[PATCH] ext4: initial copy of files from ext3

Start of the ext4 patch series.  See Documentation/filesystems/ext4.txt for
details.

This is a simple copy of the files in fs/ext3 to fs/ext4 and
/usr/incude/linux/ext3* to /usr/include/ex4*

Signed-off-by: Dave Kleikamp <shaggy@austin.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

1 files changed, 1818 insertions, 0 deletions

diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
new file mode 100644
index 000000000000..b41a7d7e20f0
--- /dev/null
+++ b/fs/ext4/balloc.c
@@ -0,0 +1,1818 @@
+/*
+ *  linux/fs/ext3/balloc.c
+ *
+ * Copyright (C) 1992, 1993, 1994, 1995
+ * Remy Card (card@masi.ibp.fr)
+ * Laboratoire MASI - Institut Blaise Pascal
+ * Universite Pierre et Marie Curie (Paris VI)
+ *
+ *  Enhanced block allocation by Stephen Tweedie (sct@redhat.com), 1993
+ *  Big-endian to little-endian byte-swapping/bitmaps by
+ *        David S. Miller (davem@caip.rutgers.edu), 1995
+ */
+
+#include <linux/time.h>
+#include <linux/capability.h>
+#include <linux/fs.h>
+#include <linux/jbd.h>
+#include <linux/ext3_fs.h>
+#include <linux/ext3_jbd.h>
+#include <linux/quotaops.h>
+#include <linux/buffer_head.h>
+
+/*
+ * balloc.c contains the blocks allocation and deallocation routines
+ */
+
+/*
+ * The free blocks are managed by bitmaps.  A file system contains several
+ * blocks groups.  Each group contains 1 bitmap block for blocks, 1 bitmap
+ * block for inodes, N blocks for the inode table and data blocks.
+ *
+ * The file system contains group descriptors which are located after the
+ * super block.  Each descriptor contains the number of the bitmap block and
+ * the free blocks count in the block.  The descriptors are loaded in memory
+ * when a file system is mounted (see ext3_read_super).
+ */
+
+
+#define in_range(b, first, len)	((b) >= (first) && (b) <= (first) + (len) - 1)
+
+/**
+ * ext3_get_group_desc() -- load group descriptor from disk
+ * @sb:			super block
+ * @block_group:	given block group
+ * @bh:			pointer to the buffer head to store the block
+ *			group descriptor
+ */
+struct ext3_group_desc * ext3_get_group_desc(struct super_block * sb,
+					     unsigned int block_group,
+					     struct buffer_head ** bh)
+{
+	unsigned long group_desc;
+	unsigned long offset;
+	struct ext3_group_desc * desc;
+	struct ext3_sb_info *sbi = EXT3_SB(sb);
+
+	if (block_group >= sbi->s_groups_count) {
+		ext3_error (sb, "ext3_get_group_desc",
+			    "block_group >= groups_count - "
+			    "block_group = %d, groups_count = %lu",
+			    block_group, sbi->s_groups_count);
+
+		return NULL;
+	}
+	smp_rmb();
+
+	group_desc = block_group >> EXT3_DESC_PER_BLOCK_BITS(sb);
+	offset = block_group & (EXT3_DESC_PER_BLOCK(sb) - 1);
+	if (!sbi->s_group_desc[group_desc]) {
+		ext3_error (sb, "ext3_get_group_desc",
+			    "Group descriptor not loaded - "
+			    "block_group = %d, group_desc = %lu, desc = %lu",
+			     block_group, group_desc, offset);
+		return NULL;
+	}
+
+	desc = (struct ext3_group_desc *) sbi->s_group_desc[group_desc]->b_data;
+	if (bh)
+		*bh = sbi->s_group_desc[group_desc];
+	return desc + offset;
+}
+
+/**
+ * read_block_bitmap()
+ * @sb:			super block
+ * @block_group:	given block group
+ *
+ * Read the bitmap for a given block_group, reading into the specified
+ * slot in the superblock's bitmap cache.
+ *
+ * Return buffer_head on success or NULL in case of failure.
+ */
+static struct buffer_head *
+read_block_bitmap(struct super_block *sb, unsigned int block_group)
+{
+	struct ext3_group_desc * desc;
+	struct buffer_head * bh = NULL;
+
+	desc = ext3_get_group_desc (sb, block_group, NULL);
+	if (!desc)
+		goto error_out;
+	bh = sb_bread(sb, le32_to_cpu(desc->bg_block_bitmap));
+	if (!bh)
+		ext3_error (sb, "read_block_bitmap",
+			    "Cannot read block bitmap - "
+			    "block_group = %d, block_bitmap = %u",
+			    block_group, le32_to_cpu(desc->bg_block_bitmap));
+error_out:
+	return bh;
+}
+/*
+ * The reservation window structure operations
+ * --------------------------------------------
+ * Operations include:
+ * dump, find, add, remove, is_empty, find_next_reservable_window, etc.
+ *
+ * We use a red-black tree to represent per-filesystem reservation
+ * windows.
+ *
+ */
+
+/**
+ * __rsv_window_dump() -- Dump the filesystem block allocation reservation map
+ * @rb_root:		root of per-filesystem reservation rb tree
+ * @verbose:		verbose mode
+ * @fn:			function which wishes to dump the reservation map
+ *
+ * If verbose is turned on, it will print the whole block reservation
+ * windows(start, end).	Otherwise, it will only print out the "bad" windows,
+ * those windows that overlap with their immediate neighbors.
+ */
+#if 1
+static void __rsv_window_dump(struct rb_root *root, int verbose,
+			      const char *fn)
+{
+	struct rb_node *n;
+	struct ext3_reserve_window_node *rsv, *prev;
+	int bad;
+
+restart:
+	n = rb_first(root);
+	bad = 0;
+	prev = NULL;
+
+	printk("Block Allocation Reservation Windows Map (%s):\n", fn);
+	while (n) {
+		rsv = list_entry(n, struct ext3_reserve_window_node, rsv_node);
+		if (verbose)
+			printk("reservation window 0x%p "
+			       "start:  %lu, end:  %lu\n",
+			       rsv, rsv->rsv_start, rsv->rsv_end);
+		if (rsv->rsv_start && rsv->rsv_start >= rsv->rsv_end) {
+			printk("Bad reservation %p (start >= end)\n",
+			       rsv);
+			bad = 1;
+		}
+		if (prev && prev->rsv_end >= rsv->rsv_start) {
+			printk("Bad reservation %p (prev->end >= start)\n",
+			       rsv);
+			bad = 1;
+		}
+		if (bad) {
+			if (!verbose) {
+				printk("Restarting reservation walk in verbose mode\n");
+				verbose = 1;
+				goto restart;
+			}
+		}
+		n = rb_next(n);
+		prev = rsv;
+	}
+	printk("Window map complete.\n");
+	if (bad)
+		BUG();
+}
+#define rsv_window_dump(root, verbose) \
+	__rsv_window_dump((root), (verbose), __FUNCTION__)
+#else
+#define rsv_window_dump(root, verbose) do {} while (0)
+#endif
+
+/**
+ * goal_in_my_reservation()
+ * @rsv:		inode's reservation window
+ * @grp_goal:		given goal block relative to the allocation block group
+ * @group:		the current allocation block group
+ * @sb:			filesystem super block
+ *
+ * Test if the given goal block (group relative) is within the file's
+ * own block reservation window range.
+ *
+ * If the reservation window is outside the goal allocation group, return 0;
+ * grp_goal (given goal block) could be -1, which means no specific
+ * goal block. In this case, always return 1.
+ * If the goal block is within the reservation window, return 1;
+ * otherwise, return 0;
+ */
+static int
+goal_in_my_reservation(struct ext3_reserve_window *rsv, ext3_grpblk_t grp_goal,
+			unsigned int group, struct super_block * sb)
+{
+	ext3_fsblk_t group_first_block, group_last_block;
+
+	group_first_block = ext3_group_first_block_no(sb, group);
+	group_last_block = group_first_block + (EXT3_BLOCKS_PER_GROUP(sb) - 1);
+
+	if ((rsv->_rsv_start > group_last_block) ||
+	    (rsv->_rsv_end < group_first_block))
+		return 0;
+	if ((grp_goal >= 0) && ((grp_goal + group_first_block < rsv->_rsv_start)
+		|| (grp_goal + group_first_block > rsv->_rsv_end)))
+		return 0;
+	return 1;
+}
+
+/**
+ * search_reserve_window()
+ * @rb_root:		root of reservation tree
+ * @goal:		target allocation block
+ *
+ * Find the reserved window which includes the goal, or the previous one
+ * if the goal is not in any window.
+ * Returns NULL if there are no windows or if all windows start after the goal.
+ */
+static struct ext3_reserve_window_node *
+search_reserve_window(struct rb_root *root, ext3_fsblk_t goal)
+{
+	struct rb_node *n = root->rb_node;
+	struct ext3_reserve_window_node *rsv;
+
+	if (!n)
+		return NULL;
+
+	do {
+		rsv = rb_entry(n, struct ext3_reserve_window_node, rsv_node);
+
+		if (goal < rsv->rsv_start)
+			n = n->rb_left;
+		else if (goal > rsv->rsv_end)
+			n = n->rb_right;
+		else
+			return rsv;
+	} while (n);
+	/*
+	 * We've fallen off the end of the tree: the goal wasn't inside
+	 * any particular node.  OK, the previous node must be to one
+	 * side of the interval containing the goal.  If it's the RHS,
+	 * we need to back up one.
+	 */
+	if (rsv->rsv_start > goal) {
+		n = rb_prev(&rsv->rsv_node);
+		rsv = rb_entry(n, struct ext3_reserve_window_node, rsv_node);
+	}
+	return rsv;
+}
+
+/**
+ * ext3_rsv_window_add() -- Insert a window to the block reservation rb tree.
+ * @sb:			super block
+ * @rsv:		reservation window to add
+ *
+ * Must be called with rsv_lock hold.
+ */
+void ext3_rsv_window_add(struct super_block *sb,
+		    struct ext3_reserve_window_node *rsv)
+{
+	struct rb_root *root = &EXT3_SB(sb)->s_rsv_window_root;
+	struct rb_node *node = &rsv->rsv_node;
+	ext3_fsblk_t start = rsv->rsv_start;
+
+	struct rb_node ** p = &root->rb_node;
+	struct rb_node * parent = NULL;
+	struct ext3_reserve_window_node *this;
+
+	while (*p)
+	{
+		parent = *p;
+		this = rb_entry(parent, struct ext3_reserve_window_node, rsv_node);
+
+		if (start < this->rsv_start)
+			p = &(*p)->rb_left;
+		else if (start > this->rsv_end)
+			p = &(*p)->rb_right;
+		else {
+			rsv_window_dump(root, 1);
+			BUG();
+		}
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+}
+
+/**
+ * ext3_rsv_window_remove() -- unlink a window from the reservation rb tree
+ * @sb:			super block
+ * @rsv:		reservation window to remove
+ *
+ * Mark the block reservation window as not allocated, and unlink it
+ * from the filesystem reservation window rb tree. Must be called with
+ * rsv_lock hold.
+ */
+static void rsv_window_remove(struct super_block *sb,
+			      struct ext3_reserve_window_node *rsv)
+{
+	rsv->rsv_start = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+	rsv->rsv_end = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+	rsv->rsv_alloc_hit = 0;
+	rb_erase(&rsv->rsv_node, &EXT3_SB(sb)->s_rsv_window_root);
+}
+
+/*
+ * rsv_is_empty() -- Check if the reservation window is allocated.
+ * @rsv:		given reservation window to check
+ *
+ * returns 1 if the end block is EXT3_RESERVE_WINDOW_NOT_ALLOCATED.
+ */
+static inline int rsv_is_empty(struct ext3_reserve_window *rsv)
+{
+	/* a valid reservation end block could not be 0 */
+	return rsv->_rsv_end == EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+}
+
+/**
+ * ext3_init_block_alloc_info()
+ * @inode:		file inode structure
+ *
+ * Allocate and initialize the	reservation window structure, and
+ * link the window to the ext3 inode structure at last
+ *
+ * The reservation window structure is only dynamically allocated
+ * and linked to ext3 inode the first time the open file
+ * needs a new block. So, before every ext3_new_block(s) call, for
+ * regular files, we should check whether the reservation window
+ * structure exists or not. In the latter case, this function is called.
+ * Fail to do so will result in block reservation being turned off for that
+ * open file.
+ *
+ * This function is called from ext3_get_blocks_handle(), also called
+ * when setting the reservation window size through ioctl before the file
+ * is open for write (needs block allocation).
+ *
+ * Needs truncate_mutex protection prior to call this function.
+ */
+void ext3_init_block_alloc_info(struct inode *inode)
+{
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	struct ext3_block_alloc_info *block_i = ei->i_block_alloc_info;
+	struct super_block *sb = inode->i_sb;
+
+	block_i = kmalloc(sizeof(*block_i), GFP_NOFS);
+	if (block_i) {
+		struct ext3_reserve_window_node *rsv = &block_i->rsv_window_node;
+
+		rsv->rsv_start = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+		rsv->rsv_end = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+
+		/*
+		 * if filesystem is mounted with NORESERVATION, the goal
+		 * reservation window size is set to zero to indicate
+		 * block reservation is off
+		 */
+		if (!test_opt(sb, RESERVATION))
+			rsv->rsv_goal_size = 0;
+		else
+			rsv->rsv_goal_size = EXT3_DEFAULT_RESERVE_BLOCKS;
+		rsv->rsv_alloc_hit = 0;
+		block_i->last_alloc_logical_block = 0;
+		block_i->last_alloc_physical_block = 0;
+	}
+	ei->i_block_alloc_info = block_i;
+}
+
+/**
+ * ext3_discard_reservation()
+ * @inode:		inode
+ *
+ * Discard(free) block reservation window on last file close, or truncate
+ * or at last iput().
+ *
+ * It is being called in three cases:
+ *	ext3_release_file(): last writer close the file
+ *	ext3_clear_inode(): last iput(), when nobody link to this file.
+ *	ext3_truncate(): when the block indirect map is about to change.
+ *
+ */
+void ext3_discard_reservation(struct inode *inode)
+{
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	struct ext3_block_alloc_info *block_i = ei->i_block_alloc_info;
+	struct ext3_reserve_window_node *rsv;
+	spinlock_t *rsv_lock = &EXT3_SB(inode->i_sb)->s_rsv_window_lock;
+
+	if (!block_i)
+		return;
+
+	rsv = &block_i->rsv_window_node;
+	if (!rsv_is_empty(&rsv->rsv_window)) {
+		spin_lock(rsv_lock);
+		if (!rsv_is_empty(&rsv->rsv_window))
+			rsv_window_remove(inode->i_sb, rsv);
+		spin_unlock(rsv_lock);
+	}
+}
+
+/**
+ * ext3_free_blocks_sb() -- Free given blocks and update quota
+ * @handle:			handle to this transaction
+ * @sb:				super block
+ * @block:			start physcial block to free
+ * @count:			number of blocks to free
+ * @pdquot_freed_blocks:	pointer to quota
+ */
+void ext3_free_blocks_sb(handle_t *handle, struct super_block *sb,
+			 ext3_fsblk_t block, unsigned long count,
+			 unsigned long *pdquot_freed_blocks)
+{
+	struct buffer_head *bitmap_bh = NULL;
+	struct buffer_head *gd_bh;
+	unsigned long block_group;
+	ext3_grpblk_t bit;
+	unsigned long i;
+	unsigned long overflow;
+	struct ext3_group_desc * desc;
+	struct ext3_super_block * es;
+	struct ext3_sb_info *sbi;
+	int err = 0, ret;
+	ext3_grpblk_t group_freed;
+
+	*pdquot_freed_blocks = 0;
+	sbi = EXT3_SB(sb);
+	es = sbi->s_es;
+	if (block < le32_to_cpu(es->s_first_data_block) ||
+	    block + count < block ||
+	    block + count > le32_to_cpu(es->s_blocks_count)) {
+		ext3_error (sb, "ext3_free_blocks",
+			    "Freeing blocks not in datazone - "
+			    "block = "E3FSBLK", count = %lu", block, count);
+		goto error_return;
+	}
+
+	ext3_debug ("freeing block(s) %lu-%lu\n", block, block + count - 1);
+
+do_more:
+	overflow = 0;
+	block_group = (block - le32_to_cpu(es->s_first_data_block)) /
+		      EXT3_BLOCKS_PER_GROUP(sb);
+	bit = (block - le32_to_cpu(es->s_first_data_block)) %
+		      EXT3_BLOCKS_PER_GROUP(sb);
+	/*
+	 * Check to see if we are freeing blocks across a group
+	 * boundary.
+	 */
+	if (bit + count > EXT3_BLOCKS_PER_GROUP(sb)) {
+		overflow = bit + count - EXT3_BLOCKS_PER_GROUP(sb);
+		count -= overflow;
+	}
+	brelse(bitmap_bh);
+	bitmap_bh = read_block_bitmap(sb, block_group);
+	if (!bitmap_bh)
+		goto error_return;
+	desc = ext3_get_group_desc (sb, block_group, &gd_bh);
+	if (!desc)
+		goto error_return;
+
+	if (in_range (le32_to_cpu(desc->bg_block_bitmap), block, count) ||
+	    in_range (le32_to_cpu(desc->bg_inode_bitmap), block, count) ||
+	    in_range (block, le32_to_cpu(desc->bg_inode_table),
+		      sbi->s_itb_per_group) ||
+	    in_range (block + count - 1, le32_to_cpu(desc->bg_inode_table),
+		      sbi->s_itb_per_group))
+		ext3_error (sb, "ext3_free_blocks",
+			    "Freeing blocks in system zones - "
+			    "Block = "E3FSBLK", count = %lu",
+			    block, count);
+
+	/*
+	 * We are about to start releasing blocks in the bitmap,
+	 * so we need undo access.
+	 */
+	/* @@@ check errors */
+	BUFFER_TRACE(bitmap_bh, "getting undo access");
+	err = ext3_journal_get_undo_access(handle, bitmap_bh);
+	if (err)
+		goto error_return;
+
+	/*
+	 * We are about to modify some metadata.  Call the journal APIs
+	 * to unshare ->b_data if a currently-committing transaction is
+	 * using it
+	 */
+	BUFFER_TRACE(gd_bh, "get_write_access");
+	err = ext3_journal_get_write_access(handle, gd_bh);
+	if (err)
+		goto error_return;
+
+	jbd_lock_bh_state(bitmap_bh);
+
+	for (i = 0, group_freed = 0; i < count; i++) {
+		/*
+		 * An HJ special.  This is expensive...
+		 */
+#ifdef CONFIG_JBD_DEBUG
+		jbd_unlock_bh_state(bitmap_bh);
+		{
+			struct buffer_head *debug_bh;
+			debug_bh = sb_find_get_block(sb, block + i);
+			if (debug_bh) {
+				BUFFER_TRACE(debug_bh, "Deleted!");
+				if (!bh2jh(bitmap_bh)->b_committed_data)
+					BUFFER_TRACE(debug_bh,
+						"No commited data in bitmap");
+				BUFFER_TRACE2(debug_bh, bitmap_bh, "bitmap");
+				__brelse(debug_bh);
+			}
+		}
+		jbd_lock_bh_state(bitmap_bh);
+#endif
+		if (need_resched()) {
+			jbd_unlock_bh_state(bitmap_bh);
+			cond_resched();
+			jbd_lock_bh_state(bitmap_bh);
+		}
+		/* @@@ This prevents newly-allocated data from being
+		 * freed and then reallocated within the same
+		 * transaction.
+		 *
+		 * Ideally we would want to allow that to happen, but to
+		 * do so requires making journal_forget() capable of
+		 * revoking the queued write of a data block, which
+		 * implies blocking on the journal lock.  *forget()
+		 * cannot block due to truncate races.
+		 *
+		 * Eventually we can fix this by making journal_forget()
+		 * return a status indicating whether or not it was able
+		 * to revoke the buffer.  On successful revoke, it is
+		 * safe not to set the allocation bit in the committed
+		 * bitmap, because we know that there is no outstanding
+		 * activity on the buffer any more and so it is safe to
+		 * reallocate it.
+		 */
+		BUFFER_TRACE(bitmap_bh, "set in b_committed_data");
+		J_ASSERT_BH(bitmap_bh,
+				bh2jh(bitmap_bh)->b_committed_data != NULL);
+		ext3_set_bit_atomic(sb_bgl_lock(sbi, block_group), bit + i,
+				bh2jh(bitmap_bh)->b_committed_data);
+
+		/*
+		 * We clear the bit in the bitmap after setting the committed
+		 * data bit, because this is the reverse order to that which
+		 * the allocator uses.
+		 */
+		BUFFER_TRACE(bitmap_bh, "clear bit");
+		if (!ext3_clear_bit_atomic(sb_bgl_lock(sbi, block_group),
+						bit + i, bitmap_bh->b_data)) {
+			jbd_unlock_bh_state(bitmap_bh);
+			ext3_error(sb, __FUNCTION__,
+				"bit already cleared for block "E3FSBLK,
+				 block + i);
+			jbd_lock_bh_state(bitmap_bh);
+			BUFFER_TRACE(bitmap_bh, "bit already cleared");
+		} else {
+			group_freed++;
+		}
+	}
+	jbd_unlock_bh_state(bitmap_bh);
+
+	spin_lock(sb_bgl_lock(sbi, block_group));
+	desc->bg_free_blocks_count =
+		cpu_to_le16(le16_to_cpu(desc->bg_free_blocks_count) +
+			group_freed);
+	spin_unlock(sb_bgl_lock(sbi, block_group));
+	percpu_counter_mod(&sbi->s_freeblocks_counter, count);
+
+	/* We dirtied the bitmap block */
+	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
+	err = ext3_journal_dirty_metadata(handle, bitmap_bh);
+
+	/* And the group descriptor block */
+	BUFFER_TRACE(gd_bh, "dirtied group descriptor block");
+	ret = ext3_journal_dirty_metadata(handle, gd_bh);
+	if (!err) err = ret;
+	*pdquot_freed_blocks += group_freed;
+
+	if (overflow && !err) {
+		block += count;
+		count = overflow;
+		goto do_more;
+	}
+	sb->s_dirt = 1;
+error_return:
+	brelse(bitmap_bh);
+	ext3_std_error(sb, err);
+	return;
+}
+
+/**
+ * ext3_free_blocks() -- Free given blocks and update quota
+ * @handle:		handle for this transaction
+ * @inode:		inode
+ * @block:		start physical block to free
+ * @count:		number of blocks to count
+ */
+void ext3_free_blocks(handle_t *handle, struct inode *inode,
+			ext3_fsblk_t block, unsigned long count)
+{
+	struct super_block * sb;
+	unsigned long dquot_freed_blocks;
+
+	sb = inode->i_sb;
+	if (!sb) {
+		printk ("ext3_free_blocks: nonexistent device");
+		return;
+	}
+	ext3_free_blocks_sb(handle, sb, block, count, &dquot_freed_blocks);
+	if (dquot_freed_blocks)
+		DQUOT_FREE_BLOCK(inode, dquot_freed_blocks);
+	return;
+}
+
+/**
+ * ext3_test_allocatable()
+ * @nr:			given allocation block group
+ * @bh:			bufferhead contains the bitmap of the given block group
+ *
+ * For ext3 allocations, we must not reuse any blocks which are
+ * allocated in the bitmap buffer's "last committed data" copy.  This
+ * prevents deletes from freeing up the page for reuse until we have
+ * committed the delete transaction.
+ *
+ * If we didn't do this, then deleting something and reallocating it as
+ * data would allow the old block to be overwritten before the
+ * transaction committed (because we force data to disk before commit).
+ * This would lead to corruption if we crashed between overwriting the
+ * data and committing the delete.
+ *
+ * @@@ We may want to make this allocation behaviour conditional on
+ * data-writes at some point, and disable it for metadata allocations or
+ * sync-data inodes.
+ */
+static int ext3_test_allocatable(ext3_grpblk_t nr, struct buffer_head *bh)
+{
+	int ret;
+	struct journal_head *jh = bh2jh(bh);
+
+	if (ext3_test_bit(nr, bh->b_data))
+		return 0;
+
+	jbd_lock_bh_state(bh);
+	if (!jh->b_committed_data)
+		ret = 1;
+	else
+		ret = !ext3_test_bit(nr, jh->b_committed_data);
+	jbd_unlock_bh_state(bh);
+	return ret;
+}
+
+/**
+ * bitmap_search_next_usable_block()
+ * @start:		the starting block (group relative) of the search
+ * @bh:			bufferhead contains the block group bitmap
+ * @maxblocks:		the ending block (group relative) of the reservation
+ *
+ * The bitmap search --- search forward alternately through the actual
+ * bitmap on disk and the last-committed copy in journal, until we find a
+ * bit free in both bitmaps.
+ */
+static ext3_grpblk_t
+bitmap_search_next_usable_block(ext3_grpblk_t start, struct buffer_head *bh,
+					ext3_grpblk_t maxblocks)
+{
+	ext3_grpblk_t next;
+	struct journal_head *jh = bh2jh(bh);
+
+	while (start < maxblocks) {
+		next = ext3_find_next_zero_bit(bh->b_data, maxblocks, start);
+		if (next >= maxblocks)
+			return -1;
+		if (ext3_test_allocatable(next, bh))
+			return next;
+		jbd_lock_bh_state(bh);
+		if (jh->b_committed_data)
+			start = ext3_find_next_zero_bit(jh->b_committed_data,
+							maxblocks, next);
+		jbd_unlock_bh_state(bh);
+	}
+	return -1;
+}
+
+/**
+ * find_next_usable_block()
+ * @start:		the starting block (group relative) to find next
+ *			allocatable block in bitmap.
+ * @bh:			bufferhead contains the block group bitmap
+ * @maxblocks:		the ending block (group relative) for the search
+ *
+ * Find an allocatable block in a bitmap.  We honor both the bitmap and
+ * its last-committed copy (if that exists), and perform the "most
+ * appropriate allocation" algorithm of looking for a free block near
+ * the initial goal; then for a free byte somewhere in the bitmap; then
+ * for any free bit in the bitmap.
+ */
+static ext3_grpblk_t
+find_next_usable_block(ext3_grpblk_t start, struct buffer_head *bh,
+			ext3_grpblk_t maxblocks)
+{
+	ext3_grpblk_t here, next;
+	char *p, *r;
+
+	if (start > 0) {
+		/*
+		 * The goal was occupied; search forward for a free
+		 * block within the next XX blocks.
+		 *
+		 * end_goal is more or less random, but it has to be
+		 * less than EXT3_BLOCKS_PER_GROUP. Aligning up to the
+		 * next 64-bit boundary is simple..
+		 */
+		ext3_grpblk_t end_goal = (start + 63) & ~63;
+		if (end_goal > maxblocks)
+			end_goal = maxblocks;
+		here = ext3_find_next_zero_bit(bh->b_data, end_goal, start);
+		if (here < end_goal && ext3_test_allocatable(here, bh))
+			return here;
+		ext3_debug("Bit not found near goal\n");
+	}
+
+	here = start;
+	if (here < 0)
+		here = 0;
+
+	p = ((char *)bh->b_data) + (here >> 3);
+	r = memscan(p, 0, (maxblocks - here + 7) >> 3);
+	next = (r - ((char *)bh->b_data)) << 3;
+
+	if (next < maxblocks && next >= start && ext3_test_allocatable(next, bh))
+		return next;
+
+	/*
+	 * The bitmap search --- search forward alternately through the actual
+	 * bitmap and the last-committed copy until we find a bit free in
+	 * both
+	 */
+	here = bitmap_search_next_usable_block(here, bh, maxblocks);
+	return here;
+}
+
+/**
+ * claim_block()
+ * @block:		the free block (group relative) to allocate
+ * @bh:			the bufferhead containts the block group bitmap
+ *
+ * We think we can allocate this block in this bitmap.  Try to set the bit.
+ * If that succeeds then check that nobody has allocated and then freed the
+ * block since we saw that is was not marked in b_committed_data.  If it _was_
+ * allocated and freed then clear the bit in the bitmap again and return
+ * zero (failure).
+ */
+static inline int
+claim_block(spinlock_t *lock, ext3_grpblk_t block, struct buffer_head *bh)
+{
+	struct journal_head *jh = bh2jh(bh);
+	int ret;
+
+	if (ext3_set_bit_atomic(lock, block, bh->b_data))
+		return 0;
+	jbd_lock_bh_state(bh);
+	if (jh->b_committed_data && ext3_test_bit(block,jh->b_committed_data)) {
+		ext3_clear_bit_atomic(lock, block, bh->b_data);
+		ret = 0;
+	} else {
+		ret = 1;
+	}
+	jbd_unlock_bh_state(bh);
+	return ret;
+}
+
+/**
+ * ext3_try_to_allocate()
+ * @sb:			superblock
+ * @handle:		handle to this transaction
+ * @group:		given allocation block group
+ * @bitmap_bh:		bufferhead holds the block bitmap
+ * @grp_goal:		given target block within the group
+ * @count:		target number of blocks to allocate
+ * @my_rsv:		reservation window
+ *
+ * Attempt to allocate blocks within a give range. Set the range of allocation
+ * first, then find the first free bit(s) from the bitmap (within the range),
+ * and at last, allocate the blocks by claiming the found free bit as allocated.
+ *
+ * To set the range of this allocation:
+ *	if there is a reservation window, only try to allocate block(s) from the
+ *	file's own reservation window;
+ *	Otherwise, the allocation range starts from the give goal block, ends at
+ *	the block group's last block.
+ *
+ * If we failed to allocate the desired block then we may end up crossing to a
+ * new bitmap.  In that case we must release write access to the old one via
+ * ext3_journal_release_buffer(), else we'll run out of credits.
+ */
+static ext3_grpblk_t
+ext3_try_to_allocate(struct super_block *sb, handle_t *handle, int group,
+			struct buffer_head *bitmap_bh, ext3_grpblk_t grp_goal,
+			unsigned long *count, struct ext3_reserve_window *my_rsv)
+{
+	ext3_fsblk_t group_first_block;
+	ext3_grpblk_t start, end;
+	unsigned long num = 0;
+
+	/* we do allocation within the reservation window if we have a window */
+	if (my_rsv) {
+		group_first_block = ext3_group_first_block_no(sb, group);
+		if (my_rsv->_rsv_start >= group_first_block)
+			start = my_rsv->_rsv_start - group_first_block;
+		else
+			/* reservation window cross group boundary */
+			start = 0;
+		end = my_rsv->_rsv_end - group_first_block + 1;
+		if (end > EXT3_BLOCKS_PER_GROUP(sb))
+			/* reservation window crosses group boundary */
+			end = EXT3_BLOCKS_PER_GROUP(sb);
+		if ((start <= grp_goal) && (grp_goal < end))
+			start = grp_goal;
+		else
+			grp_goal = -1;
+	} else {
+		if (grp_goal > 0)
+			start = grp_goal;
+		else
+			start = 0;
+		end = EXT3_BLOCKS_PER_GROUP(sb);
+	}
+
+	BUG_ON(start > EXT3_BLOCKS_PER_GROUP(sb));
+
+repeat:
+	if (grp_goal < 0 || !ext3_test_allocatable(grp_goal, bitmap_bh)) {
+		grp_goal = find_next_usable_block(start, bitmap_bh, end);
+		if (grp_goal < 0)
+			goto fail_access;
+		if (!my_rsv) {
+			int i;
+
+			for (i = 0; i < 7 && grp_goal > start &&
+					ext3_test_allocatable(grp_goal - 1,
+								bitmap_bh);
+					i++, grp_goal--)
+				;
+		}
+	}
+	start = grp_goal;
+
+	if (!claim_block(sb_bgl_lock(EXT3_SB(sb), group),
+		grp_goal, bitmap_bh)) {
+		/*
+		 * The block was allocated by another thread, or it was
+		 * allocated and then freed by another thread
+		 */
+		start++;
+		grp_goal++;
+		if (start >= end)
+			goto fail_access;
+		goto repeat;
+	}
+	num++;
+	grp_goal++;
+	while (num < *count && grp_goal < end
+		&& ext3_test_allocatable(grp_goal, bitmap_bh)
+		&& claim_block(sb_bgl_lock(EXT3_SB(sb), group),
+				grp_goal, bitmap_bh)) {
+		num++;
+		grp_goal++;
+	}
+	*count = num;
+	return grp_goal - num;
+fail_access:
+	*count = num;
+	return -1;
+}
+
+/**
+ *	find_next_reservable_window():
+ *		find a reservable space within the given range.
+ *		It does not allocate the reservation window for now:
+ *		alloc_new_reservation() will do the work later.
+ *
+ *	@search_head: the head of the searching list;
+ *		This is not necessarily the list head of the whole filesystem
+ *
+ *		We have both head and start_block to assist the search
+ *		for the reservable space. The list starts from head,
+ *		but we will shift to the place where start_block is,
+ *		then start from there, when looking for a reservable space.
+ *
+ *	@size: the target new reservation window size
+ *
+ *	@group_first_block: the first block we consider to start
+ *			the real search from
+ *
+ *	@last_block:
+ *		the maximum block number that our goal reservable space
+ *		could start from. This is normally the last block in this
+ *		group. The search will end when we found the start of next
+ *		possible reservable space is out of this boundary.
+ *		This could handle the cross boundary reservation window
+ *		request.
+ *
+ *	basically we search from the given range, rather than the whole
+ *	reservation double linked list, (start_block, last_block)
+ *	to find a free region that is of my size and has not
+ *	been reserved.
+ *
+ */
+static int find_next_reservable_window(
+				struct ext3_reserve_window_node *search_head,
+				struct ext3_reserve_window_node *my_rsv,
+				struct super_block * sb,
+				ext3_fsblk_t start_block,
+				ext3_fsblk_t last_block)
+{
+	struct rb_node *next;
+	struct ext3_reserve_window_node *rsv, *prev;
+	ext3_fsblk_t cur;
+	int size = my_rsv->rsv_goal_size;
+
+	/* TODO: make the start of the reservation window byte-aligned */
+	/* cur = *start_block & ~7;*/
+	cur = start_block;
+	rsv = search_head;
+	if (!rsv)
+		return -1;
+
+	while (1) {
+		if (cur <= rsv->rsv_end)
+			cur = rsv->rsv_end + 1;
+
+		/* TODO?
+		 * in the case we could not find a reservable space
+		 * that is what is expected, during the re-search, we could
+		 * remember what's the largest reservable space we could have
+		 * and return that one.
+		 *
+		 * For now it will fail if we could not find the reservable
+		 * space with expected-size (or more)...
+		 */
+		if (cur > last_block)
+			return -1;		/* fail */
+
+		prev = rsv;
+		next = rb_next(&rsv->rsv_node);
+		rsv = list_entry(next,struct ext3_reserve_window_node,rsv_node);
+
+		/*
+		 * Reached the last reservation, we can just append to the
+		 * previous one.
+		 */
+		if (!next)
+			break;
+
+		if (cur + size <= rsv->rsv_start) {
+			/*
+			 * Found a reserveable space big enough.  We could
+			 * have a reservation across the group boundary here
+			 */
+			break;
+		}
+	}
+	/*
+	 * we come here either :
+	 * when we reach the end of the whole list,
+	 * and there is empty reservable space after last entry in the list.
+	 * append it to the end of the list.
+	 *
+	 * or we found one reservable space in the middle of the list,
+	 * return the reservation window that we could append to.
+	 * succeed.
+	 */
+
+	if ((prev != my_rsv) && (!rsv_is_empty(&my_rsv->rsv_window)))
+		rsv_window_remove(sb, my_rsv);
+
+	/*
+	 * Let's book the whole avaliable window for now.  We will check the
+	 * disk bitmap later and then, if there are free blocks then we adjust
+	 * the window size if it's larger than requested.
+	 * Otherwise, we will remove this node from the tree next time
+	 * call find_next_reservable_window.
+	 */
+	my_rsv->rsv_start = cur;
+	my_rsv->rsv_end = cur + size - 1;
+	my_rsv->rsv_alloc_hit = 0;
+
+	if (prev != my_rsv)
+		ext3_rsv_window_add(sb, my_rsv);
+
+	return 0;
+}
+
+/**
+ *	alloc_new_reservation()--allocate a new reservation window
+ *
+ *		To make a new reservation, we search part of the filesystem
+ *		reservation list (the list that inside the group). We try to
+ *		allocate a new reservation window near the allocation goal,
+ *		or the beginning of the group, if there is no goal.
+ *
+ *		We first find a reservable space after the goal, then from
+ *		there, we check the bitmap for the first free block after
+ *		it. If there is no free block until the end of group, then the
+ *		whole group is full, we failed. Otherwise, check if the free
+ *		block is inside the expected reservable space, if so, we
+ *		succeed.
+ *		If the first free block is outside the reservable space, then
+ *		start from the first free block, we search for next available
+ *		space, and go on.
+ *
+ *	on succeed, a new reservation will be found and inserted into the list
+ *	It contains at least one free block, and it does not overlap with other
+ *	reservation windows.
+ *
+ *	failed: we failed to find a reservation window in this group
+ *
+ *	@rsv: the reservation
+ *
+ *	@grp_goal: The goal (group-relative).  It is where the search for a
+ *		free reservable space should start from.
+ *		if we have a grp_goal(grp_goal >0 ), then start from there,
+ *		no grp_goal(grp_goal = -1), we start from the first block
+ *		of the group.
+ *
+ *	@sb: the super block
+ *	@group: the group we are trying to allocate in
+ *	@bitmap_bh: the block group block bitmap
+ *
+ */
+static int alloc_new_reservation(struct ext3_reserve_window_node *my_rsv,
+		ext3_grpblk_t grp_goal, struct super_block *sb,
+		unsigned int group, struct buffer_head *bitmap_bh)
+{
+	struct ext3_reserve_window_node *search_head;
+	ext3_fsblk_t group_first_block, group_end_block, start_block;
+	ext3_grpblk_t first_free_block;
+	struct rb_root *fs_rsv_root = &EXT3_SB(sb)->s_rsv_window_root;
+	unsigned long size;
+	int ret;
+	spinlock_t *rsv_lock = &EXT3_SB(sb)->s_rsv_window_lock;
+
+	group_first_block = ext3_group_first_block_no(sb, group);
+	group_end_block = group_first_block + (EXT3_BLOCKS_PER_GROUP(sb) - 1);
+
+	if (grp_goal < 0)
+		start_block = group_first_block;
+	else
+		start_block = grp_goal + group_first_block;
+
+	size = my_rsv->rsv_goal_size;
+
+	if (!rsv_is_empty(&my_rsv->rsv_window)) {
+		/*
+		 * if the old reservation is cross group boundary
+		 * and if the goal is inside the old reservation window,
+		 * we will come here when we just failed to allocate from
+		 * the first part of the window. We still have another part
+		 * that belongs to the next group. In this case, there is no
+		 * point to discard our window and try to allocate a new one
+		 * in this group(which will fail). we should
+		 * keep the reservation window, just simply move on.
+		 *
+		 * Maybe we could shift the start block of the reservation
+		 * w