Date:	Thu, 27 Dec 2007 20:11:41 +0900
From:	Akira Fujita <a-fujita@...jp.nec.com>
To:	linux-ext4@...r.kernel.org, linux-fsdevel@...r.kernel.org
Subject: [RFC][PATCH 2/5] Allocate new contiguous blocks with mballoc

Search contiguous free blocks with mutil-block allocation
and allocate them for the temporary inode.

*This patch is applied on the top of
 ext4 git tree(linux-2.6.24-rc5).
http://repo.or.cz/r/ext4-patch-queue.git

Signed-off-by: Takashi Sato <t-sato@...jp.nec.com>
Signed-off-by: Akira Fujita <a-fujita@...jp.nec.com>
---
diff -X linux-2.6.24-rc5-defrag/Documentation/dontdiff -upNr linux-2.6.24-rc5-alloc-block/fs/ext4/defrag.c linux-2.6.24-rc5-move-data/fs/ext4/defrag.c
--- linux-2.6.24-rc5-alloc-block/fs/ext4/defrag.c	1970-01-01 09:00:00.000000000 +0900
+++ linux-2.6.24-rc5-move-data/fs/ext4/defrag.c	2007-12-25 20:47:03.000000000 +0900
@@ -0,0 +1,767 @@
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/time.h>
+#include <linux/ext4_jbd2.h>
+#include <linux/jbd2.h>
+#include <linux/highuid.h>
+#include <linux/pagemap.h>
+#include <linux/quotaops.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+#include <linux/falloc.h>
+#include <linux/ext4_fs_extents.h>
+#include <asm/uaccess.h>
+#include "group.h"
+
+/*
+ * this structure is used to gather extents from the tree via ioctl
+ */
+struct ext4_extent_buf {
+	ext4_fsblk_t start;
+	int buflen;
+	void *buffer;
+	void *cur;
+	int err;
+};
+
+/*
+ * this structure is used to collect stats info about the tree
+ */
+struct ext4_extent_tree_stats {
+	int depth;
+	int extents_num;
+	int leaf_num;
+};
+
+static int
+ext4_ext_store_extent_cb(struct inode *inode,
+			struct ext4_ext_path *path,
+			struct ext4_ext_cache *newex,
+			struct ext4_extent_buf *buf)
+{
+
+	if (newex->ec_type != EXT4_EXT_CACHE_EXTENT)
+		return EXT_CONTINUE;
+
+	if (buf->err < 0)
+		return EXT_BREAK;
+	if (buf->cur - buf->buffer + sizeof(*newex) > buf->buflen)
+		return EXT_BREAK;
+
+	if (!copy_to_user(buf->cur, newex, sizeof(*newex))) {
+		buf->err++;
+		buf->cur += sizeof(*newex);
+	} else {
+		buf->err = -EFAULT;
+		return EXT_BREAK;
+	}
+	return EXT_CONTINUE;
+}
+
+static int
+ext4_ext_collect_stats_cb(struct inode *inode,
+			struct ext4_ext_path *path,
+			struct ext4_ext_cache *ex,
+			struct ext4_extent_tree_stats *buf)
+{
+	int depth;
+
+	if (ex->ec_type != EXT4_EXT_CACHE_EXTENT)
+		return EXT_CONTINUE;
+
+	depth = ext_depth(inode);
+	buf->extents_num++;
+	if (path[depth].p_ext == EXT_FIRST_EXTENT(path[depth].p_hdr))
+		buf->leaf_num++;
+	return EXT_CONTINUE;
+}
+
+/**
+ * ext4_ext_next_extent - search for next extent and set it to "extent"
+ * @inode:	inode of the the original file
+ * @path:	this will obtain data for next extent
+ * @extent:	pointer to next extent we have just gotten
+ *
+ * This function returns 0 or 1(last_entry) if succeeded, otherwise
+ * returns -EIO
+ */
+static int
+ext4_ext_next_extent(struct inode *inode,
+		     struct ext4_ext_path *path,
+		     struct ext4_extent **extent)
+{
+	int ppos;
+	int leaf_ppos = path->p_depth;
+
+	ppos = leaf_ppos;
+	if (EXT_LAST_EXTENT(path[ppos].p_hdr) > path[ppos].p_ext) {
+		/* leaf block */
+		*extent = ++path[ppos].p_ext;
+		return 0;
+	}
+
+	while (--ppos >= 0) {
+		if (EXT_LAST_INDEX(path[ppos].p_hdr) >
+		    path[ppos].p_idx) {
+			int cur_ppos = ppos;
+
+			/* index block */
+			path[ppos].p_idx++;
+			path[ppos].p_block =
+				idx_pblock(path[ppos].p_idx);
+			if (path[ppos+1].p_bh)
+				brelse(path[ppos+1].p_bh);
+			path[ppos+1].p_bh =
+				sb_bread(inode->i_sb, path[ppos].p_block);
+			if (!path[ppos+1].p_bh)
+				return  -EIO;
+			path[ppos+1].p_hdr =
+				ext_block_hdr(path[ppos+1].p_bh);
+
+			/* halfway index block */
+			while (++cur_ppos < leaf_ppos) {
+				path[cur_ppos].p_idx =
+					EXT_FIRST_INDEX(path[cur_ppos].p_hdr);
+				path[cur_ppos].p_block =
+					idx_pblock(path[cur_ppos].p_idx);
+				if (path[cur_ppos+1].p_bh)
+					brelse(path[cur_ppos+1].p_bh);
+				path[cur_ppos+1].p_bh = sb_bread(inode->i_sb,
+					path[cur_ppos].p_block);
+				if (!path[cur_ppos+1].p_bh)
+					return  -EIO;
+				path[cur_ppos+1].p_hdr =
+					ext_block_hdr(path[cur_ppos+1].p_bh);
+			}
+
+			/* leaf block */
+			path[leaf_ppos].p_ext = *extent =
+				EXT_FIRST_EXTENT(path[leaf_ppos].p_hdr);
+			return 0;
+		}
+	}
+	/* last_extent */
+	return 1;
+}
+
+int ext4_ext_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
+			unsigned long arg)
+{
+	int err = 0;
+	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL ||
+		cmd == EXT4_IOC_FIBMAP))
+			return -EINVAL;
+
+	if (cmd == EXT4_IOC_GET_EXTENTS) {
+		struct ext4_extent_buf buf;
+
+		if (copy_from_user(&buf, (void *) arg, sizeof(buf)))
+			return -EFAULT;
+
+		buf.cur = buf.buffer;
+		buf.err = 0;
+		mutex_lock(&EXT4_I(inode)->truncate_mutex);
+		err = ext4_ext_walk_space(inode, buf.start, EXT_MAX_BLOCK,
+				(void *)ext4_ext_store_extent_cb, &buf);
+		mutex_unlock(&EXT4_I(inode)->truncate_mutex);
+		if (err == 0)
+			err = buf.err;
+	} else if (cmd == EXT4_IOC_GET_TREE_STATS) {
+		struct ext4_extent_tree_stats buf;
+
+		mutex_lock(&EXT4_I(inode)->truncate_mutex);
+		buf.depth = ext_depth(inode);
+		buf.extents_num = 0;
+		buf.leaf_num = 0;
+		err = ext4_ext_walk_space(inode, 0, EXT_MAX_BLOCK,
+				(void *)ext4_ext_collect_stats_cb, &buf);
+		mutex_unlock(&EXT4_I(inode)->truncate_mutex);
+		if (!err)
+			err = copy_to_user((void *) arg, &buf, sizeof(buf));
+	} else if (cmd == EXT4_IOC_GET_TREE_DEPTH) {
+		mutex_lock(&EXT4_I(inode)->truncate_mutex);
+		err = ext_depth(inode);
+		mutex_unlock(&EXT4_I(inode)->truncate_mutex);
+	} else if (cmd == EXT4_IOC_FIBMAP) {
+		ext4_fsblk_t __user *p = (ext4_fsblk_t __user *)arg;
+		ext4_fsblk_t block = 0;
+		struct address_space *mapping = filp->f_mapping;
+
+		if (copy_from_user(&block, (ext4_fsblk_t __user *)arg,
+					sizeof(block)))
+			return -EFAULT;
+
+		lock_kernel();
+		block = ext4_bmap(mapping, block);
+		unlock_kernel();
+
+		return put_user(block, p);
+	} else if (cmd == EXT4_IOC_DEFRAG) {
+		struct ext4_ext_defrag_data defrag;
+
+		if (copy_from_user(&defrag,
+			(struct ext4_ext_defrag_data __user *)arg,
+			sizeof(defrag)))
+			return -EFAULT;
+		err = ext4_ext_defrag(filp, defrag.start_offset,
+				defrag.defrag_size, defrag.goal, defrag.flag,
+				&defrag.ext);
+	}
+
+	return err;
+}
+
+/**
+ * ext4_ext_alloc_blocks - allocate contiguous blocks to temporary inode
+ * @dest_inode   temporary inode for multiple block allocation
+ * @org_inode    original inode
+ * @iblock       file related offset
+ * @total_blocks contiguous blocks count
+ * @goal	 block offset for allocation
+ * @phase        phase of create free space mode
+ *
+ * If succeed, fuction returns count of extent we got,
+ * otherwise returns err.
+ */
+static int ext4_ext_alloc_blocks(struct inode *dest_inode,
+		struct inode *org_inode, ext4_fsblk_t iblock,
+		ext4_fsblk_t total_blocks, ext4_fsblk_t goal, int phase)
+{
+	handle_t *handle = NULL;
+	struct ext4_ext_path *dest_path = NULL;
+	struct ext4_ext_path *org_path = NULL;
+	struct ext4_extent newex;
+	struct ext4_allocation_request ar;
+	struct buffer_head *bh = NULL;
+	struct super_block *org_sb = org_inode->i_sb;
+	ext4_fsblk_t newblock = 0;
+	ext4_fsblk_t rest = total_blocks;
+	ext4_fsblk_t alloc_total = 0;
+	unsigned long dest_grp_no, org_grp_no, org_len, goal_grp_no;
+	ext4_grpblk_t dest_blk_off, org_blk_off, goal_blk_off;
+	int org_depth = ext_depth(org_inode);
+	int metadata = 1;
+	int count = 0;
+	int credits = 0;
+	int err = 0;
+	int err2 = 0;
+	int len_cnt = 0;
+
+	ar.len = total_blocks;
+	org_len = ar.len;
+
+	/* Calculate group nubmer of org_inode block */
+	if (phase == DEFRAG_FORCE_VICTIM) {
+		org_path = ext4_ext_find_extent(org_inode, iblock, org_path);
+		if (IS_ERR(org_path)) {
+			err = PTR_ERR(org_path);
+			org_path = NULL;
+			goto out2;
+		}
+		ext4_get_group_no_and_offset(org_inode->i_sb,
+				ext_pblock(org_path[org_depth].p_ext),
+				&org_grp_no, &org_blk_off);
+		ar.excepted_group = org_grp_no;
+	} else {
+		ar.excepted_group = -1;
+	}
+
+	/* Find first extent. */
+	dest_path = ext4_ext_find_extent(dest_inode, iblock, dest_path);
+	if (IS_ERR(dest_path)) {
+		err = PTR_ERR(dest_path);
+		dest_path = NULL;
+		goto out2;
+	}
+
+	ar.inode = dest_inode;
+	ar.flags = EXT4_MB_HINT_DATA | EXT4_MB_HINT_RESERVED
+		| EXT4_MB_HINT_NOPREALLOC;
+	if (goal) {
+		ar.goal = goal;
+	} else {
+		ar.goal = ext4_ext_find_goal(dest_inode, dest_path, iblock);
+	}
+
+	ar.logical = iblock;
+	ar.lleft = 0;
+	ar.pleft = 0;
+	ar.lright = 0;
+	ar.pright = 0;
+
+	handle = ext4_journal_start(dest_inode, credits);
+	if (IS_ERR(handle)) {
+		err = PTR_ERR(handle);
+		goto out2;
+	}
+
+	while (alloc_total != total_blocks) {
+		credits = ext4_ext_calc_credits_for_insert(dest_inode,
+							dest_path);
+		handle = ext4_ext_journal_restart(handle,
+				credits + EXT4_TRANS_META_BLOCKS);
+
+		if (IS_ERR(handle))
+			return PTR_ERR(handle);
+
+		newblock = ext4_mb_new_blocks(handle, &ar, &err);
+
+		if (err) {
+			/* Failed to get the contiguous blocks */
+			goto out;
+		} else if ((ar.len != org_len) &&
+				(phase == DEFRAG_FORCE_TRY)) {
+			ext4_free_blocks(handle, org_inode, newblock,
+						ar.len, metadata);
+			/* go to force mode */
+			err = -ENOSPC;
+			goto out;
+		} else {
+			/*
+			 * If ext4_mb_new_blocks() allcates
+			 * the block which used to be the metadata block,
+			 * its dirty buffer_head causes the overwriting
+			 * with old metadata.
+			 * We should call unmap_underlying_metadata()
+			 * to clear the dirty flag.
+			 */
+			for (len_cnt = 0; len_cnt < ar.len; len_cnt++) {
+				bh = sb_find_get_block(org_sb,
+						       newblock + len_cnt);
+				unmap_underlying_metadata(org_sb->s_bdev,
+							  newblock + len_cnt);
+			}
+
+			alloc_total += ar.len;
+			ext4_get_group_no_and_offset(dest_inode->i_sb,
+				goal, &goal_grp_no, &goal_blk_off);
+			ext4_get_group_no_and_offset(dest_inode->i_sb,
+				newblock, &dest_grp_no, &dest_blk_off);
+			/* We can't allocate at the same block group */
+			switch (phase) {
+			case DEFRAG_FORCE_VICTIM:
+				if (dest_grp_no == org_grp_no) {
+					printk(KERN_ERR "defrag: Can't allocate"
+					" in same block group\n");
+					ext4_free_blocks(handle, org_inode,
+						newblock, ar.len, metadata);
+					err = -ENOSPC;
+					goto out;
+				}
+				break;
+			case DEFRAG_FORCE_GATHER:
+				/* Maybe reserved blocks are already used by
+				   other process */
+				if (dest_grp_no != goal_grp_no
+					|| alloc_total != total_blocks) {
+					printk(KERN_ERR "defrag: Already used"
+						" the specified blocks\n");
+					ext4_free_blocks(handle, org_inode,
+						newblock, ar.len, metadata);
+					err = -EIO;
+					goto out;
+				}
+				break;
+			}
+
+			newex.ee_block = cpu_to_le32(alloc_total - ar.len);
+			ext4_ext_store_pblock(&newex, newblock);
+			newex.ee_len = cpu_to_le16(ar.len);
+
+			if (!phase)
+				ar.goal = newblock + ar.len;
+			rest = rest - ar.len;
+			ar.len = rest;
+
+			err = ext4_ext_insert_extent(handle, dest_inode,
+						dest_path, &newex);
+			if (!err) {
+				count++;
+			} else {
+				ext4_free_blocks(handle, org_inode,
+					newblock, ar.len, metadata);
+				goto out;
+			}
+		}
+	}
+
+out:
+	/* Faild case: We have to remove halfway blocks */
+	if (err)
+		err2 = ext4_ext_remove_space(dest_inode, 0);
+
+	/* Successful case */
+	if (dest_path) {
+		ext4_ext_drop_refs(dest_path);
+		kfree(dest_path);
+	}
+	if (org_path) {
+		ext4_ext_drop_refs(org_path);
+		kfree(org_path);
+	}
+out2:
+	ext4_journal_stop(handle);
+
+	if (err2) {
+		return err2;
+	} else if (err) {
+		return err;
+	}
+	/* return extents count */
+	return count;
+}
+
+/**
+ * ext4_ext_new_extent_tree -  allocate contiguous blocks
+ * @inode:		inode of the original file
+ * @tmp_inode:		inode of the temporary file
+ * @path:		the structure holding some info about
+ *			original extent tree
+ * @tar_start:		starting offset to allocate in blocks
+ * @tar_blocks:		the number of blocks to allocate
+ * @iblock:		file related offset
+ * @goal:		block offset for allocaton
+ * @flag:               phase of create free space mode
+ *
+ * This function returns the value as below:
+ * 	0(succeeded)
+ *	1(not improved)
+ *	negative value(error)
+ */
+static int
+ext4_ext_new_extent_tree(struct inode *inode, struct inode *tmp_inode,
+			struct ext4_ext_path *path, ext4_fsblk_t tar_start,
+			ext4_fsblk_t tar_blocks, ext4_fsblk_t iblock,
+			ext4_fsblk_t goal, int flag)
+{
+	struct ext4_extent *ext = NULL;
+	struct ext4_extent_header *eh = NULL;
+	ext4_fsblk_t tar_end = tar_start + tar_blocks - 1;
+	int sum_org = 0, sum_tmp = 0;
+	int ret = 0, depth;
+	int last_extent = 0;
+
+	eh = ext_inode_hdr(tmp_inode);
+	eh->eh_depth = 0;
+
+	/* allocate contiguous blocks */
+	sum_tmp = ext4_ext_alloc_blocks(tmp_inode, inode, iblock,
+					tar_blocks, goal, flag);
+	if (sum_tmp < 0) {
+		ret = sum_tmp;
+		goto ERR;
+	}
+
+	depth = ext_depth(inode);
+	ext = path[depth].p_ext;
+	while (1) {
+		if (!last_extent)
+			++sum_org;
+
+		if (tar_end <= le32_to_cpu(ext->ee_block) +
+			       le32_to_cpu(ext->ee_len) - 1 ||
+			       last_extent) {
+
+			if ((sum_org == sum_tmp) && !goal) {
+				/* not improved */
+				if (!(ret =
+					ext4_ext_remove_space(tmp_inode, 0)))
+					ret = 1;
+			} else if (sum_org < sum_tmp &&
+					flag != DEFRAG_FORCE_VICTIM) {
+				/* fragment increased */
+				if (!(ret =
+					ext4_ext_remove_space(tmp_inode, 0)))
+					ret = -ENOSPC;
+				printk("defrag failed due to no space\n");
+			}
+			break;
+		}
+		if ((last_extent =
+			ext4_ext_next_extent(tmp_inode,
+				path, &ext)) < 0) {
+			ret = last_extent;
+			break;
+		}
+	}
+ERR:
+	return ret;
+}
+
+/**
+ * ext4_ext_defrag - defrag whole file
+ * @filp:            pointer to file
+ * @from:            starting offset to defrag in blocks
+ * @defrag_size:     size of defrag in blocks
+ * @goal:            block offset for allocation
+ * @flag:            phase of create free space mode
+ * @ext:             extent to be moved (only -f)
+ *
+ * This function returns the number of blocks if succeeded, otherwise
+ * returns error value
+ */
+int
+ext4_ext_defrag(struct file *filp, ext4_fsblk_t block_start,
+		ext4_fsblk_t defrag_size, ext4_fsblk_t goal,
+		int flag, struct ext4_extent_data *ext)
+{
+	struct inode *inode = filp->f_dentry->d_inode, *tmp_inode = NULL;
+	struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es;
+	struct ext4_ext_path *path = NULL, *holecheck_path = NULL;
+	struct ext4_extent *ext_prev = NULL, *ext_cur = NULL, *ext_dummy = NULL;
+	handle_t *handle;
+	ext4_fsblk_t block_end = block_start + defrag_size - 1;
+	ext4_fsblk_t seq_blocks = 0, seq_start = 0;
+	ext4_fsblk_t add_blocks = 0;
+	ext4_fsblk_t file_end = (inode->i_size - 1) >> inode->i_blkbits;
+	pgoff_t page_offset = 0;
+	pgoff_t dest_offset = 0;
+	pgoff_t seq_end_page = 0;
+	int ret = 0, depth = 0, last_extent = 0, seq_extents = 0;
+
+	/* Check goal offset if goal offset was given from userspace. */
+	if (((0 < goal) && (ext4_blocks_count(es) < goal)) && (goal != -1)) {
+		printk(KERN_ERR "defrag: incorrect goal number %llu, "
+			"you can set goal until %llu\n", goal,
+			ext4_blocks_count(es));
+		ret = -EINVAL;
+		goto ERR1;
+	}
+
+	/* Setup for fixed blocks mode */
+	if (ext->len) {
+		if (ext->len < defrag_size) {
+			printk("Cannot defrag due to the insufficient"
+			" specified free blocks\n");
+			return -EINVAL;
+		}
+		flag = DEFRAG_FORCE_GATHER;
+		goal = ext->start;
+	}
+
+	if (file_end < block_end)
+		defrag_size -= block_end - file_end;
+
+	mutex_lock(&inode->i_mutex);
+	mutex_lock(&EXT4_I(inode)->truncate_mutex);
+
+	path = ext4_ext_find_extent(inode, block_start, NULL);
+	if (IS_ERR(path)) {
+		ret = PTR_ERR(path);
+		path = NULL;
+		goto ERR2;
+	}
+
+	/* get path structure to check hole */
+	holecheck_path = ext4_ext_find_extent(inode, block_start, NULL);
+	if (IS_ERR(holecheck_path)) {
+		ret = PTR_ERR(holecheck_path);
+		holecheck_path = NULL;
+		goto ERR2;
+	}
+
+	depth = ext_depth(inode);
+	ext_cur = holecheck_path[depth].p_ext;
+	if (ext_cur == NULL)
+		goto ERR2;
+
+	/*
+	 * if block_start was within the hole, get proper extent whose ee_block
+	 * is beyond block_start
+	 */
+	if (le32_to_cpu(ext_cur->ee_block) +
+		le32_to_cpu(ext_cur->ee_len) - 1 < block_start) {
+		if ((last_extent =
+				ext4_ext_next_extent(inode, holecheck_path,
+				&ext_cur)) < 0) {
+			ret = last_extent;
+			goto ERR2;
+		}
+		if ((last_extent =
+				ext4_ext_next_extent(inode, path,
+				&ext_dummy)) < 0) {
+			ret = last_extent;
+			goto ERR2;
+		}
+	}
+	seq_extents = 1;
+	seq_start = ext_cur->ee_block;
+
+	/* no blocks existed within designated range */
+	if (le32_to_cpu(ext_cur->ee_block) > block_end) {
+		printk("nothing done due to the lack of contiguous blocks\n");
+		goto ERR2;
+	}
+
+	/* adjust start blocks */
+	add_blocks = min((ext4_fsblk_t)(ext_cur->ee_block +
+			 ext_cur->ee_len), block_end + 1) -
+		     max((ext4_fsblk_t)ext_cur->ee_block, block_start);
+
+	while (!last_extent && ext_cur->ee_block <= block_end) {
+		seq_blocks += add_blocks;
+
+		handle = ext4_journal_start(inode,
+			EXT4_DATA_TRANS_BLOCKS(inode->i_sb) +
+			EXT4_INDEX_EXTRA_TRANS_BLOCKS + 3 +
+			2 * EXT4_QUOTA_INIT_BLOCKS(inode->i_sb) + 1);
+		if (IS_ERR(handle)) {
+			ret = PTR_ERR(handle);
+			goto ERR1;
+		}
+		tmp_inode = ext4_new_inode(handle,
+			inode->i_sb->s_root->d_inode, S_IFREG);
+		if (IS_ERR(tmp_inode)) {
+			ret = -ENOMEM;
+			ext4_journal_stop(handle);
+			tmp_inode = NULL;
+			goto ERR1;
+		}
+
+		i_size_write(tmp_inode, i_size_read(inode));
+		tmp_inode->i_nlink = 0;
+		ext4_ext_tree_init(handle, tmp_inode);
+		ext4_orphan_add(handle, tmp_inode);
+		ext4_journal_stop(handle);
+
+		/* adjust tail blocks */
+		if (seq_start + seq_blocks - 1 > block_end)
+			seq_blocks = block_end - seq_start + 1;
+
+		ext_prev = ext_cur;
+		if ((last_extent =
+				ext4_ext_next_extent(inode, holecheck_path,
+				&ext_cur)) < 0) {
+			ret = last_extent;
+			break;
+		}
+		if (!last_extent)
+			seq_extents++;
+		add_blocks = le16_to_cpu(ext_cur->ee_len);
+
+		/* found hole or reached the tail of either a designated range
+		 * or the file
+		 */
+		if ((le32_to_cpu(ext_prev->ee_block) +
+				le16_to_cpu(ext_prev->ee_len) ==
+				le32_to_cpu(ext_cur->ee_block) &&
+				block_end >= le32_to_cpu(ext_cur->ee_block) &&
+				!last_extent)) {
+			if (tmp_inode) {
+				iput(tmp_inode);
+				tmp_inode = NULL;
+			}
+			continue;
+		}
+
+		/* found an isolated block */
+		if ((seq_extents == 1) && !goal) {
+			seq_start = ext_cur->ee_block;
+			goto CLEANUP;
+		}
+
+		ret = ext4_ext_new_extent_tree(inode, tmp_inode, path,
+			seq_start, seq_blocks, block_start, goal, flag);
+
+		if (ret < 0) {
+			break;
+		} else if ((ret == 1) && (!goal || (goal && !flag))) {
+			ret = 0;
+			seq_start = le32_to_cpu(ext_cur->ee_block);
+			goto CLEANUP;
+		}
+
+		page_offset = seq_start >>
+				(PAGE_CACHE_SHIFT - inode->i_blkbits);
+		seq_end_page = (seq_start + seq_blocks - 1) >>
+				(PAGE_CACHE_SHIFT - inode->i_blkbits);
+
+		dest_offset = 0;
+		seq_start = le32_to_cpu(ext_cur->ee_block);
+
+		/* Discard all preallocations.
+		 * This is provisional solution.
+		 * When true ext4_mb_return_to_preallocation() is
+		 * implemented, this will be removed.
+		 */
+		ext4_mb_discard_inode_preallocations(inode);
+
+		if (inode->i_mapping->a_ops->write_begin) {
+			while (page_offset <= seq_end_page) {
+				/* replace original branches for new branches */
+				ret = ext4_ext_defrag_partial2(tmp_inode,
+							filp, page_offset,
+							dest_offset, flag);
+				if (ret < 0)
+					goto ERR2;
+
+				page_offset++;
+				dest_offset++;
+			}
+		} else {
+			while (page_offset <= seq_end_page) {
+				/* replace original branches for new branches */
+				ret = ext4_ext_defrag_partial(tmp_inode,
+							filp, page_offset,
+							dest_offset, flag);
+				if (ret < 0)
+					goto ERR2;
+
+				page_offset++;
+				dest_offset++;
+			}
+		}
+
+		/* decrease buffer counter */
+		if (holecheck_path)
+			ext4_ext_drop_refs(holecheck_path);
+		holecheck_path =
+			ext4_ext_find_extent(inode, seq_start, holecheck_path);
+		if (IS_ERR(holecheck_path)) {
+			ret = PTR_ERR(holecheck_path);
+			holecheck_path = NULL;
+			break;
+		}
+		depth = holecheck_path->p_depth;
+
+CLEANUP:
+		/* decrease buffer counter */
+		if (path)
+			ext4_ext_drop_refs(path);
+		path = ext4_ext_find_extent(inode, seq_start, path);
+		if (IS_ERR(path)) {
+			ret = PTR_ERR(path);
+			path = NULL;
+			break;
+		}
+
+		ext_cur = holecheck_path[depth].p_ext;
+		add_blocks = le16_to_cpu(ext_cur->ee_len);
+		seq_blocks = 0;
+		dest_offset = 0;
+		seq_extents = 1;
+
+		if (tmp_inode) {
+			iput(tmp_inode);
+			tmp_inode = NULL;
+		}
+	}
+
+ERR2:
+	if (path) {
+		ext4_ext_drop_refs(path);
+		kfree(path);
+	}
+	if (holecheck_path) {
+		ext4_ext_drop_refs(holecheck_path);
+		kfree(holecheck_path);
+	}
+ERR1:
+	mutex_unlock(&EXT4_I(inode)->truncate_mutex);
+	mutex_unlock(&inode->i_mutex);
+
+	if (tmp_inode)
+		iput(tmp_inode);
+
+	return (ret ? ret : defrag_size);
+}
diff -X linux-2.6.24-rc5-defrag/Documentation/dontdiff -upNr linux-2.6.24-rc5-alloc-block/fs/ext4/extents.c linux-2.6.24-rc5-move-
data/fs/ext4/extents.c
--- linux-2.6.24-rc5-alloc-block/fs/ext4/extents.c	2007-12-25 20:57:40.000000000 +0900
+++ linux-2.6.24-rc5-move-data/fs/ext4/extents.c	2007-12-25 20:47:33.000000000 +0900
@@ -48,7 +48,7 @@
  * ext_pblock:
  * combine low and high parts of physical block number into ext4_fsblk_t
  */
-static ext4_fsblk_t ext_pblock(struct ext4_extent *ex)
+ext4_fsblk_t ext_pblock(struct ext4_extent *ex)
 {
 	ext4_fsblk_t block;
 
@@ -92,7 +92,7 @@ static void ext4_idx_store_pblock(struct
 	ix->ei_leaf_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
 }
 
-static handle_t *ext4_ext_journal_restart(handle_t *handle, int needed)
+handle_t *ext4_ext_journal_restart(handle_t *handle, int needed)
 {
 	int err;
 
@@ -142,7 +142,7 @@ static int ext4_ext_dirty(handle_t *hand
 	return err;
 }
 
-static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
+ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
 			      struct ext4_ext_path *path,
 			      ext4_lblk_t block)
 {
@@ -1948,7 +1948,7 @@ ext4_ext_more_to_rm(struct ext4_ext_path
 	return 1;
 }
 
-static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
+int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 {
 	struct super_block *sb = inode->i_sb;
 	int depth = ext_depth(inode);
diff -X linux-2.6.24-rc5-defrag/Documentation/dontdiff -upNr linux-2.6.24-rc5-alloc-block/fs/ext4/inode.c linux-2.6.24-rc5-move-data/fs/ext4/inode.c
--- linux-2.6.24-rc5-alloc-block/fs/ext4/inode.c	2007-12-25 20:58:02.000000000 +0900
+++ linux-2.6.24-rc5-move-data/fs/ext4/inode.c	2007-12-25 20:47:54.000000000 +0900
@@ -1464,7 +1464,7 @@ out:
  * So, if we see any bmap calls here on a modified, data-journaled file,
  * take extra steps to flush any blocks which might be in the cache.
  */
-static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
+sector_t ext4_bmap(struct address_space *mapping, sector_t block)
 {
 	struct inode *inode = mapping->host;
 	journal_t *journal;
diff -X linux-2.6.24-rc5-defrag/Documentation/dontdiff -upNr linux-2.6.24-rc5-alloc-block/fs/ext4/ioctl.c linux-2.6.24-rc5-move-data/fs/ext4/ioctl.c
--- linux-2.6.24-rc5-alloc-block/fs/ext4/ioctl.c	2007-12-25 20:58:29.000000000 +0900
+++ linux-2.6.24-rc5-move-data/fs/ext4/ioctl.c	2007-12-25 20:41:41.000000000 +0900
@@ -231,6 +231,19 @@ flags_err:
 
 		return err;
 	}
+	case EXT4_IOC_GET_EXTENTS:
+	case EXT4_IOC_GET_TREE_STATS:
+	case EXT4_IOC_GET_TREE_DEPTH:
+	case EXT4_IOC_FIBMAP:
+	case EXT4_IOC_DEFRAG:
+	case EXT4_IOC_GROUP_INFO:
+	case EXT4_IOC_FREE_BLOCKS_INFO:
+	case EXT4_IOC_EXTENTS_INFO:
+	case EXT4_IOC_RESERVE_BLOCK:
+	case EXT4_IOC_MOVE_VICTIM:
+	case EXT4_IOC_BLOCK_RELEASE: {
+		return ext4_ext_ioctl(inode, filp, cmd, arg);
+	}
 	case EXT4_IOC_GROUP_ADD: {
 		struct ext4_new_group_data input;
 		struct super_block *sb = inode->i_sb;
diff -X linux-2.6.24-rc5-defrag/Documentation/dontdiff -upNr linux-2.6.24-rc5-alloc-block/fs/ext4/mballoc.c linux-2.6.24-rc5-move-
data/fs/ext4/mballoc.c
--- linux-2.6.24-rc5-alloc-block/fs/ext4/mballoc.c	2007-12-25 20:59:14.000000000 +0900
+++ linux-2.6.24-rc5-move-data/fs/ext4/mballoc.c	2007-12-25 20:41:41.000000000 +0900
@@ -411,6 +411,7 @@ struct ext4_allocation_context {
 	struct page *ac_buddy_page;
 	struct ext4_prealloc_space *ac_pa;
 	struct ext4_locality_group *ac_lg;
+	long long ac_excepted_group;
 };
 
 #define AC_STATUS_CONTINUE	1
@@ -1873,6 +1874,11 @@ repeat:
 			if (group == EXT4_SB(sb)->s_groups_count)
 				group = 0;
 
+			if (ac->ac_excepted_group != -1 &&
+			    group == ac->ac_excepted_group) {
+				continue;
+			}
+
 			/* quick check to skip empty groups */
 			grp = EXT4_GROUP_INFO(ac->ac_sb, group);
 			if (grp->bb_free == 0)
@@ -4080,6 +4086,7 @@ static int ext4_mb_initialize_context(st
 	ac->ac_bitmap_page = NULL;
 	ac->ac_buddy_page = NULL;
 	ac->ac_lg = NULL;
+	ac->ac_excepted_group = ar->excepted_group;
 
 	/* we have to define context: we'll we work with a file or
 	 * locality group. this is a policy, actually */
-
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists