Merge with /home/shaggy/git/linus-clean/

Signed-off-by: Dave Kleikamp <shaggy@austin.ibm.com>
2026-03-06 15:25:10 -08:00 · 2005-07-19 13:46:53 -05:00
parent 3d9b1cdd24 f60f700876
commit 21d1ee8b37
65 changed files with 2955 additions and 1428 deletions
@@ -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 (!(*result)) {
+		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);
 }
@@ -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 int jffs2_block_check_erase(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb, uint32_t *bad_offset)
+{
+	void *ebuf;
+	uint32_t ofs;
+	size_t retlen;
+	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. Refiling\n", jeb->offset);
+		return -EAGAIN;
+	}
+
+	D1(printk(KERN_DEBUG "Verifying erase at 0x%08x\n", jeb->offset));
+
+	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;
+
+		*bad_offset = ofs;
+
+		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();
+	}
+	ret = 0;
+fail:
+	kfree(ebuf);
+	return ret;
+}
+
 static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb)
 {
 	struct jffs2_raw_node_ref *marker_ref = NULL;
-	unsigned char *ebuf;
 	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;
-		}
+	switch (jffs2_block_check_erase(c, jeb, &bad_offset)) {
+	case -EAGAIN:	goto refile;
+	case -EIO:	goto filebad;
 	}
-	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;
-
-		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;
-
-			bad_offset = ofs;
-
-			ret = c->mtd->read(c->mtd, 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 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;
-				}
-			}
-			ofs += readlen;
-			cond_resched();
-		}
-		kfree(ebuf);
-	}
-
-	bad_offset = jeb->offset;

 	/* 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;
+
+	/* Cleanmarker in oob area or no cleanmarker at all ? */
+	if (jffs2_cleanmarker_oob(c) || c->cleanmarker_size == 0) {
+
+		if (jffs2_cleanmarker_oob(c)) {
+			if (jffs2_write_nand_cleanmarker(c, jeb))
+				goto filebad;
+		}

-		if (jffs2_write_nand_cleanmarker(c, jeb))
-			goto bad2;
-			
 		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 if (c->cleanmarker_size == 0) {
-		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;
+}
@@ -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
@@ -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
@@ -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);
@@ -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 */
@@ -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().
@@ -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;
 	}
@@ -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;
@@ -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);
@@ -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;
 }

@@ -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)

 /*
@@ -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;
@@ -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. */
@@ -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);
@@ -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. */
+};
@@ -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
--- a/Show More
+++ b/Show More