+ *
+ * Clean up the dentry tree. This happens with autofs if the user
+ * space program goes away due to a SIGKILL, SIGSEGV etc.
+ */
+static void autofs4_force_release(struct autofs_sb_info *sbi)
+{
+	struct dentry *this_parent = sbi->root;
+	struct list_head *next;
+
+	spin_lock(&dcache_lock);
+repeat:
+	next = this_parent->d_subdirs.next;
+resume:
+	while (next != &this_parent->d_subdirs) {
+		struct dentry *dentry = list_entry(next, struct dentry, d_child);
+
+		/* Negative dentry - don`t care */
+		if (!simple_positive(dentry)) {
+			next = next->next;
+			continue;
+		}
+
+		if (!list_empty(&dentry->d_subdirs)) {
+			this_parent = dentry;
+			goto repeat;
+		}
+
+		next = next->next;
+		spin_unlock(&dcache_lock);
+
+		DPRINTK("dentry %p %.*s",
+			dentry, (int)dentry->d_name.len, dentry->d_name.name);
+
+		dput(dentry);
+		spin_lock(&dcache_lock);
+	}
+
+	if (this_parent != sbi->root) {
+		struct dentry *dentry = this_parent;
+
+		next = this_parent->d_child.next;
+		this_parent = this_parent->d_parent;
+		spin_unlock(&dcache_lock);
+		DPRINTK("parent dentry %p %.*s",
+			dentry, (int)dentry->d_name.len, dentry->d_name.name);
+		dput(dentry);
+		spin_lock(&dcache_lock);
+		goto resume;
+	}
+	spin_unlock(&dcache_lock);
+
+	dput(sbi->root);
+	sbi->root = NULL;
+	shrink_dcache_sb(sbi->sb);
+
+	return;
+}
+
 static void autofs4_put_super(struct super_block *sb)
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(sb);
@@ -85,6 +146,10 @@ static void autofs4_put_super(struct super_block *sb)
 	if ( !sbi->catatonic )
 		autofs4_catatonic_mode(sbi); /* Free wait queues, close pipe */
 
+	/* Clean up and release dangling references */
+	if (sbi)
+		autofs4_force_release(sbi);
+
 	kfree(sbi);
 
 	DPRINTK("shutting down");
@@ -199,6 +264,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 
 	s->s_fs_info = sbi;
 	sbi->magic = AUTOFS_SBI_MAGIC;
+	sbi->root = NULL;
 	sbi->catatonic = 0;
 	sbi->exp_timeout = 0;
 	sbi->oz_pgrp = process_group(current);
@@ -267,6 +333,13 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	sbi->pipe = pipe;
 
 	/*
+	 * Take a reference to the root dentry so we get a chance to
+	 * clean up the dentry tree on umount.
+	 * See autofs4_force_release.
+	 */
+	sbi->root = dget(root);
+
+	/*
 	 * Success! Install the root dentry now to indicate completion.
 	 */
 	s->s_root = root;
diff --git a/fs/bio.c b/fs/bio.c
index ca8f7a8..249dd6b 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -261,6 +261,7 @@ inline void __bio_clone(struct bio *bio, struct bio *bio_src)
 	 */
 	bio->bi_vcnt = bio_src->bi_vcnt;
 	bio->bi_size = bio_src->bi_size;
+	bio->bi_idx = bio_src->bi_idx;
 	bio_phys_segments(q, bio);
 	bio_hw_segments(q, bio);
 }
diff --git a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c
index 77e0591..161f156 100644
--- a/fs/ext2/ialloc.c
+++ b/fs/ext2/ialloc.c
@@ -612,6 +612,7 @@ got:
 	err = ext2_init_acl(inode, dir);
 	if (err) {
 		DQUOT_FREE_INODE(inode);
+		DQUOT_DROP(inode);
 		goto fail2;
 	}
 	mark_inode_dirty(inode);
diff --git a/fs/ext2/xattr.c b/fs/ext2/xattr.c
index 27982b5..0099462 100644
--- a/fs/ext2/xattr.c
+++ b/fs/ext2/xattr.c
@@ -823,7 +823,7 @@ cleanup:
 void
 ext2_xattr_put_super(struct super_block *sb)
 {
-	mb_cache_shrink(ext2_xattr_cache, sb->s_bdev);
+	mb_cache_shrink(sb->s_bdev);
 }
 
 
diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c
index d44431d..ca7f003 100644
--- a/fs/ext2/xip.c
+++ b/fs/ext2/xip.c
@@ -15,66 +15,79 @@
 #include "xip.h"
 
 static inline int
-__inode_direct_access(struct inode *inode, sector_t sector, unsigned long *data) {
+__inode_direct_access(struct inode *inode, sector_t sector,
+		      unsigned long *data)
+{
 	BUG_ON(!inode->i_sb->s_bdev->bd_disk->fops->direct_access);
 	return inode->i_sb->s_bdev->bd_disk->fops
 		->direct_access(inode->i_sb->s_bdev,sector,data);
 }
 
+static inline int
+__ext2_get_sector(struct inode *inode, sector_t offset, int create,
+		   sector_t *result)
+{
+	struct buffer_head tmp;
+	int rc;
+
+	memset(&tmp, 0, sizeof(struct buffer_head));
+	rc = ext2_get_block(inode, offset/ (PAGE_SIZE/512), &tmp,
+			    create);
+	*result = tmp.b_blocknr;
+
+	/* did we get a sparse block (hole in the file)? */
+	if (!tmp.b_blocknr && !rc) {
+		BUG_ON(create);
+		rc = -ENODATA;
+	}
+
+	return rc;
+}
+
 int
-ext2_clear_xip_target(struct inode *inode, int block) {
-	sector_t sector = block*(PAGE_SIZE/512);
+ext2_clear_xip_target(struct inode *inode, int block)
+{
+	sector_t sector = block * (PAGE_SIZE/512);
 	unsigned long data;
 	int rc;
 
 	rc = __inode_direct_access(inode, sector, &data);
-	if (rc)
-		return rc;
-	clear_page((void*)data);
-	return 0;
+	if (!rc)
+		clear_page((void*)data);
+	return rc;
 }
 
 void ext2_xip_verify_sb(struct super_block *sb)
 {
 	struct ext2_sb_info *sbi = EXT2_SB(sb);
 
-	if ((sbi->s_mount_opt & EXT2_MOUNT_XIP)) {
-		if ((sb->s_bdev == NULL) ||
-			sb->s_bdev->bd_disk == NULL ||
-			sb->s_bdev->bd_disk->fops == NULL ||
-			sb->s_bdev->bd_disk->fops->direct_access == NULL) {
-			sbi->s_mount_opt &= (~EXT2_MOUNT_XIP);
-			ext2_warning(sb, __FUNCTION__,
-				"ignoring xip option - not supported by bdev");
-		}
+	if ((sbi->s_mount_opt & EXT2_MOUNT_XIP) &&
+	    !sb->s_bdev->bd_disk->fops->direct_access) {
+		sbi->s_mount_opt &= (~EXT2_MOUNT_XIP);
+		ext2_warning(sb, __FUNCTION__,
+			     "ignoring xip option - not supported by bdev");
 	}
 }
 
-struct page*
-ext2_get_xip_page(struct address_space *mapping, sector_t blockno,
+struct page *
+ext2_get_xip_page(struct address_space *mapping, sector_t offset,
 		   int create)
 {
 	int rc;
 	unsigned long data;
-	struct buffer_head tmp;
+	sector_t sector;
 
-	tmp.b_state = 0;
-	tmp.b_blocknr = 0;
-	rc = ext2_get_block(mapping->host, blockno/(PAGE_SIZE/512) , &tmp,
-				create);
+	/* first, retrieve the sector number */
+	rc = __ext2_get_sector(mapping->host, offset, create, &sector);
 	if (rc)
-		return ERR_PTR(rc);
-	if (tmp.b_blocknr == 0) {
-		/* SPARSE block */
-		BUG_ON(create);
-		return ERR_PTR(-ENODATA);
-	}
+		goto error;
 
+	/* retrieve address of the target data */
 	rc = __inode_direct_access
-		(mapping->host,tmp.b_blocknr*(PAGE_SIZE/512) ,&data);
-	if (rc)
-		return ERR_PTR(rc);
+		(mapping->host, sector * (PAGE_SIZE/512), &data);
+	if (!rc)
+		return virt_to_page(data);
 
-	SetPageUptodate(virt_to_page(data));
-	return virt_to_page(data);
+ error:
+	return ERR_PTR(rc);
 }
diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c
index 1e6f3ea..6981bd0 100644
--- a/fs/ext3/ialloc.c
+++ b/fs/ext3/ialloc.c
@@ -604,12 +604,14 @@ got:
 	err = ext3_init_acl(handle, inode, dir);
 	if (err) {
 		DQUOT_FREE_INODE(inode);
+		DQUOT_DROP(inode);
 		goto fail2;
   	}
 	err = ext3_mark_inode_dirty(handle, inode);
 	if (err) {
 		ext3_std_error(sb, err);
 		DQUOT_FREE_INODE(inode);
+		DQUOT_DROP(inode);
 		goto fail2;
 	}
 
diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c
index 3f9dfa6..269c7b9 100644
--- a/fs/ext3/xattr.c
+++ b/fs/ext3/xattr.c
@@ -1106,7 +1106,7 @@ cleanup:
 void
 ext3_xattr_put_super(struct super_block *sb)
 {
-	mb_cache_shrink(ext3_xattr_cache, sb->s_bdev);
+	mb_cache_shrink(sb->s_bdev);
 }
 
 /*
diff --git a/fs/fcntl.c b/fs/fcntl.c
index 286a9f8..6fbc9d8 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -288,7 +288,7 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 		break;
 	case F_SETLK:
 	case F_SETLKW:
-		err = fcntl_setlk(filp, cmd, (struct flock __user *) arg);
+		err = fcntl_setlk(fd, filp, cmd, (struct flock __user *) arg);
 		break;
 	case F_GETOWN:
 		/*
@@ -376,7 +376,8 @@ asmlinkage long sys_fcntl64(unsigned int fd, unsigned int cmd, unsigned long arg
 			break;
 		case F_SETLK64:
 		case F_SETLKW64:
-			err = fcntl_setlk64(filp, cmd, (struct flock64 __user *) arg);
+			err = fcntl_setlk64(fd, filp, cmd,
+					(struct flock64 __user *) arg);
 			break;
 		default:
 			err = do_fcntl(fd, cmd, arg, filp);
diff --git a/fs/hostfs/hostfs.h b/fs/hostfs/hostfs.h
index c1516d0..67bca0d 100644
--- a/fs/hostfs/hostfs.h
+++ b/fs/hostfs/hostfs.h
@@ -69,6 +69,7 @@ extern int read_file(int fd, unsigned long long *offset, char *buf, int len);
 extern int write_file(int fd, unsigned long long *offset, const char *buf,
 		      int len);
 extern int lseek_file(int fd, long long offset, int whence);
+extern int fsync_file(int fd, int datasync);
 extern int file_create(char *name, int ur, int uw, int ux, int gr,
 		       int gw, int gx, int or, int ow, int ox);
 extern int set_attr(const char *file, struct hostfs_iattr *attrs);
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index 4bf43ea..b2d1820 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -15,7 +15,6 @@
 #include <linux/pagemap.h>
 #include <linux/blkdev.h>
 #include <linux/list.h>
-#include <linux/root_dev.h>
 #include <linux/statfs.h>
 #include <linux/kdev_t.h>
 #include <asm/uaccess.h>
@@ -160,8 +159,6 @@ static int read_name(struct inode *ino, char *name)
 	ino->i_size = i_size;
 	ino->i_blksize = i_blksize;
 	ino->i_blocks = i_blocks;
-	if((ino->i_sb->s_dev == ROOT_DEV) && (ino->i_uid == getuid()))
-		ino->i_uid = 0;
 	return(0);
 }
 
@@ -385,7 +382,7 @@ int hostfs_file_open(struct inode *ino, struct file *file)
 
 int hostfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 {
-	return(0);
+	return fsync_file(HOSTFS_I(dentry->d_inode)->fd, datasync);
 }
 
 static struct file_operations hostfs_file_fops = {
@@ -841,16 +838,10 @@ int hostfs_setattr(struct dentry *dentry, struct iattr *attr)
 		attrs.ia_mode = attr->ia_mode;
 	}
 	if(attr->ia_valid & ATTR_UID){
-		if((dentry->d_inode->i_sb->s_dev == ROOT_DEV) &&
-		   (attr->ia_uid == 0))
-			attr->ia_uid = getuid();
 		attrs.ia_valid |= HOSTFS_ATTR_UID;
 		attrs.ia_uid = attr->ia_uid;
 	}
 	if(attr->ia_valid & ATTR_GID){
-		if((dentry->d_inode->i_sb->s_dev == ROOT_DEV) &&
-		   (attr->ia_gid == 0))
-			attr->ia_gid = getgid();
 		attrs.ia_valid |= HOSTFS_ATTR_GID;
 		attrs.ia_gid = attr->ia_gid;
 	}
diff --git a/fs/hostfs/hostfs_user.c b/fs/hostfs/hostfs_user.c
index 4796e84..b97809d 100644
--- a/fs/hostfs/hostfs_user.c
+++ b/fs/hostfs/hostfs_user.c
@@ -153,10 +153,24 @@ int lseek_file(int fd, long long offset, int whence)
 	int ret;
 
 	ret = lseek64(fd, offset, whence);
-	if(ret < 0) return(-errno);
+	if(ret < 0)
+		return(-errno);
 	return(0);
 }
 
+int fsync_file(int fd, int datasync)
+{
+	int ret;
+	if (datasync)
+		ret = fdatasync(fd);
+	else
+		ret = fsync(fd);
+
+	if (ret < 0)
+		return -errno;
+	return 0;
+}
+
 void close_file(void *stream)
 {
 	close(*((int *) stream));
diff --git a/fs/hppfs/hppfs_kern.c b/fs/hppfs/hppfs_kern.c
index 6f553e1..ff150fe 100644
--- a/fs/hppfs/hppfs_kern.c
+++ b/fs/hppfs/hppfs_kern.c
@@ -233,7 +233,7 @@ static ssize_t read_proc(struct file *file, char *buf, ssize_t count,
 		set_fs(USER_DS);
 
 	if(ppos) *ppos = file->f_pos;
-	return(n);
+	return n;
 }
 
 static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count)
@@ -254,7 +254,7 @@ static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count)
 		err = os_read_file(fd, new_buf, cur);
 		if(err < 0){
 			printk("hppfs_read : read failed, errno = %d\n",
-			       count);
+			       err);
 			n = err;
 			goto out_free;
 		}
@@ -271,7 +271,7 @@ static ssize_t hppfs_read_file(int fd, char *buf, ssize_t count)
  out_free:
 	kfree(new_buf);
  out:
-	return(n);
+	return n;
 }
 
 static ssize_t hppfs_read(struct file *file, char *buf, size_t count,
diff --git a/fs/inode.c b/fs/inode.c
index 96364fa..e57f172 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -757,6 +757,7 @@ EXPORT_SYMBOL(igrab);
  * @head:       the head of the list to search
  * @test:	callback used for comparisons between inodes
  * @data:	opaque data pointer to pass to @test
+ * @wait:	if true wait for the inode to be unlocked, if false do not
  *
  * ifind() searches for the inode specified by @data in the inode
  * cache. This is a generalized version of ifind_fast() for file systems where
@@ -771,7 +772,7 @@ EXPORT_SYMBOL(igrab);
  */
 static inline struct inode *ifind(struct super_block *sb,
 		struct hlist_head *head, int (*test)(struct inode *, void *),
-		void *data)
+		void *data, const int wait)
 {
 	struct inode *inode;
 
@@ -780,7 +781,8 @@ static inline struct inode *ifind(struct super_block *sb,
 	if (inode) {
 		__iget(inode);
 		spin_unlock(&inode_lock);
-		wait_on_inode(inode);
+		if (likely(wait))
+			wait_on_inode(inode);
 		return inode;
 	}
 	spin_unlock(&inode_lock);
@@ -820,7 +822,7 @@ static inline struct inode *ifind_fast(struct super_block *sb,
 }
 
 /**
- * ilookup5 - search for an inode in the inode cache
+ * ilookup5_nowait - search for an inode in the inode cache
  * @sb:		super block of file system to search
  * @hashval:	hash value (usually inode number) to search for
  * @test:	callback used for comparisons between inodes
@@ -832,7 +834,38 @@ static inline struct inode *ifind_fast(struct super_block *sb,
  * identification of an inode.
  *
  * If the inode is in the cache, the inode is returned with an incremented
- * reference count.
+ * reference count.  Note, the inode lock is not waited upon so you have to be
+ * very careful what you do with the returned inode.  You probably should be
+ * using ilookup5() instead.
+ *
+ * Otherwise NULL is returned.
+ *
+ * Note, @test is called with the inode_lock held, so can't sleep.
+ */
+struct inode *ilookup5_nowait(struct super_block *sb, unsigned long hashval,
+		int (*test)(struct inode *, void *), void *data)
+{
+	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
+
+	return ifind(sb, head, test, data, 0);
+}
+
+EXPORT_SYMBOL(ilookup5_nowait);
+
+/**
+ * ilookup5 - search for an inode in the inode cache
+ * @sb:		super block of file system to search
+ * @hashval:	hash value (usually inode number) to search for
+ * @test:	callback used for comparisons between inodes
+ * @data:	opaque data pointer to pass to @test
+ *
+ * ilookup5() uses ifind() to search for the inode specified by @hashval and
+ * @data in the inode cache. This is a generalized version of ilookup() for
+ * file systems where the inode number is not sufficient for unique
+ * identification of an inode.
+ *
+ * If the inode is in the cache, the inode lock is waited upon and the inode is
+ * returned with an incremented reference count.
  *
  * Otherwise NULL is returned.
  *
@@ -843,7 +876,7 @@ struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
 {
 	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
 
-	return ifind(sb, head, test, data);
+	return ifind(sb, head, test, data, 1);
 }
 
 EXPORT_SYMBOL(ilookup5);
@@ -900,7 +933,7 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
 	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
 	struct inode *inode;
 
-	inode = ifind(sb, head, test, data);
+	inode = ifind(sb, head, test, data, 1);
 	if (inode)
 		return inode;
 	/*
diff --git a/fs/inotify.c b/fs/inotify.c
index e423bfe..a8a714e 100644
--- a/fs/inotify.c
+++ b/fs/inotify.c
@@ -29,8 +29,6 @@
 #include <linux/mount.h>
 #include <linux/namei.h>
 #include <linux/poll.h>
-#include <linux/device.h>
-#include <linux/miscdevice.h>
 #include <linux/init.h>
 #include <linux/list.h>
 #include <linux/writeback.h>
@@ -45,8 +43,8 @@ static kmem_cache_t *event_cachep;
 
 static struct vfsmount *inotify_mnt;
 
-/* These are configurable via /proc/sys/inotify */
-int inotify_max_user_devices;
+/* these are configurable via /proc/sys/fs/inotify/ */
+int inotify_max_user_instances;
 int inotify_max_user_watches;
 int inotify_max_queued_events;
 
@@ -64,8 +62,8 @@ int inotify_max_queued_events;
  * Lifetimes of the three main data structures--inotify_device, inode, and
  * inotify_watch--are managed by reference count.
  *
- * inotify_device: Lifetime is from open until release.  Additional references
- * can bump the count via get_inotify_dev() and drop the count via
+ * inotify_device: Lifetime is from inotify_init() until release.  Additional
+ * references can bump the count via get_inotify_dev() and drop the count via
  * put_inotify_dev().
  *
  * inotify_watch: Lifetime is from create_watch() to destory_watch().
@@ -77,7 +75,7 @@ int inotify_max_queued_events;
  */
 
 /*
- * struct inotify_device - represents an open instance of an inotify device
+ * struct inotify_device - represents an inotify instance
  *
  * This structure is protected by the semaphore 'sem'.
  */
@@ -125,6 +123,47 @@ struct inotify_watch {
 	u32			mask;	/* event mask for this watch */
 };
 
+#ifdef CONFIG_SYSCTL
+
+#include <linux/sysctl.h>
+
+static int zero;
+
+ctl_table inotify_table[] = {
+	{
+		.ctl_name	= INOTIFY_MAX_USER_INSTANCES,
+		.procname	= "max_user_instances",
+		.data		= &inotify_max_user_instances,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &zero,
+	},
+	{
+		.ctl_name	= INOTIFY_MAX_USER_WATCHES,
+		.procname	= "max_user_watches",
+		.data		= &inotify_max_user_watches,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &zero, 
+	},
+	{
+		.ctl_name	= INOTIFY_MAX_QUEUED_EVENTS,
+		.procname	= "max_queued_events",
+		.data		= &inotify_max_queued_events,
+		.maxlen		= sizeof(int),
+		.mode		= 0644, 
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec, 
+		.extra1		= &zero
+	},
+	{ .ctl_name = 0 }
+};
+#endif /* CONFIG_SYSCTL */
+
 static inline void get_inotify_dev(struct inotify_device *dev)
 {
 	atomic_inc(&dev->count);
@@ -332,7 +371,7 @@ static int find_inode(const char __user *dirname, struct nameidata *nd)
 	/* you can only watch an inode if you have read permissions on it */
 	error = permission(nd->dentry->d_inode, MAY_READ, NULL);
 	if (error) 
-		path_release (nd);
+		path_release(nd);
 	return error;
 }
 
@@ -348,7 +387,8 @@ static struct inotify_watch *create_watch(struct inotify_device *dev,
 	struct inotify_watch *watch;
 	int ret;
 
-	if (atomic_read(&dev->user->inotify_watches) >= inotify_max_user_watches)
+	if (atomic_read(&dev->user->inotify_watches) >=
+			inotify_max_user_watches)
 		return ERR_PTR(-ENOSPC);
 
 	watch = kmem_cache_alloc(watch_cachep, GFP_KERNEL);
@@ -744,15 +784,14 @@ static int inotify_release(struct inode *ignored, struct file *file)
 		inotify_dev_event_dequeue(dev);
 	up(&dev->sem);
 
-	/* free this device: the put matching the get in inotify_open() */
+	/* free this device: the put matching the get in inotify_init() */
 	put_inotify_dev(dev);
 
 	return 0;
 }
 
 /*
- * inotify_ignore - handle the INOTIFY_IGNORE ioctl, asking that a given wd be
- * removed from the device.
+ * inotify_ignore - remove a given wd from this inotify instance.
  *
  * Can sleep.
  */
@@ -817,42 +856,40 @@ asmlinkage long sys_inotify_init(void)
 {
 	struct inotify_device *dev;
 	struct user_struct *user;
-	int ret = -ENOTTY;
-	int fd;
-	struct file *filp;
+	struct file *filp;	
+	int fd, ret;
 
 	fd = get_unused_fd();
-	if (fd < 0) {
-		ret = fd;
-		goto out;
-	}
+	if (fd < 0)
+		return fd;
 
 	filp = get_empty_filp();
 	if (!filp) {
-		put_unused_fd(fd);
 		ret = -ENFILE;
-		goto out;
+		goto out_put_fd;
 	}
-	filp->f_op = &inotify_fops;
-	filp->f_vfsmnt = mntget(inotify_mnt);
-	filp->f_dentry = dget(inotify_mnt->mnt_root);
-	filp->f_mapping = filp->f_dentry->d_inode->i_mapping;
-	filp->f_mode = FMODE_READ;
-	filp->f_flags = O_RDONLY;
 
 	user = get_uid(current->user);
-
-	if (unlikely(atomic_read(&user->inotify_devs) >= inotify_max_user_devices)) {
+	if (unlikely(atomic_read(&user->inotify_devs) >=
+			inotify_max_user_instances)) {
 		ret = -EMFILE;
-		goto out_err;
+		goto out_free_uid;
 	}
 
 	dev = kmalloc(sizeof(struct inotify_device), GFP_KERNEL);
 	if (unlikely(!dev)) {
 		ret = -ENOMEM;
-		goto out_err;
+		goto out_free_uid;
 	}
 
+	filp->f_op = &inotify_fops;
+	filp->f_vfsmnt = mntget(inotify_mnt);
+	filp->f_dentry = dget(inotify_mnt->mnt_root);
+	filp->f_mapping = filp->f_dentry->d_inode->i_mapping;
+	filp->f_mode = FMODE_READ;
+	filp->f_flags = O_RDONLY;
+	filp->private_data = dev;
+
 	idr_init(&dev->idr);
 	INIT_LIST_HEAD(&dev->events);
 	INIT_LIST_HEAD(&dev->watches);
@@ -866,46 +903,50 @@ asmlinkage long sys_inotify_init(void)
 
 	get_inotify_dev(dev);
 	atomic_inc(&user->inotify_devs);
+	fd_install(fd, filp);
 
-	filp->private_data = dev;
-	fd_install (fd, filp);
 	return fd;
-out_err:
-	put_unused_fd (fd);
-	put_filp (filp);
+out_free_uid:
 	free_uid(user);
-out:
+	put_filp(filp);
+out_put_fd:
+	put_unused_fd(fd);
 	return ret;
 }
 
-asmlinkage long sys_inotify_add_watch(int fd, const char *path, u32 mask)
+asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
 {
 	struct inotify_watch *watch, *old;
 	struct inode *inode;
 	struct inotify_device *dev;
 	struct nameidata nd;
 	struct file *filp;
-	int ret;
+	int ret, fput_needed;
 
-	filp = fget(fd);
-	if (!filp)
+	filp = fget_light(fd, &fput_needed);
+	if (unlikely(!filp))
 		return -EBADF;
 
-	dev = filp->private_data;
+	/* verify that this is indeed an inotify instance */
+	if (unlikely(filp->f_op != &inotify_fops)) {
+		ret = -EINVAL;
+		goto fput_and_out;
+	}
 
-	ret = find_inode ((const char __user*)path, &nd);
-	if (ret)
+	ret = find_inode(path, &nd);
+	if (unlikely(ret))
 		goto fput_and_out;
 
-	/* Held in place by reference in nd */
+	/* inode held in place by reference to nd; dev by fget on fd */
 	inode = nd.dentry->d_inode;
+	dev = filp->private_data;
 
 	down(&inode->inotify_sem);
 	down(&dev->sem);
 
 	/* don't let user-space set invalid bits: we don't want flags set */
 	mask &= IN_ALL_EVENTS;
-	if (!mask) {
+	if (unlikely(!mask)) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -932,11 +973,11 @@ asmlinkage long sys_inotify_add_watch(int fd, const char *path, u32 mask)
 	list_add(&watch->i_list, &inode->inotify_watches);
 	ret = watch->wd;
 out:
-	path_release (&nd);
 	up(&dev->sem);
 	up(&inode->inotify_sem);
+	path_release(&nd);
 fput_and_out:
-	fput(filp);
+	fput_light(filp, fput_needed);
 	return ret;
 }
 
@@ -944,14 +985,23 @@ asmlinkage long sys_inotify_rm_watch(int fd, u32 wd)
 {
 	struct file *filp;
 	struct inotify_device *dev;
-	int ret;
+	int ret, fput_needed;
 
-	filp = fget(fd);
-	if (!filp)
+	filp = fget_light(fd, &fput_needed);
+	if (unlikely(!filp))
 		return -EBADF;
+
+	/* verify that this is indeed an inotify instance */
+	if (unlikely(filp->f_op != &inotify_fops)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
 	dev = filp->private_data;
-	ret = inotify_ignore (dev, wd);
-	fput(filp);
+	ret = inotify_ignore(dev, wd);
+
+out:
+	fput_light(filp, fput_needed);
 	return ret;
 }
 
@@ -969,17 +1019,24 @@ static struct file_system_type inotify_fs_type = {
 };
 
 /*
- * inotify_init - Our initialization function.  Note that we cannnot return
+ * inotify_setup - Our initialization function.  Note that we cannnot return
  * error because we have compiled-in VFS hooks.  So an (unlikely) failure here
  * must result in panic().
  */
-static int __init inotify_init(void)
+static int __init inotify_setup(void)
 {
-	register_filesystem(&inotify_fs_type);
+	int ret;
+
+	ret = register_filesystem(&inotify_fs_type);
+	if (unlikely(ret))
+		panic("inotify: register_filesystem returned %d!\n", ret);
+
 	inotify_mnt = kern_mount(&inotify_fs_type);
+	if (IS_ERR(inotify_mnt))
+		panic("inotify: kern_mount ret %ld!\n", PTR_ERR(inotify_mnt));
 
-	inotify_max_queued_events = 8192;
-	inotify_max_user_devices = 128;
+	inotify_max_queued_events = 16384;
+	inotify_max_user_instances = 128;
 	inotify_max_user_watches = 8192;
 
 	atomic_set(&inotify_cookie, 0);
@@ -991,9 +1048,7 @@ static int __init inotify_init(void)
 					 sizeof(struct inotify_kernel_event),
 					 0, SLAB_PANIC, NULL, NULL);
 
-	printk(KERN_INFO "inotify syscall\n");
-
 	return 0;
 }
 
-module_init(inotify_init);
+module_init(inotify_setup);
diff --git a/fs/jffs/intrep.c b/fs/jffs/intrep.c
index fc589dd..456d7e6 100644
--- a/fs/jffs/intrep.c
+++ b/fs/jffs/intrep.c
@@ -3397,6 +3397,9 @@ jffs_garbage_collect_thread(void *ptr)
 			siginfo_t info;
 			unsigned long signr = 0;
 
+			if (try_to_freeze())
+				continue;
+
 			spin_lock_irq(&current->sighand->siglock);
 			signr = dequeue_signal(current, &current->blocked, &info);
 			spin_unlock_irq(&current->sighand->siglock);
diff --git a/fs/jffs2/build.c b/fs/jffs2/build.c
index 3dd5394..97dc397 100644
--- a/fs/jffs2/build.c
+++ b/fs/jffs2/build.c
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: build.c,v 1.70 2005/02/28 08:21:05 dedekind Exp $
+ * $Id: build.c,v 1.71 2005/07/12 16:37:08 dedekind Exp $
  *
  */
 
@@ -336,13 +336,6 @@ int jffs2_do_mount_fs(struct jffs2_sb_info *c)
 		c->blocks[i].bad_count = 0;
 	}
 
-	init_MUTEX(&c->alloc_sem);
-	init_MUTEX(&c->erase_free_sem);
-	init_waitqueue_head(&c->erase_wait);
-	init_waitqueue_head(&c->inocache_wq);
-	spin_lock_init(&c->erase_completion_lock);
-	spin_lock_init(&c->inocache_lock);
-
 	INIT_LIST_HEAD(&c->clean_list);
 	INIT_LIST_HEAD(&c->very_dirty_list);
 	INIT_LIST_HEAD(&c->dirty_list);
diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c
index 6a4c0a3..787d84a 100644
--- a/fs/jffs2/erase.c
+++ b/fs/jffs2/erase.c
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: erase.c,v 1.76 2005/05/03 15:11:40 dedekind Exp $
+ * $Id: erase.c,v 1.80 2005/07/14 19:46:24 joern Exp $
  *
  */
 
@@ -300,100 +300,86 @@ static void jffs2_free_all_node_refs(struct jffs2_sb_info *c, struct jffs2_erase
 	jeb->last_node = NULL;
 }
 
-static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb)
+static int jffs2_block_check_erase(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb, uint32_t *bad_offset)
 {
-	struct jffs2_raw_node_ref *marker_ref = NULL;
-	unsigned char *ebuf;
+	void *ebuf;
+	uint32_t ofs;
 	size_t retlen;
-	int ret;
-	uint32_t bad_offset;
-
-	if ((!jffs2_cleanmarker_oob(c)) && (c->cleanmarker_size > 0)) {
-		marker_ref = jffs2_alloc_raw_node_ref();
-		if (!marker_ref) {
-			printk(KERN_WARNING "Failed to allocate raw node ref for clean marker\n");
-			/* Stick it back on the list from whence it came and come back later */
-			jffs2_erase_pending_trigger(c);
-			spin_lock(&c->erase_completion_lock);
-			list_add(&jeb->list, &c->erase_complete_list);
-			spin_unlock(&c->erase_completion_lock);
-			return;
-		}
-	}
+	int ret = -EIO;
+	
 	ebuf = kmalloc(PAGE_SIZE, GFP_KERNEL);
 	if (!ebuf) {
-		printk(KERN_WARNING "Failed to allocate page buffer for verifying erase at 0x%08x. Assuming it worked\n", jeb->offset);
-	} else {
-		uint32_t ofs = jeb->offset;
+		printk(KERN_WARNING "Failed to allocate page buffer for verifying erase at 0x%08x. Refiling\n", jeb->offset);
+		return -EAGAIN;
+	}
 
-		D1(printk(KERN_DEBUG "Verifying erase at 0x%08x\n", jeb->offset));
-		while(ofs < jeb->offset + c->sector_size) {
-			uint32_t readlen = min((uint32_t)PAGE_SIZE, jeb->offset + c->sector_size - ofs);
-			int i;
+	D1(printk(KERN_DEBUG "Verifying erase at 0x%08x\n", jeb->offset));
 
-			bad_offset = ofs;
+	for (ofs = jeb->offset; ofs < jeb->offset + c->sector_size; ) {
+		uint32_t readlen = min((uint32_t)PAGE_SIZE, jeb->offset + c->sector_size - ofs);
+		int i;
 
-			ret = c->mtd->read(c->mtd, ofs, readlen, &retlen, ebuf);
+		*bad_offset = ofs;
 
-			if (ret) {
-				printk(KERN_WARNING "Read of newly-erased block at 0x%08x failed: %d. Putting on bad_list\n", ofs, ret);
-				goto bad;
-			}
-			if (retlen != readlen) {
-				printk(KERN_WARNING "Short read from newly-erased block at 0x%08x. Wanted %d, got %zd\n", ofs, readlen, retlen);
-				goto bad;
-			}
-			for (i=0; i<readlen; i += sizeof(unsigned long)) {
-				/* It's OK. We know it's properly aligned */
-				unsigned long datum = *(unsigned long *)(&ebuf[i]);
-				if (datum + 1) {
-					bad_offset += i;
-					printk(KERN_WARNING "Newly-erased block contained word 0x%lx at offset 0x%08x\n", datum, bad_offset);
-				bad: 
-					if ((!jffs2_cleanmarker_oob(c)) && (c->cleanmarker_size > 0))
-						jffs2_free_raw_node_ref(marker_ref);
-					kfree(ebuf);
-				bad2:
-					spin_lock(&c->erase_completion_lock);
-					/* Stick it on a list (any list) so
-					   erase_failed can take it right off
-					   again.  Silly, but shouldn't happen
-					   often. */
-					list_add(&jeb->list, &c->erasing_list);
-					spin_unlock(&c->erase_completion_lock);
-					jffs2_erase_failed(c, jeb, bad_offset);
-					return;
-				}
+		ret = jffs2_flash_read(c, ofs, readlen, &retlen, ebuf);
+		if (ret) {
+			printk(KERN_WARNING "Read of newly-erased block at 0x%08x failed: %d. Putting on bad_list\n", ofs, ret);
+			goto fail;
+		}
+		if (retlen != readlen) {
+			printk(KERN_WARNING "Short read from newly-erased block at 0x%08x. Wanted %d, got %zd\n", ofs, readlen, retlen);
+			goto fail;
+		}
+		for (i=0; i<readlen; i += sizeof(unsigned long)) {
+			/* It's OK. We know it's properly aligned */
+			unsigned long *datum = ebuf + i;
+			if (*datum + 1) {
+				*bad_offset += i;
+				printk(KERN_WARNING "Newly-erased block contained word 0x%lx at offset 0x%08x\n", *datum, *bad_offset);
+				goto fail;
 			}
-			ofs += readlen;
-			cond_resched();
 		}
-		kfree(ebuf);
+		ofs += readlen;
+		cond_resched();
 	}
+	ret = 0;
+fail:
+	kfree(ebuf);
+	return ret;
+}
 
-	bad_offset = jeb->offset;
+static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb)
+{
+	struct jffs2_raw_node_ref *marker_ref = NULL;
+	size_t retlen;
+	int ret;
+	uint32_t bad_offset;
+
+	switch (jffs2_block_check_erase(c, jeb, &bad_offset)) {
+	case -EAGAIN:	goto refile;
+	case -EIO:	goto filebad;
+	}
 
 	/* Write the erase complete marker */	
 	D1(printk(KERN_DEBUG "Writing erased marker to block at 0x%08x\n", jeb->offset));
-	if (jffs2_cleanmarker_oob(c)) {
+	bad_offset = jeb->offset;
 
-		if (jffs2_write_nand_cleanmarker(c, jeb))
-			goto bad2;
-			
-		jeb->first_node = jeb->last_node = NULL;
+	/* Cleanmarker in oob area or no cleanmarker at all ? */
+	if (jffs2_cleanmarker_oob(c) || c->cleanmarker_size == 0) {
 
-		jeb->free_size = c->sector_size;
-		jeb->used_size = 0;
-		jeb->dirty_size = 0;
-		jeb->wasted_size = 0;
-	} else if (c->cleanmarker_size == 0) {
-		jeb->first_node = jeb->last_node = NULL;
+		if (jffs2_cleanmarker_oob(c)) {
+			if (jffs2_write_nand_cleanmarker(c, jeb))
+				goto filebad;
+		}
 
+		jeb->first_node = jeb->last_node = NULL;
 		jeb->free_size = c->sector_size;
 		jeb->used_size = 0;
 		jeb->dirty_size = 0;
 		jeb->wasted_size = 0;
+
 	} else {
+
 		struct kvec vecs[1];
 		struct jffs2_unknown_node marker = {
 			.magic =	cpu_to_je16(JFFS2_MAGIC_BITMASK),
@@ -401,21 +387,28 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb
 			.totlen =	cpu_to_je32(c->cleanmarker_size)
 		};
 
+		marker_ref = jffs2_alloc_raw_node_ref();
+		if (!marker_ref) {
+			printk(KERN_WARNING "Failed to allocate raw node ref for clean marker. Refiling\n");
+			goto refile;
+		}
+
 		marker.hdr_crc = cpu_to_je32(crc32(0, &marker, sizeof(struct jffs2_unknown_node)-4));
 
 		vecs[0].iov_base = (unsigned char *) &marker;
 		vecs[0].iov_len = sizeof(marker);
 		ret = jffs2_flash_direct_writev(c, vecs, 1, jeb->offset, &retlen);
 		
-		if (ret) {
-			printk(KERN_WARNING "Write clean marker to block at 0x%08x failed: %d\n",
-			       jeb->offset, ret);
-			goto bad2;
-		}
-		if (retlen != sizeof(marker)) {
-			printk(KERN_WARNING "Short write to newly-erased block at 0x%08x: Wanted %zd, got %zd\n",
-			       jeb->offset, sizeof(marker), retlen);
-			goto bad2;
+		if (ret || retlen != sizeof(marker)) {
+			if (ret)
+				printk(KERN_WARNING "Write clean marker to block at 0x%08x failed: %d\n",
+				       jeb->offset, ret);
+			else
+				printk(KERN_WARNING "Short write to newly-erased block at 0x%08x: Wanted %zd, got %zd\n",
+				       jeb->offset, sizeof(marker), retlen);
+
+			jffs2_free_raw_node_ref(marker_ref);
+			goto filebad;
 		}
 
 		marker_ref->next_in_ino = NULL;
@@ -444,5 +437,22 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb
 	c->nr_free_blocks++;
 	spin_unlock(&c->erase_completion_lock);
 	wake_up(&c->erase_wait);
-}
+	return;
+
+filebad:
+	spin_lock(&c->erase_completion_lock);
+	/* Stick it on a list (any list) so erase_failed can take it
+	   right off again.  Silly, but shouldn't happen often. */
+	list_add(&jeb->list, &c->erasing_list);
+	spin_unlock(&c->erase_completion_lock);
+	jffs2_erase_failed(c, jeb, bad_offset);
+	return;
 
+refile:
+	/* Stick it back on the list from whence it came and come back later */
+	jffs2_erase_pending_trigger(c);
+	spin_lock(&c->erase_completion_lock);
+	list_add(&jeb->list, &c->erase_complete_list);
+	spin_unlock(&c->erase_completion_lock);
+	return;
+}
diff --git a/fs/jffs2/nodelist.c b/fs/jffs2/nodelist.c
index c7bbdee..4991c34 100644
--- a/fs/jffs2/nodelist.c
+++ b/fs/jffs2/nodelist.c
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: nodelist.c,v 1.97 2005/07/06 15:18:41 dwmw2 Exp $
+ * $Id: nodelist.c,v 1.98 2005/07/10 15:15:32 dedekind Exp $
  *
  */
 
@@ -55,11 +55,11 @@ void jffs2_add_fd_to_list(struct jffs2_sb_info *c, struct jffs2_full_dirent *new
 	});
 }
 
-/* Put a new tmp_dnode_info into the list, keeping the list in 
-   order of increasing version
-*/
-
-static void jffs2_add_tn_to_list(struct jffs2_tmp_dnode_info *tn, struct rb_root *list)
+/* 
+ * Put a new tmp_dnode_info into the temporaty RB-tree, keeping the list in 
+ * order of increasing version.
+ */
+static void jffs2_add_tn_to_tree(struct jffs2_tmp_dnode_info *tn, struct rb_root *list)
 {
 	struct rb_node **p = &list->rb_node;
 	struct rb_node * parent = NULL;
@@ -420,7 +420,7 @@ int jffs2_get_inode_nodes(struct jffs2_sb_info *c, struct jffs2_inode_info *f,
 			D1(printk(KERN_DEBUG "dnode @%08x: ver %u, offset %04x, dsize %04x\n",
 				  ref_offset(ref), je32_to_cpu(node.i.version),
 				  je32_to_cpu(node.i.offset), je32_to_cpu(node.i.dsize)));
-			jffs2_add_tn_to_list(tn, &ret_tn);
+			jffs2_add_tn_to_tree(tn, &ret_tn);
 			break;
 
 		default:
diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h
index 7bf72e0..d900c89 100644
--- a/fs/jffs2/os-linux.h
+++ b/fs/jffs2/os-linux.h
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: os-linux.h,v 1.57 2005/07/06 12:13:09 dwmw2 Exp $
+ * $Id: os-linux.h,v 1.58 2005/07/12 02:34:35 tpoynor Exp $
  *
  */
 
@@ -86,6 +86,8 @@ static inline void jffs2_init_inode_info(struct jffs2_inode_info *f)
 #define jffs2_dataflash(c) (0)
 #define jffs2_nor_ecc_flash_setup(c) (0)
 #define jffs2_nor_ecc_flash_cleanup(c) do {} while (0)
+#define jffs2_dataflash_setup(c) (0)
+#define jffs2_dataflash_cleanup(c) do {} while (0)
 
 #else /* NAND and/or ECC'd NOR support present */
 
diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c
index 081656c..5b2a835 100644
--- a/fs/jffs2/readinode.c
+++ b/fs/jffs2/readinode.c
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: readinode.c,v 1.120 2005/07/05 21:03:07 dwmw2 Exp $
+ * $Id: readinode.c,v 1.125 2005/07/10 13:13:55 dedekind Exp $
  *
  */
 
@@ -151,6 +151,9 @@ int jffs2_add_full_dnode_to_inode(struct jffs2_sb_info *c, struct jffs2_inode_in
 
 	D1(printk(KERN_DEBUG "jffs2_add_full_dnode_to_inode(ino #%u, f %p, fn %p)\n", f->inocache->ino, f, fn));
 
+	if (unlikely(!fn->size))
+		return 0;
+
 	newfrag = jffs2_alloc_node_frag();
 	if (unlikely(!newfrag))
 		return -ENOMEM;
@@ -158,11 +161,6 @@ int jffs2_add_full_dnode_to_inode(struct jffs2_sb_info *c, struct jffs2_inode_in
 	D2(printk(KERN_DEBUG "adding node %04x-%04x @0x%08x on flash, newfrag *%p\n",
 		  fn->ofs, fn->ofs+fn->size, ref_offset(fn->raw), newfrag));
 	
-	if (unlikely(!fn->size)) {
-		jffs2_free_node_frag(newfrag);
-		return 0;
-	}
-
 	newfrag->ofs = fn->ofs;
 	newfrag->size = fn->size;
 	newfrag->node = fn;
@@ -560,7 +558,6 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c,
 		}
 	next_tn:
 		BUG_ON(rb->rb_left);
-		repl_rb = NULL;
 		if (rb->rb_parent && rb->rb_parent->rb_left == rb) {
 			/* We were then left-hand child of our parent. We need
 			   to move our own right-hand child into our place. */
diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c
index 2cf14cf..aaf9475c 100644
--- a/fs/jffs2/super.c
+++ b/fs/jffs2/super.c
@@ -7,7 +7,7 @@
  *
  * For licensing information, see the file 'LICENCE' in this directory.
  *
- * $Id: super.c,v 1.106 2005/05/18 11:37:25 dedekind Exp $
+ * $Id: super.c,v 1.107 2005/07/12 16:37:08 dedekind Exp $
  *
  */
 
@@ -140,6 +140,15 @@ static struct super_block *jffs2_get_sb_mtd(struct file_system_type *fs_type,
 	D1(printk(KERN_DEBUG "jffs2_get_sb_mtd(): New superblock for device %d (\"%s\")\n",
 		  mtd->index, mtd->name));
 
+	/* Initialize JFFS2 superblock locks, the further initialization will be
+	 * done later */
+	init_MUTEX(&c->alloc_sem);
+	init_MUTEX(&c->erase_free_sem);
+	init_waitqueue_head(&c->erase_wait);
+	init_waitqueue_head(&c->inocache_wq);
+	spin_lock_init(&c->erase_completion_lock);
+	spin_lock_init(&c->inocache_lock);
+
 	sb->s_op = &jffs2_super_operations;
 	sb->s_flags = flags | MS_NOATIME;
 
diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c
index cced2fe..c739626 100644
--- a/fs/jfs/jfs_dmap.c
+++ b/fs/jfs/jfs_dmap.c
@@ -26,36 +26,6 @@
 #include "jfs_debug.h"
 
 /*
- *	Debug code for double-checking block map
- */
-/* #define	_JFS_DEBUG_DMAP	1 */
-
-#ifdef	_JFS_DEBUG_DMAP
-#define DBINITMAP(size,ipbmap,results) \
-	DBinitmap(size,ipbmap,results)
-#define DBALLOC(dbmap,mapsize,blkno,nblocks) \
-	DBAlloc(dbmap,mapsize,blkno,nblocks)
-#define DBFREE(dbmap,mapsize,blkno,nblocks) \
-	DBFree(dbmap,mapsize,blkno,nblocks)
-#define DBALLOCCK(dbmap,mapsize,blkno,nblocks) \
-	DBAllocCK(dbmap,mapsize,blkno,nblocks)
-#define DBFREECK(dbmap,mapsize,blkno,nblocks) \
-	DBFreeCK(dbmap,mapsize,blkno,nblocks)
-
-static void DBinitmap(s64, struct inode *, u32 **);
-static void DBAlloc(uint *, s64, s64, s64);
-static void DBFree(uint *, s64, s64, s64);
-static void DBAllocCK(uint *, s64, s64, s64);
-static void DBFreeCK(uint *, s64, s64, s64);
-#else
-#define DBINITMAP(size,ipbmap,results)
-#define DBALLOC(dbmap, mapsize, blkno, nblocks)
-#define DBFREE(dbmap, mapsize, blkno, nblocks)
-#define DBALLOCCK(dbmap, mapsize, blkno, nblocks)
-#define DBFREECK(dbmap, mapsize, blkno, nblocks)
-#endif				/* _JFS_DEBUG_DMAP */
-
-/*
  *	SERIALIZATION of the Block Allocation Map.
  *
  *	the working state of the block allocation map is accessed in
@@ -105,7 +75,7 @@ static void dbAllocBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
 			int nblocks);
 static void dbSplit(dmtree_t * tp, int leafno, int splitsz, int newval);
 static void dbBackSplit(dmtree_t * tp, int leafno);
-static void dbJoin(dmtree_t * tp, int leafno, int newval);
+static int dbJoin(dmtree_t * tp, int leafno, int newval);
 static void dbAdjTree(dmtree_t * tp, int leafno, int newval);
 static int dbAdjCtl(struct bmap * bmp, s64 blkno, int newval, int alloc,
 		    int level);
@@ -128,8 +98,8 @@ static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks);
 static int dbFindBits(u32 word, int l2nb);
 static int dbFindCtl(struct bmap * bmp, int l2nb, int level, s64 * blkno);
 static int dbFindLeaf(dmtree_t * tp, int l2nb, int *leafidx);
-static void dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
-		       int nblocks);
+static int dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
+		      int nblocks);
 static int dbFreeDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
 		      int nblocks);
 static int dbMaxBud(u8 * cp);
@@ -242,7 +212,6 @@ int dbMount(struct inode *ipbmap)
 	JFS_SBI(ipbmap->i_sb)->bmap = bmp;
 
 	memset(bmp->db_active, 0, sizeof(bmp->db_active));
-	DBINITMAP(bmp->db_mapsize, ipbmap, &bmp->db_DBmap);
 
 	/*
 	 * allocate/initialize the bmap lock
@@ -407,16 +376,13 @@ int dbFree(struct inode *ip, s64 blkno, s64 nblocks)
 		 */
 		nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1)));
 
-		DBALLOCCK(bmp->db_DBmap, bmp->db_mapsize, blkno, nb);
-
 		/* free the blocks. */
 		if ((rc = dbFreeDmap(bmp, dp, blkno, nb))) {
+			jfs_error(ip->i_sb, "dbFree: error in block map\n");
 			release_metapage(mp);
 			IREAD_UNLOCK(ipbmap);
 			return (rc);
 		}
-
-		DBFREE(bmp->db_DBmap, bmp->db_mapsize, blkno, nb);
 	}
 
 	/* write the last buffer. */
@@ -775,10 +741,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 		IWRITE_LOCK(ipbmap);
 
 		rc = dbAllocAny(bmp, nblocks, l2nb, results);
-		if (rc == 0) {
-			DBALLOC(bmp->db_DBmap, bmp->db_mapsize, *results,
-				nblocks);
-		}
 
 		goto write_unlock;
 	}
@@ -836,8 +798,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 		    != -ENOSPC) {
 			if (rc == 0) {
 				*results = blkno;
-				DBALLOC(bmp->db_DBmap, bmp->db_mapsize,
-					*results, nblocks);
 				mark_metapage_dirty(mp);
 			}
 
@@ -863,11 +823,8 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 		if ((rc =
 		     dbAllocNear(bmp, dp, blkno, (int) nblocks, l2nb, results))
 		    != -ENOSPC) {
-			if (rc == 0) {
-				DBALLOC(bmp->db_DBmap, bmp->db_mapsize,
-					*results, nblocks);
+			if (rc == 0)
 				mark_metapage_dirty(mp);
-			}
 
 			release_metapage(mp);
 			goto read_unlock;
@@ -878,11 +835,8 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 		 */
 		if ((rc = dbAllocDmapLev(bmp, dp, (int) nblocks, l2nb, results))
 		    != -ENOSPC) {
-			if (rc == 0) {
-				DBALLOC(bmp->db_DBmap, bmp->db_mapsize,
-					*results, nblocks);
+			if (rc == 0)
 				mark_metapage_dirty(mp);
-			}
 
 			release_metapage(mp);
 			goto read_unlock;
@@ -896,13 +850,9 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 	 * the same allocation group as the hint.
 	 */
 	IWRITE_LOCK(ipbmap);
-	if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results))
-	    != -ENOSPC) {
-		if (rc == 0)
-			DBALLOC(bmp->db_DBmap, bmp->db_mapsize,
-				*results, nblocks);
+	if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) != -ENOSPC)
 		goto write_unlock;
-	}
+
 	IWRITE_UNLOCK(ipbmap);
 
 
@@ -918,9 +868,6 @@ int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
 	 */
 	if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) == -ENOSPC)
 		rc = dbAllocAny(bmp, nblocks, l2nb, results);
-	if (rc == 0) {
-		DBALLOC(bmp->db_DBmap, bmp->db_mapsize, *results, nblocks);
-	}
 
       write_unlock:
 	IWRITE_UNLOCK(ipbmap);
@@ -992,10 +939,9 @@ int dbAllocExact(struct inode *ip, s64 blkno, int nblocks)
 
 	IREAD_UNLOCK(ipbmap);
 
-	if (rc == 0) {
-		DBALLOC(bmp->db_DBmap, bmp->db_mapsize, blkno, nblocks);
+	if (rc == 0)
 		mark_metapage_dirty(mp);
-	}
+
 	release_metapage(mp);
 
 	return (rc);
@@ -1144,7 +1090,6 @@ static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks)
 		return -EIO;
 	}
 
-	DBALLOCCK(bmp->db_DBmap, bmp->db_mapsize, blkno, nblocks);
 	dp = (struct dmap *) mp->data;
 
 	/* try to allocate the blocks immediately following the
@@ -1155,11 +1100,9 @@ static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks)
 	IREAD_UNLOCK(ipbmap);
 
 	/* were we successful ? */
-	if (rc == 0) {
-		DBALLOC(bmp->db_DBmap, bmp->db_mapsize, extblkno,
-			addnblocks);
+	if (rc == 0)
 		write_metapage(mp);
-	} else
+	else
 		/* we were not successful */
 		release_metapage(mp);
 
@@ -2078,7 +2021,7 @@ static int dbFreeDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
 		      int nblocks)
 {
 	s8 oldroot;
-	int rc, word;
+	int rc = 0, word;
 
 	/* save the current value of the root (i.e. maximum free string)
 	 * of the dmap tree.
@@ -2086,11 +2029,11 @@ static int dbFreeDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
 	oldroot = dp->tree.stree[ROOT];
 
 	/* free the specified (blocks) bits */
-	dbFreeBits(bmp, dp, blkno, nblocks);
+	rc = dbFreeBits(bmp, dp, blkno, nblocks);
 
-	/* if the root has not changed, done. */
-	if (dp->tree.stree[ROOT] == oldroot)
-		return (0);
+	/* if error or the root has not changed, done. */
+	if (rc || (dp->tree.stree[ROOT] == oldroot))
+		return (rc);
 
 	/* root changed. bubble the change up to the dmap control pages.
 	 * if the adjustment of the upper level control pages fails,
@@ -2279,15 +2222,16 @@ static void dbAllocBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
  *      blkno	-  starting block number of the bits to be freed.
  *      nblocks	-  number of bits to be freed.
  *
- * RETURN VALUES: none
+ * RETURN VALUES: 0 for success
  *
  * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
  */
-static void dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
+static int dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
 		       int nblocks)
 {
 	int dbitno, word, rembits, nb, nwords, wbitno, nw, agno;
 	dmtree_t *tp = (dmtree_t *) & dp->tree;
+	int rc = 0;
 	int size;
 
 	/* determine the bit number and word within the dmap of the
@@ -2336,8 +2280,10 @@ static void dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
 
 			/* update the leaf for this dmap word.
 			 */
-			dbJoin(tp, word,
-			       dbMaxBud((u8 *) & dp->wmap[word]));
+			rc = dbJoin(tp, word,
+				    dbMaxBud((u8 *) & dp->wmap[word]));
+			if (rc)
+				return rc;
 
 			word += 1;
 		} else {
@@ -2368,7 +2314,9 @@ static void dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
 
 				/* update the leaf.
 				 */
-				dbJoin(tp, word, size);
+				rc = dbJoin(tp, word, size);
+				if (rc)
+					return rc;
 
 				/* get the number of dmap words handled.
 				 */
@@ -2415,6 +2363,8 @@ static void dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
 	}
 
 	BMAP_UNLOCK(bmp);
+
+	return 0;
 }
 
 
@@ -2522,7 +2472,9 @@ dbAdjCtl(struct bmap * bmp, s64 blkno, int newval, int alloc, int level)
 		}
 		dbSplit((dmtree_t *) dcp, leafno, dcp->budmin, newval);
 	} else {
-		dbJoin((dmtree_t *) dcp, leafno, newval);
+		rc = dbJoin((dmtree_t *) dcp, leafno, newval);
+		if (rc)
+			return rc;
 	}
 
 	/* check if the root of the current dmap control page changed due
@@ -2747,7 +2699,7 @@ static void dbBackSplit(dmtree_t * tp, int leafno)
  *
  * RETURN VALUES: none
  */
-static void dbJoin(dmtree_t * tp, int leafno, int newval)
+static int dbJoin(dmtree_t * tp, int leafno, int newval)
 {
 	int budsz, buddy;
 	s8 *leaf;
@@ -2787,7 +2739,9 @@ static void dbJoin(dmtree_t * tp, int leafno, int newval)
 			if (newval > leaf[buddy])
 				break;
 
-			assert(newval == leaf[buddy]);
+			/* It shouldn't be less */
+			if (newval < leaf[buddy])
+				return -EIO;
 
 			/* check which (leafno or buddy) is the left buddy.
 			 * the left buddy gets to claim the blocks resulting
@@ -2819,6 +2773,8 @@ static void dbJoin(dmtree_t * tp, int leafno, int newval)
 	/* update the leaf value.
 	 */
 	dbAdjTree(tp, leafno, newval);
+
+	return 0;
 }
 
 
@@ -3185,16 +3141,12 @@ int dbAllocBottomUp(struct inode *ip, s64 blkno, s64 nblocks)
 		 */
 		nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1)));
 
-		DBFREECK(bmp->db_DBmap, bmp->db_mapsize, blkno, nb);
-
 		/* allocate the blocks. */
 		if ((rc = dbAllocDmapBU(bmp, dp, blkno, nb))) {
 			release_metapage(mp);
 			IREAD_UNLOCK(ipbmap);
 			return (rc);
 		}
-
-		DBALLOC(bmp->db_DBmap, bmp->db_mapsize, blkno, nb);
 	}
 
 	/* write the last buffer. */
@@ -4041,223 +3993,3 @@ s64 dbMapFileSizeToMapSize(struct inode * ipbmap)
 
 	return (nblocks);
 }
-
-
-#ifdef	_JFS_DEBUG_DMAP
-/*
- *	DBinitmap()
- */
-static void DBinitmap(s64 size, struct inode *ipbmap, u32 ** results)
-{
-	int npages;
-	u32 *dbmap, *d;
-	int n;
-	s64 lblkno, cur_block;
-	struct dmap *dp;
-	struct metapage *mp;
-
-	npages = size / 32768;
-	npages += (size % 32768) ? 1 : 0;
-
-	dbmap = (u32 *) xmalloc(npages * 4096, L2PSIZE, kernel_heap);
-	if (dbmap == NULL)
-		BUG();	/* Not robust since this is only unused debug code */
-
-	for (n = 0, d = dbmap; n < npages; n++, d += 1024)
-		bzero(d, 4096);
-
-	/* Need to initialize from disk map pages
-	 */
-	for (d = dbmap, cur_block = 0; cur_block < size;
-	     cur_block += BPERDMAP, d += LPERDMAP) {
-		lblkno = BLKTODMAP(cur_block,
-				   JFS_SBI(ipbmap->i_sb)->bmap->
-				   db_l2nbperpage);
-		mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
-		if (mp == NULL) {
-			jfs_error(ipbmap->i_sb,
-				  "DBinitmap: could not read disk map page");
-			continue;
-		}
-		dp = (struct dmap *) mp->data;
-
-		for (n = 0; n < LPERDMAP; n++)
-			d[n] = le32_to_cpu(dp->wmap[n]);
-
-		release_metapage(mp);
-	}
-
-	*results = dbmap;
-}
-
-
-/*
- *	DBAlloc()
- */
-void DBAlloc(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks)
-{
-	int word, nb, bitno;
-	u32 mask;
-
-	assert(blkno > 0 && blkno < mapsize);
-	assert(nblocks > 0 && nblocks <= mapsize);
-
-	assert(blkno + nblocks <= mapsize);
-
-	dbmap += (blkno / 32);
-	while (nblocks > 0) {
-		bitno = blkno & (32 - 1);
-		nb = min(nblocks, 32 - bitno);
-
-		mask = (0xffffffff << (32 - nb) >> bitno);
-		assert((mask & *dbmap) == 0);
-		*dbmap |= mask;
-
-		dbmap++;
-		blkno += nb;
-		nblocks -= nb;
-	}
-}
-
-
-/*
- *	DBFree()
- */
-static void DBFree(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks)
-{
-	int word, nb, bitno;
-	u32 mask;
-
-	assert(blkno > 0 && blkno < mapsize);
-	assert(nblocks > 0 && nblocks <= mapsize);
-
-	assert(blkno + nblocks <= mapsize);
-
-	dbmap += (blkno / 32);
-	while (nblocks > 0) {
-		bitno = blkno & (32 - 1);
-		nb = min(nblocks, 32 - bitno);
-
-		mask = (0xffffffff << (32 - nb) >> bitno);
-		assert((mask & *dbmap) == mask);
-		*dbmap &= ~mask;
-
-		dbmap++;
-		blkno += nb;
-		nblocks -= nb;
-	}
-}
-
-
-/*
- *	DBAllocCK()
- */
-static void DBAllocCK(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks)
-{
-	int word, nb, bitno;
-	u32 mask;
-
-	assert(blkno > 0 && blkno < mapsize);
-	assert(nblocks > 0 && nblocks <= mapsize);
-
-	assert(blkno + nblocks <= mapsize);
-
-	dbmap += (blkno / 32);
-	while (nblocks > 0) {
-		bitno = blkno & (32 - 1);
-		nb = min(nblocks, 32 - bitno);
-
-		mask = (0xffffffff << (32 - nb) >> bitno);
-		assert((mask & *dbmap) == mask);
-
-		dbmap++;
-		blkno += nb;
-		nblocks -= nb;
-	}
-}
-
-
-/*
- *	DBFreeCK()
- */
-static void DBFreeCK(uint * dbmap, s64 mapsize, s64 blkno, s64 nblocks)
-{
-	int word, nb, bitno;
-	u32 mask;
-
-	assert(blkno > 0 && blkno < mapsize);
-	assert(nblocks > 0 && nblocks <= mapsize);
-
-	assert(blkno + nblocks <= mapsize);
-
-	dbmap += (blkno / 32);
-	while (nblocks > 0) {
-		bitno = blkno & (32 - 1);
-		nb = min(nblocks, 32 - bitno);
-
-		mask = (0xffffffff << (32 - nb) >> bitno);
-		assert((mask & *dbmap) == 0);
-
-		dbmap++;
-		blkno += nb;
-		nblocks -= nb;
-	}
-}
-
-
-/*
- *	dbPrtMap()
- */
-static void dbPrtMap(struct bmap * bmp)
-{
-	printk("   mapsize:   %d%d\n", bmp->db_mapsize);
-	printk("   nfree:     %d%d\n", bmp->db_nfree);
-	printk("   numag:     %d\n", bmp->db_numag);
-	printk("   agsize:    %d%d\n", bmp->db_agsize);
-	printk("   agl2size:  %d\n", bmp->db_agl2size);
-	printk("   agwidth:   %d\n", bmp->db_agwidth);
-	printk("   agstart:   %d\n", bmp->db_agstart);
-	printk("   agheigth:  %d\n", bmp->db_agheigth);
-	printk("   aglevel:   %d\n", bmp->db_aglevel);
-	printk("   maxlevel:  %d\n", bmp->db_maxlevel);
-	printk("   maxag:     %d\n", bmp->db_maxag);
-	printk("   agpref:    %d\n", bmp->db_agpref);
-	printk("   l2nbppg:   %d\n", bmp->db_l2nbperpage);
-}
-
-
-/*
- *	dbPrtCtl()
- */
-static void dbPrtCtl(struct dmapctl * dcp)
-{
-	int i, j, n;
-
-	printk("   height:    %08x\n", le32_to_cpu(dcp->height));
-	printk("   leafidx:   %08x\n", le32_to_cpu(dcp->leafidx));
-	printk("   budmin:    %08x\n", dcp->budmin);
-	printk("   nleafs:    %08x\n", le32_to_cpu(dcp->nleafs));
-	printk("   l2nleafs:  %08x\n", le32_to_cpu(dcp->l2nleafs));
-
-	printk("\n Tree:\n");
-	for (i = 0; i < CTLLEAFIND; i += 8) {
-		n = min(8, CTLLEAFIND - i);
-
-		for (j = 0; j < n; j++)
-			printf("  [%03x]: %02x", i + j,
-			       (char) dcp->stree[i + j]);
-		printf("\n");
-	}
-
-	printk("\n Tree Leaves:\n");
-	for (i = 0; i < LPERCTL; i += 8) {
-		n = min(8, LPERCTL - i);
-
-		for (j = 0; j < n; j++)
-			printf("  [%03x]: %02x",
-			       i + j,
-			       (char) dcp->stree[i + j + CTLLEAFIND]);
-		printf("\n");
-	}
-}
-#endif				/* _JFS_DEBUG_DMAP */
diff --git a/fs/jfs/jfs_dtree.c b/fs/jfs/jfs_dtree.c
index 8676aee..404f33e 100644
--- a/fs/jfs/jfs_dtree.c
+++ b/fs/jfs/jfs_dtree.c
@@ -381,9 +381,12 @@ static u32 add_index(tid_t tid, struct inode *ip, s64 bn, int slot)
 		 * It's time to move the inline table to an external
 		 * page and begin to build the xtree
 		 */
-		if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage) ||
-		    dbAlloc(ip, 0, sbi->nbperpage, &xaddr))
-			goto clean_up;	/* No space */
+		if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage))
+			goto clean_up;
+		if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) {
+			DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
+			goto clean_up;
+		}
 
 		/*
 		 * Save the table, we're going to overwrite it with the
@@ -397,13 +400,15 @@ static u32 add_index(tid_t tid, struct inode *ip, s64 bn, int slot)
 		xtInitRoot(tid, ip);
 
 		/*
-		 * Allocate the first block & add it to the xtree
+		 * Add the first block to the xtree
 		 */
 		if (xtInsert(tid, ip, 0, 0, sbi->nbperpage, &xaddr, 0)) {
 			/* This really shouldn't fail */
 			jfs_warn("add_index: xtInsert failed!");
 			memcpy(&jfs_ip->i_dirtable, temp_table,
 			       sizeof (temp_table));
+			dbFree(ip, xaddr, sbi->nbperpage);
+			DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
 			goto clean_up;
 		}
 		ip->i_size = PSIZE;
@@ -4554,202 +4559,3 @@ int dtModify(tid_t tid, struct inode *ip,
 
 	return 0;
 }
-
-#ifdef _JFS_DEBUG_DTREE
-/*
- *	dtDisplayTree()
- *
- * function: traverse forward
- */
-int dtDisplayTree(struct inode *ip)
-{
-	int rc;
-	struct metapage *mp;
-	dtpage_t *p;
-	s64 bn, pbn;
-	int index, lastindex, v, h;
-	pxd_t *xd;
-	struct btstack btstack;
-	struct btframe *btsp;
-	struct btframe *parent;
-	u8 *stbl;
-	int psize = 256;
-
-	printk("display B+-tree.\n");
-
-	/* clear stack */
-	btsp = btstack.stack;
-
-	/*
-	 * start with root
-	 *
-	 * root resides in the inode
-	 */
-	bn = 0;
-	v = h = 0;
-
-	/*
-	 * first access of each page:
-	 */
-      newPage:
-	DT_GETPAGE(ip, bn, mp, psize, p, rc);
-	if (rc)
-		return rc;
-
-	/* process entries forward from first index */
-	index = 0;
-	lastindex = p->header.nextindex - 1;
-
-	if (p->header.flag & BT_INTERNAL) {
-		/*
-		 * first access of each internal page
-		 */
-		printf("internal page ");
-		dtDisplayPage(ip, bn, p);
-
-		goto getChild;
-	} else {		/* (p->header.flag & BT_LEAF) */
-
-		/*
-		 * first access of each leaf page
-		 */
-		printf("leaf page ");
-		dtDisplayPage(ip, bn, p);
-
-		/*
-		 * process leaf page entries
-		 *
-		 for ( ; index <= lastindex; index++)
-		 {
-		 }
-		 */
-
-		/* unpin the leaf page */
-		DT_PUTPAGE(mp);
-	}
-
-	/*
-	 * go back up to the parent page
-	 */
-      getParent:
-	/* pop/restore parent entry for the current child page */
-	if ((parent = (btsp == btstack.stack ? NULL : --btsp)) == NULL)
-		/* current page must have been root */
-		return;
-
-	/*
-	 * parent page scan completed
-	 */
-	if ((index = parent->index) == (lastindex = parent->lastindex)) {
-		/* go back up to the parent page */
-		goto getParent;
-	}
-
-	/*
-	 * parent page has entries remaining
-	 */
-	/* get back the parent page */
-	bn = parent->bn;
-	/* v = parent->level; */
-	DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-	if (rc)
-		return rc;
-
-	/* get next parent entry */
-	index++;
-
-	/*
-	 * internal page: go down to child page of current entry
-	 */
-      getChild:
-	/* push/save current parent entry for the child page */
-	btsp->bn = pbn = bn;
-	btsp->index = index;
-	btsp->lastindex = lastindex;
-	/* btsp->level = v; */
-	/* btsp->node = h; */
-	++btsp;
-
-	/* get current entry for the child page */
-	stbl = DT_GETSTBL(p);
-	xd = (pxd_t *) & p->slot[stbl[index]];
-
-	/*
-	 * first access of each internal entry:
-	 */
-
-	/* get child page */
-	bn = addressPXD(xd);
-	psize = lengthPXD(xd) << ip->i_ipmnt->i_l2bsize;
-
-	printk("traverse down 0x%Lx[%d]->0x%Lx\n", pbn, index, bn);
-	v++;
-	h = index;
-
-	/* release parent page */
-	DT_PUTPAGE(mp);
-
-	/* process the child page */
-	goto newPage;
-}
-
-
-/*
- *	dtDisplayPage()
- *
- * function: display page
- */
-int dtDisplayPage(struct inode *ip, s64 bn, dtpage_t * p)
-{
-	int rc;
-	struct metapage *mp;
-	struct ldtentry *lh;
-	struct idtentry *ih;
-	pxd_t *xd;
-	int i, j;
-	u8 *stbl;
-	wchar_t name[JFS_NAME_MAX + 1];
-	struct component_name key = { 0, name };
-	int freepage = 0;
-
-	if (p == NULL) {
-		freepage = 1;
-		DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-		if (rc)
-			return rc;
-	}
-
-	/* display page control */
-	printk("bn:0x%Lx flag:0x%08x nextindex:%d\n",
-	       bn, p->header.flag, p->header.nextindex);
-
-	/* display entries */
-	stbl = DT_GETSTBL(p);
-	for (i = 0, j = 1; i < p->header.nextindex; i++, j++) {
-		dtGetKey(p, i, &key, JFS_SBI(ip->i_sb)->mntflag);
-		key.name[key.namlen] = '\0';
-		if (p->header.flag & BT_LEAF) {
-			lh = (struct ldtentry *) & p->slot[stbl[i]];
-			printf("\t[%d] %s:%d", i, key.name,
-			       le32_to_cpu(lh->inumber));
-		} else {
-			ih = (struct idtentry *) & p->slot[stbl[i]];
-			xd = (pxd_t *) ih;
-			bn = addressPXD(xd);
-			printf("\t[%d] %s:0x%Lx", i, key.name, bn);
-		}
-
-		if (j == 4) {
-			printf("\n");
-			j = 0;
-		}
-	}
-
-	printf("\n");
-
-	if (freepage)
-		DT_PUTPAGE(mp);
-
-	return 0;
-}
-#endif				/* _JFS_DEBUG_DTREE */
diff --git a/fs/jfs/jfs_dtree.h b/fs/jfs/jfs_dtree.h
index 273a801..13e4fdf 100644
--- a/fs/jfs/jfs_dtree.h
+++ b/fs/jfs/jfs_dtree.h
@@ -269,11 +269,4 @@ extern int dtModify(tid_t tid, struct inode *ip, struct component_name * key,
 		    ino_t * orig_ino, ino_t new_ino, int flag);
 
 extern int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir);
-
-#ifdef  _JFS_DEBUG_DTREE
-extern int dtDisplayTree(struct inode *ip);
-
-extern int dtDisplayPage(struct inode *ip, s64 bn, dtpage_t * p);
-#endif				/* _JFS_DEBUG_DTREE */
-
 #endif				/* !_H_JFS_DTREE */
diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
index 971af29..4021d46 100644
--- a/fs/jfs/jfs_imap.c
+++ b/fs/jfs/jfs_imap.c
@@ -87,25 +87,6 @@ static int copy_from_dinode(struct dinode *, struct inode *);
 static void copy_to_dinode(struct dinode *, struct inode *);
 
 /*
- *	debug code for double-checking inode map
- */
-/* #define	_JFS_DEBUG_IMAP	1 */
-
-#ifdef	_JFS_DEBUG_IMAP
-#define DBG_DIINIT(imap)	DBGdiInit(imap)
-#define DBG_DIALLOC(imap, ino)	DBGdiAlloc(imap, ino)
-#define DBG_DIFREE(imap, ino)	DBGdiFree(imap, ino)
-
-static void *DBGdiInit(struct inomap * imap);
-static void DBGdiAlloc(struct inomap * imap, ino_t ino);
-static void DBGdiFree(struct inomap * imap, ino_t ino);
-#else
-#define DBG_DIINIT(imap)
-#define DBG_DIALLOC(imap, ino)
-#define DBG_DIFREE(imap, ino)
-#endif				/* _JFS_DEBUG_IMAP */
-
-/*
  * NAME:        diMount()
  *
  * FUNCTION:    initialize the incore inode map control structures for
@@ -188,8 +169,6 @@ int diMount(struct inode *ipimap)
 	imap->im_ipimap = ipimap;
 	JFS_IP(ipimap)->i_imap = imap;
 
-//      DBG_DIINIT(imap);
-
 	return (0);
 }
 
@@ -1043,7 +1022,6 @@ int diFree(struct inode *ip)
 		/* update the bitmap.
 		 */
 		iagp->wmap[extno] = cpu_to_le32(bitmap);
-		DBG_DIFREE(imap, inum);
 
 		/* update the free inode counts at the iag, ag and
 		 * map level.
@@ -1231,7 +1209,6 @@ int diFree(struct inode *ip)
 		jfs_error(ip->i_sb, "diFree: the pmap does not show inode free");
 	}
 	iagp->wmap[extno] = 0;
-	DBG_DIFREE(imap, inum);
 	PXDlength(&iagp->inoext[extno], 0);
 	PXDaddress(&iagp->inoext[extno], 0);
 
@@ -1350,7 +1327,6 @@ diInitInode(struct inode *ip, int iagno, int ino, int extno, struct iag * iagp)
 	struct jfs_inode_info *jfs_ip = JFS_IP(ip);
 
 	ip->i_ino = (iagno << L2INOSPERIAG) + ino;
-	DBG_DIALLOC(JFS_IP(ipimap)->i_imap, ip->i_ino);
 	jfs_ip->ixpxd = iagp->inoext[extno];
 	jfs_ip->agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
 	jfs_ip->active_ag = -1;
@@ -3185,84 +3161,3 @@ static void copy_to_dinode(struct dinode * dip, struct inode *ip)
 	if (S_ISCHR(ip->i_mode) || S_ISBLK(ip->i_mode))
 		dip->di_rdev = cpu_to_le32(jfs_ip->dev);
 }
-
-#ifdef	_JFS_DEBUG_IMAP
-/*
- *	DBGdiInit()
- */
-static void *DBGdiInit(struct inomap * imap)
-{
-	u32 *dimap;
-	int size;
-	size = 64 * 1024;
-	if ((dimap = (u32 *) xmalloc(size, L2PSIZE, kernel_heap)) == NULL)
-		assert(0);
-	bzero((void *) dimap, size);
-	imap->im_DBGdimap = dimap;
-}
-
-/*
- *	DBGdiAlloc()
- */
-static void DBGdiAlloc(struct inomap * imap, ino_t ino)
-{
-	u32 *dimap = imap->im_DBGdimap;
-	int w, b;
-	u32 m;
-	w = ino >> 5;
-	b = ino & 31;
-	m = 0x80000000 >> b;
-	assert(w < 64 * 256);
-	if (dimap[w] & m) {
-		printk("DEBUG diAlloc: duplicate alloc ino:0x%x\n", ino);
-	}
-	dimap[w] |= m;
-}
-
-/*
- *	DBGdiFree()
- */
-static void DBGdiFree(struct inomap * imap, ino_t ino)
-{
-	u32 *dimap = imap->im_DBGdimap;
-	int w, b;
-	u32 m;
-	w = ino >> 5;
-	b = ino & 31;
-	m = 0x80000000 >> b;
-	assert(w < 64 * 256);
-	if ((dimap[w] & m) == 0) {
-		printk("DEBUG diFree: duplicate free ino:0x%x\n", ino);
-	}
-	dimap[w] &= ~m;
-}
-
-static void dump_cp(struct inomap * ipimap, char *function, int line)
-{
-	printk("\n* ********* *\nControl Page %s %d\n", function, line);
-	printk("FreeIAG %d\tNextIAG %d\n", ipimap->im_freeiag,
-	       ipimap->im_nextiag);
-	printk("NumInos %d\tNumFree %d\n",
-	       atomic_read(&ipimap->im_numinos),
-	       atomic_read(&ipimap->im_numfree));
-	printk("AG InoFree %d\tAG ExtFree %d\n",
-	       ipimap->im_agctl[0].inofree, ipimap->im_agctl[0].extfree);
-	printk("AG NumInos %d\tAG NumFree %d\n",
-	       ipimap->im_agctl[0].numinos, ipimap->im_agctl[0].numfree);
-}
-
-static void dump_iag(struct iag * iag, char *function, int line)
-{
-	printk("\n* ********* *\nIAG %s %d\n", function, line);
-	printk("IagNum %d\tIAG Free %d\n", le32_to_cpu(iag->iagnum),
-	       le32_to_cpu(iag->iagfree));
-	printk("InoFreeFwd %d\tInoFreeBack %d\n",
-	       le32_to_cpu(iag->inofreefwd),
-	       le32_to_cpu(iag->inofreeback));
-	printk("ExtFreeFwd %d\tExtFreeBack %d\n",
-	       le32_to_cpu(iag->extfreefwd),
-	       le32_to_cpu(iag->extfreeback));
-	printk("NFreeInos %d\tNFreeExts %d\n", le32_to_cpu(iag->nfreeinos),
-	       le32_to_cpu(iag->nfreeexts));
-}
-#endif				/* _JFS_DEBUG_IMAP */
diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
index 79d0762..22815e8 100644
--- a/fs/jfs/jfs_logmgr.c
+++ b/fs/jfs/jfs_logmgr.c
@@ -1030,7 +1030,8 @@ static int lmLogSync(struct jfs_log * log, int nosyncwait)
 	 * starting until all current transactions are completed
 	 * by setting syncbarrier flag.
 	 */
-	if (written > LOGSYNC_BARRIER(logsize) && logsize > 32 * LOGPSIZE) {
+	if (!test_bit(log_SYNCBARRIER, &log->flag) &&
+	    (written > LOGSYNC_BARRIER(logsize)) && log->active) {
 		set_bit(log_SYNCBARRIER, &log->flag);
 		jfs_info("log barrier on: lsn=0x%x syncpt=0x%x", lsn,
 			 log->syncpt);
diff --git a/fs/jfs/jfs_metapage.c b/fs/jfs/jfs_metapage.c
index 6c5485d..13d7e3f 100644
--- a/fs/jfs/jfs_metapage.c
+++ b/fs/jfs/jfs_metapage.c
@@ -561,7 +561,6 @@ static int metapage_releasepage(struct page *page, int gfp_mask)
 			dump_mem("page", page, sizeof(struct page));
 			dump_stack();
 		}
-		WARN_ON(mp->lsn);
 		if (mp->lsn)
 			remove_from_logsync(mp);
 		remove_metapage(page, mp);
@@ -641,7 +640,7 @@ struct metapage *__get_metapage(struct inode *inode, unsigned long lblock,
 	} else {
 		page = read_cache_page(mapping, page_index,
 			    (filler_t *)mapping->a_ops->readpage, NULL);
-		if (IS_ERR(page)) {
+		if (IS_ERR(page) || !PageUptodate(page)) {
 			jfs_err("read_cache_page failed!");
 			return NULL;
 		}
@@ -783,14 +782,6 @@ void release_metapage(struct metapage * mp)
 	if (test_bit(META_discard, &mp->flag) && !mp->count) {
 		clear_page_dirty(page);
 		ClearPageUptodate(page);
-#ifdef _NOT_YET
-		if (page->mapping) {
-		/* Remove from page cache and page cache reference */
-			remove_from_page_cache(page);
-			page_cache_release(page);
-			metapage_releasepage(page, 0);
-		}
-#endif
 	}
 #else
 	/* Try to keep metapages from using up too much memory */
diff --git a/fs/jfs/jfs_unicode.c b/fs/jfs/jfs_unicode.c
index b32208aa..f327dec 100644
--- a/fs/jfs/jfs_unicode.c
+++ b/fs/jfs/jfs_unicode.c
@@ -51,8 +51,9 @@ int jfs_strfromUCS_le(char *to, const __le16 * from,
 		}
 	} else {
 		for (i = 0; (i < len) && from[i]; i++) {
-			if (le16_to_cpu(from[i]) & 0xff00) {
-				if (warn) {
+			if (unlikely(le16_to_cpu(from[i]) & 0xff00)) {
+				to[i] = '?';
+				if (unlikely(warn)) {
 					warn--;
 					warn_again--;
 					printk(KERN_ERR
@@ -61,7 +62,7 @@ int jfs_strfromUCS_le(char *to, const __le16 * from,
 					printk(KERN_ERR
 				"mount with iocharset=utf8 to access\n");
 				}
-				to[i] = '?';
+
 			}
 			else
 				to[i] = (char) (le16_to_cpu(from[i]));
diff --git a/fs/jfs/jfs_xtree.c b/fs/jfs/jfs_xtree.c
index 31b34db..a7fe2f2 100644
--- a/fs/jfs/jfs_xtree.c
+++ b/fs/jfs/jfs_xtree.c
@@ -135,14 +135,6 @@ static int xtSearchNode(struct inode *ip,
 static int xtRelink(tid_t tid, struct inode *ip, xtpage_t * fp);
 #endif				/*  _STILL_TO_PORT */
 
-/* External references */
-
-/*
- *      debug control
- */
-/*      #define _JFS_DEBUG_XTREE        1 */
-
-
 /*
  *      xtLookup()
  *
@@ -4140,338 +4132,6 @@ s64 xtTruncate_pmap(tid_t tid, struct inode *ip, s64 committed_size)
 	return 0;
 }
 
-
-#ifdef _JFS_DEBUG_XTREE
-/*
- *      xtDisplayTree()
- *
- * function: traverse forward
- */
-int xtDisplayTree(struct inode *ip)
-{
-	int rc = 0;
-	struct metapage *mp;
-	xtpage_t *p;
-	s64 bn, pbn;
-	int index, lastindex, v, h;
-	xad_t *xad;
-	struct btstack btstack;
-	struct btframe *btsp;
-	struct btframe *parent;
-
-	printk("display B+-tree.\n");
-
-	/* clear stack */
-	btsp = btstack.stack;
-
-	/*
-	 * start with root
-	 *
-	 * root resides in the inode
-	 */
-	bn = 0;
-	v = h = 0;
-
-	/*
-	 * first access of each page:
-	 */
-      getPage:
-	XT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-	if (rc)
-		return rc;
-
-	/* process entries forward from first index */
-	index = XTENTRYSTART;
-	lastindex = le16_to_cpu(p->header.nextindex) - 1;
-
-	if (p->header.flag & BT_INTERNAL) {
-		/*
-		 * first access of each internal page
-		 */
-		goto getChild;
-	} else {		/* (p->header.flag & BT_LEAF) */
-
-		/*
-		 * first access of each leaf page
-		 */
-		printf("leaf page ");
-		xtDisplayPage(ip, bn, p);
-
-		/* unpin the leaf page */
-		XT_PUTPAGE(mp);
-	}
-
-	/*
-	 * go back up to the parent page
-	 */
-      getParent:
-	/* pop/restore parent entry for the current child page */
-	if ((parent = (btsp == btstack.stack ? NULL : --btsp)) == NULL)
-		/* current page must have been root */
-		return;
-
-	/*
-	 * parent page scan completed
-	 */
-	if ((index = parent->index) == (lastindex = parent->lastindex)) {
-		/* go back up to the parent page */
-		goto getParent;
-	}
-
-	/*
-	 * parent page has entries remaining
-	 */
-	/* get back the parent page */
-	bn = parent->bn;
-	/* v = parent->level; */
-	XT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-	if (rc)
-		return rc;
-
-	/* get next parent entry */
-	index++;
-
-	/*
-	 * internal page: go down to child page of current entry
-	 */
-      getChild:
-	/* push/save current parent entry for the child page */
-	btsp->bn = pbn = bn;
-	btsp->index = index;
-	btsp->lastindex = lastindex;
-	/* btsp->level = v; */
-	/* btsp->node = h; */
-	++btsp;
-
-	/* get child page */
-	xad = &p->xad[index];
-	bn = addressXAD(xad);
-
-	/*
-	 * first access of each internal entry:
-	 */
-	/* release parent page */
-	XT_PUTPAGE(mp);
-
-	printk("traverse down 0x%lx[%d]->0x%lx\n", (ulong) pbn, index,
-	       (ulong) bn);
-	v++;
-	h = index;
-
-	/* process the child page */
-	goto getPage;
-}
-
-
-/*
- *      xtDisplayPage()
- *
- * function: display page
- */
-int xtDisplayPage(struct inode *ip, s64 bn, xtpage_t * p)
-{
-	int rc = 0;
-	xad_t *xad;
-	s64 xaddr, xoff;
-	int xlen, i, j;
-
-	/* display page control */
-	printf("bn:0x%lx flag:0x%x nextindex:%d\n",
-	       (ulong) bn, p->header.flag,
-	       le16_to_cpu(p->header.nextindex));
-
-	/* display entries */
-	xad = &p->xad[XTENTRYSTART];
-		for (i = XTENTRYSTART, j = 1; i < le16_to_cpu(p->header.nextindex);
-		     i++, xad++, j++) {
-			xoff = offsetXAD(xad);
-			xaddr = addressXAD(xad);
-			xlen = lengthXAD(xad);
-			printf("\t[%d] 0x%lx:0x%lx(0x%x)", i, (ulong) xoff,
-			       (ulong) xaddr, xlen);
-
-			if (j == 4) {
-				printf("\n");
-				j = 0;
-		}
-	}
-
-	printf("\n");
-}
-#endif				/* _JFS_DEBUG_XTREE */
-
-
-#ifdef _JFS_WIP
-/*
- *      xtGather()
- *
- * function:
- *      traverse for allocation acquiring tlock at commit time
- *      (vs at the time of update) logging backward top down
- *
- * note:
- *      problem - establishing that all new allocation have been
- *      processed both for append and random write in sparse file
- *      at the current entry at the current subtree root page
- *
- */
-int xtGather(btree_t *t)
-{
-	int rc = 0;
-	xtpage_t *p;
-	u64 bn;
-	int index;
-	btentry_t *e;
-	struct btstack btstack;
-	struct btsf *parent;
-
-	/* clear stack */
-	BT_CLR(&btstack);
-
-	/*
-	 * start with root
-	 *
-	 * root resides in the inode
-	 */
-	bn = 0;
-	XT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-	if (rc)
-		return rc;
-
-	/* new root is NOT pointed by a new entry
-	   if (p->header.flag & NEW)
-	   allocate new page lock;
-	   write a NEWPAGE log;
-	 */
-
-      dopage:
-	/*
-	 * first access of each page:
-	 */
-	/* process entries backward from last index */
-	index = le16_to_cpu(p->header.nextindex) - 1;
-
-	if (p->header.flag & BT_LEAF) {
-		/*
-		 * first access of each leaf page
-		 */
-		/* process leaf page entries backward */
-		for (; index >= XTENTRYSTART; index--) {
-			e = &p->xad[index];
-			/*
-			 * if newpage, log NEWPAGE.
-			 *
-			 if (e->flag & XAD_NEW) {
-			 nfound =+ entry->length;
-			 update current page lock for the entry;
-			 newpage(entry);
-			 *
-			 * if moved, log move.
-			 *
-			 } else if (e->flag & XAD_MOVED) {
-			 reset flag;
-			 update current page lock for the entry;
-			 }
-			 */
-		}
-
-		/* unpin the leaf page */
-		XT_PUTPAGE(mp);
-
-		/*
-		 * go back up to the parent page
-		 */
-	      getParent:
-		/* restore parent entry for the current child page */
-		if ((parent = BT_POP(&btstack)) == NULL)
-			/* current page must have been root */
-			return 0;
-
-		if ((index = parent->index) == XTENTRYSTART) {
-			/*
-			 * parent page scan completed
-			 */
-			/* go back up to the parent page */
-			goto getParent;
-		} else {
-			/*
-			 * parent page has entries remaining
-			 */
-			/* get back the parent page */
-			bn = parent->bn;
-			XT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-			if (rc)
-				return -EIO;
-
-			/* first subroot page which
-			 * covers all new allocated blocks
-			 * itself not new/modified.
-			 * (if modified from split of descendent,
-			 * go down path of split page)
-
-			 if (nfound == nnew &&
-			 !(p->header.flag & (NEW | MOD)))
-			 exit scan;
-			 */
-
-			/* process parent page entries backward */
-			index--;
-		}
-	} else {
-		/*
-		 * first access of each internal page
-		 */
-	}
-
-	/*
-	 * internal page: go down to child page of current entry
-	 */
-
-	/* save current parent entry for the child page */
-	BT_PUSH(&btstack, bn, index);
-
-	/* get current entry for the child page */
-	e = &p->xad[index];
-
-	/*
-	 * first access of each internal entry:
-	 */
-	/*
-	 * if new entry, log btree_tnewentry.
-	 *
-	 if (e->flag & XAD_NEW)
-	 update parent page lock for the entry;
-	 */
-
-	/* release parent page */
-	XT_PUTPAGE(mp);
-
-	/* get child page */
-	bn = e->bn;
-	XT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
-	if (rc)
-		return rc;
-
-	/*
-	 * first access of each non-root page:
-	 */
-	/*
-	 * if new, log btree_newpage.
-	 *
-	 if (p->header.flag & NEW)
-	 allocate new page lock;
-	 write a NEWPAGE log (next, prev);
-	 */
-
-	/* process the child page */
-	goto dopage;
-
-      out:
-	return 0;
-}
-#endif				/* _JFS_WIP */
-
-
 #ifdef CONFIG_JFS_STATISTICS
 int jfs_xtstat_read(char *buffer, char **start, off_t offset, int length,
 		    int *eof, void *data)
diff --git a/fs/jfs/jfs_xtree.h b/fs/jfs/jfs_xtree.h
index a697842..af668a8 100644
--- a/fs/jfs/jfs_xtree.h
+++ b/fs/jfs/jfs_xtree.h
@@ -131,10 +131,4 @@ extern int xtRelocate(tid_t tid, struct inode *ip,
 extern int xtAppend(tid_t tid,
 		    struct inode *ip, int xflag, s64 xoff, int maxblocks,
 		    int *xlenp, s64 * xaddrp, int flag);
-
-#ifdef  _JFS_DEBUG_XTREE
-extern int xtDisplayTree(struct inode *ip);
-extern int xtDisplayPage(struct inode *ip, s64 bn, xtpage_t * p);
-#endif				/* _JFS_DEBUG_XTREE */
-
 #endif				/* !_H_JFS_XTREE */
diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c
index ee438d4..554ec73 100644
--- a/fs/jfs/xattr.c
+++ b/fs/jfs/xattr.c
@@ -781,7 +781,7 @@ static int can_set_xattr(struct inode *inode, const char *name,
 	if (IS_RDONLY(inode))
 		return -EROFS;
 
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode) || S_ISLNK(inode->i_mode))
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
 
 	if(strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) == 0)
@@ -790,12 +790,12 @@ static int can_set_xattr(struct inode *inode, const char *name,
 		 */
 		return can_set_system_xattr(inode, name, value, value_len);
 
-	if(strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) != 0)
+	if(strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) == 0)
 		return (capable(CAP_SYS_ADMIN) ? 0 : -EPERM);
 
 #ifdef CONFIG_JFS_SECURITY
 	if (strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN)
-	    != 0)
+	    == 0)
 		return 0;	/* Leave it to the security module */
 #endif
 		
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 6e24255..12a857c 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -331,7 +331,7 @@ static ctl_table nlm_sysctls[] = {
 		.ctl_name	= CTL_UNNUMBERED,
 		.procname	= "nlm_grace_period",
 		.data		= &nlm_grace_period,
-		.maxlen		= sizeof(int),
+		.maxlen		= sizeof(unsigned long),
 		.mode		= 0644,
 		.proc_handler	= &proc_doulongvec_minmax,
 		.extra1		= (unsigned long *) &nlm_grace_period_min,
@@ -341,7 +341,7 @@ static ctl_table nlm_sysctls[] = {
 		.ctl_name	= CTL_UNNUMBERED,
 		.procname	= "nlm_timeout",
 		.data		= &nlm_timeout,
-		.maxlen		= sizeof(int),
+		.maxlen		= sizeof(unsigned long),
 		.mode		= 0644,
 		.proc_handler	= &proc_doulongvec_minmax,
 		.extra1		= (unsigned long *) &nlm_timeout_min,
diff --git a/fs/locks.c b/fs/locks.c
index 29fa5da..11956b6 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -1591,7 +1591,8 @@ out:
 /* Apply the lock described by l to an open file descriptor.
  * This implements both the F_SETLK and F_SETLKW commands of fcntl().
  */
-int fcntl_setlk(struct file *filp, unsigned int cmd, struct flock __user *l)
+int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
+		struct flock __user *l)
 {
 	struct file_lock *file_lock = locks_alloc_lock();
 	struct flock flock;
@@ -1620,6 +1621,7 @@ int fcntl_setlk(struct file *filp, unsigned int cmd, struct flock __user *l)
 		goto out;
 	}
 
+again:
 	error = flock_to_posix_lock(filp, file_lock, &flock);
 	if (error)
 		goto out;
@@ -1648,25 +1650,33 @@ int fcntl_setlk(struct file *filp, unsigned int cmd, struct flock __user *l)
 	if (error)
 		goto out;
 
-	if (filp->f_op && filp->f_op->lock != NULL) {
+	if (filp->f_op && filp->f_op->lock != NULL)
 		error = filp->f_op->lock(filp, cmd, file_lock);
-		goto out;
-	}
+	else {
+		for (;;) {
+			error = __posix_lock_file(inode, file_lock);
+			if ((error != -EAGAIN) || (cmd == F_SETLK))
+				break;
+			error = wait_event_interruptible(file_lock->fl_wait,
+					!file_lock->fl_next);
+			if (!error)
+				continue;
 
-	for (;;) {
-		error = __posix_lock_file(inode, file_lock);
-		if ((error != -EAGAIN) || (cmd == F_SETLK))
+			locks_delete_block(file_lock);
 			break;
-		error = wait_event_interruptible(file_lock->fl_wait,
-				!file_lock->fl_next);
-		if (!error)
-			continue;
+		}
+	}
 
-		locks_delete_block(file_lock);
-		break;
+	/*
+	 * Attempt to detect a close/fcntl race and recover by
+	 * releasing the lock that was just acquired.
+	 */
+	if (!error && fcheck(fd) != filp && flock.l_type != F_UNLCK) {
+		flock.l_type = F_UNLCK;
+		goto again;
 	}
 
- out:
+out:
 	locks_free_lock(file_lock);
 	return error;
 }
@@ -1724,7 +1734,8 @@ out:
 /* Apply the lock described by l to an open file descriptor.
  * This implements both the F_SETLK and F_SETLKW commands of fcntl().
  */
-int fcntl_setlk64(struct file *filp, unsigned int cmd, struct flock64 __user *l)
+int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
+		struct flock64 __user *l)
 {
 	struct file_lock *file_lock = locks_alloc_lock();
 	struct flock64 flock;
@@ -1753,6 +1764,7 @@ int fcntl_setlk64(struct file *filp, unsigned int cmd, struct flock64 __user *l)
 		goto out;
 	}
 
+again:
 	error = flock64_to_posix_lock(filp, file_lock, &flock);
 	if (error)
 		goto out;
@@ -1781,22 +1793,30 @@ int fcntl_setlk64(struct file *filp, unsigned int cmd, struct flock64 __user *l)
 	if (error)
 		goto out;
 
-	if (filp->f_op && filp->f_op->lock != NULL) {
+	if (filp->f_op && filp->f_op->lock != NULL)
 		error = filp->f_op->lock(filp, cmd, file_lock);
-		goto out;
-	}
+	else {
+		for (;;) {
+			error = __posix_lock_file(inode, file_lock);
+			if ((error != -EAGAIN) || (cmd == F_SETLK64))
+				break;
+			error = wait_event_interruptible(file_lock->fl_wait,
+					!file_lock->fl_next);
+			if (!error)
+				continue;
 
-	for (;;) {
-		error = __posix_lock_file(inode, file_lock);
-		if ((error != -EAGAIN) || (cmd == F_SETLK64))
+			locks_delete_block(file_lock);
 			break;
-		error = wait_event_interruptible(file_lock->fl_wait,
-				!file_lock->fl_next);
-		if (!error)
-			continue;
+		}
+	}
 
-		locks_delete_block(file_lock);
-		break;
+	/*
+	 * Attempt to detect a close/fcntl race and recover by
+	 * releasing the lock that was just acquired.
+	 */
+	if (!error && fcheck(fd) != filp && flock.l_type != F_UNLCK) {
+		flock.l_type = F_UNLCK;
+		goto again;
 	}
 
 out:
@@ -1888,12 +1908,7 @@ void locks_remove_flock(struct file *filp)
 
 	while ((fl = *before) != NULL) {
 		if (fl->fl_file == filp) {
-			/*
-			 * We might have a POSIX lock that was created at the same time
-			 * the filp was closed for the last time. Just remove that too,
-			 * regardless of ownership, since nobody can own it.
-			 */
-			if (IS_FLOCK(fl) || IS_POSIX(fl)) {
+			if (IS_FLOCK(fl)) {
 				locks_delete_lock(before);
 				continue;
 			}
diff --git a/fs/mbcache.c b/fs/mbcache.c
index c7170b9..b002a08 100644
--- a/fs/mbcache.c
+++ b/fs/mbcache.c
@@ -316,11 +316,10 @@ fail:
  * currently in use cannot be freed, and thus remain in the cache. All others
  * are freed.
  *
- * @cache: which cache to shrink
  * @bdev: which device's cache entries to shrink
  */
 void
-mb_cache_shrink(struct mb_cache *cache, struct block_device *bdev)
+mb_cache_shrink(struct block_device *bdev)
 {
 	LIST_HEAD(free_list);
 	struct list_head *l, *ltmp;
diff --git a/fs/ntfs/ChangeLog b/fs/ntfs/ChangeLog
index 1d2ad15..9709fac 100644
--- a/fs/ntfs/ChangeLog
+++ b/fs/ntfs/ChangeLog
@@ -1,21 +1,18 @@
 ToDo/Notes:
 	- Find and fix bugs.
-	- Checkpoint or disable the user space journal ($UsnJrnl).
 	- In between ntfs_prepare/commit_write, need exclusion between
-	  simultaneous file extensions. Need perhaps an NInoResizeUnderway()
-	  flag which we can set in ntfs_prepare_write() and clear again in
-	  ntfs_commit_write(). Just have to be careful in readpage/writepage,
-	  as well as in truncate, that we play nice... We might need to have
-	  a data_size field in the ntfs_inode to store the real attribute
-	  length. Also need to be careful with initialized_size extention in
+	  simultaneous file extensions.  This is given to us by holding i_sem
+	  on the inode.  The only places in the kernel when a file is resized
+	  are prepare/commit write and truncate for both of which i_sem is
+	  held.  Just have to be careful in readpage/writepage and all other
+	  helpers not running under i_sem that we play nice...
+	  Also need to be careful with initialized_size extention in
 	  ntfs_prepare_write. Basically, just be _very_ careful in this code...
-	  OTOH, perhaps i_sem, which is held accross generic_file_write is
-	  sufficient for synchronisation here. We then just need to make sure
-	  ntfs_readpage/writepage/truncate interoperate properly with us.
-	  UPDATE: The above is all ok as it is due to i_sem held.  The only
-	  thing that needs to be checked is ntfs_writepage() which does not
-	  hold i_sem.  It cannot change i_size but it needs to cope with a
-	  concurrent i_size change.
+	  UPDATE: The only things that need to be checked are read/writepage
+	  which do not hold i_sem.  Note writepage cannot change i_size but it
+	  needs to cope with a concurrent i_size change, just like readpage.
+	  Also both need to cope with concurrent changes to the other sizes,
+	  i.e. initialized/allocated/compressed size, as well.
 	- Implement mft.c::sync_mft_mirror_umount().  We currently will just
 	  leave the volume dirty on umount if the final iput(vol->mft_ino)
 	  causes a write of any mirrored mft records due to the mft mirror
@@ -25,12 +22,158 @@ ToDo/Notes:
 	- Enable the code for setting the NT4 compatibility flag when we start
 	  making NTFS 1.2 specific modifications.
 
-2.1.23-WIP
+2.1.23 - Implement extension of resident files and make writing safe as well as
+	 many bug fixes, cleanups, and enhancements...
 
 	- Add printk rate limiting for ntfs_warning() and ntfs_error() when
 	  compiled without debug.  This avoids a possible denial of service
 	  attack.  Thanks to Carl-Daniel Hailfinger from SuSE for pointing this
 	  out.
+	- Fix compilation warnings on ia64.  (Randy Dunlap)
+	- Use i_size_{read,write}() instead of reading i_size by hand and cache
+	  the value where apropriate.
+	- Add size_lock to the ntfs_inode structure.  This is an rw spinlock
+	  and it locks against access to the inode sizes.  Note, ->size_lock
+	  is also accessed from irq context so you must use the _irqsave and
+	  _irqrestore lock and unlock functions, respectively.  Protect all
+	  accesses to allocated_size, initialized_size, and compressed_size.
+	- Minor optimization to fs/ntfs/super.c::ntfs_statfs() and its helpers.
+	- Implement extension of resident files in the regular file write code
+	  paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()).  At present
+	  this only works until the data attribute becomes too big for the mft
+	  record after which we abort the write returning -EOPNOTSUPP from
+	  ntfs_prepare_write().
+	- Add disable_sparse mount option together with a per volume sparse
+	  enable bit which is set appropriately and a per inode sparse disable
+	  bit which is preset on some system file inodes as appropriate.
+	- Enforce that sparse support is disabled on NTFS volumes pre 3.0.
+	- Fix a bug in fs/ntfs/runlist.c::ntfs_mapping_pairs_decompress() in
+	  the creation of the unmapped runlist element for the base attribute
+	  extent.
+	- Split ntfs_map_runlist() into ntfs_map_runlist() and a non-locking
+	  helper ntfs_map_runlist_nolock() which is used by ntfs_map_runlist().
+	  This allows us to map runlist fragments with the runlist lock already
+	  held without having to drop and reacquire it around the call.  Adapt
+	  all callers.
+	- Change ntfs_find_vcn() to ntfs_find_vcn_nolock() which takes a locked
+	  runlist.  This allows us to find runlist elements with the runlist
+	  lock already held without having to drop and reacquire it around the
+	  call.  Adapt all callers.
+	- Change time to u64 in time.h::ntfs2utc() as it otherwise generates a
+	  warning in the do_div() call on sparc32.  Thanks to Meelis Roos for
+	  the report and analysis of the warning.
+	- Fix a nasty runlist merge bug when merging two holes.
+	- Set the ntfs_inode->allocated_size to the real allocated size in the
+	  mft record for resident attributes (fs/ntfs/inode.c).
+	- Small readability cleanup to use "a" instead of "ctx->attr"
+	  everywhere (fs/ntfs/inode.c).
+	- Make fs/ntfs/namei.c::ntfs_get_{parent,dentry} static and move the
+	  definition of ntfs_export_ops from fs/ntfs/super.c to namei.c.  Also,
+	  declare ntfs_export_ops in fs/ntfs/ntfs.h.
+	- Correct sparse file handling.  The compressed values need to be
+	  checked and set in the ntfs inode as done for compressed files and
+	  the compressed size needs to be used for vfs inode->i_blocks instead
+	  of the allocated size, again, as done for compressed files.
+	- Add AT_EA in addition to AT_DATA to whitelist for being allowed to be
+	  non-resident in fs/ntfs/attrib.c::ntfs_attr_can_be_non_resident().
+	- Add fs/ntfs/attrib.c::ntfs_attr_vcn_to_lcn_nolock() used by the new
+	  write code.
+	- Fix bug in fs/ntfs/attrib.c::ntfs_find_vcn_nolock() where after
+	  dropping the read lock and taking the write lock we were not checking
+	  whether someone else did not already do the work we wanted to do.
+	- Rename fs/ntfs/attrib.c::ntfs_find_vcn_nolock() to
+	  ntfs_attr_find_vcn_nolock() and update all callers.
+	- Add fs/ntfs/attrib.[hc]::ntfs_attr_make_non_resident().
+	- Fix sign of various error return values to be negative in
+	  fs/ntfs/lcnalloc.c.
+	- Modify ->readpage and ->writepage (fs/ntfs/aops.c) so they detect and
+	  handle the case where an attribute is converted from resident to
+	  non-resident by a concurrent file write.
+	- Remove checks for NULL before calling kfree() since kfree() does the
+	  checking itself.  (Jesper Juhl)
+	- Some utilities modify the boot sector but do not update the checksum.
+	  Thus, relax the checking in fs/ntfs/super.c::is_boot_sector_ntfs() to
+	  only emit a warning when the checksum is incorrect rather than
+	  refusing the mount.  Thanks to Bernd Casimir for pointing this
+	  problem out.
+	- Update attribute definition handling.
+	- Add NTFS_MAX_CLUSTER_SIZE and NTFS_MAX_PAGES_PER_CLUSTER constants.
+	- Use NTFS_MAX_CLUSTER_SIZE in super.c instead of hard coding 0x10000.
+	- Use MAX_BUF_PER_PAGE instead of variable sized array allocation for
+	  better code generation and one less sparse warning in fs/ntfs/aops.c.
+	- Remove spurious void pointer casts from fs/ntfs/.  (Pekka Enberg)
+	- Use C99 style structure initialization after memory allocation where
+	  possible (fs/ntfs/{attrib.c,index.c,super.c}).  Thanks to Al Viro and
+	  Pekka Enberg.
+	- Stamp the transaction log ($UsnJrnl), aka user space journal, if it
+	  is active on the volume and we are mounting read-write or remounting
+	  from read-only to read-write.
+	- Fix a bug in address space operations error recovery code paths where
+	  if the runlist was not mapped at all and a mapping error occured we
+	  would leave the runlist locked on exit to the function so that the
+	  next access to the same file would try to take the lock and deadlock.
+	- Detect the case when Windows has been suspended to disk on the volume
+	  to be mounted and if this is the case do not allow (re)mounting
+	  read-write.  This is done by parsing hiberfil.sys if present.
+	- Fix several occurences of a bug where we would perform 'var & ~const'
+	  with a 64-bit variable and a int, i.e. 32-bit, constant.  This causes
+	  the higher order 32-bits of the 64-bit variable to be zeroed.  To fix
+	  this cast the 'const' to the same 64-bit type as 'var'.
+	- Change the runlist terminator of the newly allocated cluster(s) to
+	  LCN_ENOENT in ntfs_attr_make_non_resident().  Otherwise the runlist
+	  code gets confused.
+	- Add an extra parameter @last_vcn to ntfs_get_size_for_mapping_pairs()
+	  and ntfs_mapping_pairs_build() to allow the runlist encoding to be
+	  partial which is desirable when filling holes in sparse attributes.
+	  Update all callers.
+	- Change ntfs_map_runlist_nolock() to only decompress the mapping pairs
+	  if the requested vcn is inside it.  Otherwise we get into problems
+	  when we try to map an out of bounds vcn because we then try to map
+	  the already mapped runlist fragment which causes
+	  ntfs_mapping_pairs_decompress() to fail and return error.  Update
+	  ntfs_attr_find_vcn_nolock() accordingly.
+	- Fix a nasty deadlock that appeared in recent kernels.
+	  The situation: VFS inode X on a mounted ntfs volume is dirty.  For
+	  same inode X, the ntfs_inode is dirty and thus corresponding on-disk
+	  inode, i.e. mft record, which is in a dirty PAGE_CACHE_PAGE belonging
+	  to the table of inodes, i.e. $MFT, inode 0.
+	  What happens:
+	  Process 1: sys_sync()/umount()/whatever...  calls
+	  __sync_single_inode() for $MFT -> do_writepages() -> write_page for
+	  the dirty page containing the on-disk inode X, the page is now locked
+	  -> ntfs_write_mst_block() which clears PageUptodate() on the page to
+	  prevent anyone else getting hold of it whilst it does the write out.
+	  This is necessary as the on-disk inode needs "fixups" applied before
+	  the write to disk which are removed again after the write and
+	  PageUptodate is then set again.  It then analyses the page looking
+	  for dirty on-disk inodes and when it finds one it calls
+	  ntfs_may_write_mft_record() to see if it is safe to write this
+	  on-disk inode.  This then calls ilookup5() to check if the
+	  corresponding VFS inode is in icache().  This in turn calls ifind()
+	  which waits on the inode lock via wait_on_inode whilst holding the
+	  global inode_lock.
+	  Process 2: pdflush results in a call to __sync_single_inode for the
+	  same VFS inode X on the ntfs volume.  This locks the inode (I_LOCK)
+	  then calls write-inode -> ntfs_write_inode -> map_mft_record() ->
+	  read_cache_page() for the page (in page cache of table of inodes
+	  $MFT, inode 0) containing the on-disk inode.  This page has
+	  PageUptodate() clear because of Process 1 (see above) so
+	  read_cache_page() blocks when it tries to take the page lock for the
+	  page so it can call ntfs_read_page().
+	  Thus Process 1 is holding the page lock on the page containing the
+	  on-disk inode X and it is waiting on the inode X to be unlocked in
+	  ifind() so it can write the page out and then unlock the page.
+	  And Process 2 is holding the inode lock on inode X and is waiting for
+	  the page to be unlocked so it can call ntfs_readpage() or discover
+	  that Process 1 set PageUptodate() again and use the page.
+	  Thus we have a deadlock due to ifind() waiting on the inode lock.
+	  The solution: The fix is to use the newly introduced
+	  ilookup5_nowait() which does not wait on the inode's lock and hence
+	  avoids the deadlock.  This is safe as we do not care about the VFS
+	  inode and only use the fact that it is in the VFS inode cache and the
+	  fact that the vfs and ntfs inodes are one struct in memory to find
+	  the ntfs inode in memory if present.  Also, the ntfs inode has its
+	  own locking so it does not matter if the vfs inode is locked.
 
 2.1.22 - Many bug and race fixes and error handling improvements.
 
@@ -1037,7 +1180,7 @@ tng-0.0.8 - 08/03/2002 - Now using BitKeeper, http://linux-ntfs.bkbits.net/
 	- Further runlist merging work. (Richard Russon)
 	- Backwards compatibility for gcc-2.95. (Richard Russon)
 	- Update to kernel 2.5.5-pre1 and rediff the now tiny patch.
-	- Convert to new file system declaration using ->ntfs_get_sb() and
+	- Convert to new filesystem declaration using ->ntfs_get_sb() and
 	  replacing ntfs_read_super() with ntfs_fill_super().
 	- Set s_maxbytes to MAX_LFS_FILESIZE to avoid page cache page index
 	  overflow on 32-bit architectures.
@@ -1333,7 +1476,7 @@ tng-0.0.1 - The first useful version.
 	The driver is now actually useful! Yey. (-: It undoubtedly has got bugs
 	though and it doesn't implement accesssing compressed files yet. Also,
 	accessing files with attribute list attributes is not implemented yet
-	either. But for small or simple file systems it should work and allow
+	either. But for small or simple filesystems it should work and allow
 	you to list directories, use stat on directory entries and the file
 	system, open, read, mmap and llseek around in files. A big mile stone
 	has been reached!
@@ -1341,7 +1484,7 @@ tng-0.0.1 - The first useful version.
 tng-0.0.0 - Initial version tag.
 
 	Initial driver implementation. The driver can mount and umount simple
-	NTFS file systems (i.e. ones without attribute lists in the system
+	NTFS filesystems (i.e. ones without attribute lists in the system
 	files). If the mount fails there might be problems in the error handling
 	code paths, so be warned. Otherwise it seems to be loading the system
 	files nicely and the mft record read mapping/unmapping seems to be
diff --git a/fs/ntfs/Makefile b/fs/ntfs/Makefile
index 7b66381..f083f27 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=\"2.1.22\"
+EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.23\"
 
 ifeq ($(CONFIG_NTFS_DEBUG),y)
 EXTRA_CFLAGS += -DDEBUG
@@ -15,5 +15,5 @@ endif
 ifeq ($(CONFIG_NTFS_RW),y)
 EXTRA_CFLAGS += -DNTFS_RW
 
-ntfs-objs += bitmap.o lcnalloc.o logfile.o quota.o
+ntfs-objs += bitmap.o lcnalloc.o logfile.o quota.o usnjrnl.o
 endif
diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index 45d56e4..3f43bfe6 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -2,7 +2,7 @@
  * aops.c - NTFS kernel address space operations and page cache handling.
  *	    Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -66,19 +66,22 @@ static void ntfs_end_buffer_async_read(struct buffer_head *bh, int uptodate)
 	ni = NTFS_I(page->mapping->host);
 
 	if (likely(uptodate)) {
-		s64 file_ofs;
+		s64 file_ofs, initialized_size;
 
 		set_buffer_uptodate(bh);
 
 		file_ofs = ((s64)page->index << PAGE_CACHE_SHIFT) +
 				bh_offset(bh);
+		read_lock_irqsave(&ni->size_lock, flags);
+		initialized_size = ni->initialized_size;
+		read_unlock_irqrestore(&ni->size_lock, flags);
 		/* Check for the current buffer head overflowing. */
-		if (file_ofs + bh->b_size > ni->initialized_size) {
+		if (file_ofs + bh->b_size > initialized_size) {
 			char *addr;
 			int ofs = 0;
 
-			if (file_ofs < ni->initialized_size)
-				ofs = ni->initialized_size - file_ofs;
+			if (file_ofs < initialized_size)
+				ofs = initialized_size - file_ofs;
 			addr = kmap_atomic(page, KM_BIO_SRC_IRQ);
 			memset(addr + bh_offset(bh) + ofs, 0, bh->b_size - ofs);
 			flush_dcache_page(page);
@@ -132,7 +135,7 @@ static void ntfs_end_buffer_async_read(struct buffer_head *bh, int uptodate)
 					i * rec_size), rec_size);
 		flush_dcache_page(page);
 		kunmap_atomic(addr, KM_BIO_SRC_IRQ);
-		if (likely(!PageError(page) && page_uptodate))
+		if (likely(page_uptodate && !PageError(page)))
 			SetPageUptodate(page);
 	}
 	unlock_page(page);
@@ -168,6 +171,7 @@ static int ntfs_read_block(struct page *page)
 	runlist_element *rl;
 	struct buffer_head *bh, *head, *arr[MAX_BUF_PER_PAGE];
 	sector_t iblock, lblock, zblock;
+	unsigned long flags;
 	unsigned int blocksize, vcn_ofs;
 	int i, nr;
 	unsigned char blocksize_bits;
@@ -190,8 +194,10 @@ static int ntfs_read_block(struct page *page)
 	}
 
 	iblock = (s64)page->index << (PAGE_CACHE_SHIFT - blocksize_bits);
+	read_lock_irqsave(&ni->size_lock, flags);
 	lblock = (ni->allocated_size + blocksize - 1) >> blocksize_bits;
 	zblock = (ni->initialized_size + blocksize - 1) >> blocksize_bits;
+	read_unlock_irqrestore(&ni->size_lock, flags);
 
 	/* Loop through all the buffers in the page. */
 	rl = NULL;
@@ -258,7 +264,8 @@ lock_retry_remap:
 					goto lock_retry_remap;
 				rl = NULL;
 				lcn = err;
-			}
+			} else if (!rl)
+				up_read(&ni->runlist.lock);
 			/* Hard error, zero out region. */
 			bh->b_blocknr = -1;
 			SetPageError(page);
@@ -341,14 +348,15 @@ handle_zblock:
  */
 static int ntfs_readpage(struct file *file, struct page *page)
 {
-	loff_t i_size;
 	ntfs_inode *ni, *base_ni;
 	u8 *kaddr;
 	ntfs_attr_search_ctx *ctx;
 	MFT_RECORD *mrec;
+	unsigned long flags;
 	u32 attr_len;
 	int err = 0;
 
+retry_readpage:
 	BUG_ON(!PageLocked(page));
 	/*
 	 * This can potentially happen because we clear PageUptodate() during
@@ -383,9 +391,9 @@ static int ntfs_readpage(struct file *file, struct page *page)
 	 * Attribute is resident, implying it is not compressed or encrypted.
 	 * This also means the attribute is smaller than an mft record and
 	 * hence smaller than a page, so can simply zero out any pages with
-	 * index above 0.  We can also do this if the file size is 0.
+	 * index above 0.
 	 */
-	if (unlikely(page->index > 0 || !i_size_read(VFS_I(ni)))) {
+	if (unlikely(page->index > 0)) {
 		kaddr = kmap_atomic(page, KM_USER0);
 		memset(kaddr, 0, PAGE_CACHE_SIZE);
 		flush_dcache_page(page);
@@ -402,6 +410,14 @@ static int ntfs_readpage(struct file *file, struct page *page)
 		err = PTR_ERR(mrec);
 		goto err_out;
 	}
+	/*
+	 * If a parallel write made the attribute non-resident, drop the mft
+	 * record and retry the readpage.
+	 */
+	if (unlikely(NInoNonResident(ni))) {
+		unmap_mft_record(base_ni);
+		goto retry_readpage;
+	}
 	ctx = ntfs_attr_get_search_ctx(base_ni, mrec);
 	if (unlikely(!ctx)) {
 		err = -ENOMEM;
@@ -412,9 +428,10 @@ static int ntfs_readpage(struct file *file, struct page *page)
 	if (unlikely(err))
 		goto put_unm_err_out;
 	attr_len = le32_to_cpu(ctx->attr->data.resident.value_length);
-	i_size = i_size_read(VFS_I(ni));
-	if (unlikely(attr_len > i_size))
-		attr_len = i_size;
+	read_lock_irqsave(&ni->size_lock, flags);
+	if (unlikely(attr_len > ni->initialized_size))
+		attr_len = ni->initialized_size;
+	read_unlock_irqrestore(&ni->size_lock, flags);
 	kaddr = kmap_atomic(page, KM_USER0);
 	/* Copy the data to the page. */
 	memcpy(kaddr, (u8*)ctx->attr +
@@ -463,12 +480,15 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 {
 	VCN vcn;
 	LCN lcn;
+	s64 initialized_size;
+	loff_t i_size;
 	sector_t block, dblock, iblock;
 	struct inode *vi;
 	ntfs_inode *ni;
 	ntfs_volume *vol;
 	runlist_element *rl;
 	struct buffer_head *bh, *head;
+	unsigned long flags;
 	unsigned int blocksize, vcn_ofs;
 	int err;
 	BOOL need_end_writeback;
@@ -510,11 +530,16 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 	/* The first block in the page. */
 	block = (s64)page->index << (PAGE_CACHE_SHIFT - blocksize_bits);
 
+	read_lock_irqsave(&ni->size_lock, flags);
+	i_size = i_size_read(vi);
+	initialized_size = ni->initialized_size;
+	read_unlock_irqrestore(&ni->size_lock, flags);
+
 	/* The first out of bounds block for the data size. */
-	dblock = (vi->i_size + blocksize - 1) >> blocksize_bits;
+	dblock = (i_size + blocksize - 1) >> blocksize_bits;
 
 	/* The last (fully or partially) initialized block. */
-	iblock = ni->initialized_size >> blocksize_bits;
+	iblock = initialized_size >> blocksize_bits;
 
 	/*
 	 * Be very careful.  We have no exclusion from __set_page_dirty_buffers
@@ -559,7 +584,7 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 
 		/* Make sure we have enough initialized size. */
 		if (unlikely((block >= iblock) &&
-				(ni->initialized_size < vi->i_size))) {
+				(initialized_size < i_size))) {
 			/*
 			 * If this page is fully outside initialized size, zero
 			 * out all pages between the current initialized size
@@ -666,7 +691,8 @@ lock_retry_remap:
 				goto lock_retry_remap;
 			rl = NULL;
 			lcn = err;
-		}
+		} else if (!rl)
+			up_read(&ni->runlist.lock);
 		/* Failed to map the buffer, even after retrying. */
 		bh->b_blocknr = -1;
 		ntfs_error(vol->sb, "Failed to write to inode 0x%lx, "
@@ -801,17 +827,15 @@ static int ntfs_write_mst_block(struct page *page,
 	ntfs_inode *ni = NTFS_I(vi);
 	ntfs_volume *vol = ni->vol;
 	u8 *kaddr;
-	unsigned char bh_size_bits = vi->i_blkbits;
-	unsigned int bh_size = 1 << bh_size_bits;
 	unsigned int rec_size = ni->itype.index.block_size;
 	ntfs_inode *locked_nis[PAGE_CACHE_SIZE / rec_size];
 	struct buffer_head *bh, *head, *tbh, *rec_start_bh;
-	int max_bhs = PAGE_CACHE_SIZE / bh_size;
-	struct buffer_head *bhs[max_bhs];
+	struct buffer_head *bhs[MAX_BUF_PER_PAGE];
 	runlist_element *rl;
-	int i, nr_locked_nis, nr_recs, nr_bhs, bhs_per_rec, err, err2;
-	unsigned rec_size_bits;
+	int i, nr_locked_nis, nr_recs, nr_bhs, max_bhs, bhs_per_rec, err, err2;
+	unsigned bh_size, rec_size_bits;
 	BOOL sync, is_mft, page_is_dirty, rec_is_dirty;
+	unsigned char bh_size_bits;
 
 	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, page index "
 			"0x%lx.", vi->i_ino, ni->type, page->index);
@@ -826,7 +850,11 @@ static int ntfs_write_mst_block(struct page *page,
 	 */
 	BUG_ON(!(is_mft || S_ISDIR(vi->i_mode) ||
 			(NInoAttr(ni) && ni->type == AT_INDEX_ALLOCATION)));
+	bh_size_bits = vi->i_blkbits;
+	bh_size = 1 << bh_size_bits;
+	max_bhs = PAGE_CACHE_SIZE / bh_size;
 	BUG_ON(!max_bhs);
+	BUG_ON(max_bhs > MAX_BUF_PER_PAGE);
 
 	/* Were we called for sync purposes? */
 	sync = (wbc->sync_mode == WB_SYNC_ALL);
@@ -846,7 +874,7 @@ static int ntfs_write_mst_block(struct page *page,
 			(PAGE_CACHE_SHIFT - bh_size_bits);
 
 	/* The first out of bounds block for the data size. */
-	dblock = (vi->i_size + bh_size - 1) >> bh_size_bits;
+	dblock = (i_size_read(vi) + bh_size - 1) >> bh_size_bits;
 
 	rl = NULL;
 	err = err2 = nr_bhs = nr_recs = nr_locked_nis = 0;
@@ -858,6 +886,7 @@ static int ntfs_write_mst_block(struct page *page,
 		if (likely(block < rec_block)) {
 			if (unlikely(block >= dblock)) {
 				clear_buffer_dirty(bh);
+				set_buffer_uptodate(bh);
 				continue;
 			}
 			/*
@@ -938,8 +967,11 @@ lock_retry_remap:
 					if (err2 == -ENOMEM)
 						page_is_dirty = TRUE;
 					lcn = err2;
-				} else
+				} else {
 					err2 = -EIO;
+					if (!rl)
+						up_read(&ni->runlist.lock);
+				}
 				/* Hard error.  Abort writing this record. */
 				if (!err || err == -ENOMEM)
 					err = err2;
@@ -949,7 +981,8 @@ lock_retry_remap:
 						"attribute type 0x%x) because "
 						"its location on disk could "
 						"not be determined (error "
-						"code %lli).", (s64)block <<
+						"code %lli).",
+						(long long)block <<
 						bh_size_bits >>
 						vol->mft_record_size_bits,
 						ni->mft_no, ni->type,
@@ -1223,19 +1256,17 @@ done:
 static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 {
 	loff_t i_size;
-	struct inode *vi;
-	ntfs_inode *ni, *base_ni;
+	struct inode *vi = page->mapping->host;
+	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);
 	char *kaddr;
-	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *m;
+	ntfs_attr_search_ctx *ctx = NULL;
+	MFT_RECORD *m = NULL;
 	u32 attr_len;
 	int err;
 
+retry_writepage:
 	BUG_ON(!PageLocked(page));
-
-	vi = page->mapping->host;
 	i_size = i_size_read(vi);
-
 	/* Is the page fully outside i_size? (truncate in progress) */
 	if (unlikely(page->index >= (i_size + PAGE_CACHE_SIZE - 1) >>
 			PAGE_CACHE_SHIFT)) {
@@ -1248,8 +1279,6 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 		ntfs_debug("Write outside i_size - truncated?");
 		return 0;
 	}
-	ni = NTFS_I(vi);
-
 	/* NInoNonResident() == NInoIndexAllocPresent() */
 	if (NInoNonResident(ni)) {
 		/*
@@ -1326,6 +1355,14 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 		ctx = NULL;
 		goto err_out;
 	}
+	/*
+	 * If a parallel write made the attribute non-resident, drop the mft
+	 * record and retry the writepage.
+	 */
+	if (unlikely(NInoNonResident(ni))) {
+		unmap_mft_record(base_ni);
+		goto retry_writepage;
+	}
 	ctx = ntfs_attr_get_search_ctx(base_ni, m);
 	if (unlikely(!ctx)) {
 		err = -ENOMEM;
@@ -1367,15 +1404,12 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 	 */
 
 	attr_len = le32_to_cpu(ctx->attr->data.resident.value_length);
-	i_size = i_size_read(VFS_I(ni));
-	kaddr = kmap_atomic(page, KM_USER0);
+	i_size = i_size_read(vi);
 	if (unlikely(attr_len > i_size)) {
-		/* Zero out of bounds area in the mft record. */
-		memset((u8*)ctx->attr + le16_to_cpu(
-				ctx->attr->data.resident.value_offset) +
-				i_size, 0, attr_len - i_size);
 		attr_len = i_size;
+		ctx->attr->data.resident.value_length = cpu_to_le32(attr_len);
 	}
+	kaddr = kmap_atomic(page, KM_USER0);
 	/* Copy the data from the page to the mft record. */
 	memcpy((u8*)ctx->attr +
 			le16_to_cpu(ctx->attr->data.resident.value_offset),
@@ -1405,8 +1439,10 @@ err_out:
 		err = 0;
 	} else {
 		ntfs_error(vi->i_sb, "Resident attribute write failed with "
-				"error %i.  Setting page error flag.", err);
+				"error %i.", err);
 		SetPageError(page);
+		NVolSetErrors(ni->vol);
+		make_bad_inode(vi);
 	}
 	unlock_page(page);
 	if (ctx)
@@ -1425,12 +1461,15 @@ static int ntfs_prepare_nonresident_write(struct page *page,
 {
 	VCN vcn;
 	LCN lcn;
+	s64 initialized_size;
+	loff_t i_size;
 	sector_t block, ablock, iblock;
 	struct inode *vi;
 	ntfs_inode *ni;
 	ntfs_volume *vol;
 	runlist_element *rl;
 	struct buffer_head *bh, *head, *wait[2], **wait_bh = wait;
+	unsigned long flags;
 	unsigned int vcn_ofs, block_start, block_end, blocksize;
 	int err;
 	BOOL is_retry;
@@ -1462,16 +1501,20 @@ static int ntfs_prepare_nonresident_write(struct page *page,
 	/* The first block in the page. */
 	block = (s64)page->index << (PAGE_CACHE_SHIFT - blocksize_bits);
 
+	read_lock_irqsave(&ni->size_lock, flags);
 	/*
-	 * The first out of bounds block for the allocated size. No need to
+	 * The first out of bounds block for the allocated size.  No need to
 	 * round up as allocated_size is in multiples of cluster size and the
 	 * minimum cluster size is 512 bytes, which is equal to the smallest
 	 * blocksize.
 	 */
 	ablock = ni->allocated_size >> blocksize_bits;
+	i_size = i_size_read(vi);
+	initialized_size = ni->initialized_size;
+	read_unlock_irqrestore(&ni->size_lock, flags);
 
 	/* The last (fully or partially) initialized block. */
-	iblock = ni->initialized_size >> blocksize_bits;
+	iblock = initialized_size >> blocksize_bits;
 
 	/* Loop through all the buffers in the page. */
 	block_start = 0;
@@ -1518,7 +1561,7 @@ static int ntfs_prepare_nonresident_write(struct page *page,
 		 * request, i.e. block < ablock is true.
 		 */
 		if (unlikely((block >= iblock) &&
-				(ni->initialized_size < vi->i_size))) {
+				(initialized_size < i_size))) {
 			/*
 			 * If this page is fully outside initialized size, zero
 			 * out all pages between the current initialized size
@@ -1622,6 +1665,8 @@ lock_retry_remap:
 							"not supported yet. "
 							"Sorry.");
 					err = -EOPNOTSUPP;
+					if (!rl)
+						up_read(&ni->runlist.lock);
 					goto err_out;
 				} else if (!is_retry &&
 						lcn == LCN_RL_NOT_MAPPED) {
@@ -1636,7 +1681,8 @@ lock_retry_remap:
 						goto lock_retry_remap;
 					rl = NULL;
 					lcn = err;
-				}
+				} else if (!rl)
+					up_read(&ni->runlist.lock);
 				/*
 				 * Failed to map the buffer, even after
 				 * retrying.
@@ -1797,6 +1843,7 @@ static int ntfs_prepare_write(struct file *file, struct page *page,
 		unsigned from, unsigned to)
 {
 	s64 new_size;
+	loff_t i_size;
 	struct inode *vi = page->mapping->host;
 	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);
 	ntfs_volume *vol = ni->vol;
@@ -1868,14 +1915,8 @@ static int ntfs_prepare_write(struct file *file, struct page *page,
 	BUG_ON(page_has_buffers(page));
 	new_size = ((s64)page->index << PAGE_CACHE_SHIFT) + to;
 	/* If we do not need to resize the attribute allocation we are done. */
-	if (new_size <= vi->i_size)
+	if (new_size <= i_size_read(vi))
 		goto done;
-
-	// FIXME: We abort for now as this code is not safe.
-	ntfs_error(vi->i_sb, "Changing the file size is not supported yet.  "
-			"Sorry.");
-	return -EOPNOTSUPP;
-
 	/* Map, pin, and lock the (base) mft record. */
 	if (!NInoAttr(ni))
 		base_ni = ni;
@@ -1904,7 +1945,15 @@ static int ntfs_prepare_write(struct file *file, struct page *page,
 	a = ctx->attr;
 	/* The total length of the attribute value. */
 	attr_len = le32_to_cpu(a->data.resident.value_length);
-	BUG_ON(vi->i_size != attr_len);
+	/* Fix an eventual previous failure of ntfs_commit_write(). */
+	i_size = i_size_read(vi);
+	if (unlikely(attr_len > i_size)) {
+		attr_len = i_size;
+		a->data.resident.value_length = cpu_to_le32(attr_len);
+	}
+	/* If we do not need to resize the attribute allocation we are done. */
+	if (new_size <= attr_len)
+		goto done_unm;
 	/* Check if new size is allowed in $AttrDef. */
 	err = ntfs_attr_size_bounds_check(vol, ni->type, new_size);
 	if (unlikely(err)) {
@@ -1962,6 +2011,7 @@ static int ntfs_prepare_write(struct file *file, struct page *page,
 	}
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
+done_unm:
 	ntfs_attr_put_search_ctx(ctx);
 	unmap_mft_record(base_ni);
 	/*
@@ -2047,7 +2097,7 @@ static int ntfs_commit_nonresident_write(struct page *page,
 	 * now we know ntfs_prepare_write() would have failed in the write
 	 * exceeds i_size case, so this will never trigger which is fine.
 	 */
-	if (pos > vi->i_size) {
+	if (pos > i_size_read(vi)) {
 		ntfs_error(vi->i_sb, "Writing beyond the existing file size is "
 				"not supported yet.  Sorry.");
 		return -EOPNOTSUPP;
@@ -2183,9 +2233,13 @@ static int ntfs_commit_write(struct file *file, struct page *page,
 	}
 	kunmap_atomic(kaddr, KM_USER0);
 	/* Update i_size if necessary. */
-	if (vi->i_size < attr_len) {
+	if (i_size_read(vi) < attr_len) {
+		unsigned long flags;
+
+		write_lock_irqsave(&ni->size_lock, flags);
 		ni->allocated_size = ni->initialized_size = attr_len;
 		i_size_write(vi, attr_len);
+		write_unlock_irqrestore(&ni->size_lock, flags);
 	}
 	/* Mark the mft record dirty, so it gets written back. */
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
diff --git a/fs/ntfs/attrib.c b/fs/ntfs/attrib.c
index 1ff7f90..cd0f9e7 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) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -21,88 +21,217 @@
  */
 
 #include <linux/buffer_head.h>
+#include <linux/swap.h>
 
 #include "attrib.h"
 #include "debug.h"
 #include "layout.h"
+#include "lcnalloc.h"
+#include "malloc.h"
 #include "mft.h"
 #include "ntfs.h"
 #include "types.h"
 
 /**
- * ntfs_map_runlist - map (a part of) a runlist of an ntfs inode
+ * ntfs_map_runlist_nolock - map (a part of) a runlist of an ntfs inode
  * @ni:		ntfs inode for which to map (part of) a runlist
  * @vcn:	map runlist part containing this vcn
  *
  * Map the part of a runlist containing the @vcn of the ntfs inode @ni.
  *
- * Return 0 on success and -errno on error.
+ * Return 0 on success and -errno on error.  There is one special error code
+ * which is not an error as such.  This is -ENOENT.  It means that @vcn is out
+ * of bounds of the runlist.
  *
- * Locking: - The runlist must be unlocked on entry and is unlocked on return.
- *	    - This function takes the lock for writing and modifies the runlist.
+ * Locking: - The runlist must be locked for writing.
+ *	    - This function modifies the runlist.
  */
-int ntfs_map_runlist(ntfs_inode *ni, VCN vcn)
+int ntfs_map_runlist_nolock(ntfs_inode *ni, VCN vcn)
 {
+	VCN end_vcn;
 	ntfs_inode *base_ni;
+	MFT_RECORD *m;
+	ATTR_RECORD *a;
 	ntfs_attr_search_ctx *ctx;
-	MFT_RECORD *mrec;
+	runlist_element *rl;
 	int err = 0;
 
 	ntfs_debug("Mapping runlist part containing vcn 0x%llx.",
 			(unsigned long long)vcn);
-
 	if (!NInoAttr(ni))
 		base_ni = ni;
 	else
 		base_ni = ni->ext.base_ntfs_ino;
-
-	mrec = map_mft_record(base_ni);
-	if (IS_ERR(mrec))
-		return PTR_ERR(mrec);
-	ctx = ntfs_attr_get_search_ctx(base_ni, mrec);
+	m = map_mft_record(base_ni);
+	if (IS_ERR(m))
+		return PTR_ERR(m);
+	ctx = ntfs_attr_get_search_ctx(base_ni, m);
 	if (unlikely(!ctx)) {
 		err = -ENOMEM;
 		goto err_out;
 	}
 	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
 			CASE_SENSITIVE, vcn, NULL, 0, ctx);
-	if (unlikely(err))
-		goto put_err_out;
+	if (unlikely(err)) {
+		if (err == -ENOENT)
+			err = -EIO;
+		goto err_out;
+	}
+	a = ctx->attr;
+	/*
+	 * Only decompress the mapping pairs if @vcn is inside it.  Otherwise
+	 * we get into problems when we try to map an out of bounds vcn because
+	 * we then try to map the already mapped runlist fragment and
+	 * ntfs_mapping_pairs_decompress() fails.
+	 */
+	end_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn) + 1;
+	if (unlikely(!a->data.non_resident.lowest_vcn && end_vcn <= 1))
+		end_vcn = ni->allocated_size >> ni->vol->cluster_size_bits;
+	if (unlikely(vcn >= end_vcn)) {
+		err = -ENOENT;
+		goto err_out;
+	}
+	rl = ntfs_mapping_pairs_decompress(ni->vol, a, ni->runlist.rl);
+	if (IS_ERR(rl))
+		err = PTR_ERR(rl);
+	else
+		ni->runlist.rl = rl;
+err_out:
+	if (likely(ctx))
+		ntfs_attr_put_search_ctx(ctx);
+	unmap_mft_record(base_ni);
+	return err;
+}
+
+/**
+ * ntfs_map_runlist - map (a part of) a runlist of an ntfs inode
+ * @ni:		ntfs inode for which to map (part of) a runlist
+ * @vcn:	map runlist part containing this vcn
+ *
+ * Map the part of a runlist containing the @vcn of the ntfs inode @ni.
+ *
+ * Return 0 on success and -errno on error.  There is one special error code
+ * which is not an error as such.  This is -ENOENT.  It means that @vcn is out
+ * of bounds of the runlist.
+ *
+ * Locking: - The runlist must be unlocked on entry and is unlocked on return.
+ *	    - This function takes the runlist lock for writing and modifies the
+ *	      runlist.
+ */
+int ntfs_map_runlist(ntfs_inode *ni, VCN vcn)
+{
+	int err = 0;
 
 	down_write(&ni->runlist.lock);
 	/* Make sure someone else didn't do the work while we were sleeping. */
 	if (likely(ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn) <=
-			LCN_RL_NOT_MAPPED)) {
-		runlist_element *rl;
+			LCN_RL_NOT_MAPPED))
+		err = ntfs_map_runlist_nolock(ni, vcn);
+	up_write(&ni->runlist.lock);
+	return err;
+}
 
-		rl = ntfs_mapping_pairs_decompress(ni->vol, ctx->attr,
-				ni->runlist.rl);
-		if (IS_ERR(rl))
-			err = PTR_ERR(rl);
-		else
-			ni->runlist.rl = rl;
+/**
+ * ntfs_attr_vcn_to_lcn_nolock - convert a vcn into a lcn given an ntfs inode
+ * @ni:			ntfs inode of the attribute whose runlist to search
+ * @vcn:		vcn to convert
+ * @write_locked:	true if the runlist is locked for writing
+ *
+ * Find the virtual cluster number @vcn in the runlist of the ntfs attribute
+ * described by the ntfs inode @ni and return the corresponding logical cluster
+ * number (lcn).
+ *
+ * If the @vcn is not mapped yet, the attempt is made to map the attribute
+ * extent containing the @vcn and the vcn to lcn conversion is retried.
+ *
+ * If @write_locked is true the caller has locked the runlist for writing and
+ * if false for reading.
+ *
+ * Since lcns must be >= 0, we use negative return codes with special meaning:
+ *
+ * Return code	Meaning / Description
+ * ==========================================
+ *  LCN_HOLE	Hole / not allocated on disk.
+ *  LCN_ENOENT	There is no such vcn in the runlist, i.e. @vcn is out of bounds.
+ *  LCN_ENOMEM	Not enough memory to map runlist.
+ *  LCN_EIO	Critical error (runlist/file is corrupt, i/o error, etc).
+ *
+ * Locking: - The runlist must be locked on entry and is left locked on return.
+ *	    - If @write_locked is FALSE, i.e. the runlist is locked for reading,
+ *	      the lock may be dropped inside the function so you cannot rely on
+ *	      the runlist still being the same when this function returns.
+ */
+LCN ntfs_attr_vcn_to_lcn_nolock(ntfs_inode *ni, const VCN vcn,
+		const BOOL write_locked)
+{
+	LCN lcn;
+	BOOL is_retry = FALSE;
+
+	ntfs_debug("Entering for i_ino 0x%lx, vcn 0x%llx, %s_locked.",
+			ni->mft_no, (unsigned long long)vcn,
+			write_locked ? "write" : "read");
+	BUG_ON(!ni);
+	BUG_ON(!NInoNonResident(ni));
+	BUG_ON(vcn < 0);
+retry_remap:
+	/* Convert vcn to lcn.  If that fails map the runlist and retry once. */
+	lcn = ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn);
+	if (likely(lcn >= LCN_HOLE)) {
+		ntfs_debug("Done, lcn 0x%llx.", (long long)lcn);
+		return lcn;
 	}
-	up_write(&ni->runlist.lock);
+	if (lcn != LCN_RL_NOT_MAPPED) {
+		if (lcn != LCN_ENOENT)
+			lcn = LCN_EIO;
+	} else if (!is_retry) {
+		int err;
 
-put_err_out:
-	ntfs_attr_put_search_ctx(ctx);
-err_out:
-	unmap_mft_record(base_ni);
-	return err;
+		if (!write_locked) {
+			up_read(&ni->runlist.lock);
+			down_write(&ni->runlist.lock);
+			if (unlikely(ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn) !=
+					LCN_RL_NOT_MAPPED)) {
+				up_write(&ni->runlist.lock);
+				down_read(&ni->runlist.lock);
+				goto retry_remap;
+			}
+		}
+		err = ntfs_map_runlist_nolock(ni, vcn);
+		if (!write_locked) {
+			up_write(&ni->runlist.lock);
+			down_read(&ni->runlist.lock);
+		}
+		if (likely(!err)) {
+			is_retry = TRUE;
+			goto retry_remap;
+		}
+		if (err == -ENOENT)
+			lcn = LCN_ENOENT;
+		else if (err == -ENOMEM)
+			lcn = LCN_ENOMEM;
+		else
+			lcn = LCN_EIO;
+	}
+	if (lcn != LCN_ENOENT)
+		ntfs_error(ni->vol->sb, "Failed with error code %lli.",
+				(long long)lcn);
+	return lcn;
 }
 
 /**
- * ntfs_find_vcn - find a vcn in the runlist described by an ntfs inode
- * @ni:		ntfs inode describing the runlist to search
- * @vcn:	vcn to find
- * @need_write:	if false, lock for reading and if true, lock for writing
+ * ntfs_attr_find_vcn_nolock - find a vcn in the runlist of an ntfs inode
+ * @ni:			ntfs inode describing the runlist to search
+ * @vcn:		vcn to find
+ * @write_locked:	true if the runlist is locked for writing
  *
  * Find the virtual cluster number @vcn in the runlist described by the ntfs
  * inode @ni and return the address of the runlist element containing the @vcn.
- * The runlist is left locked and the caller has to unlock it.  If @need_write
- * is true, the runlist is locked for writing and if @need_write is false, the
- * runlist is locked for reading.  In the error case, the runlist is not left
- * locked.
+ *
+ * If the @vcn is not mapped yet, the attempt is made to map the attribute
+ * extent containing the @vcn and the vcn to lcn conversion is retried.
+ *
+ * If @write_locked is true the caller has locked the runlist for writing and
+ * if false for reading.
  *
  * Note you need to distinguish between the lcn of the returned runlist element
  * being >= 0 and LCN_HOLE.  In the later case you have to return zeroes on
@@ -118,34 +247,29 @@ err_out:
  *	-ENOMEM - Not enough memory to map runlist.
  *	-EIO	- Critical error (runlist/file is corrupt, i/o error, etc).
  *
- * Locking: - The runlist must be unlocked on entry.
- *	    - On failing return, the runlist is unlocked.
- *	    - On successful return, the runlist is locked.  If @need_write us
- *	      true, it is locked for writing.  Otherwise is is locked for
- *	      reading.
+ * Locking: - The runlist must be locked on entry and is left locked on return.
+ *	    - If @write_locked is FALSE, i.e. the runlist is locked for reading,
+ *	      the lock may be dropped inside the function so you cannot rely on
+ *	      the runlist still being the same when this function returns.
  */
-runlist_element *ntfs_find_vcn(ntfs_inode *ni, const VCN vcn,
-		const BOOL need_write)
+runlist_element *ntfs_attr_find_vcn_nolock(ntfs_inode *ni, const VCN vcn,
+		const BOOL write_locked)
 {
 	runlist_element *rl;
 	int err = 0;
 	BOOL is_retry = FALSE;
 
-	ntfs_debug("Entering for i_ino 0x%lx, vcn 0x%llx, lock for %sing.",
+	ntfs_debug("Entering for i_ino 0x%lx, vcn 0x%llx, %s_locked.",
 			ni->mft_no, (unsigned long long)vcn,
-			!need_write ? "read" : "writ");
+			write_locked ? "write" : "read");
 	BUG_ON(!ni);
 	BUG_ON(!NInoNonResident(ni));
 	BUG_ON(vcn < 0);
-lock_retry_remap:
-	if (!need_write)
-		down_read(&ni->runlist.lock);
-	else
-		down_write(&ni->runlist.lock);
+retry_remap:
 	rl = ni->runlist.rl;
 	if (likely(rl && vcn >= rl[0].vcn)) {
 		while (likely(rl->length)) {
-			if (likely(vcn < rl[1].vcn)) {
+			if (unlikely(vcn < rl[1].vcn)) {
 				if (likely(rl->lcn >= LCN_HOLE)) {
 					ntfs_debug("Done.");
 					return rl;
@@ -161,30 +285,41 @@ lock_retry_remap:
 				err = -EIO;
 		}
 	}
-	if (!need_write)
-		up_read(&ni->runlist.lock);
-	else
-		up_write(&ni->runlist.lock);
 	if (!err && !is_retry) {
 		/*
 		 * The @vcn is in an unmapped region, map the runlist and
 		 * retry.
 		 */
-		err = ntfs_map_runlist(ni, vcn);
+		if (!write_locked) {
+			up_read(&ni->runlist.lock);
+			down_write(&ni->runlist.lock);
+			if (unlikely(ntfs_rl_vcn_to_lcn(ni->runlist.rl, vcn) !=
+					LCN_RL_NOT_MAPPED)) {
+				up_write(&ni->runlist.lock);
+				down_read(&ni->runlist.lock);
+				goto retry_remap;
+			}
+		}
+		err = ntfs_map_runlist_nolock(ni, vcn);
+		if (!write_locked) {
+			up_write(&ni->runlist.lock);
+			down_read(&ni->runlist.lock);
+		}
 		if (likely(!err)) {
 			is_retry = TRUE;
-			goto lock_retry_remap;
+			goto retry_remap;
 		}
 		/*
-		 * -EINVAL and -ENOENT coming from a failed mapping attempt are
-		 * equivalent to i/o errors for us as they should not happen in
-		 * our code paths.
+		 * -EINVAL coming from a failed mapping attempt is equivalent
+		 * to i/o error for us as it should not happen in our code
+		 * paths.
 		 */
-		if (err == -EINVAL || err == -ENOENT)
+		if (err == -EINVAL)
 			err = -EIO;
 	} else if (!err)
 		err = -EIO;
-	ntfs_error(ni->vol->sb, "Failed with error code %i.", err);
+	if (err != -ENOENT)
+		ntfs_error(ni->vol->sb, "Failed with error code %i.", err);
 	return ERR_PTR(err);
 }
 
@@ -870,15 +1005,14 @@ int ntfs_attr_lookup(const ATTR_TYPE type, const ntfschar *name,
 static inline void ntfs_attr_init_search_ctx(ntfs_attr_search_ctx *ctx,
 		ntfs_inode *ni, MFT_RECORD *mrec)
 {
-	ctx->mrec = mrec;
-	/* Sanity checks are performed elsewhere. */
-	ctx->attr = (ATTR_RECORD*)((u8*)mrec + le16_to_cpu(mrec->attrs_offset));
-	ctx->is_first = TRUE;
-	ctx->ntfs_ino = ni;
-	ctx->al_entry = NULL;
-	ctx->base_ntfs_ino = NULL;
-	ctx->base_mrec = NULL;
-	ctx->base_attr = NULL;
+	*ctx = (ntfs_attr_search_ctx) {
+		.mrec = mrec,
+		/* Sanity checks are performed elsewhere. */
+		.attr = (ATTR_RECORD*)((u8*)mrec +
+				le16_to_cpu(mrec->attrs_offset)),
+		.is_first = TRUE,
+		.ntfs_ino = ni,
+	};
 }
 
 /**
@@ -945,6 +1079,8 @@ void ntfs_attr_put_search_ctx(ntfs_attr_search_ctx *ctx)
 	return;
 }
 
+#ifdef NTFS_RW
+
 /**
  * ntfs_attr_find_in_attrdef - find an attribute in the $AttrDef system file
  * @vol:	ntfs volume to which the attribute belongs
@@ -1024,27 +1160,21 @@ int ntfs_attr_size_bounds_check(const ntfs_volume *vol, const ATTR_TYPE type,
  * Check whether the attribute of @type on the ntfs volume @vol is allowed to
  * be non-resident.  This information is obtained from $AttrDef system file.
  *
- * Return 0 if the attribute is allowed to be non-resident, -EPERM if not, or
+ * Return 0 if the attribute is allowed to be non-resident, -EPERM if not, and
  * -ENOENT if the attribute is not listed in $AttrDef.
  */
 int ntfs_attr_can_be_non_resident(const ntfs_volume *vol, const ATTR_TYPE type)
 {
 	ATTR_DEF *ad;
 
-	/*
-	 * $DATA is always allowed to be non-resident even if $AttrDef does not
-	 * specify this in the flags of the $DATA attribute definition record.
-	 */
-	if (type == AT_DATA)
-		return 0;
 	/* Find the attribute definition record in $AttrDef. */
 	ad = ntfs_attr_find_in_attrdef(vol, type);
 	if (unlikely(!ad))
 		return -ENOENT;
 	/* Check the flags and return the result. */
-	if (ad->flags & CAN_BE_NON_RESIDENT)
-		return 0;
-	return -EPERM;
+	if (ad->flags & ATTR_DEF_RESIDENT)
+		return -EPERM;
+	return 0;
 }
 
 /**
@@ -1067,9 +1197,9 @@ int ntfs_attr_can_be_non_resident(const ntfs_volume *vol, const ATTR_TYPE type)
  */
 int ntfs_attr_can_be_resident(const ntfs_volume *vol, const ATTR_TYPE type)
 {
-	if (type != AT_INDEX_ALLOCATION && type != AT_EA)
-		return 0;
-	return -EPERM;
+	if (type == AT_INDEX_ALLOCATION || type == AT_EA)
+		return -EPERM;
+	return 0;
 }
 
 /**
@@ -1117,6 +1247,328 @@ int ntfs_attr_record_resize(MFT_RECORD *m, ATTR_RECORD *a, u32 new_size)
 }
 
 /**
+ * ntfs_attr_make_non_resident - convert a resident to a non-resident attribute
+ * @ni:		ntfs inode describing the attribute to convert
+ *
+ * Convert the resident ntfs attribute described by the ntfs inode @ni to a
+ * non-resident one.
+ *
+ * Return 0 on success and -errno on error.  The following error return codes
+ * are defined:
+ *	-EPERM	- The attribute is not allowed to be non-resident.
+ *	-ENOMEM	- Not enough memory.
+ *	-ENOSPC	- Not enough disk space.
+ *	-EINVAL	- Attribute not defined on the volume.
+ *	-EIO	- I/o error or other error.
+ * Note that -ENOSPC is also returned in the case that there is not enough
+ * space in the mft record to do the conversion.  This can happen when the mft
+ * record is already very full.  The caller is responsible for trying to make
+ * space in the mft record and trying again.  FIXME: Do we need a separate
+ * error return code for this kind of -ENOSPC or is it always worth trying
+ * again in case the attribute may then fit in a resident state so no need to
+ * make it non-resident at all?  Ho-hum...  (AIA)
+ *
+ * NOTE to self: No changes in the attribute list are required to move from
+ *		 a resident to a non-resident attribute.
+ *
+ * Locking: - The caller must hold i_sem on the inode.
+ */
+int ntfs_attr_make_non_resident(ntfs_inode *ni)
+{
+	s64 new_size;
+	struct inode *vi = VFS_I(ni);
+	ntfs_volume *vol = ni->vol;
+	ntfs_inode *base_ni;
+	MFT_RECORD *m;
+	ATTR_RECORD *a;
+	ntfs_attr_search_ctx *ctx;
+	struct page *page;
+	runlist_element *rl;
+	u8 *kaddr;
+	unsigned long flags;
+	int mp_size, mp_ofs, name_ofs, arec_size, err, err2;
+	u32 attr_size;
+	u8 old_res_attr_flags;
+
+	/* Check that the attribute is allowed to be non-resident. */
+	err = ntfs_attr_can_be_non_resident(vol, ni->type);
+	if (unlikely(err)) {
+		if (err == -EPERM)
+			ntfs_debug("Attribute is not allowed to be "
+					"non-resident.");
+		else
+			ntfs_debug("Attribute not defined on the NTFS "
+					"volume!");
+		return err;
+	}
+	/*
+	 * The size needs to be aligned to a cluster boundary for allocation
+	 * purposes.
+	 */
+	new_size = (i_size_read(vi) + vol->cluster_size - 1) &
+			~(vol->cluster_size - 1);
+	if (new_size > 0) {
+		runlist_element *rl2;
+
+		/*
+		 * Will need the page later and since the page lock nests
+		 * outside all ntfs locks, we need to get the page now.
+		 */
+		page = find_or_create_page(vi->i_mapping, 0,
+				mapping_gfp_mask(vi->i_mapping));
+		if (unlikely(!page))
+			return -ENOMEM;
+		/* Start by allocating clusters to hold the attribute value. */
+		rl = ntfs_cluster_alloc(vol, 0, new_size >>
+				vol->cluster_size_bits, -1, DATA_ZONE);
+		if (IS_ERR(rl)) {
+			err = PTR_ERR(rl);
+			ntfs_debug("Failed to allocate cluster%s, error code "
+					"%i.", (new_size >>
+					vol->cluster_size_bits) > 1 ? "s" : "",
+					err);
+			goto page_err_out;
+		}
+		/* Change the runlist terminator to LCN_ENOENT. */
+		rl2 = rl;
+		while (rl2->length)
+			rl2++;
+		BUG_ON(rl2->lcn != LCN_RL_NOT_MAPPED);
+		rl2->lcn = LCN_ENOENT;
+	} else {
+		rl = NULL;
+		page = NULL;
+	}
+	/* Determine the size of the mapping pairs array. */
+	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl, 0, -1);
+	if (unlikely(mp_size < 0)) {
+		err = mp_size;
+		ntfs_debug("Failed to get size for mapping pairs array, error "
+				"code %i.", err);
+		goto rl_err_out;
+	}
+	down_write(&ni->runlist.lock);
+	if (!NInoAttr(ni))
+		base_ni = ni;
+	else
+		base_ni = ni->ext.base_ntfs_ino;
+	m = map_mft_record(base_ni);
+	if (IS_ERR(m)) {
+		err = PTR_ERR(m);
+		m = NULL;
+		ctx = NULL;
+		goto err_out;
+	}
+	ctx = ntfs_attr_get_search_ctx(base_ni, m);
+	if (unlikely(!ctx)) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
+			CASE_SENSITIVE, 0, NULL, 0, ctx);
+	if (unlikely(err)) {
+		if (err == -ENOENT)
+			err = -EIO;
+		goto err_out;
+	}
+	m = ctx->mrec;
+	a = ctx->attr;
+	BUG_ON(NInoNonResident(ni));
+	BUG_ON(a->non_resident);
+	/*
+	 * Calculate new offsets for the name and the mapping pairs array.
+	 * We assume the attribute is not compressed or sparse.
+	 */
+	name_ofs = (offsetof(ATTR_REC,
+			data.non_resident.compressed_size) + 7) & ~7;
+	mp_ofs = (name_ofs + a->name_length * sizeof(ntfschar) + 7) & ~7;
+	/*
+	 * Determine the size of the resident part of the now non-resident
+	 * attribute record.
+	 */
+	arec_size = (mp_ofs + mp_size + 7) & ~7;
+	/*
+	 * If the page is not uptodate bring it uptodate by copying from the
+	 * attribute value.
+	 */
+	attr_size = le32_to_cpu(a->data.resident.value_length);
+	BUG_ON(attr_size != i_size_read(vi));
+	if (page && !PageUptodate(page)) {
+		kaddr = kmap_atomic(page, KM_USER0);
+		memcpy(kaddr, (u8*)a +
+				le16_to_cpu(a->data.resident.value_offset),
+				attr_size);
+		memset(kaddr + attr_size, 0, PAGE_CACHE_SIZE - attr_size);
+		kunmap_atomic(kaddr, KM_USER0);
+		flush_dcache_page(page);
+		SetPageUptodate(page);
+	}
+	/* Backup the attribute flag. */
+	old_res_attr_flags = a->data.resident.flags;
+	/* Resize the resident part of the attribute record. */
+	err = ntfs_attr_record_resize(m, a, arec_size);
+	if (unlikely(err))
+		goto err_out;
+	/*
+	 * Convert the resident part of the attribute record to describe a
+	 * non-resident attribute.
+	 */
+	a->non_resident = 1;
+	/* Move the attribute name if it exists and update the offset. */
+	if (a->name_length)
+		memmove((u8*)a + name_ofs, (u8*)a + le16_to_cpu(a->name_offset),
+				a->name_length * sizeof(ntfschar));
+	a->name_offset = cpu_to_le16(name_ofs);
+	/*
+	 * FIXME: For now just clear all of these as we do not support them
+	 * when writing.
+	 */
+	a->flags &= cpu_to_le16(0xffff & ~le16_to_cpu(ATTR_IS_SPARSE |
+			ATTR_IS_ENCRYPTED | ATTR_COMPRESSION_MASK));
+	/* Setup the fields specific to non-resident attributes. */
+	a->data.non_resident.lowest_vcn = 0;
+	a->data.non_resident.highest_vcn = cpu_to_sle64((new_size - 1) >>
+			vol->cluster_size_bits);
+	a->data.non_resident.mapping_pairs_offset = cpu_to_le16(mp_ofs);
+	a->data.non_resident.compression_unit = 0;
+	memset(&a->data.non_resident.reserved, 0,
+			sizeof(a->data.non_resident.reserved));
+	a->data.non_resident.allocated_size = cpu_to_sle64(new_size);
+	a->data.non_resident.data_size =
+			a->data.non_resident.initialized_size =
+			cpu_to_sle64(attr_size);
+	/* Generate the mapping pairs array into the attribute record. */
+	err = ntfs_mapping_pairs_build(vol, (u8*)a + mp_ofs,
+			arec_size - mp_ofs, rl, 0, -1, NULL);
+	if (unlikely(err)) {
+		ntfs_debug("Failed to build mapping pairs, error code %i.",
+				err);
+		goto undo_err_out;
+	}
+	/* Setup the in-memory attribute structure to be non-resident. */
+	/*
+	 * FIXME: For now just clear all of these as we do not support them
+	 * when writing.
+	 */
+	NInoClearSparse(ni);
+	NInoClearEncrypted(ni);
+	NInoClearCompressed(ni);
+	ni->runlist.rl = rl;
+	write_lock_irqsave(&ni->size_lock, flags);
+	ni->allocated_size = new_size;
+	write_unlock_irqrestore(&ni->size_lock, flags);
+	/*
+	 * This needs to be last since the address space operations ->readpage
+	 * and ->writepage can run concurrently with us as they are not
+	 * serialized on i_sem.  Note, we are not allowed to fail once we flip
+	 * this switch, which is another reason to do this last.
+	 */
+	NInoSetNonResident(ni);
+	/* Mark the mft record dirty, so it gets written back. */
+	flush_dcache_mft_record_page(ctx->ntfs_ino);
+	mark_mft_record_dirty(ctx->ntfs_ino);
+	ntfs_attr_put_search_ctx(ctx);
+	unmap_mft_record(base_ni);
+	up_write(&ni->runlist.lock);
+	if (page) {
+		set_page_dirty(page);
+		unlock_page(page);
+		mark_page_accessed(page);
+		page_cache_release(page);
+	}
+	ntfs_debug("Done.");
+	return 0;
+undo_err_out:
+	/* Convert the attribute back into a resident attribute. */
+	a->non_resident = 0;
+	/* Move the attribute name if it exists and update the offset. */
+	name_ofs = (offsetof(ATTR_RECORD, data.resident.reserved) +
+			sizeof(a->data.resident.reserved) + 7) & ~7;
+	if (a->name_length)
+		memmove((u8*)a + name_ofs, (u8*)a + le16_to_cpu(a->name_offset),
+				a->name_length * sizeof(ntfschar));
+	mp_ofs = (name_ofs + a->name_length * sizeof(ntfschar) + 7) & ~7;
+	a->name_offset = cpu_to_le16(name_ofs);
+	arec_size = (mp_ofs + attr_size + 7) & ~7;
+	/* Resize the resident part of the attribute record. */
+	err2 = ntfs_attr_record_resize(m, a, arec_size);
+	if (unlikely(err2)) {
+		/*
+		 * This cannot happen (well if memory corruption is at work it
+		 * could happen in theory), but deal with it as well as we can.
+		 * If the old size is too small, truncate the attribute,
+		 * otherwise simply give it a larger allocated size.
+		 * FIXME: Should check whether chkdsk complains when the
+		 * allocated size is much bigger than the resident value size.
+		 */
+		arec_size = le32_to_cpu(a->length);
+		if ((mp_ofs + attr_size) > arec_size) {
+			err2 = attr_size;
+			attr_size = arec_size - mp_ofs;
+			ntfs_error(vol->sb, "Failed to undo partial resident "
+					"to non-resident attribute "
+					"conversion.  Truncating inode 0x%lx, "
+					"attribute type 0x%x from %i bytes to "
+					"%i bytes to maintain metadata "
+					"consistency.  THIS MEANS YOU ARE "
+					"LOSING %i BYTES DATA FROM THIS %s.",
+					vi->i_ino,
+					(unsigned)le32_to_cpu(ni->type),
+					err2, attr_size, err2 - attr_size,
+					((ni->type == AT_DATA) &&
+					!ni->name_len) ? "FILE": "ATTRIBUTE");
+			write_lock_irqsave(&ni->size_lock, flags);
+			ni->initialized_size = attr_size;
+			i_size_write(vi, attr_size);
+			write_unlock_irqrestore(&ni->size_lock, flags);
+		}
+	}
+	/* Setup the fields specific to resident attributes. */
+	a->data.resident.value_length = cpu_to_le32(attr_size);
+	a->data.resident.value_offset = cpu_to_le16(mp_ofs);
+	a->data.resident.flags = old_res_attr_flags;
+	memset(&a->data.resident.reserved, 0,
+			sizeof(a->data.resident.reserved));
+	/* Copy the data from the page back to the attribute value. */
+	if (page) {
+		kaddr = kmap_atomic(page, KM_USER0);
+		memcpy((u8*)a + mp_ofs, kaddr, attr_size);
+		kunmap_atomic(kaddr, KM_USER0);
+	}
+	/* Setup the allocated size in the ntfs inode in case it changed. */
+	write_lock_irqsave(&ni->size_lock, flags);
+	ni->allocated_size = arec_size - mp_ofs;
+	write_unlock_irqrestore(&ni->size_lock, flags);
+	/* Mark the mft record dirty, so it gets written back. */
+	flush_dcache_mft_record_page(ctx->ntfs_ino);
+	mark_mft_record_dirty(ctx->ntfs_ino);
+err_out:
+	if (ctx)
+		ntfs_attr_put_search_ctx(ctx);
+	if (m)
+		unmap_mft_record(base_ni);
+	ni->runlist.rl = NULL;
+	up_write(&ni->runlist.lock);
+rl_err_out:
+	if (rl) {
+		if (ntfs_cluster_free_from_rl(vol, rl) < 0) {
+			ntfs_error(vol->sb, "Failed to release allocated "
+					"cluster(s) in error code path.  Run "
+					"chkdsk to recover the lost "
+					"cluster(s).");
+			NVolSetErrors(vol);
+		}
+		ntfs_free(rl);
+page_err_out:
+		unlock_page(page);
+		page_cache_release(page);
+	}
+	if (err == -EINVAL)
+		err = -EIO;
+	return err;
+}
+
+/**
  * ntfs_attr_set - fill (a part of) an attribute with a byte
  * @ni:		ntfs inode describing the attribute to fill
  * @ofs:	offset inside the attribute at which to start to fill
@@ -1127,6 +1579,10 @@ int ntfs_attr_record_resize(MFT_RECORD *m, ATTR_RECORD *a, u32 new_size)
  * byte offset @ofs inside the attribute with the constant byte @val.
  *
  * This function is effectively like memset() applied to an ntfs attribute.
+ * Note thie function actually only operates on the page cache pages belonging
+ * to the ntfs attribute and it marks them dirty after doing the memset().
+ * Thus it relies on the vm dirty page write code paths to cause the modified
+ * pages to be written to the mft record/disk.
  *
  * Return 0 on success and -errno on error.  An error code of -ESPIPE means
  * that @ofs + @cnt were outside the end of the attribute and no write was
@@ -1155,7 +1611,7 @@ int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt, const u8 val)
 	end = ofs + cnt;
 	end_ofs = end & ~PAGE_CACHE_MASK;
 	/* If the end is outside the inode size return -ESPIPE. */
-	if (unlikely(end > VFS_I(ni)->i_size)) {
+	if (unlikely(end > i_size_read(VFS_I(ni)))) {
 		ntfs_error(vol->sb, "Request exceeds end of attribute.");
 		return -ESPIPE;
 	}
@@ -1256,3 +1712,5 @@ done:
 	ntfs_debug("Done.");
 	return 0;
 }
+
+#endif /* NTFS_RW */
diff --git a/fs/ntfs/attrib.h b/fs/ntfs/attrib.h
index e0c2c6c..0e4ac6d 100644
--- a/fs/ntfs/attrib.h
+++ b/fs/ntfs/attrib.h
@@ -2,7 +2,7 @@
  * attrib.h - Defines for attribute handling in NTFS Linux kernel driver.
  *	      Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -60,10 +60,14 @@ typedef struct {
 	ATTR_RECORD *base_attr;
 } ntfs_attr_search_ctx;
 
+extern int ntfs_map_runlist_nolock(ntfs_inode *ni, VCN vcn);
 extern int ntfs_map_runlist(ntfs_inode *ni, VCN vcn);
 
-extern runlist_element *ntfs_find_vcn(ntfs_inode *ni, const VCN vcn,
-		const BOOL need_write);
+extern LCN ntfs_attr_vcn_to_lcn_nolock(ntfs_inode *ni, const VCN vcn,
+		const BOOL write_locked);
+
+extern runlist_element *ntfs_attr_find_vcn_nolock(ntfs_inode *ni,
+		const VCN vcn, const BOOL write_locked);
 
 int ntfs_attr_lookup(const ATTR_TYPE type, const ntfschar *name,
 		const u32 name_len, const IGNORE_CASE_BOOL ic,
@@ -85,6 +89,8 @@ extern ntfs_attr_search_ctx *ntfs_attr_get_search_ctx(ntfs_inode *ni,
 		MFT_RECORD *mrec);
 extern void ntfs_attr_put_search_ctx(ntfs_attr_search_ctx *ctx);
 
+#ifdef NTFS_RW
+
 extern int ntfs_attr_size_bounds_check(const ntfs_volume *vol,
 		const ATTR_TYPE type, const s64 size);
 extern int ntfs_attr_can_be_non_resident(const ntfs_volume *vol,
@@ -94,7 +100,11 @@ extern int ntfs_attr_can_be_resident(const ntfs_volume *vol,
 
 extern int ntfs_attr_record_resize(MFT_RECORD *m, ATTR_RECORD *a, u32 new_size);
 
+extern int ntfs_attr_make_non_resident(ntfs_inode *ni);
+
 extern int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt,
 		const u8 val);
 
+#endif /* NTFS_RW */
+
 #endif /* _LINUX_NTFS_ATTRIB_H */
diff --git a/fs/ntfs/compress.c b/fs/ntfs/compress.c
index ee5ae70..6d265cf 100644
--- a/fs/ntfs/compress.c
+++ b/fs/ntfs/compress.c
@@ -96,13 +96,14 @@ void free_compression_buffers(void)
 /**
  * zero_partial_compressed_page - zero out of bounds compressed page region
  */
-static void zero_partial_compressed_page(ntfs_inode *ni, struct page *page)
+static void zero_partial_compressed_page(struct page *page,
+		const s64 initialized_size)
 {
 	u8 *kp = page_address(page);
 	unsigned int kp_ofs;
 
 	ntfs_debug("Zeroing page region outside initialized size.");
-	if (((s64)page->index << PAGE_CACHE_SHIFT) >= ni->initialized_size) {
+	if (((s64)page->index << PAGE_CACHE_SHIFT) >= initialized_size) {
 		/*
 		 * FIXME: Using clear_page() will become wrong when we get
 		 * PAGE_CACHE_SIZE != PAGE_SIZE but for now there is no problem.
@@ -110,7 +111,7 @@ static void zero_partial_compressed_page(ntfs_inode *ni, struct page *page)
 		clear_page(kp);
 		return;
 	}
-	kp_ofs = ni->initialized_size & ~PAGE_CACHE_MASK;
+	kp_ofs = initialized_size & ~PAGE_CACHE_MASK;
 	memset(kp + kp_ofs, 0, PAGE_CACHE_SIZE - kp_ofs);
 	return;
 }
@@ -118,12 +119,12 @@ static void zero_partial_compressed_page(ntfs_inode *ni, struct page *page)
 /**
  * handle_bounds_compressed_page - test for&handle out of bounds compressed page
  */
-static inline void handle_bounds_compressed_page(ntfs_inode *ni,
-		struct page *page)
+static inline void handle_bounds_compressed_page(struct page *page,
+		const loff_t i_size, const s64 initialized_size)
 {
-	if ((page->index >= (ni->initialized_size >> PAGE_CACHE_SHIFT)) &&
-			(ni->initialized_size < VFS_I(ni)->i_size))
-		zero_partial_compressed_page(ni, page);
+	if ((page->index >= (initialized_size >> PAGE_CACHE_SHIFT)) &&
+			(initialized_size < i_size))
+		zero_partial_compressed_page(page, initialized_size);
 	return;
 }
 
@@ -138,6 +139,8 @@ static inline void handle_bounds_compressed_page(ntfs_inode *ni,
  * @xpage_done:		set to 1 if xpage was completed successfully (IN/OUT)
  * @cb_start:		compression block to decompress (IN)
  * @cb_size:		size of compression block @cb_start in bytes (IN)
+ * @i_size:		file size when we started the read (IN)
+ * @initialized_size:	initialized file size when we started the read (IN)
  *
  * The caller must have disabled preemption. ntfs_decompress() reenables it when
  * the critical section is finished.
@@ -165,7 +168,8 @@ static inline void handle_bounds_compressed_page(ntfs_inode *ni,
 static int ntfs_decompress(struct page *dest_pages[], int *dest_index,
 		int *dest_ofs, const int dest_max_index, const int dest_max_ofs,
 		const int xpage, char *xpage_done, u8 *const cb_start,
-		const u32 cb_size)
+		const u32 cb_size, const loff_t i_size,
+		const s64 initialized_size)
 {
 	/*
 	 * Pointers into the compressed data, i.e. the compression block (cb),
@@ -219,9 +223,6 @@ return_error:
 		spin_unlock(&ntfs_cb_lock);
 		/* Second stage: finalize completed pages. */
 		if (nr_completed_pages > 0) {
-			struct page *page = dest_pages[completed_pages[0]];
-			ntfs_inode *ni = NTFS_I(page->mapping->host);
-
 			for (i = 0; i < nr_completed_pages; i++) {
 				int di = completed_pages[i];
 
@@ -230,7 +231,8 @@ return_error:
 				 * If we are outside the initialized size, zero
 				 * the out of bounds page range.
 				 */
-				handle_bounds_compressed_page(ni, dp);
+				handle_bounds_compressed_page(dp, i_size,
+						initialized_size);
 				flush_dcache_page(dp);
 				kunmap(dp);
 				SetPageUptodate(dp);
@@ -478,12 +480,14 @@ return_overflow:
  */
 int ntfs_read_compressed_block(struct page *page)
 {
+	loff_t i_size;
+	s64 initialized_size;
 	struct address_space *mapping = page->mapping;
 	ntfs_inode *ni = NTFS_I(mapping->host);
 	ntfs_volume *vol = ni->vol;
 	struct super_block *sb = vol->sb;
 	runlist_element *rl;
-	unsigned long block_size = sb->s_blocksize;
+	unsigned long flags, block_size = sb->s_blocksize;
 	unsigned char block_size_bits = sb->s_blocksize_bits;
 	u8 *cb, *cb_pos, *cb_end;
 	struct buffer_head **bhs;
@@ -552,8 +556,12 @@ int ntfs_read_compressed_block(struct page *page)
 	 * The remaining pages need to be allocated and inserted into the page
 	 * cache, alignment guarantees keep all the below much simpler. (-8
 	 */
-	max_page = ((VFS_I(ni)->i_size + PAGE_CACHE_SIZE - 1) >>
-			PAGE_CACHE_SHIFT) - offset;
+	read_lock_irqsave(&ni->size_lock, flags);
+	i_size = i_size_read(VFS_I(ni));
+	initialized_size = ni->initialized_size;
+	read_unlock_irqrestore(&ni->size_lock, flags);
+	max_page = ((i_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
+			offset;
 	if (nr_pages < max_page)
 		max_page = nr_pages;
 	for (i = 0; i < max_page; i++, offset++) {
@@ -824,7 +832,8 @@ lock_retry_remap:
 				 * If we are outside the initialized size, zero
 				 * the out of bounds page range.
 				 */
-				handle_bounds_compressed_page(ni, page);
+				handle_bounds_compressed_page(page, i_size,
+						initialized_size);
 				flush_dcache_page(page);
 				kunmap(page);
 				SetPageUptodate(page);
@@ -847,7 +856,8 @@ lock_retry_remap:
 		ntfs_debug("Found compressed compression block.");
 		err = ntfs_decompress(pages, &cur_page, &cur_ofs,
 				cb_max_page, cb_max_ofs, xpage, &xpage_done,
-				cb_pos,	cb_size - (cb_pos - cb));
+				cb_pos,	cb_size - (cb_pos - cb), i_size,
+				initialized_size);
 		/*
 		 * We can sleep from now on, lock already dropped by
 		 * ntfs_decompress().
diff --git a/fs/ntfs/debug.c b/fs/ntfs/debug.c
index 6fb6bb5..807150e 100644
--- a/fs/ntfs/debug.c
+++ b/fs/ntfs/debug.c
@@ -164,14 +164,17 @@ void ntfs_debug_dump_runlist(const runlist_element *rl)
 			if (index > -LCN_ENOENT - 1)
 				index = 3;
 			printk(KERN_DEBUG "%-16Lx %s %-16Lx%s\n",
-					(rl + i)->vcn, lcn_str[index],
-					(rl + i)->length, (rl + i)->length ?
-					"" : " (runlist end)");
+					(long long)(rl + i)->vcn, lcn_str[index],
+					(long long)(rl + i)->length,
+					(rl + i)->length ? "" :
+						" (runlist end)");
 		} else
 			printk(KERN_DEBUG "%-16Lx %-16Lx  %-16Lx%s\n",
-					(rl + i)->vcn, (rl + i)->lcn,
-					(rl + i)->length, (rl + i)->length ?
-					"" : " (runlist end)");
+					(long long)(rl + i)->vcn,
+					(long long)(rl + i)->lcn,
+					(long long)(rl + i)->length,
+					(rl + i)->length ? "" :
+						" (runlist end)");
 		if (!(rl + i)->length)
 			break;
 	}
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
index 9357756..4677947 100644
--- a/fs/ntfs/dir.c
+++ b/fs/ntfs/dir.c
@@ -1,7 +1,7 @@
 /**
  * dir.c - NTFS kernel directory operations. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -183,8 +183,7 @@ found_it:
 				name->len = 0;
 				*res = name;
 			} else {
-				if (name)
-					kfree(name);
+				kfree(name);
 				*res = NULL;
 			}
 			mref = le64_to_cpu(ie->data.dir.indexed_file);
@@ -444,8 +443,7 @@ found_it2:
 				name->len = 0;
 				*res = name;
 			} else {
-				if (name)
-					kfree(name);
+				kfree(name);
 				*res = NULL;
 			}
 			mref = le64_to_cpu(ie->data.dir.indexed_file);
@@ -610,7 +608,7 @@ dir_err_out:
 // TODO: (AIA)
 // The algorithm embedded in this code will be required for the time when we
 // want to support adding of entries to directories, where we require correct
-// collation of file names in order not to cause corruption of the file system.
+// collation of file names in order not to cause corruption of the filesystem.
 
 /**
  * ntfs_lookup_inode_by_name - find an inode in a directory given its name
@@ -1101,7 +1099,7 @@ static inline int ntfs_filldir(ntfs_volume *vol, loff_t fpos,
 static int ntfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 {
 	s64 ia_pos, ia_start, prev_ia_pos, bmp_pos;
-	loff_t fpos;
+	loff_t fpos, i_size;
 	struct inode *bmp_vi, *vdir = filp->f_dentry->d_inode;
 	struct super_block *sb = vdir->i_sb;
 	ntfs_inode *ndir = NTFS_I(vdir);
@@ -1122,7 +1120,8 @@ static int ntfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 			vdir->i_ino, fpos);
 	rc = err = 0;
 	/* Are we at end of dir yet? */
-	if (fpos >= vdir->i_size + vol->mft_record_size)
+	i_size = i_size_read(vdir);
+	if (fpos >= i_size + vol->mft_record_size)
 		goto done;
 	/* Emulate . and .. for all directories. */
 	if (!fpos) {
@@ -1264,7 +1263,7 @@ skip_index_root:
 	bmp_mapping = bmp_vi->i_mapping;
 	/* Get the starting bitmap bit position and sanity check it. */
 	bmp_pos = ia_pos >> ndir->itype.index.block_size_bits;
-	if (unlikely(bmp_pos >> 3 >= bmp_vi->i_size)) {
+	if (unlikely(bmp_pos >> 3 >= i_size_read(bmp_vi))) {
 		ntfs_error(sb, "Current index allocation position exceeds "
 				"index bitmap size.");
 		goto err_out;
@@ -1301,7 +1300,7 @@ find_next_index_buffer:
 			goto get_next_bmp_page;
 		}
 		/* If we have reached the end of the bitmap, we are done. */
-		if (unlikely(((bmp_pos + cur_bmp_pos) >> 3) >= vdir->i_size))
+		if (unlikely(((bmp_pos + cur_bmp_pos) >> 3) >= i_size))
 			goto unm_EOD;
 		ia_pos = (bmp_pos + cur_bmp_pos) <<
 				ndir->itype.index.block_size_bits;
@@ -1309,7 +1308,8 @@ find_next_index_buffer:
 	ntfs_debug("Handling index buffer 0x%llx.",
 			(unsigned long long)bmp_pos + cur_bmp_pos);
 	/* If the current index buffer is in the same page we reuse the page. */
-	if ((prev_ia_pos & PAGE_CACHE_MASK) != (ia_pos & PAGE_CACHE_MASK)) {
+	if ((prev_ia_pos & (s64)PAGE_CACHE_MASK) !=
+			(ia_pos & (s64)PAGE_CACHE_MASK)) {
 		prev_ia_pos = ia_pos;
 		if (likely(ia_page != NULL)) {
 			unlock_page(ia_page);
@@ -1441,7 +1441,7 @@ unm_EOD:
 	ntfs_unmap_page(bmp_page);
 EOD:
 	/* We are finished, set fpos to EOD. */
-	fpos = vdir->i_size + vol->mft_record_size;
+	fpos = i_size + vol->mft_record_size;
 abort:
 	kfree(name);
 done:
@@ -1461,10 +1461,8 @@ err_out:
 		unlock_page(ia_page);
 		ntfs_unmap_page(ia_page);
 	}
-	if (ir)
-		kfree(ir);
-	if (name)
-		kfree(name);
+	kfree(ir);
+	kfree(name);
 	if (ctx)
 		ntfs_attr_put_search_ctx(ctx);
 	if (m)
@@ -1495,7 +1493,7 @@ err_out:
 static int ntfs_dir_open(struct inode *vi, struct file *filp)
 {
 	if (sizeof(unsigned long) < 8) {
-		if (vi->i_size > MAX_LFS_FILESIZE)
+		if (i_size_read(vi) > MAX_LFS_FILESIZE)
 			return -EFBIG;
 	}
 	return 0;
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index db8713e..e0f530c 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -47,7 +47,7 @@
 static int ntfs_file_open(struct inode *vi, struct file *filp)
 {
 	if (sizeof(unsigned long) < 8) {
-		if (vi->i_size > MAX_LFS_FILESIZE)
+		if (i_size_read(vi) > MAX_LFS_FILESIZE)
 			return -EFBIG;
 	}
 	return generic_file_open(vi, filp);
diff --git a/fs/ntfs/index.c b/fs/ntfs/index.c
index 71bd2cd..11fd530 100644
--- a/fs/ntfs/index.c
+++ b/fs/ntfs/index.c
@@ -1,7 +1,7 @@
 /*
  * index.c - NTFS kernel index handling.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2004 Anton Altaparmakov
+ * Copyright (c) 2004-2005 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
@@ -39,18 +39,8 @@ ntfs_index_context *ntfs_index_ctx_get(ntfs_inode *idx_ni)
 	ntfs_index_context *ictx;
 
 	ictx = kmem_cache_alloc(ntfs_index_ctx_cache, SLAB_NOFS);
-	if (ictx) {
-		ictx->idx_ni = idx_ni;
-		ictx->entry = NULL;
-		ictx->data = NULL;
-		ictx->data_len = 0;
-		ictx->is_in_root = 0;
-		ictx->ir = NULL;
-		ictx->actx = NULL;
-		ictx->base_ni = NULL;
-		ictx->ia = NULL;
-		ictx->page = NULL;
-	}
+	if (ictx)
+		*ictx = (ntfs_index_context){ .idx_ni = idx_ni };
 	return ictx;
 }
 
diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c
index 31840ba..886214a 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) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 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
@@ -174,7 +174,7 @@ struct inode *ntfs_iget(struct super_block *sb, unsigned long mft_no)
 
 	vi = iget5_locked(sb, mft_no, (test_t)ntfs_test_inode,
 			(set_t)ntfs_init_locked_inode, &na);
-	if (!vi)
+	if (unlikely(!vi))
 		return ERR_PTR(-ENOMEM);
 
 	err = 0;
@@ -188,7 +188,7 @@ struct inode *ntfs_iget(struct super_block *sb, unsigned long mft_no)
 	 * There is no point in keeping bad inodes around if the failure was
 	 * due to ENOMEM. We want to be able to retry again later.
 	 */
-	if (err == -ENOMEM) {
+	if (unlikely(err == -ENOMEM)) {
 		iput(vi);
 		vi = ERR_PTR(err);
 	}
@@ -235,7 +235,7 @@ struct inode *ntfs_attr_iget(struct inode *base_vi, ATTR_TYPE type,
 
 	vi = iget5_locked(base_vi->i_sb, na.mft_no, (test_t)ntfs_test_inode,
 			(set_t)ntfs_init_locked_inode, &na);
-	if (!vi)
+	if (unlikely(!vi))
 		return ERR_PTR(-ENOMEM);
 
 	err = 0;
@@ -250,7 +250,7 @@ struct inode *ntfs_attr_iget(struct inode *base_vi, ATTR_TYPE type,
 	 * simplifies things in that we never need to check for bad attribute
 	 * inodes elsewhere.
 	 */
-	if (err) {
+	if (unlikely(err)) {
 		iput(vi);
 		vi = ERR_PTR(err);
 	}
@@ -290,7 +290,7 @@ struct inode *ntfs_index_iget(struct inode *base_vi, ntfschar *name,
 
 	vi = iget5_locked(base_vi->i_sb, na.mft_no, (test_t)ntfs_test_inode,
 			(set_t)ntfs_init_locked_inode, &na);
-	if (!vi)
+	if (unlikely(!vi))
 		return ERR_PTR(-ENOMEM);
 
 	err = 0;
@@ -305,7 +305,7 @@ struct inode *ntfs_index_iget(struct inode *base_vi, ntfschar *name,
 	 * simplifies things in that we never need to check for bad index
 	 * inodes elsewhere.
 	 */
-	if (err) {
+	if (unlikely(err)) {
 		iput(vi);
 		vi = ERR_PTR(err);
 	}
@@ -317,8 +317,7 @@ struct inode *ntfs_alloc_big_inode(struct super_block *sb)
 	ntfs_inode *ni;
 
 	ntfs_debug("Entering.");
-	ni = (ntfs_inode *)kmem_cache_alloc(ntfs_big_inode_cache,
-			SLAB_NOFS);
+	ni = kmem_cache_alloc(ntfs_big_inode_cache, SLAB_NOFS);
 	if (likely(ni != NULL)) {
 		ni->state = 0;
 		return VFS_I(ni);
@@ -343,7 +342,7 @@ static inline ntfs_inode *ntfs_alloc_extent_inode(void)
 	ntfs_inode *ni;
 
 	ntfs_debug("Entering.");
-	ni = (ntfs_inode *)kmem_cache_alloc(ntfs_inode_cache, SLAB_NOFS);
+	ni = kmem_cache_alloc(ntfs_inode_cache, SLAB_NOFS);
 	if (likely(ni != NULL)) {
 		ni->state = 0;
 		return ni;
@@ -376,6 +375,7 @@ static void ntfs_destroy_extent_inode(ntfs_inode *ni)
 void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
 {
 	ntfs_debug("Entering.");
+	rwlock_init(&ni->size_lock);
 	ni->initialized_size = ni->allocated_size = 0;
 	ni->seq_no = 0;
 	atomic_set(&ni->count, 1);
@@ -524,6 +524,7 @@ static int ntfs_read_locked_inode(struct inode *vi)
 	ntfs_volume *vol = NTFS_SB(vi->i_sb);
 	ntfs_inode *ni;
 	MFT_RECORD *m;
+	ATTR_RECORD *a;
 	STANDARD_INFORMATION *si;
 	ntfs_attr_search_ctx *ctx;
 	int err = 0;
@@ -632,9 +633,10 @@ static int ntfs_read_locked_inode(struct inode *vi)
 		}
 		goto unm_err_out;
 	}
+	a = ctx->attr;
 	/* Get the standard information attribute value. */
-	si = (STANDARD_INFORMATION*)((char*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
+	si = (STANDARD_INFORMATION*)((u8*)a +
+			le16_to_cpu(a->data.resident.value_offset));
 
 	/* Transfer information from the standard information into vi. */
 	/*
@@ -673,15 +675,16 @@ static int ntfs_read_locked_inode(struct inode *vi)
 			goto skip_attr_list_load;
 		ntfs_debug("Attribute list found in inode 0x%lx.", vi->i_ino);
 		NInoSetAttrList(ni);
-		if (ctx->attr->flags & ATTR_IS_ENCRYPTED ||
-				ctx->attr->flags & ATTR_COMPRESSION_MASK ||
-				ctx->attr->flags & ATTR_IS_SPARSE) {
+		a = ctx->attr;
+		if (a->flags & ATTR_IS_ENCRYPTED ||
+				a->flags & ATTR_COMPRESSION_MASK ||
+				a->flags & ATTR_IS_SPARSE) {
 			ntfs_error(vi->i_sb, "Attribute list attribute is "
 					"compressed/encrypted/sparse.");
 			goto unm_err_out;
 		}
 		/* Now allocate memory for the attribute list. */
-		ni->attr_list_size = (u32)ntfs_attr_size(ctx->attr);
+		ni->attr_list_size = (u32)ntfs_attr_size(a);
 		ni->attr_list = ntfs_malloc_nofs(ni->attr_list_size);
 		if (!ni->attr_list) {
 			ntfs_error(vi->i_sb, "Not enough memory to allocate "
@@ -689,9 +692,9 @@ static int ntfs_read_locked_inode(struct inode *vi)
 			err = -ENOMEM;
 			goto unm_err_out;
 		}
-		if (ctx->attr->non_resident) {
+		if (a->non_resident) {
 			NInoSetAttrListNonResident(ni);
-			if (ctx->attr->data.non_resident.lowest_vcn) {
+			if (a->data.non_resident.lowest_vcn) {
 				ntfs_error(vi->i_sb, "Attribute list has non "
 						"zero lowest_vcn.");
 				goto unm_err_out;
@@ -701,7 +704,7 @@ static int ntfs_read_locked_inode(struct inode *vi)
 			 * exclusive access to the inode at this time.
 			 */
 			ni->attr_list_rl.rl = ntfs_mapping_pairs_decompress(vol,
-					ctx->attr, NULL);
+					a, NULL);
 			if (IS_ERR(ni->attr_list_rl.rl)) {
 				err = PTR_ERR(ni->attr_list_rl.rl);
 				ni->attr_list_rl.rl = NULL;
@@ -712,27 +715,26 @@ static int ntfs_read_locked_inode(struct inode *vi)
 			/* Now load the attribute list. */
 			if ((err = load_attribute_list(vol, &ni->attr_list_rl,
 					ni->attr_list, ni->attr_list_size,
-					sle64_to_cpu(ctx->attr->data.
-					non_resident.initialized_size)))) {
+					sle64_to_cpu(a->data.non_resident.
+					initialized_size)))) {
 				ntfs_error(vi->i_sb, "Failed to load "
 						"attribute list attribute.");
 				goto unm_err_out;
 			}
-		} else /* if (!ctx.attr->non_resident) */ {
-			if ((u8*)ctx->attr + le16_to_cpu(
-					ctx->attr->data.resident.value_offset) +
-					le32_to_cpu(
-					ctx->attr->data.resident.value_length) >
+		} else /* if (!a->non_resident) */ {
+			if ((u8*)a + le16_to_cpu(a->data.resident.value_offset)
+					+ le32_to_cpu(
+					a->data.resident.value_length) >
 					(u8*)ctx->mrec + vol->mft_record_size) {
 				ntfs_error(vi->i_sb, "Corrupt attribute list "
 						"in inode.");
 				goto unm_err_out;
 			}
 			/* Now copy the attribute list. */
-			memcpy(ni->attr_list, (u8*)ctx->attr + le16_to_cpu(
-					ctx->attr->data.resident.value_offset),
+			memcpy(ni->attr_list, (u8*)a + le16_to_cpu(
+					a->data.resident.value_offset),
 					le32_to_cpu(
-					ctx->attr->data.resident.value_length));
+					a->data.resident.value_length));
 		}
 	}
 skip_attr_list_load:
@@ -741,10 +743,11 @@ skip_attr_list_load:
 	 * in ntfs_ino->attr_list and it is ntfs_ino->attr_list_size bytes.
 	 */
 	if (S_ISDIR(vi->i_mode)) {
+		loff_t bvi_size;
 		struct inode *bvi;
 		ntfs_inode *bni;
 		INDEX_ROOT *ir;
-		char *ir_end, *index_end;
+		u8 *ir_end, *index_end;
 
 		/* It is a directory, find index root attribute. */
 		ntfs_attr_reinit_search_ctx(ctx);
@@ -760,17 +763,16 @@ skip_attr_list_load:
 			}
 			goto unm_err_out;
 		}
+		a = ctx->attr;
 		/* Set up the state. */
-		if (unlikely(ctx->attr->non_resident)) {
+		if (unlikely(a->non_resident)) {
 			ntfs_error(vol->sb, "$INDEX_ROOT attribute is not "
 					"resident.");
 			goto unm_err_out;
 		}
 		/* Ensure the attribute name is placed before the value. */
-		if (unlikely(ctx->attr->name_length &&
-				(le16_to_cpu(ctx->attr->name_offset) >=
-				le16_to_cpu(ctx->attr->data.resident.
-				value_offset)))) {
+		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+				le16_to_cpu(a->data.resident.value_offset)))) {
 			ntfs_error(vol->sb, "$INDEX_ROOT attribute name is "
 					"placed after the attribute value.");
 			goto unm_err_out;
@@ -781,28 +783,27 @@ skip_attr_list_load:
 		 * encrypted. However index root cannot be both compressed and
 		 * encrypted.
 		 */
-		if (ctx->attr->flags & ATTR_COMPRESSION_MASK)
+		if (a->flags & ATTR_COMPRESSION_MASK)
 			NInoSetCompressed(ni);
-		if (ctx->attr->flags & ATTR_IS_ENCRYPTED) {
-			if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
+		if (a->flags & ATTR_IS_ENCRYPTED) {
+			if (a->flags & ATTR_COMPRESSION_MASK) {
 				ntfs_error(vi->i_sb, "Found encrypted and "
 						"compressed attribute.");
 				goto unm_err_out;
 			}
 			NInoSetEncrypted(ni);
 		}
-		if (ctx->attr->flags & ATTR_IS_SPARSE)
+		if (a->flags & ATTR_IS_SPARSE)
 			NInoSetSparse(ni);
-		ir = (INDEX_ROOT*)((char*)ctx->attr + le16_to_cpu(
-				ctx->attr->data.resident.value_offset));
-		ir_end = (char*)ir + le32_to_cpu(
-				ctx->attr->data.resident.value_length);
-		if (ir_end > (char*)ctx->mrec + vol->mft_record_size) {
+		ir = (INDEX_ROOT*)((u8*)a +
+				le16_to_cpu(a->data.resident.value_offset));
+		ir_end = (u8*)ir + le32_to_cpu(a->data.resident.value_length);
+		if (ir_end > (u8*)ctx->mrec + vol->mft_record_size) {
 			ntfs_error(vi->i_sb, "$INDEX_ROOT attribute is "
 					"corrupt.");
 			goto unm_err_out;
 		}
-		index_end = (char*)&ir->index +
+		index_end = (u8*)&ir->index +
 				le32_to_cpu(ir->index.index_length);
 		if (index_end > ir_end) {
 			ntfs_error(vi->i_sb, "Directory index is corrupt.");
@@ -889,7 +890,8 @@ skip_attr_list_load:
 						"attribute.");
 			goto unm_err_out;
 		}
-		if (!ctx->attr->non_resident) {
+		a = ctx->attr;
+		if (!a->non_resident) {
 			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
 					"is resident.");
 			goto unm_err_out;
@@ -898,42 +900,40 @@ skip_attr_list_load:
 		 * Ensure the attribute name is placed before the mapping pairs
 		 * array.
 		 */
-		if (unlikely(ctx->attr->name_length &&
-				(le16_to_cpu(ctx->attr->name_offset) >=
-				le16_to_cpu(ctx->attr->data.non_resident.
-				mapping_pairs_offset)))) {
+		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+				le16_to_cpu(
+				a->data.non_resident.mapping_pairs_offset)))) {
 			ntfs_error(vol->sb, "$INDEX_ALLOCATION attribute name "
 					"is placed after the mapping pairs "
 					"array.");
 			goto unm_err_out;
 		}
-		if (ctx->attr->flags & ATTR_IS_ENCRYPTED) {
+		if (a->flags & ATTR_IS_ENCRYPTED) {
 			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
 					"is encrypted.");
 			goto unm_err_out;
 		}
-		if (ctx->attr->flags & ATTR_IS_SPARSE) {
+		if (a->flags & ATTR_IS_SPARSE) {
 			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
 					"is sparse.");
 			goto unm_err_out;
 		}
-		if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
+		if (a->flags & ATTR_COMPRESSION_MASK) {
 			ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute "
 					"is compressed.");
 			goto unm_err_out;
 		}
-		if (ctx->attr->data.non_resident.lowest_vcn) {
+		if (a->data.non_resident.lowest_vcn) {
 			ntfs_error(vi->i_sb, "First extent of "
 					"$INDEX_ALLOCATION attribute has non "
 					"zero lowest_vcn.");
 			goto unm_err_out;
 		}
-		vi->i_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.data_size);
+		vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
 		ni->initialized_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.initialized_size);
+				a->data.non_resident.initialized_size);
 		ni->allocated_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.allocated_size);
+				a->data.non_resident.allocated_size);
 		/*
 		 * We are done with the mft record, so we release it. Otherwise
 		 * we would deadlock in ntfs_attr_iget().
@@ -958,11 +958,12 @@ skip_attr_list_load:
 			goto unm_err_out;
 		}
 		/* Consistency check bitmap size vs. index allocation size. */
-		if ((bvi->i_size << 3) < (vi->i_size >>
+		bvi_size = i_size_read(bvi);
+		if ((bvi_size << 3) < (vi->i_size >>
 				ni->itype.index.block_size_bits)) {
 			ntfs_error(vi->i_sb, "Index bitmap too small (0x%llx) "
 					"for index allocation (0x%llx).",
-					bvi->i_size << 3, vi->i_size);
+					bvi_size << 3, vi->i_size);
 			goto unm_err_out;
 		}
 skip_large_dir_stuff:
@@ -1010,87 +1011,92 @@ skip_large_dir_stuff:
 			ntfs_error(vi->i_sb, "$DATA attribute is missing.");
 			goto unm_err_out;
 		}
+		a = ctx->attr;
 		/* Setup the state. */
-		if (ctx->attr->non_resident) {
+		if (a->non_resident) {
 			NInoSetNonResident(ni);
-			if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
-				NInoSetCompressed(ni);
-				if (vol->cluster_size > 4096) {
-					ntfs_error(vi->i_sb, "Found "
-						"compressed data but "
-						"compression is disabled due "
-						"to cluster size (%i) > 4kiB.",
-						vol->cluster_size);
-					goto unm_err_out;
-				}
-				if ((ctx->attr->flags & ATTR_COMPRESSION_MASK)
-						!= ATTR_IS_COMPRESSED) {
-					ntfs_error(vi->i_sb, "Found "
-						"unknown compression method or "
-						"corrupt file.");
-					goto unm_err_out;
+			if (a->flags & (ATTR_COMPRESSION_MASK |
+					ATTR_IS_SPARSE)) {
+				if (a->flags & ATTR_COMPRESSION_MASK) {
+					NInoSetCompressed(ni);
+					if (vol->cluster_size > 4096) {
+						ntfs_error(vi->i_sb, "Found "
+							"compressed data but "
+							"compression is "
+							"disabled due to "
+							"cluster size (%i) > "
+							"4kiB.",
+							vol->cluster_size);
+						goto unm_err_out;
+					}
+					if ((a->flags & ATTR_COMPRESSION_MASK)
+							!= ATTR_IS_COMPRESSED) {
+						ntfs_error(vi->i_sb, "Found "
+							"unknown compression "
+							"method or corrupt "
+							"file.");
+						goto unm_err_out;
+					}
 				}
-				ni->itype.compressed.block_clusters = 1U <<
-						ctx->attr->data.non_resident.
-						compression_unit;
-				if (ctx->attr->data.non_resident.
-						compression_unit != 4) {
+				if (a->flags & ATTR_IS_SPARSE)
+					NInoSetSparse(ni);
+				if (a->data.non_resident.compression_unit !=
+						4) {
 					ntfs_error(vi->i_sb, "Found "
 						"nonstandard compression unit "
 						"(%u instead of 4).  Cannot "
 						"handle this.",
-						ctx->attr->data.non_resident.
+						a->data.non_resident.
 						compression_unit);
 					err = -EOPNOTSUPP;
 					goto unm_err_out;
 				}
+				ni->itype.compressed.block_clusters = 1U <<
+						a->data.non_resident.
+						compression_unit;
 				ni->itype.compressed.block_size = 1U << (
-						ctx->attr->data.non_resident.
+						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_size) - 1;
+				ni->itype.compressed.size = sle64_to_cpu(
+						a->data.non_resident.
+						compressed_size);
 			}
-			if (ctx->attr->flags & ATTR_IS_ENCRYPTED) {
-				if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
+			if (a->flags & ATTR_IS_ENCRYPTED) {
+				if (a->flags & ATTR_COMPRESSION_MASK) {
 					ntfs_error(vi->i_sb, "Found encrypted "
 							"and compressed data.");
 					goto unm_err_out;
 				}
 				NInoSetEncrypted(ni);
 			}
-			if (ctx->attr->flags & ATTR_IS_SPARSE)
-				NInoSetSparse(ni);
-			if (ctx->attr->data.non_resident.lowest_vcn) {
+			if (a->data.non_resident.lowest_vcn) {
 				ntfs_error(vi->i_sb, "First extent of $DATA "
 						"attribute has non zero "
 						"lowest_vcn.");
 				goto unm_err_out;
 			}
-			/* Setup all the sizes. */
 			vi->i_size = sle64_to_cpu(
-					ctx->attr->data.non_resident.data_size);
+					a->data.non_resident.data_size);
 			ni->initialized_size = sle64_to_cpu(
-					ctx->attr->data.non_resident.
-					initialized_size);
+					a->data.non_resident.initialized_size);
 			ni->allocated_size = sle64_to_cpu(
-					ctx->attr->data.non_resident.
-					allocated_size);
-			if (NInoCompressed(ni)) {
-				ni->itype.compressed.size = sle64_to_cpu(
-						ctx->attr->data.non_resident.
-						compressed_size);
-			}
+					a->data.non_resident.allocated_size);
 		} else { /* Resident attribute. */
-			/*
-			 * Make all sizes equal for simplicity in read code
-			 * paths. FIXME: Need to keep this in mind when
-			 * converting to non-resident attribute in write code
-			 * path. (Probably only affects truncate().)
-			 */
-			vi->i_size = ni->initialized_size = ni->allocated_size =
-					le32_to_cpu(
-					ctx->attr->data.resident.value_length);
+			vi->i_size = ni->initialized_size = le32_to_cpu(
+					a->data.resident.value_length);
+			ni->allocated_size = le32_to_cpu(a->length) -
+					le16_to_cpu(
+					a->data.resident.value_offset);
+			if (vi->i_size > ni->allocated_size) {
+				ntfs_error(vi->i_sb, "Resident data attribute "
+						"is corrupt (size exceeds "
+						"allocation).");
+				goto unm_err_out;
+			}
 		}
 no_data_attr_special_case:
 		/* We are done with the mft record, so we release it. */
@@ -1117,11 +1123,10 @@ no_data_attr_special_case:
 	 * sizes of all non-resident attributes present to give us the Linux
 	 * correct size that should go into i_blocks (after division by 512).
 	 */
-	if (S_ISDIR(vi->i_mode) || !NInoCompressed(ni))
-		vi->i_blocks = ni->allocated_size >> 9;
-	else
+	if (S_ISREG(vi->i_mode) && (NInoCompressed(ni) || NInoSparse(ni)))
 		vi->i_blocks = ni->itype.compressed.size >> 9;
-
+	else
+		vi->i_blocks = ni->allocated_size >> 9;
 	ntfs_debug("Done.");
 	return 0;
 
@@ -1166,6 +1171,7 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 	ntfs_volume *vol = NTFS_SB(vi->i_sb);
 	ntfs_inode *ni, *base_ni;
 	MFT_RECORD *m;
+	ATTR_RECORD *a;
 	ntfs_attr_search_ctx *ctx;
 	int err = 0;
 
@@ -1200,24 +1206,21 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 		err = -ENOMEM;
 		goto unm_err_out;
 	}
-
 	/* Find the attribute. */
 	err = ntfs_attr_lookup(ni->type, ni->name, ni->name_len,
 			CASE_SENSITIVE, 0, NULL, 0, ctx);
 	if (unlikely(err))
 		goto unm_err_out;
-
-	if (!ctx->attr->non_resident) {
+	a = ctx->attr;
+	if (!a->non_resident) {
 		/* Ensure the attribute name is placed before the value. */
-		if (unlikely(ctx->attr->name_length &&
-				(le16_to_cpu(ctx->attr->name_offset) >=
-				le16_to_cpu(ctx->attr->data.resident.
-				value_offset)))) {
+		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+				le16_to_cpu(a->data.resident.value_offset)))) {
 			ntfs_error(vol->sb, "Attribute name is placed after "
 					"the attribute value.");
 			goto unm_err_out;
 		}
-		if (NInoMstProtected(ni) || ctx->attr->flags) {
+		if (NInoMstProtected(ni) || a->flags) {
 			ntfs_error(vi->i_sb, "Found mst protected attribute "
 					"or attribute with non-zero flags but "
 					"the attribute is resident.  Please "
@@ -1225,85 +1228,95 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 					"linux-ntfs-dev@lists.sourceforge.net");
 			goto unm_err_out;
 		}
-		/*
-		 * Resident attribute. Make all sizes equal for simplicity in
-		 * read code paths.
-		 */
-		vi->i_size = ni->initialized_size = ni->allocated_size =
-			le32_to_cpu(ctx->attr->data.resident.value_length);
+		vi->i_size = ni->initialized_size = le32_to_cpu(
+				a->data.resident.value_length);
+		ni->allocated_size = le32_to_cpu(a->length) -
+				le16_to_cpu(a->data.resident.value_offset);
+		if (vi->i_size > ni->allocated_size) {
+			ntfs_error(vi->i_sb, "Resident attribute is corrupt "
+					"(size exceeds allocation).");
+			goto unm_err_out;
+		}
 	} else {
 		NInoSetNonResident(ni);
 		/*
 		 * Ensure the attribute name is placed before the mapping pairs
 		 * array.
 		 */
-		if (unlikely(ctx->attr->name_length &&
-				(le16_to_cpu(ctx->attr->name_offset) >=
-				le16_to_cpu(ctx->attr->data.non_resident.
-				mapping_pairs_offset)))) {
+		if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+				le16_to_cpu(
+				a->data.non_resident.mapping_pairs_offset)))) {
 			ntfs_error(vol->sb, "Attribute name is placed after "
 					"the mapping pairs array.");
 			goto unm_err_out;
 		}
-		if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
+		if (a->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_SPARSE)) {
+			if (a->flags & ATTR_COMPRESSION_MASK) {
+				NInoSetCompressed(ni);
+				if ((ni->type != AT_DATA) || (ni->type ==
+						AT_DATA && ni->name_len)) {
+					ntfs_error(vi->i_sb, "Found compressed "
+							"non-data or named "
+							"data attribute.  "
+							"Please report you "
+							"saw this message to "
+							"linux-ntfs-dev@lists."
+							"sourceforge.net");
+					goto unm_err_out;
+				}
+				if (vol->cluster_size > 4096) {
+					ntfs_error(vi->i_sb, "Found compressed "
+							"attribute but "
+							"compression is "
+							"disabled due to "
+							"cluster size (%i) > "
+							"4kiB.",
+							vol->cluster_size);
+					goto unm_err_out;
+				}
+				if ((a->flags & ATTR_COMPRESSION_MASK) !=
+						ATTR_IS_COMPRESSED) {
+					ntfs_error(vi->i_sb, "Found unknown "
+							"compression method.");
+					goto unm_err_out;
+				}
+			}
 			if (NInoMstProtected(ni)) {
 				ntfs_error(vi->i_sb, "Found mst protected "
 						"attribute but the attribute "
-						"is compressed.  Please report "
-						"you saw this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				goto unm_err_out;
-			}
-			NInoSetCompressed(ni);
-			if ((ni->type != AT_DATA) || (ni->type == AT_DATA &&
-					ni->name_len)) {
-				ntfs_error(vi->i_sb, "Found compressed "
-						"non-data or named data "
-						"attribute.  Please report "
-						"you saw this message to "
+						"is %s.  Please report you "
+						"saw this message to "
 						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				goto unm_err_out;
-			}
-			if (vol->cluster_size > 4096) {
-				ntfs_error(vi->i_sb, "Found compressed "
-						"attribute but compression is "
-						"disabled due to cluster size "
-						"(%i) > 4kiB.",
-						vol->cluster_size);
+						"sourceforge.net",
+						NInoCompressed(ni) ?
+						"compressed" : "sparse");
 				goto unm_err_out;
 			}
-			if ((ctx->attr->flags & ATTR_COMPRESSION_MASK)
-					!= ATTR_IS_COMPRESSED) {
-				ntfs_error(vi->i_sb, "Found unknown "
-						"compression method.");
-				goto unm_err_out;
-			}
-			ni->itype.compressed.block_clusters = 1U <<
-					ctx->attr->data.non_resident.
-					compression_unit;
-			if (ctx->attr->data.non_resident.compression_unit !=
-					4) {
+			if (a->flags & ATTR_IS_SPARSE)
+				NInoSetSparse(ni);
+			if (a->data.non_resident.compression_unit != 4) {
 				ntfs_error(vi->i_sb, "Found nonstandard "
 						"compression unit (%u instead "
 						"of 4).  Cannot handle this.",
-						ctx->attr->data.non_resident.
+						a->data.non_resident.
 						compression_unit);
 				err = -EOPNOTSUPP;
 				goto unm_err_out;
 			}
+			ni->itype.compressed.block_clusters = 1U <<
+					a->data.non_resident.compression_unit;
 			ni->itype.compressed.block_size = 1U << (
-					ctx->attr->data.non_resident.
-					compression_unit +
+					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_size) - 1;
+			ni->itype.compressed.size = sle64_to_cpu(
+					a->data.non_resident.compressed_size);
 		}
-		if (ctx->attr->flags & ATTR_IS_ENCRYPTED) {
-			if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
-				ntfs_error(vi->i_sb, "Found encrypted "
-						"and compressed data.");
+		if (a->flags & ATTR_IS_ENCRYPTED) {
+			if (a->flags & ATTR_COMPRESSION_MASK) {
+				ntfs_error(vi->i_sb, "Found encrypted and "
+						"compressed data.");
 				goto unm_err_out;
 			}
 			if (NInoMstProtected(ni)) {
@@ -1317,37 +1330,17 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 			}
 			NInoSetEncrypted(ni);
 		}
-		if (ctx->attr->flags & ATTR_IS_SPARSE) {
-			if (NInoMstProtected(ni)) {
-				ntfs_error(vi->i_sb, "Found mst protected "
-						"attribute but the attribute "
-						"is sparse.  Please report "
-						"you saw this message to "
-						"linux-ntfs-dev@lists."
-						"sourceforge.net");
-				goto unm_err_out;
-			}
-			NInoSetSparse(ni);
-		}
-		if (ctx->attr->data.non_resident.lowest_vcn) {
+		if (a->data.non_resident.lowest_vcn) {
 			ntfs_error(vi->i_sb, "First extent of attribute has "
 					"non-zero lowest_vcn.");
 			goto unm_err_out;
 		}
-		/* Setup all the sizes. */
-		vi->i_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.data_size);
+		vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
 		ni->initialized_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.initialized_size);
+				a->data.non_resident.initialized_size);
 		ni->allocated_size = sle64_to_cpu(
-				ctx->attr->data.non_resident.allocated_size);
-		if (NInoCompressed(ni)) {
-			ni->itype.compressed.size = sle64_to_cpu(
-					ctx->attr->data.non_resident.
-					compressed_size);
-		}
+				a->data.non_resident.allocated_size);
 	}
-
 	/* Setup the operations for this attribute inode. */
 	vi->i_op = NULL;
 	vi->i_fop = NULL;
@@ -1355,12 +1348,10 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 		vi->i_mapping->a_ops = &ntfs_mst_aops;
 	else
 		vi->i_mapping->a_ops = &ntfs_aops;
-
-	if (!NInoCompressed(ni))
-		vi->i_blocks = ni->allocated_size >> 9;
-	else
+	if (NInoCompressed(ni) || NInoSparse(ni))
 		vi->i_blocks = ni->itype.compressed.size >> 9;
-
+	else
+		vi->i_blocks = ni->allocated_size >> 9;
 	/*
 	 * Make sure the base inode doesn't go away and attach it to the
 	 * attribute inode.
@@ -1429,10 +1420,12 @@ err_out:
  */
 static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 {
+	loff_t bvi_size;
 	ntfs_volume *vol = NTFS_SB(vi->i_sb);
 	ntfs_inode *ni, *base_ni, *bni;
 	struct inode *bvi;
 	MFT_RECORD *m;
+	ATTR_RECORD *a;
 	ntfs_attr_search_ctx *ctx;
 	INDEX_ROOT *ir;
 	u8 *ir_end, *index_end;
@@ -1474,30 +1467,28 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 					"missing.");
 		goto unm_err_out;
 	}
+	a = ctx->attr;
 	/* Set up the state. */
-	if (unlikely(ctx->attr->non_resident)) {
+	if (unlikely(a->non_resident)) {
 		ntfs_error(vol->sb, "$INDEX_ROOT attribute is not resident.");
 		goto unm_err_out;
 	}
 	/* Ensure the attribute name is placed before the value. */
-	if (unlikely(ctx->attr->name_length &&
-			(le16_to_cpu(ctx->attr->name_offset) >=
-			le16_to_cpu(ctx->attr->data.resident.
-			value_offset)))) {
+	if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+			le16_to_cpu(a->data.resident.value_offset)))) {
 		ntfs_error(vol->sb, "$INDEX_ROOT attribute name is placed "
 				"after the attribute value.");
 		goto unm_err_out;
 	}
 	/* Compressed/encrypted/sparse index root is not allowed. */
-	if (ctx->attr->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_ENCRYPTED |
+	if (a->flags & (ATTR_COMPRESSION_MASK | ATTR_IS_ENCRYPTED |
 			ATTR_IS_SPARSE)) {
 		ntfs_error(vi->i_sb, "Found compressed/encrypted/sparse index "
 				"root attribute.");
 		goto unm_err_out;
 	}
-	ir = (INDEX_ROOT*)((u8*)ctx->attr +
-			le16_to_cpu(ctx->attr->data.resident.value_offset));
-	ir_end = (u8*)ir + le32_to_cpu(ctx->attr->data.resident.value_length);
+	ir = (INDEX_ROOT*)((u8*)a + le16_to_cpu(a->data.resident.value_offset));
+	ir_end = (u8*)ir + le32_to_cpu(a->data.resident.value_length);
 	if (ir_end > (u8*)ctx->mrec + vol->mft_record_size) {
 		ntfs_error(vi->i_sb, "$INDEX_ROOT attribute is corrupt.");
 		goto unm_err_out;
@@ -1570,7 +1561,7 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 					"$INDEX_ALLOCATION attribute.");
 		goto unm_err_out;
 	}
-	if (!ctx->attr->non_resident) {
+	if (!a->non_resident) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
 				"resident.");
 		goto unm_err_out;
@@ -1578,37 +1569,36 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 	/*
 	 * Ensure the attribute name is placed before the mapping pairs array.
 	 */
-	if (unlikely(ctx->attr->name_length && (le16_to_cpu(
-			ctx->attr->name_offset) >= le16_to_cpu(
-			ctx->attr->data.non_resident.mapping_pairs_offset)))) {
+	if (unlikely(a->name_length && (le16_to_cpu(a->name_offset) >=
+			le16_to_cpu(
+			a->data.non_resident.mapping_pairs_offset)))) {
 		ntfs_error(vol->sb, "$INDEX_ALLOCATION attribute name is "
 				"placed after the mapping pairs array.");
 		goto unm_err_out;
 	}
-	if (ctx->attr->flags & ATTR_IS_ENCRYPTED) {
+	if (a->flags & ATTR_IS_ENCRYPTED) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
 				"encrypted.");
 		goto unm_err_out;
 	}
-	if (ctx->attr->flags & ATTR_IS_SPARSE) {
+	if (a->flags & ATTR_IS_SPARSE) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is sparse.");
 		goto unm_err_out;
 	}
-	if (ctx->attr->flags & ATTR_COMPRESSION_MASK) {
+	if (a->flags & ATTR_COMPRESSION_MASK) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
 				"compressed.");
 		goto unm_err_out;
 	}
-	if (ctx->attr->data.non_resident.lowest_vcn) {
+	if (a->data.non_resident.lowest_vcn) {
 		ntfs_error(vi->i_sb, "First extent of $INDEX_ALLOCATION "
 				"attribute has non zero lowest_vcn.");
 		goto unm_err_out;
 	}
-	vi->i_size = sle64_to_cpu(ctx->attr->data.non_resident.data_size);
+	vi->i_size = sle64_to_cpu(a->data.non_resident.data_size);
 	ni->initialized_size = sle64_to_cpu(
-			ctx->attr->data.non_resident.initialized_size);
-	ni->allocated_size = sle64_to_cpu(
-			ctx->attr->data.non_resident.allocated_size);
+			a->data.non_resident.initialized_size);
+	ni->allocated_size = sle64_to_cpu(a->data.non_resident.allocated_size);
 	/*
 	 * We are done with the mft record, so we release it.  Otherwise
 	 * we would deadlock in ntfs_attr_iget().
@@ -1632,10 +1622,10 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 		goto iput_unm_err_out;
 	}
 	/* Consistency check bitmap size vs. index allocation size. */
-	if ((bvi->i_size << 3) < (vi->i_size >>
-			ni->itype.index.block_size_bits)) {
+	bvi_size = i_size_read(bvi);
+	if ((bvi_size << 3) < (vi->i_size >> ni->itype.index.block_size_bits)) {
 		ntfs_error(vi->i_sb, "Index bitmap too small (0x%llx) for "
-				"index allocation (0x%llx).", bvi->i_size << 3,
+				"index allocation (0x%llx).", bvi_size << 3,
 				vi->i_size);
 		goto iput_unm_err_out;
 	}
@@ -1646,7 +1636,6 @@ skip_large_index_stuff:
 	vi->i_fop = NULL;
 	vi->i_mapping->a_ops = &ntfs_mst_aops;
 	vi->i_blocks = ni->allocated_size >> 9;
-
 	/*
 	 * Make sure the base inode doesn't go away and attach it to the
 	 * index inode.
@@ -1712,7 +1701,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 	struct buffer_head *bh;
 	ntfs_inode *ni;
 	MFT_RECORD *m = NULL;
-	ATTR_RECORD *attr;
+	ATTR_RECORD *a;
 	ntfs_attr_search_ctx *ctx;
 	unsigned int i, nr_blocks;
 	int err;
@@ -1727,10 +1716,10 @@ int ntfs_read_inode_mount(struct inode *vi)
 	/* Setup the data attribute. It is special as it is mst protected. */
 	NInoSetNonResident(ni);
 	NInoSetMstProtected(ni);
+	NInoSetSparseDisabled(ni);
 	ni->type = AT_DATA;
 	ni->name = NULL;
 	ni->name_len = 0;
-
 	/*
 	 * This sets up our little cheat allowing us to reuse the async read io
 	 * completion handler for directories.
@@ -1808,9 +1797,10 @@ int ntfs_read_inode_mount(struct inode *vi)
 
 		ntfs_debug("Attribute list attribute found in $MFT.");
 		NInoSetAttrList(ni);
-		if (ctx->attr->flags & ATTR_IS_ENCRYPTED ||
-				ctx->attr->flags & ATTR_COMPRESSION_MASK ||
-				ctx->attr->flags & ATTR_IS_SPARSE) {
+		a = ctx->attr;
+		if (a->flags & ATTR_IS_ENCRYPTED ||
+				a->flags & ATTR_COMPRESSION_MASK ||
+				a->flags & ATTR_IS_SPARSE) {
 			ntfs_error(sb, "Attribute list attribute is "
 					"compressed/encrypted/sparse. Not "
 					"allowed. $MFT is corrupt. You should "
@@ -1818,16 +1808,16 @@ int ntfs_read_inode_mount(struct inode *vi)
 			goto put_err_out;
 		}
 		/* Now allocate memory for the attribute list. */
-		ni->attr_list_size = (u32)ntfs_attr_size(ctx->attr);
+		ni->attr_list_size = (u32)ntfs_attr_size(a);
 		ni->attr_list = ntfs_malloc_nofs(ni->attr_list_size);
 		if (!ni->attr_list) {
 			ntfs_error(sb, "Not enough memory to allocate buffer "
 					"for attribute list.");
 			goto put_err_out;
 		}
-		if (ctx->attr->non_resident) {
+		if (a->non_resident) {
 			NInoSetAttrListNonResident(ni);
-			if (ctx->attr->data.non_resident.lowest_vcn) {
+			if (a->data.non_resident.lowest_vcn) {
 				ntfs_error(sb, "Attribute list has non zero "
 						"lowest_vcn. $MFT is corrupt. "
 						"You should run chkdsk.");
@@ -1835,7 +1825,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 			}
 			/* Setup the runlist. */
 			ni->attr_list_rl.rl = ntfs_mapping_pairs_decompress(vol,
-					ctx->attr, NULL);
+					a, NULL);
 			if (IS_ERR(ni->attr_list_rl.rl)) {
 				err = PTR_ERR(ni->attr_list_rl.rl);
 				ni->attr_list_rl.rl = NULL;
@@ -1847,7 +1837,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 			/* Now load the attribute list. */
 			if ((err = load_attribute_list(vol, &ni->attr_list_rl,
 					ni->attr_list, ni->attr_list_size,
-					sle64_to_cpu(ctx->attr->data.
+					sle64_to_cpu(a->data.
 					non_resident.initialized_size)))) {
 				ntfs_error(sb, "Failed to load attribute list "
 						"attribute with error code %i.",
@@ -1855,20 +1845,20 @@ int ntfs_read_inode_mount(struct inode *vi)
 				goto put_err_out;
 			}
 		} else /* if (!ctx.attr->non_resident) */ {
-			if ((u8*)ctx->attr + le16_to_cpu(
-					ctx->attr->data.resident.value_offset) +
+			if ((u8*)a + le16_to_cpu(
+					a->data.resident.value_offset) +
 					le32_to_cpu(
-					ctx->attr->data.resident.value_length) >
+					a->data.resident.value_length) >
 					(u8*)ctx->mrec + vol->mft_record_size) {
 				ntfs_error(sb, "Corrupt attribute list "
 						"attribute.");
 				goto put_err_out;
 			}
 			/* Now copy the attribute list. */
-			memcpy(ni->attr_list, (u8*)ctx->attr + le16_to_cpu(
-					ctx->attr->data.resident.value_offset),
+			memcpy(ni->attr_list, (u8*)a + le16_to_cpu(
+					a->data.resident.value_offset),
 					le32_to_cpu(
-					ctx->attr->data.resident.value_length));
+					a->data.resident.value_length));
 		}
 		/* The attribute list is now setup in memory. */
 		/*
@@ -1934,25 +1924,25 @@ int ntfs_read_inode_mount(struct inode *vi)
 	ntfs_attr_reinit_search_ctx(ctx);
 
 	/* Now load all attribute extents. */
-	attr = NULL;
+	a = NULL;
 	next_vcn = last_vcn = highest_vcn = 0;
 	while (!(err = ntfs_attr_lookup(AT_DATA, NULL, 0, 0, next_vcn, NULL, 0,
 			ctx))) {
 		runlist_element *nrl;
 
 		/* Cache the current attribute. */
-		attr = ctx->attr;
+		a = ctx->attr;
 		/* $MFT must be non-resident. */
-		if (!attr->non_resident) {
+		if (!a->non_resident) {
 			ntfs_error(sb, "$MFT must be non-resident but a "
 					"resident extent was found. $MFT is "
 					"corrupt. Run chkdsk.");
 			goto put_err_out;
 		}
 		/* $MFT must be uncompressed and unencrypted. */
-		if (attr->flags & ATTR_COMPRESSION_MASK ||
-				attr->flags & ATTR_IS_ENCRYPTED ||
-				attr->flags & ATTR_IS_SPARSE) {
+		if (a->flags & ATTR_COMPRESSION_MASK ||
+				a->flags & ATTR_IS_ENCRYPTED ||
+				a->flags & ATTR_IS_SPARSE) {
 			ntfs_error(sb, "$MFT must be uncompressed, "
 					"non-sparse, and unencrypted but a "
 					"compressed/sparse/encrypted extent "
@@ -1966,7 +1956,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 		 * as we have exclusive access to the inode at this time and we
 		 * are a mount in progress task, too.
 		 */
-		nrl = ntfs_mapping_pairs_decompress(vol, attr, ni->runlist.rl);
+		nrl = ntfs_mapping_pairs_decompress(vol, a, ni->runlist.rl);
 		if (IS_ERR(nrl)) {
 			ntfs_error(sb, "ntfs_mapping_pairs_decompress() "
 					"failed with error code %ld.  $MFT is "
@@ -1977,7 +1967,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 
 		/* Are we in the first extent? */
 		if (!next_vcn) {
-			if (attr->data.non_resident.lowest_vcn) {
+			if (a->data.non_resident.lowest_vcn) {
 				ntfs_error(sb, "First extent of $DATA "
 						"attribute has non zero "
 						"lowest_vcn. $MFT is corrupt. "
@@ -1986,15 +1976,15 @@ int ntfs_read_inode_mount(struct inode *vi)
 			}
 			/* Get the last vcn in the $DATA attribute. */
 			last_vcn = sle64_to_cpu(
-					attr->data.non_resident.allocated_size)
+					a->data.non_resident.allocated_size)
 					>> vol->cluster_size_bits;
 			/* Fill in the inode size. */
 			vi->i_size = sle64_to_cpu(
-					attr->data.non_resident.data_size);
-			ni->initialized_size = sle64_to_cpu(attr->data.
-					non_resident.initialized_size);
+					a->data.non_resident.data_size);
+			ni->initialized_size = sle64_to_cpu(
+					a->data.non_resident.initialized_size);
 			ni->allocated_size = sle64_to_cpu(
-					attr->data.non_resident.allocated_size);
+					a->data.non_resident.allocated_size);
 			/*
 			 * Verify the number of mft records does not exceed
 			 * 2^32 - 1.
@@ -2051,7 +2041,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 		}
 
 		/* Get the lowest vcn for the next extent. */
-		highest_vcn = sle64_to_cpu(attr->data.non_resident.highest_vcn);
+		highest_vcn = sle64_to_cpu(a->data.non_resident.highest_vcn);
 		next_vcn = highest_vcn + 1;
 
 		/* Only one extent or error, which we catch below. */
@@ -2060,7 +2050,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 
 		/* Avoid endless loops due to corruption. */
 		if (next_vcn < sle64_to_cpu(
-				attr->data.non_resident.lowest_vcn)) {
+				a->data.non_resident.lowest_vcn)) {
 			ntfs_error(sb, "$MFT has corrupt attribute list "
 					"attribute. Run chkdsk.");
 			goto put_err_out;
@@ -2071,7 +2061,7 @@ int ntfs_read_inode_mount(struct inode *vi)
 				"$MFT is corrupt. Run chkdsk.");
 		goto put_err_out;
 	}
-	if (!attr) {
+	if (!a) {
 		ntfs_error(sb, "$MFT/$DATA attribute not found. $MFT is "
 				"corrupt. Run chkdsk.");
 		goto put_err_out;
@@ -2275,6 +2265,8 @@ int ntfs_show_options(struct seq_file *sf, struct vfsmount *mnt)
 		seq_printf(sf, ",case_sensitive");
 	if (NVolShowSystemFiles(vol))
 		seq_printf(sf, ",show_sys_files");
+	if (!NVolSparseEnabled(vol))
+		seq_printf(sf, ",disable_sparse");
 	for (i = 0; on_errors_arr[i].val; i++) {
 		if (on_errors_arr[i].val & vol->on_errors)
 			seq_printf(sf, ",errors=%s", on_errors_arr[i].str);
@@ -2311,6 +2303,7 @@ int ntfs_truncate(struct inode *vi)
 	ntfs_volume *vol = ni->vol;
 	ntfs_attr_search_ctx *ctx;
 	MFT_RECORD *m;
+	ATTR_RECORD *a;
 	const char *te = "  Leaving file length out of sync with i_size.";
 	int err;
 
@@ -2347,14 +2340,15 @@ int ntfs_truncate(struct inode *vi)
 					vi->i_ino, err);
 		goto err_out;
 	}
+	a = ctx->attr;
 	/* If the size has not changed there is nothing to do. */
-	if (ntfs_attr_size(ctx->attr) == i_size_read(vi))
+	if (ntfs_attr_size(a) == i_size_read(vi))
 		goto done;
 	// TODO: Implement the truncate...
 	ntfs_error(vi->i_sb, "Inode size has changed but this is not "
 			"implemented yet.  Resetting inode size to old value. "
 			" This is most likely a bug in the ntfs driver!");
-	i_size_write(vi, ntfs_attr_size(ctx->attr)); 
+	i_size_write(vi, ntfs_attr_size(a)); 
 done:
 	ntfs_attr_put_search_ctx(ctx);
 	unmap_mft_record(ni);
@@ -2515,18 +2509,18 @@ int ntfs_write_inode(struct inode *vi, int sync)
 	nt = utc2ntfs(vi->i_mtime);
 	if (si->last_data_change_time != nt) {
 		ntfs_debug("Updating mtime for inode 0x%lx: old = 0x%llx, "
-				"new = 0x%llx", vi->i_ino,
+				"new = 0x%llx", vi->i_ino, (long long)
 				sle64_to_cpu(si->last_data_change_time),
-				sle64_to_cpu(nt));
+				(long long)sle64_to_cpu(nt));
 		si->last_data_change_time = nt;
 		modified = TRUE;
 	}
 	nt = utc2ntfs(vi->i_ctime);
 	if (si->last_mft_change_time != nt) {
 		ntfs_debug("Updating ctime for inode 0x%lx: old = 0x%llx, "
-				"new = 0x%llx", vi->i_ino,
+				"new = 0x%llx", vi->i_ino, (long long)
 				sle64_to_cpu(si->last_mft_change_time),
-				sle64_to_cpu(nt));
+				(long long)sle64_to_cpu(nt));
 		si->last_mft_change_time = nt;
 		modified = TRUE;
 	}
@@ -2534,8 +2528,8 @@ int ntfs_write_inode(struct inode *vi, int sync)
 	if (si->last_access_time != nt) {
 		ntfs_debug("Updating atime for inode 0x%lx: old = 0x%llx, "
 				"new = 0x%llx", vi->i_ino,
-				sle64_to_cpu(si->last_access_time),
-				sle64_to_cpu(nt));
+				(long long)sle64_to_cpu(si->last_access_time),
+				(long long)sle64_to_cpu(nt));
 		si->last_access_time = nt;
 		modified = TRUE;
 	}
diff --git a/fs/ntfs/inode.h b/fs/ntfs/inode.h
index 9958045..3de5c02 100644
--- a/fs/ntfs/inode.h
+++ b/fs/ntfs/inode.h
@@ -2,7 +2,7 @@
  * inode.h - Defines for inode structures NTFS Linux kernel driver. Part of
  *	     the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -44,6 +44,7 @@ typedef struct _ntfs_inode ntfs_inode;
  * fields already provided in the VFS inode.
  */
 struct _ntfs_inode {
+	rwlock_t size_lock;	/* Lock serializing access to inode sizes. */
 	s64 initialized_size;	/* Copy from the attribute record. */
 	s64 allocated_size;	/* Copy from the attribute record. */
 	unsigned long state;	/* NTFS specific flags describing this inode.
@@ -109,7 +110,7 @@ struct _ntfs_inode {
 			u8 block_size_bits; 	/* Log2 of the above. */
 			u8 vcn_size_bits;	/* Log2 of the above. */
 		} index;
-		struct { /* It is a compressed file or an attribute inode. */
+		struct { /* It is a compressed/sparse file/attribute inode. */
 			s64 size;		/* Copy of compressed_size from
 						   $DATA. */
 			u32 block_size;		/* Size of a compression block
@@ -165,6 +166,7 @@ typedef enum {
 	NI_Sparse,		/* 1: Unnamed data attr is sparse (f).
 				   1: Create sparse files by default (d).
 				   1: Attribute is sparse (a). */
+	NI_SparseDisabled,	/* 1: May not create sparse regions. */
 	NI_TruncateFailed,	/* 1: Last ntfs_truncate() call failed. */
 } ntfs_inode_state_bits;
 
@@ -217,6 +219,7 @@ NINO_FNS(IndexAllocPresent)
 NINO_FNS(Compressed)
 NINO_FNS(Encrypted)
 NINO_FNS(Sparse)
+NINO_FNS(SparseDisabled)
 NINO_FNS(TruncateFailed)
 
 /*
diff --git a/fs/ntfs/layout.h b/fs/ntfs/layout.h
index 47b3389..609ad17 100644
--- a/fs/ntfs/layout.h
+++ b/fs/ntfs/layout.h
@@ -2,7 +2,7 @@
  * layout.h - All NTFS associated on-disk structures. Part of the Linux-NTFS
  *	      project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -547,26 +547,44 @@ enum {
 	COLLATION_NTOFS_ULONG		= const_cpu_to_le32(0x10),
 	COLLATION_NTOFS_SID		= const_cpu_to_le32(0x11),
 	COLLATION_NTOFS_SECURITY_HASH	= const_cpu_to_le32(0x12),
-	COLLATION_NTOFS_ULONGS		= const_cpu_to_le32(0x13)
+	COLLATION_NTOFS_ULONGS		= const_cpu_to_le32(0x13),
 };
 
 typedef le32 COLLATION_RULE;
 
 /*
  * The flags (32-bit) describing attribute properties in the attribute
- * definition structure.  FIXME: This information is from Regis's information
- * and, according to him, it is not certain and probably incomplete.
- * The INDEXABLE flag is fairly certainly correct as only the file name
- * attribute has this flag set and this is the only attribute indexed in NT4.
+ * definition structure.  FIXME: This information is based on Regis's
+ * information and, according to him, it is not certain and probably
+ * incomplete.  The INDEXABLE flag is fairly certainly correct as only the file
+ * name attribute has this flag set and this is the only attribute indexed in
+ * NT4.
  */
 enum {
-	INDEXABLE	    = const_cpu_to_le32(0x02), /* Attribute can be
-							  indexed. */
-	NEED_TO_REGENERATE  = const_cpu_to_le32(0x40), /* Need to regenerate
-							  during regeneration
-							  phase. */
-	CAN_BE_NON_RESIDENT = const_cpu_to_le32(0x80), /* Attribute can be
-							  non-resident. */
+	ATTR_DEF_INDEXABLE	= const_cpu_to_le32(0x02), /* Attribute can be
+					indexed. */
+	ATTR_DEF_MULTIPLE	= const_cpu_to_le32(0x04), /* Attribute type
+					can be present multiple times in the
+					mft records of an inode. */
+	ATTR_DEF_NOT_ZERO	= const_cpu_to_le32(0x08), /* Attribute value
+					must contain at least one non-zero
+					byte. */
+	ATTR_DEF_INDEXED_UNIQUE	= const_cpu_to_le32(0x10), /* Attribute must be
+					indexed and the attribute value must be
+					unique for the attribute type in all of
+					the mft records of an inode. */
+	ATTR_DEF_NAMED_UNIQUE	= const_cpu_to_le32(0x20), /* Attribute must be
+					named and the name must be unique for
+					the attribute type in all of the mft
+					records of an inode. */
+	ATTR_DEF_RESIDENT	= const_cpu_to_le32(0x40), /* Attribute must be
+					resident. */
+	ATTR_DEF_ALWAYS_LOG	= const_cpu_to_le32(0x80), /* Always log
+					modifications to this attribute,
+					regardless of whether it is resident or
+					non-resident.  Without this, only log
+					modifications if the attribute is
+					resident. */
 };
 
 typedef le32 ATTR_DEF_FLAGS;
@@ -749,10 +767,11 @@ typedef struct {
 				record header aligned to 8-byte boundary. */
 /* 34*/			u8 compression_unit; /* The compression unit expressed
 				as the log to the base 2 of the number of
-				clusters in a compression unit. 0 means not
-				compressed. (This effectively limits the
+				clusters in a compression unit.  0 means not
+				compressed.  (This effectively limits the
 				compression unit size to be a power of two
-				clusters.) WinNT4 only uses a value of 4. */
+				clusters.)  WinNT4 only uses a value of 4.
+				Sparse files also have this set to 4. */
 /* 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.*/
@@ -772,10 +791,10 @@ typedef struct {
 				data_size. */
 /* sizeof(uncompressed attr) = 64*/
 /* 64*/			sle64 compressed_size;	/* Byte size of the attribute
-				value after compression. Only present when
-				compressed. Always is a multiple of the
-				cluster size. Represents the actual amount of
-				disk space being used on the disk. */
+				value after compression.  Only present when
+				compressed or sparse.  Always is a multiple of
+				the cluster size.  Represents the actual amount
+				of disk space being used on the disk. */
 /* sizeof(compressed attr) = 72*/
 		} __attribute__ ((__packed__)) non_resident;
 	} __attribute__ ((__packed__)) data;
@@ -834,7 +853,7 @@ enum {
 	/* 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
 	   index, quota index, one of the security indexes or the encrypting
-	   file system related indexes). */
+	   filesystem related indexes). */
 };
 
 typedef le32 FILE_ATTR_FLAGS;
@@ -917,20 +936,12 @@ typedef struct {
 		/* 56*/	le64 quota_charged;	/* Byte size of the charge to
 				the quota for all streams of the file. Note: Is
 				zero if quotas are disabled. */
-		/* 64*/	le64 usn;		/* Last update sequence number
-				of the file. This is a direct index into the
-				change (aka usn) journal file. It is zero if
-				the usn journal is disabled.
-				NOTE: To disable the journal need to delete
-				the journal file itself and to then walk the
-				whole mft and set all Usn entries in all mft
-				records to zero! (This can take a while!)
-				The journal is FILE_Extend/$UsnJrnl. Win2k
-				will recreate the journal and initiate
-				logging if necessary when mounting the
-				partition. This, in contrast to disabling the
-				journal is a very fast process, so the user
-				won't even notice it. */
+		/* 64*/	leUSN usn;		/* Last update sequence number
+				of the file.  This is a direct index into the
+				transaction log file ($UsnJrnl).  It is zero if
+				the usn journal is disabled or this file has
+				not been subject to logging yet.  See usnjrnl.h
+				for details. */
 		} __attribute__ ((__packed__)) v3;
 	/* sizeof() = 72 bytes (NTFS 3.x) */
 	} __attribute__ ((__packed__)) ver;
@@ -1893,7 +1904,7 @@ enum {
 	VOLUME_FLAGS_MASK		= const_cpu_to_le16(0x803f),
 
 	/* To make our life easier when checking if we must mount read-only. */
-	VOLUME_MUST_MOUNT_RO_MASK	= const_cpu_to_le16(0x8037),
+	VOLUME_MUST_MOUNT_RO_MASK	= const_cpu_to_le16(0x8027),
 } __attribute__ ((__packed__));
 
 typedef le16 VOLUME_FLAGS;
diff --git a/fs/ntfs/lcnalloc.c b/fs/ntfs/lcnalloc.c
index 23fd911..a4bc076 100644
--- a/fs/ntfs/lcnalloc.c
+++ b/fs/ntfs/lcnalloc.c
@@ -1,7 +1,7 @@
 /*
  * lcnalloc.c - Cluster (de)allocation code.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2004 Anton Altaparmakov
+ * Copyright (c) 2004-2005 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
@@ -60,7 +60,7 @@ int ntfs_cluster_free_from_rl_nolock(ntfs_volume *vol,
 		if (rl->lcn < 0)
 			continue;
 		err = ntfs_bitmap_clear_run(lcnbmp_vi, rl->lcn, rl->length);
-		if (unlikely(err && (!ret || ret == ENOMEM) && ret != err))
+		if (unlikely(err && (!ret || ret == -ENOMEM) && ret != err))
 			ret = err;
 	}
 	ntfs_debug("Done.");
@@ -140,6 +140,7 @@ runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
 	LCN zone_start, zone_end, bmp_pos, bmp_initial_pos, last_read_pos, lcn;
 	LCN prev_lcn = 0, prev_run_len = 0, mft_zone_size;
 	s64 clusters;
+	loff_t i_size;
 	struct inode *lcnbmp_vi;
 	runlist_element *rl = NULL;
 	struct address_space *mapping;
@@ -249,6 +250,7 @@ runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
 	clusters = count;
 	rlpos = rlsize = 0;
 	mapping = lcnbmp_vi->i_mapping;
+	i_size = i_size_read(lcnbmp_vi);
 	while (1) {
 		ntfs_debug("Start of outer while loop: done_zones 0x%x, "
 				"search_zone %i, pass %i, zone_start 0x%llx, "
@@ -263,7 +265,7 @@ runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
 		last_read_pos = bmp_pos >> 3;
 		ntfs_debug("last_read_pos 0x%llx.",
 				(unsigned long long)last_read_pos);
-		if (last_read_pos > lcnbmp_vi->i_size) {
+		if (last_read_pos > i_size) {
 			ntfs_debug("End of attribute reached.  "
 					"Skipping to zone_pass_done.");
 			goto zone_pass_done;
@@ -287,11 +289,11 @@ runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
 		buf_size = last_read_pos & ~PAGE_CACHE_MASK;
 		buf = page_address(page) + buf_size;
 		buf_size = PAGE_CACHE_SIZE - buf_size;
-		if (unlikely(last_read_pos + buf_size > lcnbmp_vi->i_size))
-			buf_size = lcnbmp_vi->i_size - last_read_pos;
+		if (unlikely(last_read_pos + buf_size > i_size))
+			buf_size = i_size - last_read_pos;
 		buf_size <<= 3;
 		lcn = bmp_pos & 7;
-		bmp_pos &= ~7;
+		bmp_pos &= ~(LCN)7;
 		ntfs_debug("Before inner while loop: buf_size %i, lcn 0x%llx, "
 				"bmp_pos 0x%llx, need_writeback %i.", buf_size,
 				(unsigned long long)lcn,
@@ -309,7 +311,7 @@ runlist_element *ntfs_cluster_alloc(ntfs_volume *vol, const VCN start_vcn,
 					(unsigned int)*byte);
 			/* Skip full bytes. */
 			if (*byte == 0xff) {
-				lcn = (lcn + 8) & ~7;
+				lcn = (lcn + 8) & ~(LCN)7;
 				ntfs_debug("Continuing while loop 1.");
 				continue;
 			}
@@ -691,7 +693,7 @@ switch_to_data1_zone:		search_zone = 2;
 		if (zone == MFT_ZONE || mft_zone_size <= 0) {
 			ntfs_debug("No free clusters left, going to out.");
 			/* Really no more space left on device. */
-			err = ENOSPC;
+			err = -ENOSPC;
 			goto out;
 		} /* zone == DATA_ZONE && mft_zone_size > 0 */
 		ntfs_debug("Shrinking mft zone.");
@@ -755,13 +757,13 @@ out:
 	if (rl) {
 		int err2;
 
-		if (err == ENOSPC)
+		if (err == -ENOSPC)
 			ntfs_debug("Not enough space to complete allocation, "
-					"err ENOSPC, first free lcn 0x%llx, "
+					"err -ENOSPC, first free lcn 0x%llx, "
 					"could allocate up to 0x%llx "
 					"clusters.",
 					(unsigned long long)rl[0].lcn,
-					(unsigned long long)count - clusters);
+					(unsigned long long)(count - clusters));
 		/* Deallocate all allocated clusters. */
 		ntfs_debug("Attempting rollback...");
 		err2 = ntfs_cluster_free_from_rl_nolock(vol, rl);
@@ -773,10 +775,10 @@ out:
 		}
 		/* Free the runlist. */
 		ntfs_free(rl);
-	} else if (err == ENOSPC)
-		ntfs_debug("No space left at all, err = ENOSPC, "
-				"first free lcn = 0x%llx.",
-				(unsigned long long)vol->data1_zone_pos);
+	} else if (err == -ENOSPC)
+		ntfs_debug("No space left at all, err = -ENOSPC, first free "
+				"lcn = 0x%llx.",
+				(long long)vol->data1_zone_pos);
 	up_write(&vol->lcnbmp_lock);
 	return ERR_PTR(err);
 }
@@ -846,8 +848,8 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 
 	total_freed = real_freed = 0;
 
-	/* This returns with ni->runlist locked for reading on success. */
-	rl = ntfs_find_vcn(ni, start_vcn, FALSE);
+	down_read(&ni->runlist.lock);
+	rl = ntfs_attr_find_vcn_nolock(ni, start_vcn, FALSE);
 	if (IS_ERR(rl)) {
 		if (!is_rollback)
 			ntfs_error(vol->sb, "Failed to find first runlist "
@@ -861,7 +863,7 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 			ntfs_error(vol->sb, "First runlist element has "
 					"invalid lcn, aborting.");
 		err = -EIO;
-		goto unl_err_out;
+		goto err_out;
 	}
 	/* Find the starting cluster inside the run that needs freeing. */
 	delta = start_vcn - rl->vcn;
@@ -879,7 +881,7 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 			if (!is_rollback)
 				ntfs_error(vol->sb, "Failed to clear first run "
 						"(error %i), aborting.", err);
-			goto unl_err_out;
+			goto err_out;
 		}
 		/* We have freed @to_free real clusters. */
 		real_freed = to_free;
@@ -899,30 +901,15 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 		if (unlikely(rl->lcn < LCN_HOLE)) {
 			VCN vcn;
 
-			/*
-			 * Attempt to map runlist, dropping runlist lock for
-			 * the duration.
-			 */
+			/* Attempt to map runlist. */
 			vcn = rl->vcn;
-			up_read(&ni->runlist.lock);
-			err = ntfs_map_runlist(ni, vcn);
-			if (err) {
-				if (!is_rollback)
-					ntfs_error(vol->sb, "Failed to map "
-							"runlist fragment.");
-				if (err == -EINVAL || err == -ENOENT)
-					err = -EIO;
-				goto err_out;
-			}
-			/*
-			 * This returns with ni->runlist locked for reading on
-			 * success.
-			 */
-			rl = ntfs_find_vcn(ni, vcn, FALSE);
+			rl = ntfs_attr_find_vcn_nolock(ni, vcn, FALSE);
 			if (IS_ERR(rl)) {
 				err = PTR_ERR(rl);
 				if (!is_rollback)
-					ntfs_error(vol->sb, "Failed to find "
+					ntfs_error(vol->sb, "Failed to map "
+							"runlist fragment or "
+							"failed to find "
 							"subsequent runlist "
 							"element.");
 				goto err_out;
@@ -935,7 +922,7 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 							(unsigned long long)
 							rl->lcn);
 				err = -EIO;
-				goto unl_err_out;
+				goto err_out;
 			}
 		}
 		/* The number of clusters in this run that need freeing. */
@@ -951,7 +938,7 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 				if (!is_rollback)
 					ntfs_error(vol->sb, "Failed to clear "
 							"subsequent run.");
-				goto unl_err_out;
+				goto err_out;
 			}
 			/* We have freed @to_free real clusters. */
 			real_freed += to_free;
@@ -972,9 +959,8 @@ s64 __ntfs_cluster_free(struct inode *vi, const VCN start_vcn, s64 count,
 	/* We are done.  Return the number of actually freed clusters. */
 	ntfs_debug("Done.");
 	return real_freed;
-unl_err_out:
-	up_read(&ni->runlist.lock);
 err_out:
+	up_read(&ni->runlist.lock);
 	if (is_rollback)
 		return err;
 	/* If no real clusters were freed, no need to rollback. */
diff --git a/fs/ntfs/logfile.c b/fs/ntfs/logfile.c
index 5e280ab..8edb8e2 100644
--- a/fs/ntfs/logfile.c
+++ b/fs/ntfs/logfile.c
@@ -1,7 +1,7 @@
 /*
  * logfile.c - NTFS kernel journal handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2002-2004 Anton Altaparmakov
+ * Copyright (c) 2002-2005 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
@@ -410,7 +410,7 @@ err_out:
 }
 
 /**
- * ntfs_ckeck_logfile - check in the journal if the volume is consistent
+ * ntfs_check_logfile - check the journal for consistency
  * @log_vi:	struct inode of loaded journal $LogFile to check
  *
  * Check the $LogFile journal for consistency and return TRUE if it is
@@ -443,7 +443,7 @@ BOOL ntfs_check_logfile(struct inode *log_vi)
 	/* An empty $LogFile must have been clean before it got emptied. */
 	if (NVolLogFileEmpty(vol))
 		goto is_empty;
-	size = log_vi->i_size;
+	size = i_size_read(log_vi);
 	/* Make sure the file doesn't exceed the maximum allowed size. */
 	if (size > MaxLogFileSize)
 		size = MaxLogFileSize;
@@ -464,7 +464,7 @@ BOOL ntfs_check_logfile(struct inode *log_vi)
 	 * optimize log_page_size and log_page_bits into constants.
 	 */
 	log_page_bits = generic_ffs(log_page_size) - 1;
-	size &= ~(log_page_size - 1);
+	size &= ~(s64)(log_page_size - 1);
 	/*
 	 * Ensure the log file is big enough to store at least the two restart
 	 * pages and the minimum number of log record pages.
@@ -689,7 +689,8 @@ BOOL ntfs_empty_logfile(struct inode *log_vi)
 	if (!NVolLogFileEmpty(vol)) {
 		int err;
 		
-		err = ntfs_attr_set(NTFS_I(log_vi), 0, log_vi->i_size, 0xff);
+		err = ntfs_attr_set(NTFS_I(log_vi), 0, i_size_read(log_vi),
+				0xff);
 		if (unlikely(err)) {
 			ntfs_error(vol->sb, "Failed to fill $LogFile with "
 					"0xff bytes (error code %i).", err);
diff --git a/fs/ntfs/mft.c b/fs/ntfs/mft.c
index dfa85ac..ac9ff39 100644
--- a/fs/ntfs/mft.c
+++ b/fs/ntfs/mft.c
@@ -1,7 +1,7 @@
 /**
  * mft.c - NTFS kernel mft record operations. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -45,6 +45,7 @@
  */
 static inline MFT_RECORD *map_mft_record_page(ntfs_inode *ni)
 {
+	loff_t i_size;
 	ntfs_volume *vol = ni->vol;
 	struct inode *mft_vi = vol->mft_ino;
 	struct page *page;
@@ -60,13 +61,14 @@ static inline MFT_RECORD *map_mft_record_page(ntfs_inode *ni)
 	index = ni->mft_no << vol->mft_record_size_bits >> PAGE_CACHE_SHIFT;
 	ofs = (ni->mft_no << vol->mft_record_size_bits) & ~PAGE_CACHE_MASK;
 
+	i_size = i_size_read(mft_vi);
 	/* The maximum valid index into the page cache for $MFT's data. */
-	end_index = mft_vi->i_size >> PAGE_CACHE_SHIFT;
+	end_index = i_size >> PAGE_CACHE_SHIFT;
 
 	/* If the wanted index is out of bounds the mft record doesn't exist. */
 	if (unlikely(index >= end_index)) {
-		if (index > end_index || (mft_vi->i_size & ~PAGE_CACHE_MASK) <
-				ofs + vol->mft_record_size) {
+		if (index > end_index || (i_size & ~PAGE_CACHE_MASK) < ofs +
+				vol->mft_record_size) {
 			page = ERR_PTR(-ENOENT);
 			ntfs_error(vol->sb, "Attemt to read mft record 0x%lx, "
 					"which is beyond the end of the mft.  "
@@ -285,7 +287,7 @@ MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
 			}
 			unmap_mft_record(ni);
 			ntfs_error(base_ni->vol->sb, "Found stale extent mft "
-					"reference! Corrupt file system. "
+					"reference! Corrupt filesystem. "
 					"Run chkdsk.");
 			return ERR_PTR(-EIO);
 		}
@@ -316,7 +318,7 @@ map_err_out:
 	/* Verify the sequence number if it is present. */
 	if (seq_no && (le16_to_cpu(m->sequence_number) != seq_no)) {
 		ntfs_error(base_ni->vol->sb, "Found stale extent mft "
-				"reference! Corrupt file system. Run chkdsk.");
+				"reference! Corrupt filesystem. Run chkdsk.");
 		destroy_ni = TRUE;
 		m = ERR_PTR(-EIO);
 		goto unm_err_out;
@@ -946,20 +948,23 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	na.name_len = 0;
 	na.type = AT_UNUSED;
 	/*
-	 * For inode 0, i.e. $MFT itself, we cannot use ilookup5() from here or
-	 * we deadlock because the inode is already locked by the kernel
-	 * (fs/fs-writeback.c::__sync_single_inode()) and ilookup5() waits
-	 * until the inode is unlocked before returning it and it never gets
-	 * unlocked because ntfs_should_write_mft_record() never returns.  )-:
-	 * Fortunately, we have inode 0 pinned in icache for the duration of
-	 * the mount so we can access it directly.
+	 * Optimize inode 0, i.e. $MFT itself, since we have it in memory and
+	 * we get here for it rather often.
 	 */
 	if (!mft_no) {
 		/* Balance the below iput(). */
 		vi = igrab(mft_vi);
 		BUG_ON(vi != mft_vi);
-	} else
-		vi = ilookup5(sb, mft_no, (test_t)ntfs_test_inode, &na);
+	} else {
+		/*
+		 * Have to use ilookup5_nowait() since ilookup5() waits for the
+		 * inode lock which causes ntfs to deadlock when a concurrent
+		 * inode write via the inode dirty code paths and the page
+		 * dirty code path of the inode dirty code path when writing
+		 * $MFT occurs.
+		 */
+		vi = ilookup5_nowait(sb, mft_no, (test_t)ntfs_test_inode, &na);
+	}
 	if (vi) {
 		ntfs_debug("Base inode 0x%lx is in icache.", mft_no);
 		/* The inode is in icache. */
@@ -1014,7 +1019,13 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	na.mft_no = MREF_LE(m->base_mft_record);
 	ntfs_debug("Mft record 0x%lx is an extent record.  Looking for base "
 			"inode 0x%lx in icache.", mft_no, na.mft_no);
-	vi = ilookup5(sb, na.mft_no, (test_t)ntfs_test_inode, &na);
+	if (!na.mft_no) {
+		/* Balance the below iput(). */
+		vi = igrab(mft_vi);
+		BUG_ON(vi != mft_vi);
+	} else
+		vi = ilookup5_nowait(sb, na.mft_no, (test_t)ntfs_test_inode,
+				&na);
 	if (!vi) {
 		/*
 		 * The base inode is not in icache, write this extent mft
@@ -1121,6 +1132,7 @@ static int ntfs_mft_bitmap_find_and_alloc_free_rec_nolock(ntfs_volume *vol,
 		ntfs_inode *base_ni)
 {
 	s64 pass_end, ll, data_pos, pass_start, ofs, bit;
+	unsigned long flags;
 	struct address_space *mftbmp_mapping;
 	u8 *buf, *byte;
 	struct page *page;
@@ -1134,9 +1146,13 @@ static int ntfs_mft_bitmap_find_and_alloc_free_rec_nolock(ntfs_volume *vol,
 	 * Set the end of the pass making sure we do not overflow the mft
 	 * bitmap.
 	 */
+	read_lock_irqsave(&NTFS_I(vol->mft_ino)->size_lock, flags);
 	pass_end = NTFS_I(vol->mft_ino)->allocated_size >>
 			vol->mft_record_size_bits;
+	read_unlock_irqrestore(&NTFS_I(vol->mft_ino)->size_lock, flags);
+	read_lock_irqsave(&NTFS_I(vol->mftbmp_ino)->size_lock, flags);
 	ll = NTFS_I(vol->mftbmp_ino)->initialized_size << 3;
+	read_unlock_irqrestore(&NTFS_I(vol->mftbmp_ino)->size_lock, flags);
 	if (pass_end > ll)
 		pass_end = ll;
 	pass = 1;
@@ -1263,6 +1279,7 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 {
 	LCN lcn;
 	s64 ll;
+	unsigned long flags;
 	struct page *page;
 	ntfs_inode *mft_ni, *mftbmp_ni;
 	runlist_element *rl, *rl2 = NULL;
@@ -1284,17 +1301,20 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 	/*
 	 * Determine the last lcn of the mft bitmap.  The allocated size of the
 	 * mft bitmap cannot be zero so we are ok to do this.
-	 * ntfs_find_vcn() returns the runlist locked on success.
 	 */
-	rl = ntfs_find_vcn(mftbmp_ni, (mftbmp_ni->allocated_size - 1) >>
-			vol->cluster_size_bits, TRUE);
+	down_write(&mftbmp_ni->runlist.lock);
+	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
+	ll = mftbmp_ni->allocated_size;
+	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+	rl = ntfs_attr_find_vcn_nolock(mftbmp_ni,
+			(ll - 1) >> vol->cluster_size_bits, TRUE);
 	if (unlikely(IS_ERR(rl) || !rl->length || rl->lcn < 0)) {
+		up_write(&mftbmp_ni->runlist.lock);
 		ntfs_error(vol->sb, "Failed to determine last allocated "
 				"cluster of mft bitmap attribute.");
-		if (!IS_ERR(rl)) {
-			up_write(&mftbmp_ni->runlist.lock);
+		if (!IS_ERR(rl))
 			ret = -EIO;
-		} else
+		else
 			ret = PTR_ERR(rl);
 		return ret;
 	}
@@ -1396,7 +1416,7 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 	BUG_ON(ll < rl2->vcn);
 	BUG_ON(ll >= rl2->vcn + rl2->length);
 	/* Get the size for the new mapping pairs array for this extent. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll);
+	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
 	if (unlikely(mp_size <= 0)) {
 		ntfs_error(vol->sb, "Get size for mapping pairs failed for "
 				"mft bitmap attribute extent.");
@@ -1418,6 +1438,8 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 		// TODO: Deal with this by moving this extent to a new mft
 		// record or by starting a new extent in a new mft record or by
 		// moving other attributes out of this mft record.
+		// Note: It will need to be a special mft record and if none of
+		// those are available it gets rather complicated...
 		ntfs_error(vol->sb, "Not enough space in this mft record to "
 				"accomodate extended mft bitmap attribute "
 				"extent.  Cannot handle this yet.");
@@ -1428,7 +1450,7 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 	/* Generate the mapping pairs array directly into the attr record. */
 	ret = ntfs_mapping_pairs_build(vol, (u8*)a +
 			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, rl2, ll, NULL);
+			mp_size, rl2, ll, -1, NULL);
 	if (unlikely(ret)) {
 		ntfs_error(vol->sb, "Failed to build mapping pairs array for "
 				"mft bitmap attribute.");
@@ -1458,9 +1480,11 @@ static int ntfs_mft_bitmap_extend_allocation_nolock(ntfs_volume *vol)
 		}
 		a = ctx->attr;
 	}
+	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
 	mftbmp_ni->allocated_size += vol->cluster_size;
 	a->data.non_resident.allocated_size =
 			cpu_to_sle64(mftbmp_ni->allocated_size);
+	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
 	/* Ensure the changes make it to disk. */
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
@@ -1476,7 +1500,9 @@ restore_undo_alloc:
 			0, ctx)) {
 		ntfs_error(vol->sb, "Failed to find last attribute extent of "
 				"mft bitmap attribute.%s", es);
+		write_lock_irqsave(&mftbmp_ni->size_lock, flags);
 		mftbmp_ni->allocated_size += vol->cluster_size;
+		write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
 		ntfs_attr_put_search_ctx(ctx);
 		unmap_mft_record(mft_ni);
 		up_write(&mftbmp_ni->runlist.lock);
@@ -1512,7 +1538,7 @@ undo_alloc:
 				a->data.non_resident.mapping_pairs_offset),
 				old_alen - le16_to_cpu(
 				a->data.non_resident.mapping_pairs_offset),
-				rl2, ll, NULL)) {
+				rl2, ll, -1, NULL)) {
 			ntfs_error(vol->sb, "Failed to restore mapping pairs "
 					"array.%s", es);
 			NVolSetErrors(vol);
@@ -1550,6 +1576,7 @@ undo_alloc:
 static int ntfs_mft_bitmap_extend_initialized_nolock(ntfs_volume *vol)
 {
 	s64 old_data_size, old_initialized_size;
+	unsigned long flags;
 	struct inode *mftbmp_vi;
 	ntfs_inode *mft_ni, *mftbmp_ni;
 	ntfs_attr_search_ctx *ctx;
@@ -1583,7 +1610,8 @@ static int ntfs_mft_bitmap_extend_initialized_nolock(ntfs_volume *vol)
 		goto put_err_out;
 	}
 	a = ctx->attr;
-	old_data_size = mftbmp_vi->i_size;
+	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
+	old_data_size = i_size_read(mftbmp_vi);
 	old_initialized_size = mftbmp_ni->initialized_size;
 	/*
 	 * We can simply update the initialized_size before filling the space
@@ -1593,11 +1621,12 @@ static int ntfs_mft_bitmap_extend_initialized_nolock(ntfs_volume *vol)
 	mftbmp_ni->initialized_size += 8;
 	a->data.non_resident.initialized_size =
 			cpu_to_sle64(mftbmp_ni->initialized_size);
-	if (mftbmp_ni->initialized_size > mftbmp_vi->i_size) {
-		mftbmp_vi->i_size = mftbmp_ni->initialized_size;
+	if (mftbmp_ni->initialized_size > old_data_size) {
+		i_size_write(mftbmp_vi, mftbmp_ni->initialized_size);
 		a->data.non_resident.data_size =
-				cpu_to_sle64(mftbmp_vi->i_size);
+				cpu_to_sle64(mftbmp_ni->initialized_size);
 	}
+	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
 	/* Ensure the changes make it to disk. */
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
@@ -1636,22 +1665,28 @@ unm_err_out:
 		goto err_out;
 	}
 	a = ctx->attr;
+	write_lock_irqsave(&mftbmp_ni->size_lock, flags);
 	mftbmp_ni->initialized_size = old_initialized_size;
 	a->data.non_resident.initialized_size =
 			cpu_to_sle64(old_initialized_size);
-	if (mftbmp_vi->i_size != old_data_size) {
-		mftbmp_vi->i_size = old_data_size;
+	if (i_size_read(mftbmp_vi) != old_data_size) {
+		i_size_write(mftbmp_vi, old_data_size);
 		a->data.non_resident.data_size = cpu_to_sle64(old_data_size);
 	}
+	write_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
 	ntfs_attr_put_search_ctx(ctx);
 	unmap_mft_record(mft_ni);
+#ifdef DEBUG
+	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
 	ntfs_debug("Restored status of mftbmp: allocated_size 0x%llx, "
 			"data_size 0x%llx, initialized_size 0x%llx.",
 			(long long)mftbmp_ni->allocated_size,
-			(long long)mftbmp_vi->i_size,
+			(long long)i_size_read(mftbmp_vi),
 			(long long)mftbmp_ni->initialized_size);
+	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+#endif /* DEBUG */
 err_out:
 	return ret;
 }
@@ -1679,7 +1714,8 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 {
 	LCN lcn;
 	VCN old_last_vcn;
-	s64 min_nr, nr, ll = 0;
+	s64 min_nr, nr, ll;
+	unsigned long flags;
 	ntfs_inode *mft_ni;
 	runlist_element *rl, *rl2;
 	ntfs_attr_search_ctx *ctx = NULL;
@@ -1695,23 +1731,25 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 	 * Determine the preferred allocation location, i.e. the last lcn of
 	 * the mft data attribute.  The allocated size of the mft data
 	 * attribute cannot be zero so we are ok to do this.
-	 * ntfs_find_vcn() returns the runlist locked on success.
 	 */
-	rl = ntfs_find_vcn(mft_ni, (mft_ni->allocated_size - 1) >>
-			vol->cluster_size_bits, TRUE);
+	down_write(&mft_ni->runlist.lock);
+	read_lock_irqsave(&mft_ni->size_lock, flags);
+	ll = mft_ni->allocated_size;
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
+	rl = ntfs_attr_find_vcn_nolock(mft_ni,
+			(ll - 1) >> vol->cluster_size_bits, TRUE);
 	if (unlikely(IS_ERR(rl) || !rl->length || rl->lcn < 0)) {
+		up_write(&mft_ni->runlist.lock);
 		ntfs_error(vol->sb, "Failed to determine last allocated "
 				"cluster of mft data attribute.");
-		if (!IS_ERR(rl)) {
-			up_write(&mft_ni->runlist.lock);
+		if (!IS_ERR(rl))
 			ret = -EIO;
-		} else
+		else
 			ret = PTR_ERR(rl);
 		return ret;
 	}
 	lcn = rl->lcn + rl->length;
-	ntfs_debug("Last lcn of mft data attribute is 0x%llx.",
-			(long long)lcn);
+	ntfs_debug("Last lcn of mft data attribute is 0x%llx.", (long long)lcn);
 	/* Minimum allocation is one mft record worth of clusters. */
 	min_nr = vol->mft_record_size >> vol->cluster_size_bits;
 	if (!min_nr)
@@ -1721,12 +1759,13 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 	if (!nr)
 		nr = min_nr;
 	/* Ensure we do not go above 2^32-1 mft records. */
-	if (unlikely((mft_ni->allocated_size +
-			(nr << vol->cluster_size_bits)) >>
+	read_lock_irqsave(&mft_ni->size_lock, flags);
+	ll = mft_ni->allocated_size;
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
+	if (unlikely((ll + (nr << vol->cluster_size_bits)) >>
 			vol->mft_record_size_bits >= (1ll << 32))) {
 		nr = min_nr;
-		if (unlikely((mft_ni->allocated_size +
-				(nr << vol->cluster_size_bits)) >>
+		if (unlikely((ll + (nr << vol->cluster_size_bits)) >>
 				vol->mft_record_size_bits >= (1ll << 32))) {
 			ntfs_warning(vol->sb, "Cannot allocate mft record "
 					"because the maximum number of inodes "
@@ -1772,7 +1811,7 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 		return PTR_ERR(rl);
 	}
 	mft_ni->runlist.rl = rl;
-	ntfs_debug("Allocated %lli clusters.", nr);
+	ntfs_debug("Allocated %lli clusters.", (long long)nr);
 	/* Find the last run in the new runlist. */
 	for (; rl[1].length; rl++)
 		;
@@ -1808,7 +1847,7 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 	BUG_ON(ll < rl2->vcn);
 	BUG_ON(ll >= rl2->vcn + rl2->length);
 	/* Get the size for the new mapping pairs array for this extent. */
-	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll);
+	mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
 	if (unlikely(mp_size <= 0)) {
 		ntfs_error(vol->sb, "Get size for mapping pairs failed for "
 				"mft data attribute extent.");
@@ -1832,7 +1871,11 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 		// moving other attributes out of this mft record.
 		// Note: Use the special reserved mft records and ensure that
 		// this extent is not required to find the mft record in
-		// question.
+		// question.  If no free special records left we would need to
+		// move an existing record away, insert ours in its place, and
+		// then place the moved record into the newly allocated space
+		// and we would then need to update all references to this mft
+		// record appropriately.  This is rather complicated...
 		ntfs_error(vol->sb, "Not enough space in this mft record to "
 				"accomodate extended mft data attribute "
 				"extent.  Cannot handle this yet.");
@@ -1843,7 +1886,7 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 	/* Generate the mapping pairs array directly into the attr record. */
 	ret = ntfs_mapping_pairs_build(vol, (u8*)a +
 			le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
-			mp_size, rl2, ll, NULL);
+			mp_size, rl2, ll, -1, NULL);
 	if (unlikely(ret)) {
 		ntfs_error(vol->sb, "Failed to build mapping pairs array of "
 				"mft data attribute.");
@@ -1875,9 +1918,11 @@ static int ntfs_mft_data_extend_allocation_nolock(ntfs_volume *vol)
 		}
 		a = ctx->attr;
 	}
+	write_lock_irqsave(&mft_ni->size_lock, flags);
 	mft_ni->allocated_size += nr << vol->cluster_size_bits;
 	a->data.non_resident.allocated_size =
 			cpu_to_sle64(mft_ni->allocated_size);
+	write_unlock_irqrestore(&mft_ni->size_lock, flags);
 	/* Ensure the changes make it to disk. */
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
@@ -1892,7 +1937,9 @@ restore_undo_alloc:
 			CASE_SENSITIVE, rl[1].vcn, NULL, 0, ctx)) {
 		ntfs_error(vol->sb, "Failed to find last attribute extent of "
 				"mft data attribute.%s", es);
+		write_lock_irqsave(&mft_ni->size_lock, flags);
 		mft_ni->allocated_size += nr << vol->cluster_size_bits;
+		write_unlock_irqrestore(&mft_ni->size_lock, flags);
 		ntfs_attr_put_search_ctx(ctx);
 		unmap_mft_record(mft_ni);
 		up_write(&mft_ni->runlist.lock);
@@ -1921,7 +1968,7 @@ undo_alloc:
 				a->data.non_resident.mapping_pairs_offset),
 				old_alen - le16_to_cpu(
 				a->data.non_resident.mapping_pairs_offset),
-				rl2, ll, NULL)) {
+				rl2, ll, -1, NULL)) {
 			ntfs_error(vol->sb, "Failed to restore mapping pairs "
 					"array.%s", es);
 			NVolSetErrors(vol);
@@ -1991,7 +2038,7 @@ static int ntfs_mft_record_layout(const ntfs_volume *vol, const s64 mft_no,
 				"reports this as corruption, please email "
 				"linux-ntfs-dev@lists.sourceforge.net stating "
 				"that you saw this message and that the "
-				"modified file system created was corrupt.  "
+				"modified filesystem created was corrupt.  "
 				"Thank you.");
 	}
 	/* Set the update sequence number to 1. */
@@ -2036,6 +2083,7 @@ static int ntfs_mft_record_layout(const ntfs_volume *vol, const s64 mft_no,
  */
 static int ntfs_mft_record_format(const ntfs_volume *vol, const s64 mft_no)
 {
+	loff_t i_size;
 	struct inode *mft_vi = vol->mft_ino;
 	struct page *page;
 	MFT_RECORD *m;
@@ -2051,10 +2099,11 @@ static int ntfs_mft_record_format(const ntfs_volume *vol, const s64 mft_no)
 	index = mft_no << vol->mft_record_size_bits >> PAGE_CACHE_SHIFT;
 	ofs = (mft_no << vol->mft_record_size_bits) & ~PAGE_CACHE_MASK;
 	/* The maximum valid index into the page cache for $MFT's data. */
-	end_index = mft_vi->i_size >> PAGE_CACHE_SHIFT;
+	i_size = i_size_read(mft_vi);
+	end_index = i_size >> PAGE_CACHE_SHIFT;
 	if (unlikely(index >= end_index)) {
 		if (unlikely(index > end_index || ofs + vol->mft_record_size >=
-				(mft_vi->i_size & ~PAGE_CACHE_MASK))) {
+				(i_size & ~PAGE_CACHE_MASK))) {
 			ntfs_error(vol->sb, "Tried to format non-existing mft "
 					"record 0x%llx.", (long long)mft_no);
 			return -ENOENT;
@@ -2188,6 +2237,7 @@ ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
 		ntfs_inode *base_ni, MFT_RECORD **mrec)
 {
 	s64 ll, bit, old_data_initialized, old_data_size;
+	unsigned long flags;
 	struct inode *vi;
 	struct page *page;
 	ntfs_inode *mft_ni, *mftbmp_ni, *ni;
@@ -2237,9 +2287,13 @@ ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
 	 * the first 24 mft records as they are special and whilst they may not
 	 * be in use, we do not allocate from them.
 	 */
+	read_lock_irqsave(&mft_ni->size_lock, flags);
 	ll = mft_ni->initialized_size >> vol->mft_record_size_bits;
-	if (mftbmp_ni->initialized_size << 3 > ll &&
-			mftbmp_ni->initialized_size > 3) {
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
+	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
+	old_data_initialized = mftbmp_ni->initialized_size;
+	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+	if (old_data_initialized << 3 > ll && old_data_initialized > 3) {
 		bit = ll;
 		if (bit < 24)
 			bit = 24;
@@ -2254,15 +2308,18 @@ ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
 	 * mft record that we can allocate.
 	 * Note: The smallest mft record we allocate is mft record 24.
 	 */
-	bit = mftbmp_ni->initialized_size << 3;
+	bit = old_data_initialized << 3;
 	if (unlikely(bit >= (1ll << 32)))
 		goto max_err_out;
+	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
+	old_data_size = mftbmp_ni->allocated_size;
 	ntfs_debug("Status of mftbmp before extension: allocated_size 0x%llx, "
 			"data_size 0x%llx, initialized_size 0x%llx.",
-			(long long)mftbmp_ni->allocated_size,
-			(long long)vol->mftbmp_ino->i_size,
-			(long long)mftbmp_ni->initialized_size);
-	if (mftbmp_ni->initialized_size + 8 > mftbmp_ni->allocated_size) {
+			(long long)old_data_size,
+			(long long)i_size_read(vol->mftbmp_ino),
+			(long long)old_data_initialized);
+	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+	if (old_data_initialized + 8 > old_data_size) {
 		/* Need to extend bitmap by one more cluster. */
 		ntfs_debug("mftbmp: initialized_size + 8 > allocated_size.");
 		err = ntfs_mft_bitmap_extend_allocation_nolock(vol);
@@ -2270,12 +2327,16 @@ ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
 			up_write(&vol->mftbmp_lock);
 			goto err_out;
 		}
+#ifdef DEBUG
+		read_lock_irqsave(&mftbmp_ni->size_lock, flags);
 		ntfs_debug("Status of mftbmp after allocation extension: "
 				"allocated_size 0x%llx, data_size 0x%llx, "
 				"initialized_size 0x%llx.",
 				(long long)mftbmp_ni->allocated_size,
-				(long long)vol->mftbmp_ino->i_size,
+				(long long)i_size_read(vol->mftbmp_ino),
 				(long long)mftbmp_ni->initialized_size);
+		read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+#endif /* DEBUG */
 	}
 	/*
 	 * We now have sufficient allocated space, extend the initialized_size
@@ -2287,12 +2348,16 @@ ntfs_inode *ntfs_mft_record_alloc(ntfs_volume *vol, const int mode,
 		up_write(&vol->mftbmp_lock);
 		goto err_out;
 	}
+#ifdef DEBUG
+	read_lock_irqsave(&mftbmp_ni->size_lock, flags);
 	ntfs_debug("Status of mftbmp after initialized extention: "
 			"allocated_size 0x%llx, data_size 0x%llx, "
 			"initialized_size 0x%llx.",
 			(long long)mftbmp_ni->allocated_size,
-			(long long)vol->mftbmp_ino->i_size,
+			(long long)i_size_read(vol->mftbmp_ino),
 			(long long)mftbmp_ni->initialized_size);
+	read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
+#endif /* DEBUG */
 	ntfs_debug("Found free record (#3), bit 0x%llx.", (long long)bit);
 found_free_rec:
 	/* @bit is the found free mft record, allocate it in the mft bitmap. */
@@ -2314,7 +2379,10 @@ have_alloc_rec:
 	 * parallel allocation could allocate the same mft record as this one.
 	 */
 	ll = (bit + 1) << vol->mft_record_size_bits;
-	if (ll <= mft_ni->initialized_size) {
+	read_lock_irqsave(&mft_ni->size_lock, flags);
+	old_data_initialized = mft_ni->initialized_size;
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
+	if (ll <= old_data_initialized) {
 		ntfs_debug("Allocated mft record already initialized.");
 		goto mft_rec_already_initialized;
 	}
@@ -2325,26 +2393,30 @@ have_alloc_rec:
 	 * actually traversed more than once when a freshly formatted volume is
 	 * first written to so it optimizes away nicely in the common case.
 	 */
+	read_lock_irqsave(&mft_ni->size_lock, flags);
 	ntfs_debug("Status of mft data before extension: "
 			"allocated_size 0x%llx, data_size 0x%llx, "
 			"initialized_size 0x%llx.",
 			(long long)mft_ni->allocated_size,
-			(long long)vol->mft_ino->i_size,
+			(long long)i_size_read(vol->mft_ino),
 			(long long)mft_ni->initialized_size);
 	while (ll > mft_ni->allocated_size) {
+		read_unlock_irqrestore(&mft_ni->size_lock, flags);
 		err = ntfs_mft_data_extend_allocation_nolock(vol);
 		if (unlikely(err)) {
 			ntfs_error(vol->sb, "Failed to extend mft data "
 					"allocation.");
 			goto undo_mftbmp_alloc_nolock;
 		}
+		read_lock_irqsave(&mft_ni->size_lock, flags);
 		ntfs_debug("Status of mft data after allocation extension: "
 				"allocated_size 0x%llx, data_size 0x%llx, "
 				"initialized_size 0x%llx.",
 				(long long)mft_ni->allocated_size,
-				(long long)vol->mft_ino->i_size,
+				(long long)i_size_read(vol->mft_ino),
 				(long long)mft_ni->initialized_size);
 	}
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
 	/*
 	 * Extend mft data initialized size (and data size of course) to reach
 	 * the allocated mft record, formatting the mft records allong the way.
@@ -2352,6 +2424,7 @@ have_alloc_rec:
 	 * needed by ntfs_mft_record_format().  We will update the attribute
 	 * record itself in one fell swoop later on.
 	 */
+	write_lock_irqsave(&mft_ni->size_lock, flags);
 	old_data_initialized = mft_ni->initialized_size;
 	old_data_size = vol->mft_ino->i_size;
 	while (ll > mft_ni->initialized_size) {
@@ -2360,8 +2433,9 @@ have_alloc_rec:
 		new_initialized_size = mft_ni->initialized_size +
 				vol->mft_record_size;
 		mft_no = mft_ni->initialized_size >> vol->mft_record_size_bits;
-		if (new_initialized_size > vol->mft_ino->i_size)
-			vol->mft_ino->i_size = new_initialized_size;
+		if (new_initialized_size > i_size_read(vol->mft_ino))
+			i_size_write(vol->mft_ino, new_initialized_size);
+		write_unlock_irqrestore(&mft_ni->size_lock, flags);
 		ntfs_debug("Initializing mft record 0x%llx.",
 				(long long)mft_no);
 		err = ntfs_mft_record_format(vol, mft_no);
@@ -2369,8 +2443,10 @@ have_alloc_rec:
 			ntfs_error(vol->sb, "Failed to format mft record.");
 			goto undo_data_init;
 		}
+		write_lock_irqsave(&mft_ni->size_lock, flags);
 		mft_ni->initialized_size = new_initialized_size;
 	}
+	write_unlock_irqrestore(&mft_ni->size_lock, flags);
 	record_formatted = TRUE;
 	/* Update the mft data attribute record to reflect the new sizes. */
 	m = map_mft_record(mft_ni);
@@ -2396,22 +2472,27 @@ have_alloc_rec:
 		goto undo_data_init;
 	}
 	a = ctx->attr;
+	read_lock_irqsave(&mft_ni->size_lock, flags);
 	a->data.non_resident.initialized_size =
 			cpu_to_sle64(mft_ni->initialized_size);
-	a->data.non_resident.data_size = cpu_to_sle64(vol->mft_ino->i_size);
+	a->data.non_resident.data_size =
+			cpu_to_sle64(i_size_read(vol->mft_ino));
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
 	/* Ensure the changes make it to disk. */
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
 	mark_mft_record_dirty(ctx->ntfs_ino);
 	ntfs_attr_put_search_ctx(ctx);
 	unmap_mft_record(mft_ni);
+	read_lock_irqsave(&mft_ni->size_lock, flags);
 	ntfs_debug("Status of mft data after mft record initialization: "
 			"allocated_size 0x%llx, data_size 0x%llx, "
 			"initialized_size 0x%llx.",
 			(long long)mft_ni->allocated_size,
-			(long long)vol->mft_ino->i_size,
+			(long long)i_size_read(vol->mft_ino),
 			(long long)mft_ni->initialized_size);
-	BUG_ON(vol->mft_ino->i_size > mft_ni->allocated_size);
-	BUG_ON(mft_ni->initialized_size > vol->mft_ino->i_size);
+	BUG_ON(i_size_read(vol->mft_ino) > mft_ni->allocated_size);
+	BUG_ON(mft_ni->initialized_size > i_size_read(vol->mft_ino));
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
 mft_rec_already_initialized:
 	/*
 	 * We can finally drop the mft bitmap lock as the mft data attribute
@@ -2652,8 +2733,10 @@ mft_rec_already_initialized:
 	*mrec = m;
 	return ni;
 undo_data_init:
+	write_lock_irqsave(&mft_ni->size_lock, flags);
 	mft_ni->initialized_size = old_data_initialized;
-	vol->mft_ino->i_size = old_data_size;
+	i_size_write(vol->mft_ino, old_data_size);
+	write_unlock_irqrestore(&mft_ni->size_lock, flags);
 	goto undo_mftbmp_alloc_nolock;
 undo_mftbmp_alloc:
 	down_write(&vol->mftbmp_lock);
diff --git a/fs/ntfs/namei.c b/fs/ntfs/namei.c
index 7c7e13b..351dbc3 100644
--- a/fs/ntfs/namei.c
+++ b/fs/ntfs/namei.c
@@ -153,8 +153,7 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 			ntfs_error(vol->sb, "ntfs_iget(0x%lx) failed with "
 					"error code %li.", dent_ino,
 					PTR_ERR(dent_inode));
-		if (name)
-			kfree(name);
+		kfree(name);
 		/* Return the error code. */
 		return (struct dentry *)dent_inode;
 	}
@@ -380,7 +379,7 @@ struct inode_operations ntfs_dir_inode_ops = {
  * Return the dentry of the parent directory on success or the error code on
  * error (IS_ERR() is true).
  */
-struct dentry *ntfs_get_parent(struct dentry *child_dent)
+static struct dentry *ntfs_get_parent(struct dentry *child_dent)
 {
 	struct inode *vi = child_dent->d_inode;
 	ntfs_inode *ni = NTFS_I(vi);
@@ -465,7 +464,7 @@ try_next:
  *
  * Return the dentry on success or the error code on error (IS_ERR() is true).
  */
-struct dentry *ntfs_get_dentry(struct super_block *sb, void *fh)
+static struct dentry *ntfs_get_dentry(struct super_block *sb, void *fh)
 {
 	struct inode *vi;
 	struct dentry *dent;
@@ -496,3 +495,30 @@ struct dentry *ntfs_get_dentry(struct super_block *sb, void *fh)
 	ntfs_debug("Done for inode 0x%lx, generation 0x%x.", ino, gen);
 	return dent;
 }
+
+/**
+ * Export operations allowing NFS exporting of mounted NTFS partitions.
+ *
+ * We use the default ->decode_fh() and ->encode_fh() for now.  Note that they
+ * use 32 bits to store the inode number which is an unsigned long so on 64-bit
+ * architectures is usually 64 bits so it would all fail horribly on huge
+ * volumes.  I guess we need to define our own encode and decode fh functions
+ * that store 64-bit inode numbers at some point but for now we will ignore the
+ * problem...
+ *
+ * We also use the default ->get_name() helper (used by ->decode_fh() via
+ * fs/exportfs/expfs.c::find_exported_dentry()) as that is completely fs
+ * independent.
+ *
+ * The default ->get_parent() just returns -EACCES so we have to provide our
+ * own and the default ->get_dentry() is incompatible with NTFS due to not
+ * allowing the inode number 0 which is used in NTFS for the system file $MFT
+ * and due to using iget() whereas NTFS needs ntfs_iget().
+ */
+struct export_operations ntfs_export_ops = {
+	.get_parent	= ntfs_get_parent,	/* Find the parent of a given
+						   directory. */
+	.get_dentry	= ntfs_get_dentry,	/* Find a dentry for the inode
+						   given a file handle
+						   sub-fragment. */
+};
diff --git a/fs/ntfs/ntfs.h b/fs/ntfs/ntfs.h
index 720ffb7..446b501 100644
--- a/fs/ntfs/ntfs.h
+++ b/fs/ntfs/ntfs.h
@@ -2,7 +2,7 @@
  * ntfs.h - Defines for NTFS Linux kernel driver. Part of the Linux-NTFS
  *	    project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (C) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -31,6 +31,7 @@
 #include <linux/fs.h>
 #include <linux/nls.h>
 #include <linux/smp.h>
+#include <linux/pagemap.h>
 
 #include "types.h"
 #include "volume.h"
@@ -41,6 +42,9 @@ typedef enum {
 	NTFS_BLOCK_SIZE_BITS	= 9,
 	NTFS_SB_MAGIC		= 0x5346544e,	/* 'NTFS' */
 	NTFS_MAX_NAME_LEN	= 255,
+	NTFS_MAX_ATTR_NAME_LEN	= 255,
+	NTFS_MAX_CLUSTER_SIZE	= 64 * 1024,	/* 64kiB */
+	NTFS_MAX_PAGES_PER_CLUSTER = NTFS_MAX_CLUSTER_SIZE / PAGE_CACHE_SIZE,
 } NTFS_CONSTANTS;
 
 /* Global variables. */
@@ -65,6 +69,8 @@ extern struct inode_operations ntfs_dir_inode_ops;
 extern struct  file_operations ntfs_empty_file_ops;
 extern struct inode_operations ntfs_empty_inode_ops;
 
+extern struct export_operations ntfs_export_ops;
+
 /**
  * NTFS_SB - return the ntfs volume given a vfs super block
  * @sb:		VFS super block
diff --git a/fs/ntfs/runlist.c b/fs/ntfs/runlist.c
index 8438fb1..758855b 100644
--- a/fs/ntfs/runlist.c
+++ b/fs/ntfs/runlist.c
@@ -1,7 +1,7 @@
 /**
  * runlist.c - NTFS runlist handling code.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -59,7 +59,7 @@ static inline void ntfs_rl_mc(runlist_element *dstbase, int dst,
  *
  * As the runlists grow, more memory will be required.  To prevent the
  * kernel having to allocate and reallocate large numbers of small bits of
- * memory, this function returns and entire page of memory.
+ * memory, this function returns an entire page of memory.
  *
  * It is up to the caller to serialize access to the runlist @rl.
  *
@@ -113,8 +113,11 @@ static inline BOOL ntfs_are_rl_mergeable(runlist_element *dst,
 	BUG_ON(!dst);
 	BUG_ON(!src);
 
-	if ((dst->lcn < 0) || (src->lcn < 0))     /* Are we merging holes? */
+	if ((dst->lcn < 0) || (src->lcn < 0)) {   /* Are we merging holes? */
+		if (dst->lcn == LCN_HOLE && src->lcn == LCN_HOLE)
+			return TRUE;
 		return FALSE;
+	}
 	if ((dst->lcn + dst->length) != src->lcn) /* Are the runs contiguous? */
 		return FALSE;
 	if ((dst->vcn + dst->length) != src->vcn) /* Are the runs misaligned? */
@@ -855,30 +858,42 @@ mpa_err:
 	if (!attr->data.non_resident.lowest_vcn) {
 		VCN max_cluster;
 
-		max_cluster = (sle64_to_cpu(
+		max_cluster = ((sle64_to_cpu(
 				attr->data.non_resident.allocated_size) +
 				vol->cluster_size - 1) >>
-				vol->cluster_size_bits;
+				vol->cluster_size_bits) - 1;
 		/*
-		 * If there is a difference between the highest_vcn and the
-		 * highest cluster, the runlist is either corrupt or, more
-		 * likely, there are more extents following this one.
+		 * A highest_vcn of zero means this is a single extent
+		 * attribute so simply terminate the runlist with LCN_ENOENT).
 		 */
-		if (deltaxcn < --max_cluster) {
-			ntfs_debug("More extents to follow; deltaxcn = 0x%llx, "
-					"max_cluster = 0x%llx",
-					(unsigned long long)deltaxcn,
-					(unsigned long long)max_cluster);
-			rl[rlpos].vcn = vcn;
-			vcn += rl[rlpos].length = max_cluster - deltaxcn;
-			rl[rlpos].lcn = LCN_RL_NOT_MAPPED;
-			rlpos++;
-		} else if (unlikely(deltaxcn > max_cluster)) {
-			ntfs_error(vol->sb, "Corrupt attribute. deltaxcn = "
-					"0x%llx, max_cluster = 0x%llx",
-					(unsigned long long)deltaxcn,
-					(unsigned long long)max_cluster);
-			goto mpa_err;
+		if (deltaxcn) {
+			/*
+			 * If there is a difference between the highest_vcn and
+			 * the highest cluster, the runlist is either corrupt
+			 * or, more likely, there are more extents following
+			 * this one.
+			 */
+			if (deltaxcn < max_cluster) {
+				ntfs_debug("More extents to follow; deltaxcn "
+						"= 0x%llx, max_cluster = "
+						"0x%llx",
+						(unsigned long long)deltaxcn,
+						(unsigned long long)
+						max_cluster);
+				rl[rlpos].vcn = vcn;
+				vcn += rl[rlpos].length = max_cluster -
+						deltaxcn;
+				rl[rlpos].lcn = LCN_RL_NOT_MAPPED;
+				rlpos++;
+			} else if (unlikely(deltaxcn > max_cluster)) {
+				ntfs_error(vol->sb, "Corrupt attribute.  "
+						"deltaxcn = 0x%llx, "
+						"max_cluster = 0x%llx",
+						(unsigned long long)deltaxcn,
+						(unsigned long long)
+						max_cluster);
+				goto mpa_err;
+			}
 		}
 		rl[rlpos].lcn = LCN_ENOENT;
 	} else /* Not the base extent. There may be more extents to follow. */
@@ -918,17 +933,18 @@ err_out:
  *
  * It is up to the caller to serialize access to the runlist @rl.
  *
- * Since lcns must be >= 0, we use negative return values with special meaning:
+ * Since lcns must be >= 0, we use negative return codes with special meaning:
  *
- * Return value			Meaning / Description
+ * Return code		Meaning / Description
  * ==================================================
- *  -1 = LCN_HOLE		Hole / not allocated on disk.
- *  -2 = LCN_RL_NOT_MAPPED	This is part of the runlist which has not been
- *				inserted into the runlist yet.
- *  -3 = LCN_ENOENT		There is no such vcn in the attribute.
+ *  LCN_HOLE		Hole / not allocated on disk.
+ *  LCN_RL_NOT_MAPPED	This is part of the runlist which has not been
+ *			inserted into the runlist yet.
+ *  LCN_ENOENT		There is no such vcn in the attribute.
  *
  * Locking: - The caller must have locked the runlist (for reading or writing).
- *	    - This function does not touch the lock.
+ *	    - This function does not touch the lock, nor does it modify the
+ *	      runlist.
  */
 LCN ntfs_rl_vcn_to_lcn(const runlist_element *rl, const VCN vcn)
 {
@@ -964,6 +980,39 @@ LCN ntfs_rl_vcn_to_lcn(const runlist_element *rl, const VCN vcn)
 	return LCN_ENOENT;
 }
 
+#ifdef NTFS_RW
+
+/**
+ * ntfs_rl_find_vcn_nolock - find a vcn in a runlist
+ * @rl:		runlist to search
+ * @vcn:	vcn to find
+ *
+ * Find the virtual cluster number @vcn in the runlist @rl and return the
+ * address of the runlist element containing the @vcn on success.
+ *
+ * Return NULL if @rl is NULL or @vcn is in an unmapped part/out of bounds of
+ * the runlist.
+ *
+ * Locking: The runlist must be locked on entry.
+ */
+runlist_element *ntfs_rl_find_vcn_nolock(runlist_element *rl, const VCN vcn)
+{
+	BUG_ON(vcn < 0);
+	if (unlikely(!rl || vcn < rl[0].vcn))
+		return NULL;
+	while (likely(rl->length)) {
+		if (unlikely(vcn < rl[1].vcn)) {
+			if (likely(rl->lcn >= LCN_HOLE))
+				return rl;
+			return NULL;
+		}
+		rl++;
+	}
+	if (likely(rl->lcn == LCN_ENOENT))
+		return rl;
+	return NULL;
+}
+
 /**
  * ntfs_get_nr_significant_bytes - get number of bytes needed to store a number
  * @n:		number for which to get the number of bytes for
@@ -999,10 +1048,17 @@ static inline int ntfs_get_nr_significant_bytes(const s64 n)
  * ntfs_get_size_for_mapping_pairs - get bytes needed for mapping pairs array
  * @vol:	ntfs volume (needed for the ntfs version)
  * @rl:		locked runlist to determine the size of the mapping pairs of
- * @start_vcn:	vcn at which to start the mapping pairs array
+ * @first_vcn:	first vcn which to include in the mapping pairs array
+ * @last_vcn:	last vcn which to include in the mapping pairs array
  *
  * Walk the locked runlist @rl and calculate the size in bytes of the mapping
- * pairs array corresponding to the runlist @rl, starting at vcn @start_vcn.
+ * pairs array corresponding to the runlist @rl, starting at vcn @first_vcn and
+ * finishing with vcn @last_vcn.
+ *
+ * A @last_vcn of -1 means end of runlist and in that case the size of the
+ * mapping pairs array corresponding to the runlist starting at vcn @first_vcn
+ * and finishing at the end of the runlist is determined.
+ *
  * This for example allows us to allocate a buffer of the right size when
  * building the mapping pairs array.
  *
@@ -1018,34 +1074,50 @@ static inline int ntfs_get_nr_significant_bytes(const s64 n)
  *	    remains locked throughout, and is left locked upon return.
  */
 int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
-		const runlist_element *rl, const VCN start_vcn)
+		const runlist_element *rl, const VCN first_vcn,
+		const VCN last_vcn)
 {
 	LCN prev_lcn;
 	int rls;
+	BOOL the_end = FALSE;
 
-	BUG_ON(start_vcn < 0);
+	BUG_ON(first_vcn < 0);
+	BUG_ON(last_vcn < -1);
+	BUG_ON(last_vcn >= 0 && first_vcn > last_vcn);
 	if (!rl) {
-		BUG_ON(start_vcn);
+		BUG_ON(first_vcn);
+		BUG_ON(last_vcn > 0);
 		return 1;
 	}
-	/* Skip to runlist element containing @start_vcn. */
-	while (rl->length && start_vcn >= rl[1].vcn)
+	/* Skip to runlist element containing @first_vcn. */
+	while (rl->length && first_vcn >= rl[1].vcn)
 		rl++;
-	if ((!rl->length && start_vcn > rl->vcn) || start_vcn < rl->vcn)
+	if (unlikely((!rl->length && first_vcn > rl->vcn) ||
+			first_vcn < rl->vcn))
 		return -EINVAL;
 	prev_lcn = 0;
 	/* Always need the termining zero byte. */
 	rls = 1;
 	/* Do the first partial run if present. */
-	if (start_vcn > rl->vcn) {
-		s64 delta;
+	if (first_vcn > rl->vcn) {
+		s64 delta, length = rl->length;
 
 		/* We know rl->length != 0 already. */
-		if (rl->length < 0 || rl->lcn < LCN_HOLE)
+		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
 			goto err_out;
-		delta = start_vcn - rl->vcn;
+		/*
+		 * If @stop_vcn is given and finishes inside this run, cap the
+		 * run length.
+		 */
+		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
+			s64 s1 = last_vcn + 1;
+			if (unlikely(rl[1].vcn > s1))
+				length = s1 - rl->vcn;
+			the_end = TRUE;
+		}
+		delta = first_vcn - rl->vcn;
 		/* Header byte + length. */
-		rls += 1 + ntfs_get_nr_significant_bytes(rl->length - delta);
+		rls += 1 + ntfs_get_nr_significant_bytes(length - delta);
 		/*
 		 * If the logical cluster number (lcn) denotes a hole and we
 		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
@@ -1053,9 +1125,9 @@ int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
 		 * Note: this assumes that on NTFS 1.2-, holes are stored with
 		 * an lcn of -1 and not a delta_lcn of -1 (unless both are -1).
 		 */
-		if (rl->lcn >= 0 || vol->major_ver < 3) {
+		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
 			prev_lcn = rl->lcn;
-			if (rl->lcn >= 0)
+			if (likely(rl->lcn >= 0))
 				prev_lcn += delta;
 			/* Change in lcn. */
 			rls += ntfs_get_nr_significant_bytes(prev_lcn);
@@ -1064,11 +1136,23 @@ int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
 		rl++;
 	}
 	/* Do the full runs. */
-	for (; rl->length; rl++) {
-		if (rl->length < 0 || rl->lcn < LCN_HOLE)
+	for (; rl->length && !the_end; rl++) {
+		s64 length = rl->length;
+
+		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
 			goto err_out;
+		/*
+		 * If @stop_vcn is given and finishes inside this run, cap the
+		 * run length.
+		 */
+		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
+			s64 s1 = last_vcn + 1;
+			if (unlikely(rl[1].vcn > s1))
+				length = s1 - rl->vcn;
+			the_end = TRUE;
+		}
 		/* Header byte + length. */
-		rls += 1 + ntfs_get_nr_significant_bytes(rl->length);
+		rls += 1 + ntfs_get_nr_significant_bytes(length);
 		/*
 		 * If the logical cluster number (lcn) denotes a hole and we
 		 * are on NTFS 3.0+, we don't store it at all, i.e. we need
@@ -1076,7 +1160,7 @@ int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
 		 * Note: this assumes that on NTFS 1.2-, holes are stored with
 		 * an lcn of -1 and not a delta_lcn of -1 (unless both are -1).
 		 */
-		if (rl->lcn >= 0 || vol->major_ver < 3) {
+		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
 			/* Change in lcn. */
 			rls += ntfs_get_nr_significant_bytes(rl->lcn -
 					prev_lcn);
@@ -1119,7 +1203,7 @@ static inline int ntfs_write_significant_bytes(s8 *dst, const s8 *dst_max,
 
 	i = 0;
 	do {
-		if (dst > dst_max)
+		if (unlikely(dst > dst_max))
 			goto err_out;
 		*dst++ = l & 0xffll;
 		l >>= 8;
@@ -1128,12 +1212,12 @@ static inline int ntfs_write_significant_bytes(s8 *dst, const s8 *dst_max,
 	j = (n >> 8 * (i - 1)) & 0xff;
 	/* If the sign bit is wrong, we need an extra byte. */
 	if (n < 0 && j >= 0) {
-		if (dst > dst_max)
+		if (unlikely(dst > dst_max))
 			goto err_out;
 		i++;
 		*dst = (s8)-1;
 	} else if (n > 0 && j < 0) {
-		if (dst > dst_max)
+		if (unlikely(dst > dst_max))
 			goto err_out;
 		i++;
 		*dst = (s8)0;
@@ -1149,13 +1233,18 @@ err_out:
  * @dst:	destination buffer to which to write the mapping pairs array
  * @dst_len:	size of destination buffer @dst in bytes
  * @rl:		locked runlist for which to build the mapping pairs array
- * @start_vcn:	vcn at which to start the mapping pairs array
+ * @first_vcn:	first vcn which to include in the mapping pairs array
+ * @last_vcn:	last vcn which to include in the mapping pairs array
  * @stop_vcn:	first vcn outside destination buffer on success or -ENOSPC
  *
  * Create the mapping pairs array from the locked runlist @rl, starting at vcn
- * @start_vcn and save the array in @dst.  @dst_len is the size of @dst in
- * bytes and it should be at least equal to the value obtained by calling
- * ntfs_get_size_for_mapping_pairs().
+ * @first_vcn and finishing with vcn @last_vcn and save the array in @dst.
+ * @dst_len is the size of @dst in bytes and it should be at least equal to the
+ * value obtained by calling ntfs_get_size_for_mapping_pairs().
+ *
+ * A @last_vcn of -1 means end of runlist and in that case the mapping pairs
+ * array corresponding to the runlist starting at vcn @first_vcn and finishing
+ * at the end of the runlist is created.
  *
  * If @rl is NULL, just write a single terminator byte to @dst.
  *
@@ -1164,7 +1253,7 @@ err_out:
  * been filled with all the mapping pairs that will fit, thus it can be treated
  * as partial success, in that a new attribute extent needs to be created or
  * the next extent has to be used and the mapping pairs build has to be
- * continued with @start_vcn set to *@stop_vcn.
+ * continued with @first_vcn set to *@stop_vcn.
  *
  * Return 0 on success and -errno on error.  The following error codes are
  * defined:
@@ -1178,27 +1267,32 @@ err_out:
  */
 int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 		const int dst_len, const runlist_element *rl,
-		const VCN start_vcn, VCN *const stop_vcn)
+		const VCN first_vcn, const VCN last_vcn, VCN *const stop_vcn)
 {
 	LCN prev_lcn;
 	s8 *dst_max, *dst_next;
 	int err = -ENOSPC;
+	BOOL the_end = FALSE;
 	s8 len_len, lcn_len;
 
-	BUG_ON(start_vcn < 0);
+	BUG_ON(first_vcn < 0);
+	BUG_ON(last_vcn < -1);
+	BUG_ON(last_vcn >= 0 && first_vcn > last_vcn);
 	BUG_ON(dst_len < 1);
 	if (!rl) {
-		BUG_ON(start_vcn);
+		BUG_ON(first_vcn);
+		BUG_ON(last_vcn > 0);
 		if (stop_vcn)
 			*stop_vcn = 0;
 		/* Terminator byte. */
 		*dst = 0;
 		return 0;
 	}
-	/* Skip to runlist element containing @start_vcn. */
-	while (rl->length && start_vcn >= rl[1].vcn)
+	/* Skip to runlist element containing @first_vcn. */
+	while (rl->length && first_vcn >= rl[1].vcn)
 		rl++;
-	if ((!rl->length && start_vcn > rl->vcn) || start_vcn < rl->vcn)
+	if (unlikely((!rl->length && first_vcn > rl->vcn) ||
+			first_vcn < rl->vcn))
 		return -EINVAL;
 	/*
 	 * @dst_max is used for bounds checking in
@@ -1207,17 +1301,27 @@ int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 	dst_max = dst + dst_len - 1;
 	prev_lcn = 0;
 	/* Do the first partial run if present. */
-	if (start_vcn > rl->vcn) {
-		s64 delta;
+	if (first_vcn > rl->vcn) {
+		s64 delta, length = rl->length;
 
 		/* We know rl->length != 0 already. */
-		if (rl->length < 0 || rl->lcn < LCN_HOLE)
+		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
 			goto err_out;
-		delta = start_vcn - rl->vcn;
+		/*
+		 * If @stop_vcn is given and finishes inside this run, cap the
+		 * run length.
+		 */
+		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
+			s64 s1 = last_vcn + 1;
+			if (unlikely(rl[1].vcn > s1))
+				length = s1 - rl->vcn;
+			the_end = TRUE;
+		}
+		delta = first_vcn - rl->vcn;
 		/* Write length. */
 		len_len = ntfs_write_significant_bytes(dst + 1, dst_max,
-				rl->length - delta);
-		if (len_len < 0)
+				length - delta);
+		if (unlikely(len_len < 0))
 			goto size_err;
 		/*
 		 * If the logical cluster number (lcn) denotes a hole and we
@@ -1228,19 +1332,19 @@ int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 		 * case on NT4. - We assume that we just need to write the lcn
 		 * change until someone tells us otherwise... (AIA)
 		 */
-		if (rl->lcn >= 0 || vol->major_ver < 3) {
+		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
 			prev_lcn = rl->lcn;
-			if (rl->lcn >= 0)
+			if (likely(rl->lcn >= 0))
 				prev_lcn += delta;
 			/* Write change in lcn. */
 			lcn_len = ntfs_write_significant_bytes(dst + 1 +
 					len_len, dst_max, prev_lcn);
-			if (lcn_len < 0)
+			if (unlikely(lcn_len < 0))
 				goto size_err;
 		} else
 			lcn_len = 0;
 		dst_next = dst + len_len + lcn_len + 1;
-		if (dst_next > dst_max)
+		if (unlikely(dst_next > dst_max))
 			goto size_err;
 		/* Update header byte. */
 		*dst = lcn_len << 4 | len_len;
@@ -1250,13 +1354,25 @@ int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 		rl++;
 	}
 	/* Do the full runs. */
-	for (; rl->length; rl++) {
-		if (rl->length < 0 || rl->lcn < LCN_HOLE)
+	for (; rl->length && !the_end; rl++) {
+		s64 length = rl->length;
+
+		if (unlikely(length < 0 || rl->lcn < LCN_HOLE))
 			goto err_out;
+		/*
+		 * If @stop_vcn is given and finishes inside this run, cap the
+		 * run length.
+		 */
+		if (unlikely(last_vcn >= 0 && rl[1].vcn > last_vcn)) {
+			s64 s1 = last_vcn + 1;
+			if (unlikely(rl[1].vcn > s1))
+				length = s1 - rl->vcn;
+			the_end = TRUE;
+		}
 		/* Write length. */
 		len_len = ntfs_write_significant_bytes(dst + 1, dst_max,
-				rl->length);
-		if (len_len < 0)
+				length);
+		if (unlikely(len_len < 0))
 			goto size_err;
 		/*
 		 * If the logical cluster number (lcn) denotes a hole and we
@@ -1267,17 +1383,17 @@ int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 		 * case on NT4. - We assume that we just need to write the lcn
 		 * change until someone tells us otherwise... (AIA)
 		 */
-		if (rl->lcn >= 0 || vol->major_ver < 3) {
+		if (likely(rl->lcn >= 0 || vol->major_ver < 3)) {
 			/* Write change in lcn. */
 			lcn_len = ntfs_write_significant_bytes(dst + 1 +
 					len_len, dst_max, rl->lcn - prev_lcn);
-			if (lcn_len < 0)
+			if (unlikely(lcn_len < 0))
 				goto size_err;
 			prev_lcn = rl->lcn;
 		} else
 			lcn_len = 0;
 		dst_next = dst + len_len + lcn_len + 1;
-		if (dst_next > dst_max)
+		if (unlikely(dst_next > dst_max))
 			goto size_err;
 		/* Update header byte. */
 		*dst = lcn_len << 4 | len_len;
@@ -1436,3 +1552,5 @@ int ntfs_rl_truncate_nolock(const ntfs_volume *vol, runlist *const runlist,
 	ntfs_debug("Done.");
 	return 0;
 }
+
+#endif /* NTFS_RW */
diff --git a/fs/ntfs/runlist.h b/fs/ntfs/runlist.h
index 7107fde..aa0ee65 100644
--- a/fs/ntfs/runlist.h
+++ b/fs/ntfs/runlist.h
@@ -2,7 +2,7 @@
  * runlist.h - Defines for runlist handling in NTFS Linux kernel driver.
  *	       Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -66,6 +66,8 @@ typedef enum {
 	LCN_HOLE		= -1,	/* Keep this as highest value or die! */
 	LCN_RL_NOT_MAPPED	= -2,
 	LCN_ENOENT		= -3,
+	LCN_ENOMEM		= -4,
+	LCN_EIO			= -5,
 } LCN_SPECIAL_VALUES;
 
 extern runlist_element *ntfs_runlists_merge(runlist_element *drl,
@@ -76,14 +78,22 @@ extern runlist_element *ntfs_mapping_pairs_decompress(const ntfs_volume *vol,
 
 extern LCN ntfs_rl_vcn_to_lcn(const runlist_element *rl, const VCN vcn);
 
+#ifdef NTFS_RW
+
+extern runlist_element *ntfs_rl_find_vcn_nolock(runlist_element *rl,
+		const VCN vcn);
+
 extern int ntfs_get_size_for_mapping_pairs(const ntfs_volume *vol,
-		const runlist_element *rl, const VCN start_vcn);
+		const runlist_element *rl, const VCN first_vcn,
+		const VCN last_vcn);
 
 extern int ntfs_mapping_pairs_build(const ntfs_volume *vol, s8 *dst,
 		const int dst_len, const runlist_element *rl,
-		const VCN start_vcn, VCN *const stop_vcn);
+		const VCN first_vcn, const VCN last_vcn, VCN *const stop_vcn);
 
 extern int ntfs_rl_truncate_nolock(const ntfs_volume *vol,
 		runlist *const runlist, const s64 new_length);
 
+#endif /* NTFS_RW */
+
 #endif /* _LINUX_NTFS_RUNLIST_H */
diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c
index 212a3d0..41aa8eb 100644
--- a/fs/ntfs/super.c
+++ b/fs/ntfs/super.c
@@ -1,7 +1,7 @@
 /*
  * super.c - NTFS kernel super block handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2001,2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -34,14 +34,16 @@
 #include "sysctl.h"
 #include "logfile.h"
 #include "quota.h"
+#include "usnjrnl.h"
 #include "dir.h"
 #include "debug.h"
 #include "index.h"
 #include "aops.h"
+#include "layout.h"
 #include "malloc.h"
 #include "ntfs.h"
 
-/* Number of mounted file systems which have compression enabled. */
+/* Number of mounted filesystems which have compression enabled. */
 static unsigned long ntfs_nr_compression_users;
 
 /* A global default upcase table and a corresponding reference count. */
@@ -102,7 +104,7 @@ static BOOL parse_options(ntfs_volume *vol, char *opt)
 	gid_t gid = (gid_t)-1;
 	mode_t fmask = (mode_t)-1, dmask = (mode_t)-1;
 	int mft_zone_multiplier = -1, on_errors = -1;
-	int show_sys_files = -1, case_sensitive = -1;
+	int show_sys_files = -1, case_sensitive = -1, disable_sparse = -1;
 	struct nls_table *nls_map = NULL, *old_nls;
 
 	/* I am lazy... (-8 */
@@ -162,6 +164,7 @@ static BOOL parse_options(ntfs_volume *vol, char *opt)
 		else NTFS_GETOPT_WITH_DEFAULT("sloppy", sloppy, TRUE)
 		else NTFS_GETOPT_BOOL("show_sys_files", show_sys_files)
 		else NTFS_GETOPT_BOOL("case_sensitive", case_sensitive)
+		else NTFS_GETOPT_BOOL("disable_sparse", disable_sparse)
 		else NTFS_GETOPT_OPTIONS_ARRAY("errors", on_errors,
 				on_errors_arr)
 		else if (!strcmp(p, "posix") || !strcmp(p, "show_inodes"))
@@ -291,6 +294,21 @@ no_mount_options:
 		else
 			NVolClearCaseSensitive(vol);
 	}
+	if (disable_sparse != -1) {
+		if (disable_sparse)
+			NVolClearSparseEnabled(vol);
+		else {
+			if (!NVolSparseEnabled(vol) &&
+					vol->major_ver && vol->major_ver < 3)
+				ntfs_warning(vol->sb, "Not enabling sparse "
+						"support due to NTFS volume "
+						"version %i.%i (need at least "
+						"version 3.0).", vol->major_ver,
+						vol->minor_ver);
+			else
+				NVolSetSparseEnabled(vol);
+		}
+	}
 	return TRUE;
 needs_arg:
 	ntfs_error(vol->sb, "The %s option requires an argument.", p);
@@ -480,6 +498,12 @@ static int ntfs_remount(struct super_block *sb, int *flags, char *opt)
 			NVolSetErrors(vol);
 			return -EROFS;
 		}
+		if (!ntfs_stamp_usnjrnl(vol)) {
+			ntfs_error(sb, "Failed to stamp transation log "
+					"($UsnJrnl)%s", es);
+			NVolSetErrors(vol);
+			return -EROFS;
+		}
 	} else if (!(sb->s_flags & MS_RDONLY) && (*flags & MS_RDONLY)) {
 		/* Remounting read-only. */
 		if (!NVolErrors(vol)) {
@@ -516,16 +540,19 @@ static BOOL is_boot_sector_ntfs(const struct super_block *sb,
 {
 	/*
 	 * Check that checksum == sum of u32 values from b to the checksum
-	 * field. If checksum is zero, no checking is done.
+	 * field.  If checksum is zero, no checking is done.  We will work when
+	 * the checksum test fails, since some utilities update the boot sector
+	 * ignoring the checksum which leaves the checksum out-of-date.  We
+	 * report a warning if this is the case.
 	 */
-	if ((void*)b < (void*)&b->checksum && b->checksum) {
+	if ((void*)b < (void*)&b->checksum && b->checksum && !silent) {
 		le32 *u;
 		u32 i;
 
 		for (i = 0, u = (le32*)b; u < (le32*)(&b->checksum); ++u)
 			i += le32_to_cpup(u);
 		if (le32_to_cpu(b->checksum) != i)
-			goto not_ntfs;
+			ntfs_warning(sb, "Invalid boot sector checksum.");
 	}
 	/* Check OEMidentifier is "NTFS    " */
 	if (b->oem_id != magicNTFS)
@@ -541,9 +568,9 @@ static BOOL is_boot_sector_ntfs(const struct super_block *sb,
 	default:
 		goto not_ntfs;
 	}
-	/* Check the cluster size is not above 65536 bytes. */
+	/* Check the cluster size is not above the maximum (64kiB). */
 	if ((u32)le16_to_cpu(b->bpb.bytes_per_sector) *
-			b->bpb.sectors_per_cluster > 0x10000)
+			b->bpb.sectors_per_cluster > NTFS_MAX_CLUSTER_SIZE)
 		goto not_ntfs;
 	/* Check reserved/unused fields are really zero. */
 	if (le16_to_cpu(b->bpb.reserved_sectors) ||
@@ -575,7 +602,7 @@ static BOOL is_boot_sector_ntfs(const struct super_block *sb,
 	 * many BIOSes will refuse to boot from a bootsector if the magic is
 	 * incorrect, so we emit a warning.
 	 */
-	if (!silent && b->end_of_sector_marker != cpu_to_le16(0xaa55))
+	if (!silent && b->end_of_sector_marker != const_cpu_to_le16(0xaa55))
 		ntfs_warning(sb, "Invalid end of sector marker.");
 	return TRUE;
 not_ntfs:
@@ -967,6 +994,7 @@ static BOOL load_and_init_mft_mirror(ntfs_volume *vol)
 	tmp_ni = NTFS_I(tmp_ino);
 	/* The $MFTMirr, like the $MFT is multi sector transfer protected. */
 	NInoSetMstProtected(tmp_ni);
+	NInoSetSparseDisabled(tmp_ni);
 	/*
 	 * Set up our little cheat allowing us to reuse the async read io
 	 * completion handler for directories.
@@ -990,12 +1018,12 @@ static BOOL load_and_init_mft_mirror(ntfs_volume *vol)
  */
 static BOOL check_mft_mirror(ntfs_volume *vol)
 {
-	unsigned long index;
 	struct super_block *sb = vol->sb;
 	ntfs_inode *mirr_ni;
 	struct page *mft_page, *mirr_page;
 	u8 *kmft, *kmirr;
 	runlist_element *rl, rl2[2];
+	pgoff_t index;
 	int mrecs_per_page, i;
 
 	ntfs_debug("Entering.");
@@ -1122,11 +1150,130 @@ static BOOL load_and_check_logfile(ntfs_volume *vol)
 		/* ntfs_check_logfile() will have displayed error output. */
 		return FALSE;
 	}
+	NInoSetSparseDisabled(NTFS_I(tmp_ino));
 	vol->logfile_ino = tmp_ino;
 	ntfs_debug("Done.");
 	return TRUE;
 }
 
+#define NTFS_HIBERFIL_HEADER_SIZE	4096
+
+/**
+ * check_windows_hibernation_status - check if Windows is suspended on a volume
+ * @vol:	ntfs super block of device to check
+ *
+ * Check if Windows is hibernated on the ntfs volume @vol.  This is done by
+ * looking for the file hiberfil.sys in the root directory of the volume.  If
+ * the file is not present Windows is definitely not suspended.
+ *
+ * If hiberfil.sys exists and is less than 4kiB in size it means Windows is
+ * definitely suspended (this volume is not the system volume).  Caveat:  on a
+ * system with many volumes it is possible that the < 4kiB check is bogus but
+ * for now this should do fine.
+ *
+ * If hiberfil.sys exists and is larger than 4kiB in size, we need to read the
+ * hiberfil header (which is the first 4kiB).  If this begins with "hibr",
+ * Windows is definitely suspended.  If it is completely full of zeroes,
+ * Windows is definitely not hibernated.  Any other case is treated as if
+ * Windows is suspended.  This caters for the above mentioned caveat of a
+ * system with many volumes where no "hibr" magic would be present and there is
+ * no zero header.
+ *
+ * Return 0 if Windows is not hibernated on the volume, >0 if Windows is
+ * hibernated on the volume, and -errno on error.
+ */
+static int check_windows_hibernation_status(ntfs_volume *vol)
+{
+	MFT_REF mref;
+	struct inode *vi;
+	ntfs_inode *ni;
+	struct page *page;
+	u32 *kaddr, *kend;
+	ntfs_name *name = NULL;
+	int ret = 1;
+	static const ntfschar hiberfil[13] = { const_cpu_to_le16('h'),
+			const_cpu_to_le16('i'), const_cpu_to_le16('b'),
+			const_cpu_to_le16('e'), const_cpu_to_le16('r'),
+			const_cpu_to_le16('f'), const_cpu_to_le16('i'),
+			const_cpu_to_le16('l'), const_cpu_to_le16('.'),
+			const_cpu_to_le16('s'), const_cpu_to_le16('y'),
+			const_cpu_to_le16('s'), 0 };
+
+	ntfs_debug("Entering.");
+	/*
+	 * Find the inode number for the hibernation file by looking up the
+	 * filename hiberfil.sys in the root directory.
+	 */
+	down(&vol->root_ino->i_sem);
+	mref = ntfs_lookup_inode_by_name(NTFS_I(vol->root_ino), hiberfil, 12,
+			&name);
+	up(&vol->root_ino->i_sem);
+	if (IS_ERR_MREF(mref)) {
+		ret = MREF_ERR(mref);
+		/* If the file does not exist, Windows is not hibernated. */
+		if (ret == -ENOENT) {
+			ntfs_debug("hiberfil.sys not present.  Windows is not "
+					"hibernated on the volume.");
+			return 0;
+		}
+		/* A real error occured. */
+		ntfs_error(vol->sb, "Failed to find inode number for "
+				"hiberfil.sys.");
+		return ret;
+	}
+	/* We do not care for the type of match that was found. */
+	kfree(name);
+	/* Get the inode. */
+	vi = ntfs_iget(vol->sb, MREF(mref));
+	if (IS_ERR(vi) || is_bad_inode(vi)) {
+		if (!IS_ERR(vi))
+			iput(vi);
+		ntfs_error(vol->sb, "Failed to load hiberfil.sys.");
+		return IS_ERR(vi) ? PTR_ERR(vi) : -EIO;
+	}
+	if (unlikely(i_size_read(vi) < NTFS_HIBERFIL_HEADER_SIZE)) {
+		ntfs_debug("hiberfil.sys is smaller than 4kiB (0x%llx).  "
+				"Windows is hibernated on the volume.  This "
+				"is not the system volume.", i_size_read(vi));
+		goto iput_out;
+	}
+	ni = NTFS_I(vi);
+	page = ntfs_map_page(vi->i_mapping, 0);
+	if (IS_ERR(page)) {
+		ntfs_error(vol->sb, "Failed to read from hiberfil.sys.");
+		ret = PTR_ERR(page);
+		goto iput_out;
+	}
+	kaddr = (u32*)page_address(page);
+	if (*(le32*)kaddr == const_cpu_to_le32(0x72626968)/*'hibr'*/) {
+		ntfs_debug("Magic \"hibr\" found in hiberfil.sys.  Windows is "
+				"hibernated on the volume.  This is the "
+				"system volume.");
+		goto unm_iput_out;
+	}
+	kend = kaddr + NTFS_HIBERFIL_HEADER_SIZE/sizeof(*kaddr);
+	do {
+		if (unlikely(*kaddr)) {
+			ntfs_debug("hiberfil.sys is larger than 4kiB "
+					"(0x%llx), does not contain the "
+					"\"hibr\" magic, and does not have a "
+					"zero header.  Windows is hibernated "
+					"on the volume.  This is not the "
+					"system volume.", i_size_read(vi));
+			goto unm_iput_out;
+		}
+	} while (++kaddr < kend);
+	ntfs_debug("hiberfil.sys contains a zero header.  Windows is not "
+			"hibernated on the volume.  This is the system "
+			"volume.");
+	ret = 0;
+unm_iput_out:
+	ntfs_unmap_page(page);
+iput_out:
+	iput(vi);
+	return ret;
+}
+
 /**
  * load_and_init_quota - load and setup the quota file for a volume if present
  * @vol:	ntfs super block describing device whose quota file to load
@@ -1175,8 +1322,7 @@ static BOOL load_and_init_quota(ntfs_volume *vol)
 		return FALSE;
 	}
 	/* We do not care for the type of match that was found. */
-	if (name)
-		kfree(name);
+	kfree(name);
 	/* Get the inode. */
 	tmp_ino = ntfs_iget(vol->sb, MREF(mref));
 	if (IS_ERR(tmp_ino) || is_bad_inode(tmp_ino)) {
@@ -1198,6 +1344,167 @@ static BOOL load_and_init_quota(ntfs_volume *vol)
 }
 
 /**
+ * load_and_init_usnjrnl - load and setup the transaction log if present
+ * @vol:	ntfs super block describing device whose usnjrnl file to load
+ *
+ * Return TRUE on success or FALSE on error.
+ *
+ * If $UsnJrnl is not present or in the process of being disabled, we set
+ * NVolUsnJrnlStamped() and return success.
+ *
+ * If the $UsnJrnl $DATA/$J attribute has a size equal to the lowest valid usn,
+ * i.e. transaction logging has only just been enabled or the journal has been
+ * stamped and nothing has been logged since, we also set NVolUsnJrnlStamped()
+ * and return success.
+ */
+static BOOL load_and_init_usnjrnl(ntfs_volume *vol)
+{
+	MFT_REF mref;
+	struct inode *tmp_ino;
+	ntfs_inode *tmp_ni;
+	struct page *page;
+	ntfs_name *name = NULL;
+	USN_HEADER *uh;
+	static const ntfschar UsnJrnl[9] = { const_cpu_to_le16('$'),
+			const_cpu_to_le16('U'), const_cpu_to_le16('s'),
+			const_cpu_to_le16('n'), const_cpu_to_le16('J'),
+			const_cpu_to_le16('r'), const_cpu_to_le16('n'),
+			const_cpu_to_le16('l'), 0 };
+	static ntfschar Max[5] = { const_cpu_to_le16('$'),
+			const_cpu_to_le16('M'), const_cpu_to_le16('a'),
+			const_cpu_to_le16('x'), 0 };
+	static ntfschar J[3] = { const_cpu_to_le16('$'),
+			const_cpu_to_le16('J'), 0 };
+
+	ntfs_debug("Entering.");
+	/*
+	 * Find the inode number for the transaction log file by looking up the
+	 * filename $UsnJrnl in the extended system files directory $Extend.
+	 */
+	down(&vol->extend_ino->i_sem);
+	mref = ntfs_lookup_inode_by_name(NTFS_I(vol->extend_ino), UsnJrnl, 8,
+			&name);
+	up(&vol->extend_ino->i_sem);
+	if (IS_ERR_MREF(mref)) {
+		/*
+		 * If the file does not exist, transaction logging is disabled,
+		 * just return success.
+		 */
+		if (MREF_ERR(mref) == -ENOENT) {
+			ntfs_debug("$UsnJrnl not present.  Volume does not "
+					"have transaction logging enabled.");
+not_enabled:
+			/*
+			 * No need to try to stamp the transaction log if
+			 * transaction logging is not enabled.
+			 */
+			NVolSetUsnJrnlStamped(vol);
+			return TRUE;
+		}
+		/* A real error occured. */
+		ntfs_error(vol->sb, "Failed to find inode number for "
+				"$UsnJrnl.");
+		return FALSE;
+	}
+	/* We do not care for the type of match that was found. */
+	kfree(name);
+	/* Get the inode. */
+	tmp_ino = ntfs_iget(vol->sb, MREF(mref));
+	if (unlikely(IS_ERR(tmp_ino) || is_bad_inode(tmp_ino))) {
+		if (!IS_ERR(tmp_ino))
+			iput(tmp_ino);
+		ntfs_error(vol->sb, "Failed to load $UsnJrnl.");
+		return FALSE;
+	}
+	vol->usnjrnl_ino = tmp_ino;
+	/*
+	 * If the transaction log is in the process of being deleted, we can
+	 * ignore it.
+	 */
+	if (unlikely(vol->vol_flags & VOLUME_DELETE_USN_UNDERWAY)) {
+		ntfs_debug("$UsnJrnl in the process of being disabled.  "
+				"Volume does not have transaction logging "
+				"enabled.");
+		goto not_enabled;
+	}
+	/* Get the $DATA/$Max attribute. */
+	tmp_ino = ntfs_attr_iget(vol->usnjrnl_ino, AT_DATA, Max, 4);
+	if (IS_ERR(tmp_ino)) {
+		ntfs_error(vol->sb, "Failed to load $UsnJrnl/$DATA/$Max "
+				"attribute.");
+		return FALSE;
+	}
+	vol->usnjrnl_max_ino = tmp_ino;
+	if (unlikely(i_size_read(tmp_ino) < sizeof(USN_HEADER))) {
+		ntfs_error(vol->sb, "Found corrupt $UsnJrnl/$DATA/$Max "
+				"attribute (size is 0x%llx but should be at "
+				"least 0x%x bytes).", i_size_read(tmp_ino),
+				sizeof(USN_HEADER));
+		return FALSE;
+	}
+	/* Get the $DATA/$J attribute. */
+	tmp_ino = ntfs_attr_iget(vol->usnjrnl_ino, AT_DATA, J, 2);
+	if (IS_ERR(tmp_ino)) {
+		ntfs_error(vol->sb, "Failed to load $UsnJrnl/$DATA/$J "
+				"attribute.");
+		return FALSE;
+	}
+	vol->usnjrnl_j_ino = tmp_ino;
+	/* Verify $J is non-resident and sparse. */
+	tmp_ni = NTFS_I(vol->usnjrnl_j_ino);
+	if (unlikely(!NInoNonResident(tmp_ni) || !NInoSparse(tmp_ni))) {
+		ntfs_error(vol->sb, "$UsnJrnl/$DATA/$J attribute is resident "
+				"and/or not sparse.");
+		return FALSE;
+	}
+	/* Read the USN_HEADER from $DATA/$Max. */
+	page = ntfs_map_page(vol->usnjrnl_max_ino->i_mapping, 0);
+	if (IS_ERR(page)) {
+		ntfs_error(vol->sb, "Failed to read from $UsnJrnl/$DATA/$Max "
+				"attribute.");
+		return FALSE;
+	}
+	uh = (USN_HEADER*)page_address(page);
+	/* Sanity check the $Max. */
+	if (unlikely(sle64_to_cpu(uh->allocation_delta) >
+			sle64_to_cpu(uh->maximum_size))) {
+		ntfs_error(vol->sb, "Allocation delta (0x%llx) exceeds "
+				"maximum size (0x%llx).  $UsnJrnl is corrupt.",
+				(long long)sle64_to_cpu(uh->allocation_delta),
+				(long long)sle64_to_cpu(uh->maximum_size));
+		ntfs_unmap_page(page);
+		return FALSE;
+	}
+	/*
+	 * If the transaction log has been stamped and nothing has been written
+	 * to it since, we do not need to stamp it.
+	 */
+	if (unlikely(sle64_to_cpu(uh->lowest_valid_usn) >=
+			i_size_read(vol->usnjrnl_j_ino))) {
+		if (likely(sle64_to_cpu(uh->lowest_valid_usn) ==
+				i_size_read(vol->usnjrnl_j_ino))) {
+			ntfs_unmap_page(page);
+			ntfs_debug("$UsnJrnl is enabled but nothing has been "
+					"logged since it was last stamped.  "
+					"Treating this as if the volume does "
+					"not have transaction logging "
+					"enabled.");
+			goto not_enabled;
+		}
+		ntfs_error(vol->sb, "$UsnJrnl has lowest valid usn (0x%llx) "
+				"which is out of bounds (0x%llx).  $UsnJrnl "
+				"is corrupt.",
+				(long long)sle64_to_cpu(uh->lowest_valid_usn),
+				i_size_read(vol->usnjrnl_j_ino));
+		ntfs_unmap_page(page);
+		return FALSE;
+	}
+	ntfs_unmap_page(page);
+	ntfs_debug("Done.");
+	return TRUE;
+}
+
+/**
  * load_and_init_attrdef - load the attribute definitions table for a volume
  * @vol:	ntfs super block describing device whose attrdef to load
  *
@@ -1205,10 +1512,11 @@ static BOOL load_and_init_quota(ntfs_volume *vol)
  */
 static BOOL load_and_init_attrdef(ntfs_volume *vol)
 {
+	loff_t i_size;
 	struct super_block *sb = vol->sb;
 	struct inode *ino;
 	struct page *page;
-	unsigned long index, max_index;
+	pgoff_t index, max_index;
 	unsigned int size;
 
 	ntfs_debug("Entering.");
@@ -1219,14 +1527,16 @@ static BOOL load_and_init_attrdef(ntfs_volume *vol)
 			iput(ino);
 		goto failed;
 	}
+	NInoSetSparseDisabled(NTFS_I(ino));
 	/* The size of FILE_AttrDef must be above 0 and fit inside 31 bits. */
-	if (!ino->i_size || ino->i_size > 0x7fffffff)
+	i_size = i_size_read(ino);
+	if (i_size <= 0 || i_size > 0x7fffffff)
 		goto iput_failed;
-	vol->attrdef = (ATTR_DEF*)ntfs_malloc_nofs(ino->i_size);
+	vol->attrdef = (ATTR_DEF*)ntfs_malloc_nofs(i_size);
 	if (!vol->attrdef)
 		goto iput_failed;
 	index = 0;
-	max_index = ino->i_size >> PAGE_CACHE_SHIFT;
+	max_index = i_size >> PAGE_CACHE_SHIFT;
 	size = PAGE_CACHE_SIZE;
 	while (index < max_index) {
 		/* Read the attrdef table and copy it into the linear buffer. */
@@ -1239,12 +1549,12 @@ read_partial_attrdef_page:
 		ntfs_unmap_page(page);
 	};
 	if (size == PAGE_CACHE_SIZE) {
-		size = ino->i_size & ~PAGE_CACHE_MASK;
+		size = i_size & ~PAGE_CACHE_MASK;
 		if (size)
 			goto read_partial_attrdef_page;
 	}
-	vol->attrdef_size = ino->i_size;
-	ntfs_debug("Read %llu bytes from $AttrDef.", ino->i_size);
+	vol->attrdef_size = i_size;
+	ntfs_debug("Read %llu bytes from $AttrDef.", i_size);
 	iput(ino);
 	return TRUE;
 free_iput_failed:
@@ -1267,10 +1577,11 @@ failed:
  */
 static BOOL load_and_init_upcase(ntfs_volume *vol)
 {
+	loff_t i_size;
 	struct super_block *sb = vol->sb;
 	struct inode *ino;
 	struct page *page;
-	unsigned long index, max_index;
+	pgoff_t index, max_index;
 	unsigned int size;
 	int i, max;
 
@@ -1286,14 +1597,15 @@ static BOOL load_and_init_upcase(ntfs_volume *vol)
 	 * The upcase size must not be above 64k Unicode characters, must not
 	 * be zero and must be a multiple of sizeof(ntfschar).
 	 */
-	if (!ino->i_size || ino->i_size & (sizeof(ntfschar) - 1) ||
-			ino->i_size > 64ULL * 1024 * sizeof(ntfschar))
+	i_size = i_size_read(ino);
+	if (!i_size || i_size & (sizeof(ntfschar) - 1) ||
+			i_size > 64ULL * 1024 * sizeof(ntfschar))
 		goto iput_upcase_failed;
-	vol->upcase = (ntfschar*)ntfs_malloc_nofs(ino->i_size);
+	vol->upcase = (ntfschar*)ntfs_malloc_nofs(i_size);
 	if (!vol->upcase)
 		goto iput_upcase_failed;
 	index = 0;
-	max_index = ino->i_size >> PAGE_CACHE_SHIFT;
+	max_index = i_size >> PAGE_CACHE_SHIFT;
 	size = PAGE_CACHE_SIZE;
 	while (index < max_index) {
 		/* Read the upcase table and copy it into the linear buffer. */
@@ -1306,13 +1618,13 @@ read_partial_upcase_page:
 		ntfs_unmap_page(page);
 	};
 	if (size == PAGE_CACHE_SIZE) {
-		size = ino->i_size & ~PAGE_CACHE_MASK;
+		size = i_size & ~PAGE_CACHE_MASK;
 		if (size)
 			goto read_partial_upcase_page;
 	}
-	vol->upcase_len = ino->i_size >> UCHAR_T_SIZE_BITS;
+	vol->upcase_len = i_size >> UCHAR_T_SIZE_BITS;
 	ntfs_debug("Read %llu bytes from $UpCase (expected %zu bytes).",
-			ino->i_size, 64 * 1024 * sizeof(ntfschar));
+			i_size, 64 * 1024 * sizeof(ntfschar));
 	iput(ino);
 	down(&ntfs_lock);
 	if (!default_upcase) {
@@ -1376,6 +1688,9 @@ static BOOL load_system_files(ntfs_volume *vol)
 	MFT_RECORD *m;
 	VOLUME_INFORMATION *vi;
 	ntfs_attr_search_ctx *ctx;
+#ifdef NTFS_RW
+	int err;
+#endif /* NTFS_RW */
 
 	ntfs_debug("Entering.");
 #ifdef NTFS_RW
@@ -1435,7 +1750,8 @@ static BOOL load_system_files(ntfs_volume *vol)
 			iput(vol->lcnbmp_ino);
 		goto bitmap_failed;
 	}
-	if ((vol->nr_clusters + 7) >> 3 > vol->lcnbmp_ino->i_size) {
+	NInoSetSparseDisabled(NTFS_I(vol->lcnbmp_ino));
+	if ((vol->nr_clusters + 7) >> 3 > i_size_read(vol->lcnbmp_ino)) {
 		iput(vol->lcnbmp_ino);
 bitmap_failed:
 		ntfs_error(sb, "Failed to load $Bitmap.");
@@ -1486,6 +1802,12 @@ get_ctx_vol_failed:
 	unmap_mft_record(NTFS_I(vol->vol_ino));
 	printk(KERN_INFO "NTFS volume version %i.%i.\n", vol->major_ver,
 			vol->minor_ver);
+	if (vol->major_ver < 3 && NVolSparseEnabled(vol)) {
+		ntfs_warning(vol->sb, "Disabling sparse support due to NTFS "
+				"volume version %i.%i (need at least version "
+				"3.0).", vol->major_ver, vol->minor_ver);
+		NVolClearSparseEnabled(vol);
+	}
 #ifdef NTFS_RW
 	/* Make sure that no unsupported volume flags are set. */
 	if (vol->vol_flags & VOLUME_MUST_MOUNT_RO_MASK) {
@@ -1545,6 +1867,50 @@ get_ctx_vol_failed:
 		/* This will prevent a read-write remount. */
 		NVolSetErrors(vol);
 	}
+#endif /* NTFS_RW */
+	/* Get the root directory inode so we can do path lookups. */
+	vol->root_ino = ntfs_iget(sb, FILE_root);
+	if (IS_ERR(vol->root_ino) || is_bad_inode(vol->root_ino)) {
+		if (!IS_ERR(vol->root_ino))
+			iput(vol->root_ino);
+		ntfs_error(sb, "Failed to load root directory.");
+		goto iput_logfile_err_out;
+	}
+#ifdef NTFS_RW
+	/*
+	 * Check if Windows is suspended to disk on the target volume.  If it
+	 * is hibernated, we must not write *anything* to the disk so set
+	 * NVolErrors() without setting the dirty volume flag and mount
+	 * read-only.  This will prevent read-write remounting and it will also
+	 * prevent all writes.
+	 */
+	err = check_windows_hibernation_status(vol);
+	if (unlikely(err)) {
+		static const char *es1a = "Failed to determine if Windows is "
+				"hibernated";
+		static const char *es1b = "Windows is hibernated";
+		static const char *es2 = ".  Run chkdsk.";
+		const char *es1;
+
+		es1 = err < 0 ? es1a : es1b;
+		/* If a read-write mount, convert it to a read-only mount. */
+		if (!(sb->s_flags & MS_RDONLY)) {
+			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
+					ON_ERRORS_CONTINUE))) {
+				ntfs_error(sb, "%s and neither on_errors="
+						"continue nor on_errors="
+						"remount-ro was specified%s",
+						es1, es2);
+				goto iput_root_err_out;
+			}
+			sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
+			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
+		} else
+			ntfs_warning(sb, "%s.  Will not be able to remount "
+					"read-write%s", es1, es2);
+		/* This will prevent a read-write remount. */
+		NVolSetErrors(vol);
+	}
 	/* If (still) a read-write mount, mark the volume dirty. */
 	if (!(sb->s_flags & MS_RDONLY) &&
 			ntfs_set_volume_flags(vol, VOLUME_IS_DIRTY)) {
@@ -1558,7 +1924,7 @@ get_ctx_vol_failed:
 			ntfs_error(sb, "%s and neither on_errors=continue nor "
 					"on_errors=remount-ro was specified%s",
 					es1, es2);
-			goto iput_logfile_err_out;
+			goto iput_root_err_out;
 		}
 		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
 		sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
@@ -1585,7 +1951,7 @@ get_ctx_vol_failed:
 			ntfs_error(sb, "%s and neither on_errors=continue nor "
 					"on_errors=remount-ro was specified%s",
 					es1, es2);
-			goto iput_logfile_err_out;
+			goto iput_root_err_out;
 		}
 		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
 		sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
@@ -1604,23 +1970,15 @@ get_ctx_vol_failed:
 			ntfs_error(sb, "%s and neither on_errors=continue nor "
 					"on_errors=remount-ro was specified%s",
 					es1, es2);
-			goto iput_logfile_err_out;
+			goto iput_root_err_out;
 		}
 		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
 		sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
 		NVolSetErrors(vol);
 	}
 #endif /* NTFS_RW */
-	/* Get the root directory inode. */
-	vol->root_ino = ntfs_iget(sb, FILE_root);
-	if (IS_ERR(vol->root_ino) || is_bad_inode(vol->root_ino)) {
-		if (!IS_ERR(vol->root_ino))
-			iput(vol->root_ino);
-		ntfs_error(sb, "Failed to load root directory.");
-		goto iput_logfile_err_out;
-	}
 	/* If on NTFS versions before 3.0, we are done. */
-	if (vol->major_ver < 3)
+	if (unlikely(vol->major_ver < 3))
 		return TRUE;
 	/* NTFS 3.0+ specific initialization. */
 	/* Get the security descriptors inode. */
@@ -1631,7 +1989,7 @@ get_ctx_vol_failed:
 		ntfs_error(sb, "Failed to load $Secure.");
 		goto iput_root_err_out;
 	}
-	// FIXME: Initialize security.
+	// TODO: Initialize security.
 	/* Get the extended system files' directory inode. */
 	vol->extend_ino = ntfs_iget(sb, FILE_Extend);
 	if (IS_ERR(vol->extend_ino) || is_bad_inode(vol->extend_ino)) {
@@ -1682,10 +2040,60 @@ get_ctx_vol_failed:
 		sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
 		NVolSetErrors(vol);
 	}
-	// TODO: Delete or checkpoint the $UsnJrnl if it exists.
+	/*
+	 * Find the transaction log file ($UsnJrnl), load it if present, check
+	 * it, and set it up.
+	 */
+	if (!load_and_init_usnjrnl(vol)) {
+		static const char *es1 = "Failed to load $UsnJrnl";
+		static const char *es2 = ".  Run chkdsk.";
+
+		/* If a read-write mount, convert it to a read-only mount. */
+		if (!(sb->s_flags & MS_RDONLY)) {
+			if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
+					ON_ERRORS_CONTINUE))) {
+				ntfs_error(sb, "%s and neither on_errors="
+						"continue nor on_errors="
+						"remount-ro was specified%s",
+						es1, es2);
+				goto iput_usnjrnl_err_out;
+			}
+			sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
+			ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
+		} else
+			ntfs_warning(sb, "%s.  Will not be able to remount "
+					"read-write%s", es1, es2);
+		/* This will prevent a read-write remount. */
+		NVolSetErrors(vol);
+	}
+	/* If (still) a read-write mount, stamp the transaction log. */
+	if (!(sb->s_flags & MS_RDONLY) && !ntfs_stamp_usnjrnl(vol)) {
+		static const char *es1 = "Failed to stamp transaction log "
+				"($UsnJrnl)";
+		static const char *es2 = ".  Run chkdsk.";
+
+		/* Convert to a read-only mount. */
+		if (!(vol->on_errors & (ON_ERRORS_REMOUNT_RO |
+				ON_ERRORS_CONTINUE))) {
+			ntfs_error(sb, "%s and neither on_errors=continue nor "
+					"on_errors=remount-ro was specified%s",
+					es1, es2);
+			goto iput_usnjrnl_err_out;
+		}
+		ntfs_error(sb, "%s.  Mounting read-only%s", es1, es2);
+		sb->s_flags |= MS_RDONLY | MS_NOATIME | MS_NODIRATIME;
+		NVolSetErrors(vol);
+	}
 #endif /* NTFS_RW */
 	return TRUE;
 #ifdef NTFS_RW
+iput_usnjrnl_err_out:
+	if (vol->usnjrnl_j_ino)
+		iput(vol->usnjrnl_j_ino);
+	if (vol->usnjrnl_max_ino)
+		iput(vol->usnjrnl_max_ino);
+	if (vol->usnjrnl_ino)
+		iput(vol->usnjrnl_ino);
 iput_quota_err_out:
 	if (vol->quota_q_ino)
 		iput(vol->quota_q_ino);
@@ -1759,6 +2167,12 @@ static void ntfs_put_super(struct super_block *sb)
 
 	/* NTFS 3.0+ specific. */
 	if (vol->major_ver >= 3) {
+		if (vol->usnjrnl_j_ino)
+			ntfs_commit_inode(vol->usnjrnl_j_ino);
+		if (vol->usnjrnl_max_ino)
+			ntfs_commit_inode(vol->usnjrnl_max_ino);
+		if (vol->usnjrnl_ino)
+			ntfs_commit_inode(vol->usnjrnl_ino);
 		if (vol->quota_q_ino)
 			ntfs_commit_inode(vol->quota_q_ino);
 		if (vol->quota_ino)
@@ -1814,6 +2228,18 @@ static void ntfs_put_super(struct super_block *sb)
 	/* NTFS 3.0+ specific clean up. */
 	if (vol->major_ver >= 3) {
 #ifdef NTFS_RW
+		if (vol->usnjrnl_j_ino) {
+			iput(vol->usnjrnl_j_ino);
+			vol->usnjrnl_j_ino = NULL;
+		}
+		if (vol->usnjrnl_max_ino) {
+			iput(vol->usnjrnl_max_ino);
+			vol->usnjrnl_max_ino = NULL;
+		}
+		if (vol->usnjrnl_ino) {
+			iput(vol->usnjrnl_ino);
+			vol->usnjrnl_ino = NULL;
+		}
 		if (vol->quota_q_ino) {
 			iput(vol->quota_q_ino);
 			vol->quota_q_ino = NULL;
@@ -1959,8 +2385,7 @@ static s64 get_nr_free_clusters(ntfs_volume *vol)
 	struct address_space *mapping = vol->lcnbmp_ino->i_mapping;
 	filler_t *readpage = (filler_t*)mapping->a_ops->readpage;
 	struct page *page;
-	unsigned long index, max_index;
-	unsigned int max_size;
+	pgoff_t index, max_index;
 
 	ntfs_debug("Entering.");
 	/* Serialize accesses to the cluster bitmap. */
@@ -1972,11 +2397,10 @@ static s64 get_nr_free_clusters(ntfs_volume *vol)
 	 */
 	max_index = (((vol->nr_clusters + 7) >> 3) + PAGE_CACHE_SIZE - 1) >>
 			PAGE_CACHE_SHIFT;
-	/* Use multiples of 4 bytes. */
-	max_size = PAGE_CACHE_SIZE >> 2;
-	ntfs_debug("Reading $Bitmap, max_index = 0x%lx, max_size = 0x%x.",
-			max_index, max_size);
-	for (index = 0UL; index < max_index; index++) {
+	/* Use multiples of 4 bytes, thus max_size is PAGE_CACHE_SIZE / 4. */
+	ntfs_debug("Reading $Bitmap, max_index = 0x%lx, max_size = 0x%lx.",
+			max_index, PAGE_CACHE_SIZE / 4);
+	for (index = 0; index < max_index; index++) {
 		unsigned int i;
 		/*
 		 * Read the page from page cache, getting it from backing store
@@ -2008,7 +2432,7 @@ static s64 get_nr_free_clusters(ntfs_volume *vol)
 		 * the result as all out of range bytes are set to zero by
 		 * ntfs_readpage().
 		 */
-	  	for (i = 0; i < max_size; i++)
+	  	for (i = 0; i < PAGE_CACHE_SIZE / 4; i++)
 			nr_free -= (s64)hweight32(kaddr[i]);
 		kunmap_atomic(kaddr, KM_USER0);
 		page_cache_release(page);
@@ -2031,6 +2455,8 @@ static s64 get_nr_free_clusters(ntfs_volume *vol)
 /**
  * __get_nr_free_mft_records - return the number of free inodes on a volume
  * @vol:	ntfs volume for which to obtain free inode count
+ * @nr_free:	number of mft records in filesystem
+ * @max_index:	maximum number of pages containing set bits
  *
  * Calculate the number of free mft records (inodes) on the mounted NTFS
  * volume @vol. We actually calculate the number of mft records in use instead
@@ -2043,32 +2469,20 @@ static s64 get_nr_free_clusters(ntfs_volume *vol)
  *
  * NOTE: Caller must hold mftbmp_lock rw_semaphore for reading or writing.
  */
-static unsigned long __get_nr_free_mft_records(ntfs_volume *vol)
+static unsigned long __get_nr_free_mft_records(ntfs_volume *vol,
+		s64 nr_free, const pgoff_t max_index)
 {
-	s64 nr_free;
 	u32 *kaddr;
 	struct address_space *mapping = vol->mftbmp_ino->i_mapping;
 	filler_t *readpage = (filler_t*)mapping->a_ops->readpage;
 	struct page *page;
-	unsigned long index, max_index;
-	unsigned int max_size;
+	pgoff_t index;
 
 	ntfs_debug("Entering.");
-	/* Number of mft records in file system (at this point in time). */
-	nr_free = vol->mft_ino->i_size >> vol->mft_record_size_bits;
-	/*
-	 * Convert the maximum number of set bits into bytes rounded up, then
-	 * convert into multiples of PAGE_CACHE_SIZE, rounding up so that if we
-	 * have one full and one partial page max_index = 2.
-	 */
-	max_index = ((((NTFS_I(vol->mft_ino)->initialized_size >>
-			vol->mft_record_size_bits) + 7) >> 3) +
-			PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
-	/* Use multiples of 4 bytes. */
-	max_size = PAGE_CACHE_SIZE >> 2;
+	/* Use multiples of 4 bytes, thus max_size is PAGE_CACHE_SIZE / 4. */
 	ntfs_debug("Reading $MFT/$BITMAP, max_index = 0x%lx, max_size = "
-			"0x%x.", max_index, max_size);
-	for (index = 0UL; index < max_index; index++) {
+			"0x%lx.", max_index, PAGE_CACHE_SIZE / 4);
+	for (index = 0; index < max_index; index++) {
 		unsigned int i;
 		/*
 		 * Read the page from page cache, getting it from backing store
@@ -2100,7 +2514,7 @@ static unsigned long __get_nr_free_mft_records(ntfs_volume *vol)
 		 * the result as all out of range bytes are set to zero by
 		 * ntfs_readpage().
 		 */
-	  	for (i = 0; i < max_size; i++)
+	  	for (i = 0; i < PAGE_CACHE_SIZE / 4; i++)
 			nr_free -= (s64)hweight32(kaddr[i]);
 		kunmap_atomic(kaddr, KM_USER0);
 		page_cache_release(page);
@@ -2134,8 +2548,11 @@ static unsigned long __get_nr_free_mft_records(ntfs_volume *vol)
  */
 static int ntfs_statfs(struct super_block *sb, struct kstatfs *sfs)
 {
-	ntfs_volume *vol = NTFS_SB(sb);
 	s64 size;
+	ntfs_volume *vol = NTFS_SB(sb);
+	ntfs_inode *mft_ni = NTFS_I(vol->mft_ino);
+	pgoff_t max_index;
+	unsigned long flags;
 
 	ntfs_debug("Entering.");
 	/* Type of filesystem. */
@@ -2143,13 +2560,13 @@ static int ntfs_statfs(struct super_block *sb, struct kstatfs *sfs)
 	/* Optimal transfer block size. */
 	sfs->f_bsize  = PAGE_CACHE_SIZE;
 	/*
-	 * Total data blocks in file system in units of f_bsize and since
+	 * Total data blocks in filesystem in units of f_bsize and since
 	 * inodes are also stored in data blocs ($MFT is a file) this is just
 	 * the total clusters.
 	 */
 	sfs->f_blocks = vol->nr_clusters << vol->cluster_size_bits >>
 				PAGE_CACHE_SHIFT;
-	/* Free data blocks in file system in units of f_bsize. */
+	/* Free data blocks in filesystem in units of f_bsize. */
 	size	      = get_nr_free_clusters(vol) << vol->cluster_size_bits >>
 				PAGE_CACHE_SHIFT;
 	if (size < 0LL)
@@ -2158,17 +2575,27 @@ static int ntfs_statfs(struct super_block *sb, struct kstatfs *sfs)
 	sfs->f_bavail = sfs->f_bfree = size;
 	/* Serialize accesses to the inode bitmap. */
 	down_read(&vol->mftbmp_lock);
-	/* Number of inodes in file system (at this point in time). */
-	sfs->f_files = vol->mft_ino->i_size >> vol->mft_record_size_bits;
+	read_lock_irqsave(&mft_ni->size_lock, flags);
+	size = i_size_read(vol->mft_ino) >> vol->mft_record_size_bits;
+	/*
+	 * Convert the maximum number of set bits into bytes rounded up, then
+	 * convert into multiples of PAGE_CACHE_SIZE, rounding up so that if we
+	 * have one full and one partial page max_index = 2.
+	 */
+	max_index = ((((mft_ni->initialized_size >> vol->mft_record_size_bits)
+			+ 7) >> 3) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+	read_unlock_irqrestore(&mft_ni->size_lock, flags);
+	/* Number of inodes in filesystem (at this point in time). */
+	sfs->f_files = size;
 	/* Free inodes in fs (based on current total count). */
-	sfs->f_ffree = __get_nr_free_mft_records(vol);
+	sfs->f_ffree = __get_nr_free_mft_records(vol, size, max_index);
 	up_read(&vol->mftbmp_lock);
 	/*
 	 * File system id. This is extremely *nix flavour dependent and even
 	 * within Linux itself all fs do their own thing. I interpret this to
 	 * mean a unique id associated with the mounted fs and not the id
-	 * associated with the file system driver, the latter is already given
-	 * by the file system type in sfs->f_type. Thus we use the 64-bit
+	 * associated with the filesystem driver, the latter is already given
+	 * by the filesystem type in sfs->f_type. Thus we use the 64-bit
 	 * volume serial number splitting it into two 32-bit parts. We enter
 	 * the least significant 32-bits in f_fsid[0] and the most significant
 	 * 32-bits in f_fsid[1].
@@ -2219,53 +2646,19 @@ static struct super_operations ntfs_sops = {
 						   proc. */
 };
 
-
 /**
- * Declarations for NTFS specific export operations (fs/ntfs/namei.c).
- */
-extern struct dentry *ntfs_get_parent(struct dentry *child_dent);
-extern struct dentry *ntfs_get_dentry(struct super_block *sb, void *fh);
-
-/**
- * Export operations allowing NFS exporting of mounted NTFS partitions.
- *
- * We use the default ->decode_fh() and ->encode_fh() for now.  Note that they
- * use 32 bits to store the inode number which is an unsigned long so on 64-bit
- * architectures is usually 64 bits so it would all fail horribly on huge
- * volumes.  I guess we need to define our own encode and decode fh functions
- * that store 64-bit inode numbers at some point but for now we will ignore the
- * problem...
- *
- * We also use the default ->get_name() helper (used by ->decode_fh() via
- * fs/exportfs/expfs.c::find_exported_dentry()) as that is completely fs
- * independent.
- *
- * The default ->get_parent() just returns -EACCES so we have to provide our
- * own and the default ->get_dentry() is incompatible with NTFS due to not
- * allowing the inode number 0 which is used in NTFS for the system file $MFT
- * and due to using iget() whereas NTFS needs ntfs_iget().
- */
-static struct export_operations ntfs_export_ops = {
-	.get_parent	= ntfs_get_parent,	/* Find the parent of a given
-						   directory. */
-	.get_dentry	= ntfs_get_dentry,	/* Find a dentry for the inode
-						   given a file handle
-						   sub-fragment. */
-};
-
-/**
- * ntfs_fill_super - mount an ntfs files system
- * @sb:		super block of ntfs file system to mount
+ * ntfs_fill_super - mount an ntfs filesystem
+ * @sb:		super block of ntfs filesystem to mount
  * @opt:	string containing the mount options
  * @silent:	silence error output
  *
  * ntfs_fill_super() is called by the VFS to mount the device described by @sb
- * with the mount otions in @data with the NTFS file system.
+ * with the mount otions in @data with the NTFS filesystem.
  *
  * If @silent is true, remain silent even if errors are detected. This is used
- * during bootup, when the kernel tries to mount the root file system with all
- * registered file systems one after the other until one succeeds. This implies
- * that all file systems except the correct one will quite correctly and
+ * during bootup, when the kernel tries to mount the root filesystem with all
+ * registered filesystems one after the other until one succeeds. This implies
+ * that all filesystems except the correct one will quite correctly and
  * expectedly return an error, but nobody wants to see error messages when in
  * fact this is what is supposed to happen.
  *
@@ -2292,39 +2685,25 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 		return -ENOMEM;
 	}
 	/* Initialize ntfs_volume structure. */
-	memset(vol, 0, sizeof(ntfs_volume));
-	vol->sb = sb;
-	vol->upcase = NULL;
-	vol->attrdef = NULL;
-	vol->mft_ino = NULL;
-	vol->mftbmp_ino = NULL;
+	*vol = (ntfs_volume) {
+		.sb = sb,
+		/*
+		 * Default is group and other don't have any access to files or
+		 * directories while owner has full access. Further, files by
+		 * default are not executable but directories are of course
+		 * browseable.
+		 */
+		.fmask = 0177,
+		.dmask = 0077,
+	};
 	init_rwsem(&vol->mftbmp_lock);
-#ifdef NTFS_RW
-	vol->mftmirr_ino = NULL;
-	vol->logfile_ino = NULL;
-#endif /* NTFS_RW */
-	vol->lcnbmp_ino = NULL;
 	init_rwsem(&vol->lcnbmp_lock);
-	vol->vol_ino = NULL;
-	vol->root_ino = NULL;
-	vol->secure_ino = NULL;
-	vol->extend_ino = NULL;
-#ifdef NTFS_RW
-	vol->quota_ino = NULL;
-	vol->quota_q_ino = NULL;
-#endif /* NTFS_RW */
-	vol->nls_map = NULL;
-
-	/*
-	 * Default is group and other don't have any access to files or
-	 * directories while owner has full access. Further, files by default
-	 * are not executable but directories are of course browseable.
-	 */
-	vol->fmask = 0177;
-	vol->dmask = 0077;
 
 	unlock_kernel();
 
+	/* By default, enable sparse support. */
+	NVolSetSparseEnabled(vol);
+
 	/* Important to get the mount options dealt with now. */
 	if (!parse_options(vol, (char*)opt))
 		goto err_out_now;
@@ -2347,7 +2726,8 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 	}
 
 	/* Get the size of the device in units of NTFS_BLOCK_SIZE bytes. */
-	vol->nr_blocks = sb->s_bdev->bd_inode->i_size >> NTFS_BLOCK_SIZE_BITS;
+	vol->nr_blocks = i_size_read(sb->s_bdev->bd_inode) >>
+			NTFS_BLOCK_SIZE_BITS;
 
 	/* Read the boot sector and return unlocked buffer head to it. */
 	if (!(bh = read_ntfs_boot_sector(sb, silent))) {
@@ -2476,6 +2856,18 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 	/* NTFS 3.0+ specific clean up. */
 	if (vol->major_ver >= 3) {
 #ifdef NTFS_RW
+		if (vol->usnjrnl_j_ino) {
+			iput(vol->usnjrnl_j_ino);
+			vol->usnjrnl_j_ino = NULL;
+		}
+		if (vol->usnjrnl_max_ino) {
+			iput(vol->usnjrnl_max_ino);
+			vol->usnjrnl_max_ino = NULL;
+		}
+		if (vol->usnjrnl_ino) {
+			iput(vol->usnjrnl_ino);
+			vol->usnjrnl_ino = NULL;
+		}
 		if (vol->quota_q_ino) {
 			iput(vol->quota_q_ino);
 			vol->quota_q_ino = NULL;
@@ -2581,7 +2973,7 @@ err_out_now:
  */
 kmem_cache_t *ntfs_name_cache;
 
-/* Slab caches for efficient allocation/deallocation of of inodes. */
+/* Slab caches for efficient allocation/deallocation of inodes. */
 kmem_cache_t *ntfs_inode_cache;
 kmem_cache_t *ntfs_big_inode_cache;
 
@@ -2705,7 +3097,7 @@ static int __init init_ntfs_fs(void)
 		ntfs_debug("NTFS driver registered successfully.");
 		return 0; /* Success! */
 	}
-	printk(KERN_CRIT "NTFS: Failed to register NTFS file system driver!\n");
+	printk(KERN_CRIT "NTFS: Failed to register NTFS filesystem driver!\n");
 
 sysctl_err_out:
 	kmem_cache_destroy(ntfs_big_inode_cache);
@@ -2719,7 +3111,7 @@ actx_err_out:
 	kmem_cache_destroy(ntfs_index_ctx_cache);
 ictx_err_out:
 	if (!err) {
-		printk(KERN_CRIT "NTFS: Aborting NTFS file system driver "
+		printk(KERN_CRIT "NTFS: Aborting NTFS filesystem driver "
 				"registration...\n");
 		err = -ENOMEM;
 	}
@@ -2759,7 +3151,7 @@ static void __exit exit_ntfs_fs(void)
 }
 
 MODULE_AUTHOR("Anton Altaparmakov <aia21@cantab.net>");
-MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2004 Anton Altaparmakov");
+MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2005 Anton Altaparmakov");
 MODULE_VERSION(NTFS_VERSION);
 MODULE_LICENSE("GPL");
 #ifdef DEBUG
diff --git a/fs/ntfs/sysctl.c b/fs/ntfs/sysctl.c
index 75067e4..1c23138 100644
--- a/fs/ntfs/sysctl.c
+++ b/fs/ntfs/sysctl.c
@@ -3,7 +3,7 @@
  *	      the Linux-NTFS project. Adapted from the old NTFS driver,
  *	      Copyright (C) 1997 Martin von Löwis, Régis Duchesne
  *
- * Copyright (c) 2002-2004 Anton Altaparmakov
+ * Copyright (c) 2002-2005 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
@@ -67,7 +67,7 @@ int ntfs_sysctl(int add)
 			return -ENOMEM;
 #ifdef CONFIG_PROC_FS
 		/*
-		 * If the proc file system is in use and we are a module, need
+		 * If the proc filesystem is in use and we are a module, need
 		 * to set the owner of our proc entry to our module. In the
 		 * non-modular case, THIS_MODULE is NULL, so this is ok.
 		 */
diff --git a/fs/ntfs/sysctl.h b/fs/ntfs/sysctl.h
index df749cc..c8064ca 100644
--- a/fs/ntfs/sysctl.h
+++ b/fs/ntfs/sysctl.h
@@ -26,7 +26,7 @@
 
 #include <linux/config.h>
 
-#if (DEBUG && CONFIG_SYSCTL)
+#if defined(DEBUG) && defined(CONFIG_SYSCTL)
 
 extern int ntfs_sysctl(int add);
 
diff --git a/fs/ntfs/time.h b/fs/ntfs/time.h
index a09a51d..0123398 100644
--- a/fs/ntfs/time.h
+++ b/fs/ntfs/time.h
@@ -1,7 +1,7 @@
 /*
  * time.h - NTFS time conversion functions.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 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
@@ -87,7 +87,7 @@ static inline struct timespec ntfs2utc(const sle64 time)
 	struct timespec ts;
 
 	/* Subtract the NTFS time offset. */
-	s64 t = sle64_to_cpu(time) - NTFS_TIME_OFFSET;
+	u64 t = (u64)(sle64_to_cpu(time) - NTFS_TIME_OFFSET);
 	/*
 	 * Convert the time to 1-second intervals and the remainder to
 	 * 1-nano-second intervals.
diff --git a/fs/ntfs/types.h b/fs/ntfs/types.h
index 08a55aa..6e4a7e3 100644
--- a/fs/ntfs/types.h
+++ b/fs/ntfs/types.h
@@ -2,7 +2,7 @@
  * types.h - Defines for NTFS Linux kernel driver specific types.
  *	     Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 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
@@ -53,6 +53,14 @@ typedef sle64 leLCN;
 typedef s64 LSN;
 typedef sle64 leLSN;
 
+/*
+ * The NTFS transaction log $UsnJrnl uses usn which are signed 64-bit values.
+ * We define our own type USN, to allow for type checking and better code
+ * readability.
+ */
+typedef s64 USN;
+typedef sle64 leUSN;
+
 typedef enum {
 	FALSE = 0,
 	TRUE = 1
diff --git a/fs/ntfs/unistr.c b/fs/ntfs/unistr.c
index 560b0ea..19c42e2 100644
--- a/fs/ntfs/unistr.c
+++ b/fs/ntfs/unistr.c
@@ -264,7 +264,7 @@ int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
 
 	/* We don't trust outside sources. */
 	if (ins) {
-		ucs = (ntfschar*)kmem_cache_alloc(ntfs_name_cache, SLAB_NOFS);
+		ucs = kmem_cache_alloc(ntfs_name_cache, SLAB_NOFS);
 		if (ucs) {
 			for (i = o = 0; i < ins_len; i += wc_len) {
 				wc_len = nls->char2uni(ins + i, ins_len - i,
diff --git a/fs/ntfs/usnjrnl.c b/fs/ntfs/usnjrnl.c
new file mode 100644
index 0000000..7777324
--- /dev/null
+++ b/fs/ntfs/usnjrnl.c
@@ -0,0 +1,84 @@
+/*
+ * usnjrnl.h - NTFS kernel transaction log ($UsnJrnl) handling.  Part of the
+ *	       Linux-NTFS project.
+ *
+ * Copyright (c) 2005 Anton Altaparmakov
+ *
+ * This program/include file is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as published
+ * by the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program/include file is distributed in the hope that it will be
+ * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program (in the main directory of the Linux-NTFS
+ * distribution in the file COPYING); if not, write to the Free Software
+ * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifdef NTFS_RW
+
+#include <linux/fs.h>
+#include <linux/highmem.h>
+#include <linux/mm.h>
+
+#include "aops.h"
+#include "debug.h"
+#include "endian.h"
+#include "time.h"
+#include "types.h"
+#include "usnjrnl.h"
+#include "volume.h"
+
+/**
+ * ntfs_stamp_usnjrnl - stamp the transaction log ($UsnJrnl) on an ntfs volume
+ * @vol:	ntfs volume on which to stamp the transaction log
+ *
+ * Stamp the transaction log ($UsnJrnl) on the ntfs volume @vol and return
+ * TRUE on success and FALSE on error.
+ *
+ * This function assumes that the transaction log has already been loaded and
+ * consistency checked by a call to fs/ntfs/super.c::load_and_init_usnjrnl().
+ */
+BOOL ntfs_stamp_usnjrnl(ntfs_volume *vol)
+{
+	ntfs_debug("Entering.");
+	if (likely(!NVolUsnJrnlStamped(vol))) {
+		sle64 stamp;
+		struct page *page;
+		USN_HEADER *uh;
+
+		page = ntfs_map_page(vol->usnjrnl_max_ino->i_mapping, 0);
+		if (IS_ERR(page)) {
+			ntfs_error(vol->sb, "Failed to read from "
+					"$UsnJrnl/$DATA/$Max attribute.");
+			return FALSE;
+		}
+		uh = (USN_HEADER*)page_address(page);
+		stamp = get_current_ntfs_time();
+		ntfs_debug("Stamping transaction log ($UsnJrnl): old "
+				"journal_id 0x%llx, old lowest_valid_usn "
+				"0x%llx, new journal_id 0x%llx, new "
+				"lowest_valid_usn 0x%llx.",
+				(long long)sle64_to_cpu(uh->journal_id),
+				(long long)sle64_to_cpu(uh->lowest_valid_usn),
+				(long long)sle64_to_cpu(stamp),
+				i_size_read(vol->usnjrnl_j_ino));
+		uh->lowest_valid_usn =
+				cpu_to_sle64(i_size_read(vol->usnjrnl_j_ino));
+		uh->journal_id = stamp;
+		flush_dcache_page(page);
+		set_page_dirty(page);
+		ntfs_unmap_page(page);
+		/* Set the flag so we do not have to do it again on remount. */
+		NVolSetUsnJrnlStamped(vol);
+	}
+	ntfs_debug("Done.");
+	return TRUE;
+}
+
+#endif /* NTFS_RW */
diff --git a/fs/ntfs/usnjrnl.h b/fs/ntfs/usnjrnl.h
new file mode 100644
index 0000000..ff988b0
--- /dev/null
+++ b/fs/ntfs/usnjrnl.h
@@ -0,0 +1,205 @@
+/*
+ * usnjrnl.h - Defines for NTFS kernel transaction log ($UsnJrnl) handling.
+ *	       Part of the Linux-NTFS project.
+ *
+ * Copyright (c) 2005 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
+ * by the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program/include file is distributed in the hope that it will be
+ * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program (in the main directory of the Linux-NTFS
+ * distribution in the file COPYING); if not, write to the Free Software
+ * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef _LINUX_NTFS_USNJRNL_H
+#define _LINUX_NTFS_USNJRNL_H
+
+#ifdef NTFS_RW
+
+#include "types.h"
+#include "endian.h"
+#include "layout.h"
+#include "volume.h"
+
+/*
+ * Transaction log ($UsnJrnl) organization:
+ *
+ * The transaction log records whenever a file is modified in any way.  So for
+ * example it will record that file "blah" was written to at a particular time
+ * but not what was written.  If will record that a file was deleted or
+ * created, that a file was truncated, etc.  See below for all the reason
+ * codes used.
+ *
+ * The transaction log is in the $Extend directory which is in the root
+ * directory of each volume.  If it is not present it means transaction
+ * logging is disabled.  If it is present it means transaction logging is
+ * either enabled or in the process of being disabled in which case we can
+ * ignore it as it will go away as soon as Windows gets its hands on it.
+ *
+ * To determine whether the transaction logging is enabled or in the process
+ * of being disabled, need to check the volume flags in the
+ * $VOLUME_INFORMATION attribute in the $Volume system file (which is present
+ * in the root directory and has a fixed mft record number, see layout.h).
+ * If the flag VOLUME_DELETE_USN_UNDERWAY is set it means the transaction log
+ * is in the process of being disabled and if this flag is clear it means the
+ * transaction log is enabled.
+ *
+ * The transaction log consists of two parts; the $DATA/$Max attribute as well
+ * as the $DATA/$J attribute.  $Max is a header describing the transaction
+ * log whilst $J is the transaction log data itself as a sequence of variable
+ * sized USN_RECORDs (see below for all the structures).
+ *
+ * We do not care about transaction logging at this point in time but we still
+ * need to let windows know that the transaction log is out of date.  To do
+ * this we need to stamp the transaction log.  This involves setting the
+ * lowest_valid_usn field in the $DATA/$Max attribute to the usn to be used
+ * for the next added USN_RECORD to the $DATA/$J attribute as well as
+ * generating a new journal_id in $DATA/$Max.
+ *
+ * The journal_id is as of the current version (2.0) of the transaction log
+ * simply the 64-bit timestamp of when the journal was either created or last
+ * stamped.
+ *
+ * To determine the next usn there are two ways.  The first is to parse
+ * $DATA/$J and to find the last USN_RECORD in it and to add its record_length
+ * to its usn (which is the byte offset in the $DATA/$J attribute).  The
+ * second is simply to take the data size of the attribute.  Since the usns
+ * are simply byte offsets into $DATA/$J, this is exactly the next usn.  For
+ * obvious reasons we use the second method as it is much simpler and faster.
+ *
+ * As an aside, note that to actually disable the transaction log, one would
+ * need to set the VOLUME_DELETE_USN_UNDERWAY flag (see above), then go
+ * through all the mft records on the volume and set the usn field in their
+ * $STANDARD_INFORMATION attribute to zero.  Once that is done, one would need
+ * to delete the transaction log file, i.e. \$Extent\$UsnJrnl, and finally,
+ * one would need to clear the VOLUME_DELETE_USN_UNDERWAY flag.
+ *
+ * Note that if a volume is unmounted whilst the transaction log is being
+ * disabled, the process will continue the next time the volume is mounted.
+ * This is why we can safely mount read-write when we see a transaction log
+ * in the process of being deleted.
+ */
+
+/* Some $UsnJrnl related constants. */
+#define UsnJrnlMajorVer		2
+#define UsnJrnlMinorVer		0
+
+/*
+ * $DATA/$Max attribute.  This is (always?) resident and has a fixed size of
+ * 32 bytes.  It contains the header describing the transaction log.
+ */
+typedef struct {
+/*Ofs*/
+/*   0*/sle64 maximum_size;	/* The maximum on-disk size of the $DATA/$J
+				   attribute. */
+/*   8*/sle64 allocation_delta;	/* Number of bytes by which to increase the
+				   size of the $DATA/$J attribute. */
+/*0x10*/sle64 journal_id;	/* Current id of the transaction log. */
+/*0x18*/leUSN lowest_valid_usn;	/* Lowest valid usn in $DATA/$J for the
+				   current journal_id. */
+/* sizeof() = 32 (0x20) bytes */
+} __attribute__ ((__packed__)) USN_HEADER;
+
+/*
+ * Reason flags (32-bit).  Cumulative flags describing the change(s) to the
+ * file since it was last opened.  I think the names speak for themselves but
+ * if you disagree check out the descriptions in the Linux NTFS project NTFS
+ * documentation: http://linux-ntfs.sourceforge.net/ntfs/files/usnjrnl.html
+ */
+enum {
+	USN_REASON_DATA_OVERWRITE	= const_cpu_to_le32(0x00000001),
+	USN_REASON_DATA_EXTEND		= const_cpu_to_le32(0x00000002),
+	USN_REASON_DATA_TRUNCATION	= const_cpu_to_le32(0x00000004),
+	USN_REASON_NAMED_DATA_OVERWRITE	= const_cpu_to_le32(0x00000010),
+	USN_REASON_NAMED_DATA_EXTEND	= const_cpu_to_le32(0x00000020),
+	USN_REASON_NAMED_DATA_TRUNCATION= const_cpu_to_le32(0x00000040),
+	USN_REASON_FILE_CREATE		= const_cpu_to_le32(0x00000100),
+	USN_REASON_FILE_DELETE		= const_cpu_to_le32(0x00000200),
+	USN_REASON_EA_CHANGE		= const_cpu_to_le32(0x00000400),
+	USN_REASON_SECURITY_CHANGE	= const_cpu_to_le32(0x00000800),
+	USN_REASON_RENAME_OLD_NAME	= const_cpu_to_le32(0x00001000),
+	USN_REASON_RENAME_NEW_NAME	= const_cpu_to_le32(0x00002000),
+	USN_REASON_INDEXABLE_CHANGE	= const_cpu_to_le32(0x00004000),
+	USN_REASON_BASIC_INFO_CHANGE	= const_cpu_to_le32(0x00008000),
+	USN_REASON_HARD_LINK_CHANGE	= const_cpu_to_le32(0x00010000),
+	USN_REASON_COMPRESSION_CHANGE	= const_cpu_to_le32(0x00020000),
+	USN_REASON_ENCRYPTION_CHANGE	= const_cpu_to_le32(0x00040000),
+	USN_REASON_OBJECT_ID_CHANGE	= const_cpu_to_le32(0x00080000),
+	USN_REASON_REPARSE_POINT_CHANGE	= const_cpu_to_le32(0x00100000),
+	USN_REASON_STREAM_CHANGE	= const_cpu_to_le32(0x00200000),
+	USN_REASON_CLOSE		= const_cpu_to_le32(0x80000000),
+};
+
+typedef le32 USN_REASON_FLAGS;
+
+/*
+ * Source info flags (32-bit).  Information about the source of the change(s)
+ * to the file.  For detailed descriptions of what these mean, see the Linux
+ * NTFS project NTFS documentation:
+ *	http://linux-ntfs.sourceforge.net/ntfs/files/usnjrnl.html
+ */
+enum {
+	USN_SOURCE_DATA_MANAGEMENT	  = const_cpu_to_le32(0x00000001),
+	USN_SOURCE_AUXILIARY_DATA	  = const_cpu_to_le32(0x00000002),
+	USN_SOURCE_REPLICATION_MANAGEMENT = const_cpu_to_le32(0x00000004),
+};
+
+typedef le32 USN_SOURCE_INFO_FLAGS;
+
+/*
+ * $DATA/$J attribute.  This is always non-resident, is marked as sparse, and
+ * is of variabled size.  It consists of a sequence of variable size
+ * USN_RECORDS.  The minimum allocated_size is allocation_delta as
+ * specified in $DATA/$Max.  When the maximum_size specified in $DATA/$Max is
+ * exceeded by more than allocation_delta bytes, allocation_delta bytes are
+ * allocated and appended to the $DATA/$J attribute and an equal number of
+ * bytes at the beginning of the attribute are freed and made sparse.  Note the
+ * making sparse only happens at volume checkpoints and hence the actual
+ * $DATA/$J size can exceed maximum_size + allocation_delta temporarily.
+ */
+typedef struct {
+/*Ofs*/
+/*   0*/le32 length;		/* Byte size of this record (8-byte
+				   aligned). */
+/*   4*/le16 major_ver;		/* Major version of the transaction log used
+				   for this record. */
+/*   6*/le16 minor_ver;		/* Minor version of the transaction log used
+				   for this record. */
+/*   8*/leMFT_REF mft_reference;/* The mft reference of the file (or
+				   directory) described by this record. */
+/*0x10*/leMFT_REF parent_directory;/* The mft reference of the parent
+				   directory of the file described by this
+				   record. */
+/*0x18*/leUSN usn;		/* The usn of this record.  Equals the offset
+				   within the $DATA/$J attribute. */
+/*0x20*/sle64 time;		/* Time when this record was created. */
+/*0x28*/USN_REASON_FLAGS reason;/* Reason flags (see above). */
+/*0x2c*/USN_SOURCE_INFO_FLAGS source_info;/* Source info flags (see above). */
+/*0x30*/le32 security_id;	/* File security_id copied from
+				   $STANDARD_INFORMATION. */
+/*0x34*/FILE_ATTR_FLAGS file_attributes;	/* File attributes copied from
+				   $STANDARD_INFORMATION or $FILE_NAME (not
+				   sure which). */
+/*0x38*/le16 file_name_size;	/* Size of the file name in bytes. */
+/*0x3a*/le16 file_name_offset;	/* Offset to the file name in bytes from the
+				   start of this record. */
+/*0x3c*/ntfschar file_name[0];	/* Use when creating only.  When reading use
+				   file_name_offset to determine the location
+				   of the name. */
+/* sizeof() = 60 (0x3c) bytes */
+} __attribute__ ((__packed__)) USN_RECORD;
+
+extern BOOL ntfs_stamp_usnjrnl(ntfs_volume *vol);
+
+#endif /* NTFS_RW */
+
+#endif /* _LINUX_NTFS_USNJRNL_H */
diff --git a/fs/ntfs/volume.h b/fs/ntfs/volume.h
index 4b97fa8..375cd20 100644
--- a/fs/ntfs/volume.h
+++ b/fs/ntfs/volume.h
@@ -2,7 +2,7 @@
  * volume.h - Defines for volume structures in NTFS Linux kernel driver. Part
  *	      of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2005 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -54,7 +54,7 @@ typedef struct {
 	mode_t dmask;			/* The mask for directory
 					   permissions. */
 	u8 mft_zone_multiplier;		/* Initial mft zone multiplier. */
-	u8 on_errors;			/* What to do on file system errors. */
+	u8 on_errors;			/* What to do on filesystem errors. */
 	/* NTFS bootsector provided information. */
 	u16 sector_size;		/* in bytes */
 	u8 sector_size_bits;		/* log2(sector_size) */
@@ -125,6 +125,10 @@ typedef struct {
 	/* $Quota stuff is NTFS3.0+ specific.  Unused/NULL otherwise. */
 	struct inode *quota_ino;	/* The VFS inode of $Quota. */
 	struct inode *quota_q_ino;	/* Attribute inode for $Quota/$Q. */
+	/* $UsnJrnl stuff is NTFS3.0+ specific.  Unused/NULL otherwise. */
+	struct inode *usnjrnl_ino;	/* The VFS inode of $UsnJrnl. */
+	struct inode *usnjrnl_max_ino;	/* Attribute inode for $UsnJrnl/$Max. */
+	struct inode *usnjrnl_j_ino;	/* Attribute inode for $UsnJrnl/$J. */
 #endif /* NTFS_RW */
 	struct nls_table *nls_map;
 } ntfs_volume;
@@ -141,6 +145,8 @@ typedef enum {
 				      file names in WIN32 namespace. */
 	NV_LogFileEmpty,	/* 1: $LogFile journal is empty. */
 	NV_QuotaOutOfDate,	/* 1: $Quota is out of date. */
+	NV_UsnJrnlStamped,	/* 1: $UsnJrnl has been stamped. */
+	NV_SparseEnabled,	/* 1: May create sparse files. */
 } ntfs_volume_flags;
 
 /*
@@ -167,5 +173,7 @@ NVOL_FNS(ShowSystemFiles)
 NVOL_FNS(CaseSensitive)
 NVOL_FNS(LogFileEmpty)
 NVOL_FNS(QuotaOutOfDate)
+NVOL_FNS(UsnJrnlStamped)
+NVOL_FNS(SparseEnabled)
 
 #endif /* _LINUX_NTFS_VOLUME_H */
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index 1aaf2c7..d9f614a 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -1980,7 +1980,17 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
       out_inserted_sd:
 	inode->i_nlink = 0;
 	th->t_trans_id = 0;	/* so the caller can't use this handle later */
-	iput(inode);
+
+	/* If we were inheriting an ACL, we need to release the lock so that
+	 * iput doesn't deadlock in reiserfs_delete_xattrs. The locking
+	 * code really needs to be reworked, but this will take care of it
+	 * for now. -jeffm */
+	if (REISERFS_I(dir)->i_acl_default) {
+		reiserfs_write_unlock_xattrs(dir->i_sb);
+		iput(inode);
+		reiserfs_write_lock_xattrs(dir->i_sb);
+	} else
+		iput(inode);
 	return err;
 }
 
diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c
index c66c27e..ca7989b 100644
--- a/fs/reiserfs/journal.c
+++ b/fs/reiserfs/journal.c
@@ -556,14 +556,14 @@ static inline void insert_journal_hash(struct reiserfs_journal_cnode **table,
 }
 
 /* lock the current transaction */
-inline static void lock_journal(struct super_block *p_s_sb)
+static inline void lock_journal(struct super_block *p_s_sb)
 {
 	PROC_INFO_INC(p_s_sb, journal.lock_journal);
 	down(&SB_JOURNAL(p_s_sb)->j_lock);
 }
 
 /* unlock the current transaction */
-inline static void unlock_journal(struct super_block *p_s_sb)
+static inline void unlock_journal(struct super_block *p_s_sb)
 {
 	up(&SB_JOURNAL(p_s_sb)->j_lock);
 }
diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c
index e386d3d..87ac9dc 100644
--- a/fs/reiserfs/xattr.c
+++ b/fs/reiserfs/xattr.c
@@ -39,7 +39,6 @@
 #include <linux/xattr.h>
 #include <linux/reiserfs_xattr.h>
 #include <linux/reiserfs_acl.h>
-#include <linux/mbcache.h>
 #include <asm/uaccess.h>
 #include <asm/checksum.h>
 #include <linux/smp_lock.h>
diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index 335288b9..4013d79 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -437,8 +437,8 @@ int sysfs_chmod_file(struct kobject *kobj, struct attribute *attr, mode_t mode)
 {
 	struct dentry *dir = kobj->dentry;
 	struct dentry *victim;
-	struct sysfs_dirent *sd;
-	umode_t umode = (mode & S_IALLUGO) | S_IFREG;
+	struct inode * inode;
+	struct iattr newattrs;
 	int res = -ENOENT;
 
 	down(&dir->d_inode->i_sem);
@@ -446,13 +446,15 @@ int sysfs_chmod_file(struct kobject *kobj, struct attribute *attr, mode_t mode)
 	if (!IS_ERR(victim)) {
 		if (victim->d_inode &&
 		    (victim->d_parent->d_inode == dir->d_inode)) {
-			sd = victim->d_fsdata;
-			attr->mode = mode;
-			sd->s_mode = umode;
-			victim->d_inode->i_mode = umode;
-			dput(victim);
-			res = 0;
+			inode = victim->d_inode;
+			down(&inode->i_sem);
+			newattrs.ia_mode = (mode & S_IALLUGO) |
+						(inode->i_mode & ~S_IALLUGO);
+			newattrs.ia_valid = ATTR_MODE | ATTR_CTIME;
+			res = notify_change(victim, &newattrs);
+			up(&inode->i_sem);
 		}
+		dput(victim);
 	}
 	up(&dir->d_inode->i_sem);
 
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index 8de13ba..d727dc9 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -85,7 +85,7 @@ int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 
 		if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
 			mode &= ~S_ISGID;
-		sd_iattr->ia_mode = mode;
+		sd_iattr->ia_mode = sd->s_mode = mode;
 	}
 
 	return error;