Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2:
  nilfs2: use mark_buffer_dirty to mark btnode or meta data dirty
  nilfs2: always set back pointer to host inode in mapping->host
  nilfs2: get rid of NILFS_I_NILFS
  nilfs2: use list_first_entry
  nilfs2: use empty_aops for gc-inodes
  nilfs2: implement resize ioctl
  nilfs2: add truncation routine of segment usage file
  nilfs2: add routine to move secondary super block
  nilfs2: add ioctl which limits range of segment to be allocated
  nilfs2: zero fill unused portion of super root block
  nilfs2: super root size should change depending on inode size
  nilfs2: get rid of private page allocator
  nilfs2: merge list_del()/list_add_tail() to list_move_tail()

fs/nilfs2/alloc.c

@@ -489,8 +489,8 @@ int nilfs_palloc_prepare_alloc_entry(struct inode *inode,
 void nilfs_palloc_commit_alloc_entry(struct inode *inode,
 				     struct nilfs_palloc_req *req)
 {
-	nilfs_mdt_mark_buffer_dirty(req->pr_bitmap_bh);
-	nilfs_mdt_mark_buffer_dirty(req->pr_desc_bh);
+	mark_buffer_dirty(req->pr_bitmap_bh);
+	mark_buffer_dirty(req->pr_desc_bh);
 	nilfs_mdt_mark_dirty(inode);
 
 	brelse(req->pr_bitmap_bh);
@@ -527,8 +527,8 @@ void nilfs_palloc_commit_free_entry(struct inode *inode,
 	kunmap(req->pr_bitmap_bh->b_page);
 	kunmap(req->pr_desc_bh->b_page);
 
-	nilfs_mdt_mark_buffer_dirty(req->pr_desc_bh);
-	nilfs_mdt_mark_buffer_dirty(req->pr_bitmap_bh);
+	mark_buffer_dirty(req->pr_desc_bh);
+	mark_buffer_dirty(req->pr_bitmap_bh);
 	nilfs_mdt_mark_dirty(inode);
 
 	brelse(req->pr_bitmap_bh);
@@ -683,8 +683,8 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems)
 		kunmap(bitmap_bh->b_page);
 		kunmap(desc_bh->b_page);
 
-		nilfs_mdt_mark_buffer_dirty(desc_bh);
-		nilfs_mdt_mark_buffer_dirty(bitmap_bh);
+		mark_buffer_dirty(desc_bh);
+		mark_buffer_dirty(bitmap_bh);
 		nilfs_mdt_mark_dirty(inode);
 
 		brelse(bitmap_bh);

fs/nilfs2/bmap.c

@@ -34,7 +34,9 @@
 
 struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *bmap)
 {
-	return NILFS_I_NILFS(bmap->b_inode)->ns_dat;
+	struct the_nilfs *nilfs = bmap->b_inode->i_sb->s_fs_info;
+
+	return nilfs->ns_dat;
 }
 
 static int nilfs_bmap_convert_error(struct nilfs_bmap *bmap,

fs/nilfs2/btnode.c

@@ -34,12 +34,6 @@
 #include "page.h"
 #include "btnode.h"
 
-void nilfs_btnode_cache_init(struct address_space *btnc,
-			     struct backing_dev_info *bdi)
-{
-	nilfs_mapping_init(btnc, bdi);
-}
-
 void nilfs_btnode_cache_clear(struct address_space *btnc)
 {
 	invalidate_mapping_pages(btnc, 0, -1);
@@ -62,7 +56,7 @@ nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr)
 		BUG();
 	}
 	memset(bh->b_data, 0, 1 << inode->i_blkbits);
-	bh->b_bdev = NILFS_I_NILFS(inode)->ns_bdev;
+	bh->b_bdev = inode->i_sb->s_bdev;
 	bh->b_blocknr = blocknr;
 	set_buffer_mapped(bh);
 	set_buffer_uptodate(bh);
@@ -94,10 +88,11 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr,
 	if (pblocknr == 0) {
 		pblocknr = blocknr;
 		if (inode->i_ino != NILFS_DAT_INO) {
-			struct inode *dat = NILFS_I_NILFS(inode)->ns_dat;
+			struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
 
 			/* blocknr is a virtual block number */
-			err = nilfs_dat_translate(dat, blocknr, &pblocknr);
+			err = nilfs_dat_translate(nilfs->ns_dat, blocknr,
+						  &pblocknr);
 			if (unlikely(err)) {
 				brelse(bh);
 				goto out_locked;
@@ -120,7 +115,7 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr,
 		goto found;
 	}
 	set_buffer_mapped(bh);
-	bh->b_bdev = NILFS_I_NILFS(inode)->ns_bdev;
+	bh->b_bdev = inode->i_sb->s_bdev;
 	bh->b_blocknr = pblocknr; /* set block address for read */
 	bh->b_end_io = end_buffer_read_sync;
 	get_bh(bh);
@@ -259,7 +254,7 @@ void nilfs_btnode_commit_change_key(struct address_space *btnc,
 				       "invalid oldkey %lld (newkey=%lld)",
 				       (unsigned long long)oldkey,
 				       (unsigned long long)newkey);
-		nilfs_btnode_mark_dirty(obh);
+		mark_buffer_dirty(obh);
 
 		spin_lock_irq(&btnc->tree_lock);
 		radix_tree_delete(&btnc->page_tree, oldkey);
@@ -271,7 +266,7 @@ void nilfs_btnode_commit_change_key(struct address_space *btnc,
 		unlock_page(opage);
 	} else {
 		nilfs_copy_buffer(nbh, obh);
-		nilfs_btnode_mark_dirty(nbh);
+		mark_buffer_dirty(nbh);
 
 		nbh->b_blocknr = newkey;
 		ctxt->bh = nbh;

fs/nilfs2/btnode.h

@@ -37,7 +37,6 @@ struct nilfs_btnode_chkey_ctxt {
 	struct buffer_head *newbh;
 };
 
-void nilfs_btnode_cache_init(struct address_space *, struct backing_dev_info *);
 void nilfs_btnode_cache_clear(struct address_space *);
 struct buffer_head *nilfs_btnode_create_block(struct address_space *btnc,
 					      __u64 blocknr);
@@ -51,7 +50,4 @@ void nilfs_btnode_commit_change_key(struct address_space *,
 void nilfs_btnode_abort_change_key(struct address_space *,
 				   struct nilfs_btnode_chkey_ctxt *);
 
-#define nilfs_btnode_mark_dirty(bh)	nilfs_mark_buffer_dirty(bh)
-
-
 #endif	/* _NILFS_BTNODE_H */

fs/nilfs2/btree.c

@@ -714,7 +714,7 @@ static void nilfs_btree_promote_key(struct nilfs_bmap *btree,
 				nilfs_btree_get_nonroot_node(path, level),
 				path[level].bp_index, key);
 			if (!buffer_dirty(path[level].bp_bh))
-				nilfs_btnode_mark_dirty(path[level].bp_bh);
+				mark_buffer_dirty(path[level].bp_bh);
 		} while ((path[level].bp_index == 0) &&
 			 (++level < nilfs_btree_height(btree) - 1));
 	}
@@ -739,7 +739,7 @@ static void nilfs_btree_do_insert(struct nilfs_bmap *btree,
 		nilfs_btree_node_insert(node, path[level].bp_index,
 					*keyp, *ptrp, ncblk);
 		if (!buffer_dirty(path[level].bp_bh))
-			nilfs_btnode_mark_dirty(path[level].bp_bh);
+			mark_buffer_dirty(path[level].bp_bh);
 
 		if (path[level].bp_index == 0)
 			nilfs_btree_promote_key(btree, path, level + 1,
@@ -777,9 +777,9 @@ static void nilfs_btree_carry_left(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_left(left, node, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	nilfs_btree_promote_key(btree, path, level + 1,
 				nilfs_btree_node_get_key(node, 0));
@@ -823,9 +823,9 @@ static void nilfs_btree_carry_right(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_right(node, right, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	path[level + 1].bp_index++;
 	nilfs_btree_promote_key(btree, path, level + 1,
@@ -870,9 +870,9 @@ static void nilfs_btree_split(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_right(node, right, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	newkey = nilfs_btree_node_get_key(right, 0);
 	newptr = path[level].bp_newreq.bpr_ptr;
@@ -919,7 +919,7 @@ static void nilfs_btree_grow(struct nilfs_bmap *btree,
 	nilfs_btree_node_set_level(root, level + 1);
 
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	path[level].bp_bh = path[level].bp_sib_bh;
 	path[level].bp_sib_bh = NULL;
@@ -1194,7 +1194,7 @@ static void nilfs_btree_do_delete(struct nilfs_bmap *btree,
 		nilfs_btree_node_delete(node, path[level].bp_index,
 					keyp, ptrp, ncblk);
 		if (!buffer_dirty(path[level].bp_bh))
-			nilfs_btnode_mark_dirty(path[level].bp_bh);
+			mark_buffer_dirty(path[level].bp_bh);
 		if (path[level].bp_index == 0)
 			nilfs_btree_promote_key(btree, path, level + 1,
 				nilfs_btree_node_get_key(node, 0));
@@ -1226,9 +1226,9 @@ static void nilfs_btree_borrow_left(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_right(left, node, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	nilfs_btree_promote_key(btree, path, level + 1,
 				nilfs_btree_node_get_key(node, 0));
@@ -1258,9 +1258,9 @@ static void nilfs_btree_borrow_right(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_left(node, right, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	path[level + 1].bp_index++;
 	nilfs_btree_promote_key(btree, path, level + 1,
@@ -1289,7 +1289,7 @@ static void nilfs_btree_concat_left(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_left(left, node, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_sib_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
+		mark_buffer_dirty(path[level].bp_sib_bh);
 
 	nilfs_btnode_delete(path[level].bp_bh);
 	path[level].bp_bh = path[level].bp_sib_bh;
@@ -1315,7 +1315,7 @@ static void nilfs_btree_concat_right(struct nilfs_bmap *btree,
 	nilfs_btree_node_move_left(node, right, n, ncblk, ncblk);
 
 	if (!buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 
 	nilfs_btnode_delete(path[level].bp_sib_bh);
 	path[level].bp_sib_bh = NULL;
@@ -1709,7 +1709,7 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *btree,
 		nilfs_btree_node_init(node, 0, 1, n, ncblk, keys, ptrs);
 		nilfs_btree_node_insert(node, n, key, dreq->bpr_ptr, ncblk);
 		if (!buffer_dirty(bh))
-			nilfs_btnode_mark_dirty(bh);
+			mark_buffer_dirty(bh);
 		if (!nilfs_bmap_dirty(btree))
 			nilfs_bmap_set_dirty(btree);
 
@@ -1787,7 +1787,7 @@ static int nilfs_btree_propagate_p(struct nilfs_bmap *btree,
 {
 	while ((++level < nilfs_btree_height(btree) - 1) &&
 	       !buffer_dirty(path[level].bp_bh))
-		nilfs_btnode_mark_dirty(path[level].bp_bh);
+		mark_buffer_dirty(path[level].bp_bh);
 
 	return 0;
 }
@@ -2229,7 +2229,7 @@ static int nilfs_btree_mark(struct nilfs_bmap *btree, __u64 key, int level)
 	}
 
 	if (!buffer_dirty(bh))
-		nilfs_btnode_mark_dirty(bh);
+		mark_buffer_dirty(bh);
 	brelse(bh);
 	if (!nilfs_bmap_dirty(btree))
 		nilfs_bmap_set_dirty(btree);

fs/nilfs2/cpfile.c

@@ -216,14 +216,14 @@ int nilfs_cpfile_get_checkpoint(struct inode *cpfile,
 		if (!nilfs_cpfile_is_in_first(cpfile, cno))
 			nilfs_cpfile_block_add_valid_checkpoints(cpfile, cp_bh,
 								 kaddr, 1);
-		nilfs_mdt_mark_buffer_dirty(cp_bh);
+		mark_buffer_dirty(cp_bh);
 
 		kaddr = kmap_atomic(header_bh->b_page, KM_USER0);
 		header = nilfs_cpfile_block_get_header(cpfile, header_bh,
 						       kaddr);
 		le64_add_cpu(&header->ch_ncheckpoints, 1);
 		kunmap_atomic(kaddr, KM_USER0);
-		nilfs_mdt_mark_buffer_dirty(header_bh);
+		mark_buffer_dirty(header_bh);
 		nilfs_mdt_mark_dirty(cpfile);
 	}
 
@@ -326,7 +326,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 		}
 		if (nicps > 0) {
 			tnicps += nicps;
-			nilfs_mdt_mark_buffer_dirty(cp_bh);
+			mark_buffer_dirty(cp_bh);
 			nilfs_mdt_mark_dirty(cpfile);
 			if (!nilfs_cpfile_is_in_first(cpfile, cno)) {
 				count =
@@ -358,7 +358,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile,
 		header = nilfs_cpfile_block_get_header(cpfile, header_bh,
 						       kaddr);
 		le64_add_cpu(&header->ch_ncheckpoints, -(u64)tnicps);
-		nilfs_mdt_mark_buffer_dirty(header_bh);
+		mark_buffer_dirty(header_bh);
 		nilfs_mdt_mark_dirty(cpfile);
 		kunmap_atomic(kaddr, KM_USER0);
 	}
@@ -671,10 +671,10 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno)
 	le64_add_cpu(&header->ch_nsnapshots, 1);
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(prev_bh);
-	nilfs_mdt_mark_buffer_dirty(curr_bh);
-	nilfs_mdt_mark_buffer_dirty(cp_bh);
-	nilfs_mdt_mark_buffer_dirty(header_bh);
+	mark_buffer_dirty(prev_bh);
+	mark_buffer_dirty(curr_bh);
+	mark_buffer_dirty(cp_bh);
+	mark_buffer_dirty(header_bh);
 	nilfs_mdt_mark_dirty(cpfile);
 
 	brelse(prev_bh);
@@ -774,10 +774,10 @@ static int nilfs_cpfile_clear_snapshot(struct inode *cpfile, __u64 cno)
 	le64_add_cpu(&header->ch_nsnapshots, -1);
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(next_bh);
-	nilfs_mdt_mark_buffer_dirty(prev_bh);
-	nilfs_mdt_mark_buffer_dirty(cp_bh);
-	nilfs_mdt_mark_buffer_dirty(header_bh);
+	mark_buffer_dirty(next_bh);
+	mark_buffer_dirty(prev_bh);
+	mark_buffer_dirty(cp_bh);
+	mark_buffer_dirty(header_bh);
 	nilfs_mdt_mark_dirty(cpfile);
 
 	brelse(prev_bh);

fs/nilfs2/dat.c

@@ -54,7 +54,7 @@ static int nilfs_dat_prepare_entry(struct inode *dat,
 static void nilfs_dat_commit_entry(struct inode *dat,
 				   struct nilfs_palloc_req *req)
 {
-	nilfs_mdt_mark_buffer_dirty(req->pr_entry_bh);
+	mark_buffer_dirty(req->pr_entry_bh);
 	nilfs_mdt_mark_dirty(dat);
 	brelse(req->pr_entry_bh);
 }
@@ -361,7 +361,7 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr)
 	entry->de_blocknr = cpu_to_le64(blocknr);
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(entry_bh);
+	mark_buffer_dirty(entry_bh);
 	nilfs_mdt_mark_dirty(dat);
 
 	brelse(entry_bh);

fs/nilfs2/file.c

@@ -111,7 +111,6 @@ static int nilfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 	nilfs_transaction_commit(inode->i_sb);
 
  mapped:
-	SetPageChecked(page);
 	wait_on_page_writeback(page);
 	return VM_FAULT_LOCKED;
 }

fs/nilfs2/gcinode.c

@@ -48,9 +48,6 @@
 #include "dat.h"
 #include "ifile.h"
 
-static const struct address_space_operations def_gcinode_aops = {
-};
-
 /*
  * nilfs_gccache_submit_read_data() - add data buffer and submit read request
  * @inode - gc inode
@@ -87,9 +84,9 @@ int nilfs_gccache_submit_read_data(struct inode *inode, sector_t blkoff,
 		goto out;
 
 	if (pbn == 0) {
-		struct inode *dat_inode = NILFS_I_NILFS(inode)->ns_dat;
-					  /* use original dat, not gc dat. */
-		err = nilfs_dat_translate(dat_inode, vbn, &pbn);
+		struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
+
+		err = nilfs_dat_translate(nilfs->ns_dat, vbn, &pbn);
 		if (unlikely(err)) { /* -EIO, -ENOMEM, -ENOENT */
 			brelse(bh);
 			goto failed;
@@ -103,7 +100,7 @@ int nilfs_gccache_submit_read_data(struct inode *inode, sector_t blkoff,
 	}
 
 	if (!buffer_mapped(bh)) {
-		bh->b_bdev = NILFS_I_NILFS(inode)->ns_bdev;
+		bh->b_bdev = inode->i_sb->s_bdev;
 		set_buffer_mapped(bh);
 	}
 	bh->b_blocknr = pbn;
@@ -160,15 +157,11 @@ int nilfs_gccache_wait_and_mark_dirty(struct buffer_head *bh)
 	if (buffer_dirty(bh))
 		return -EEXIST;
 
-	if (buffer_nilfs_node(bh)) {
-		if (nilfs_btree_broken_node_block(bh)) {
-			clear_buffer_uptodate(bh);
-			return -EIO;
-		}
-		nilfs_btnode_mark_dirty(bh);
-	} else {
-		nilfs_mark_buffer_dirty(bh);
+	if (buffer_nilfs_node(bh) && nilfs_btree_broken_node_block(bh)) {
+		clear_buffer_uptodate(bh);
+		return -EIO;
 	}
+	mark_buffer_dirty(bh);
 	return 0;
 }
 
@@ -178,7 +171,7 @@ int nilfs_init_gcinode(struct inode *inode)
 
 	inode->i_mode = S_IFREG;
 	mapping_set_gfp_mask(inode->i_mapping, GFP_NOFS);
-	inode->i_mapping->a_ops = &def_gcinode_aops;
+	inode->i_mapping->a_ops = &empty_aops;
 	inode->i_mapping->backing_dev_info = inode->i_sb->s_bdi;
 
 	ii->i_flags = 0;

fs/nilfs2/ifile.c

@@ -80,7 +80,7 @@ int nilfs_ifile_create_inode(struct inode *ifile, ino_t *out_ino,
 		return ret;
 	}
 	nilfs_palloc_commit_alloc_entry(ifile, &req);
-	nilfs_mdt_mark_buffer_dirty(req.pr_entry_bh);
+	mark_buffer_dirty(req.pr_entry_bh);
 	nilfs_mdt_mark_dirty(ifile);
 	*out_ino = (ino_t)req.pr_entry_nr;
 	*out_bh = req.pr_entry_bh;
@@ -128,7 +128,7 @@ int nilfs_ifile_delete_inode(struct inode *ifile, ino_t ino)
 	raw_inode->i_flags = 0;
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(req.pr_entry_bh);
+	mark_buffer_dirty(req.pr_entry_bh);
 	brelse(req.pr_entry_bh);
 
 	nilfs_palloc_commit_free_entry(ifile, &req);

fs/nilfs2/inode.c

@@ -74,14 +74,14 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff,
 		    struct buffer_head *bh_result, int create)
 {
 	struct nilfs_inode_info *ii = NILFS_I(inode);
+	struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
 	__u64 blknum = 0;
 	int err = 0, ret;
-	struct inode *dat = NILFS_I_NILFS(inode)->ns_dat;
 	unsigned maxblocks = bh_result->b_size >> inode->i_blkbits;
 
-	down_read(&NILFS_MDT(dat)->mi_sem);
+	down_read(&NILFS_MDT(nilfs->ns_dat)->mi_sem);
 	ret = nilfs_bmap_lookup_contig(ii->i_bmap, blkoff, &blknum, maxblocks);
-	up_read(&NILFS_MDT(dat)->mi_sem);
+	up_read(&NILFS_MDT(nilfs->ns_dat)->mi_sem);
 	if (ret >= 0) {	/* found */
 		map_bh(bh_result, inode->i_sb, blknum);
 		if (ret > 0)
@@ -596,6 +596,16 @@ void nilfs_write_inode_common(struct inode *inode,
 	raw_inode->i_flags = cpu_to_le32(ii->i_flags);
 	raw_inode->i_generation = cpu_to_le32(inode->i_generation);
 
+	if (NILFS_ROOT_METADATA_FILE(inode->i_ino)) {
+		struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
+
+		/* zero-fill unused portion in the case of super root block */
+		raw_inode->i_xattr = 0;
+		raw_inode->i_pad = 0;
+		memset((void *)raw_inode + sizeof(*raw_inode), 0,
+		       nilfs->ns_inode_size - sizeof(*raw_inode));
+	}
+
 	if (has_bmap)
 		nilfs_bmap_write(ii->i_bmap, raw_inode);
 	else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
@@ -872,8 +882,7 @@ int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty)
 			return -EINVAL; /* NILFS_I_DIRTY may remain for
 					   freeing inode */
 		}
-		list_del(&ii->i_dirty);
-		list_add_tail(&ii->i_dirty, &nilfs->ns_dirty_files);
+		list_move_tail(&ii->i_dirty, &nilfs->ns_dirty_files);
 		set_bit(NILFS_I_QUEUED, &ii->i_state);
 	}
 	spin_unlock(&nilfs->ns_inode_lock);
@@ -892,7 +901,7 @@ int nilfs_mark_inode_dirty(struct inode *inode)
 		return err;
 	}
 	nilfs_update_inode(inode, ibh);
-	nilfs_mdt_mark_buffer_dirty(ibh);
+	mark_buffer_dirty(ibh);
 	nilfs_mdt_mark_dirty(NILFS_I(inode)->i_root->ifile);
 	brelse(ibh);
 	return 0;
@@ -931,7 +940,7 @@ void nilfs_dirty_inode(struct inode *inode)
 int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		 __u64 start, __u64 len)
 {
-	struct the_nilfs *nilfs = NILFS_I_NILFS(inode);
+	struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
 	__u64 logical = 0, phys = 0, size = 0;
 	__u32 flags = 0;
 	loff_t isize;

fs/nilfs2/ioctl.c

@@ -698,6 +698,63 @@ static int nilfs_ioctl_sync(struct inode *inode, struct file *filp,
 	return 0;
 }
 
+static int nilfs_ioctl_resize(struct inode *inode, struct file *filp,
+			      void __user *argp)
+{
+	__u64 newsize;
+	int ret = -EPERM;
+
+	if (!capable(CAP_SYS_ADMIN))
+		goto out;
+
+	ret = mnt_want_write(filp->f_path.mnt);
+	if (ret)
+		goto out;
+
+	ret = -EFAULT;
+	if (copy_from_user(&newsize, argp, sizeof(newsize)))
+		goto out_drop_write;
+
+	ret = nilfs_resize_fs(inode->i_sb, newsize);
+
+out_drop_write:
+	mnt_drop_write(filp->f_path.mnt);
+out:
+	return ret;
+}
+
+static int nilfs_ioctl_set_alloc_range(struct inode *inode, void __user *argp)
+{
+	struct the_nilfs *nilfs = inode->i_sb->s_fs_info;
+	__u64 range[2];
+	__u64 minseg, maxseg;
+	unsigned long segbytes;
+	int ret = -EPERM;
+
+	if (!capable(CAP_SYS_ADMIN))
+		goto out;
+
+	ret = -EFAULT;
+	if (copy_from_user(range, argp, sizeof(__u64[2])))
+		goto out;
+
+	ret = -ERANGE;
+	if (range[1] > i_size_read(inode->i_sb->s_bdev->bd_inode))
+		goto out;
+
+	segbytes = nilfs->ns_blocks_per_segment * nilfs->ns_blocksize;
+
+	minseg = range[0] + segbytes - 1;
+	do_div(minseg, segbytes);
+	maxseg = NILFS_SB2_OFFSET_BYTES(range[1]);
+	do_div(maxseg, segbytes);
+	maxseg--;
+
+	ret = nilfs_sufile_set_alloc_range(nilfs->ns_sufile, minseg, maxseg);
+out:
+	return ret;
+}
+
 static int nilfs_ioctl_get_info(struct inode *inode, struct file *filp,
 				unsigned int cmd, void __user *argp,
 				size_t membsz,
@@ -763,6 +820,10 @@ long nilfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 		return nilfs_ioctl_clean_segments(inode, filp, cmd, argp);
 	case NILFS_IOCTL_SYNC:
 		return nilfs_ioctl_sync(inode, filp, cmd, argp);
+	case NILFS_IOCTL_RESIZE:
+		return nilfs_ioctl_resize(inode, filp, argp);
+	case NILFS_IOCTL_SET_ALLOC_RANGE:
+		return nilfs_ioctl_set_alloc_range(inode, argp);
 	default:
 		return -ENOTTY;
 	}

fs/nilfs2/mdt.c

@@ -66,7 +66,7 @@ nilfs_mdt_insert_new_block(struct inode *inode, unsigned long block,
 	kunmap_atomic(kaddr, KM_USER0);
 
 	set_buffer_uptodate(bh);
-	nilfs_mark_buffer_dirty(bh);
+	mark_buffer_dirty(bh);
 	nilfs_mdt_mark_dirty(inode);
 	return 0;
 }
@@ -355,7 +355,7 @@ int nilfs_mdt_mark_block_dirty(struct inode *inode, unsigned long block)
 	err = nilfs_mdt_read_block(inode, block, 0, &bh);
 	if (unlikely(err))
 		return err;
-	nilfs_mark_buffer_dirty(bh);
+	mark_buffer_dirty(bh);
 	nilfs_mdt_mark_dirty(inode);
 	brelse(bh);
 	return 0;
@@ -450,9 +450,9 @@ int nilfs_mdt_setup_shadow_map(struct inode *inode,
 
 	INIT_LIST_HEAD(&shadow->frozen_buffers);
 	address_space_init_once(&shadow->frozen_data);
-	nilfs_mapping_init(&shadow->frozen_data, bdi);
+	nilfs_mapping_init(&shadow->frozen_data, inode, bdi);
 	address_space_init_once(&shadow->frozen_btnodes);
-	nilfs_mapping_init(&shadow->frozen_btnodes, bdi);
+	nilfs_mapping_init(&shadow->frozen_btnodes, inode, bdi);
 	mi->mi_shadow = shadow;
 	return 0;
 }

fs/nilfs2/mdt.h

@@ -64,11 +64,6 @@ static inline struct nilfs_mdt_info *NILFS_MDT(const struct inode *inode)
 	return inode->i_private;
 }
 
-static inline struct the_nilfs *NILFS_I_NILFS(struct inode *inode)
-{
-	return inode->i_sb->s_fs_info;
-}
-
 /* Default GFP flags using highmem */
 #define NILFS_MDT_GFP      (__GFP_WAIT | __GFP_IO | __GFP_HIGHMEM)
 
@@ -93,8 +88,6 @@ int nilfs_mdt_freeze_buffer(struct inode *inode, struct buffer_head *bh);
 struct buffer_head *nilfs_mdt_get_frozen_buffer(struct inode *inode,
 						struct buffer_head *bh);
 
-#define nilfs_mdt_mark_buffer_dirty(bh)	nilfs_mark_buffer_dirty(bh)
-
 static inline void nilfs_mdt_mark_dirty(struct inode *inode)
 {
 	if (!test_bit(NILFS_I_DIRTY, &NILFS_I(inode)->i_state))
@@ -108,7 +101,7 @@ static inline void nilfs_mdt_clear_dirty(struct inode *inode)
 
 static inline __u64 nilfs_mdt_cno(struct inode *inode)
 {
-	return NILFS_I_NILFS(inode)->ns_cno;
+	return ((struct the_nilfs *)inode->i_sb->s_fs_info)->ns_cno;
 }
 
 #define nilfs_mdt_bgl_lock(inode, bg) \

fs/nilfs2/nilfs.h

@@ -80,12 +80,6 @@ static inline struct inode *NILFS_BTNC_I(struct address_space *btnc)
 	return &ii->vfs_inode;
 }
 
-static inline struct inode *NILFS_AS_I(struct address_space *mapping)
-{
-	return (mapping->host) ? :
-		container_of(mapping, struct inode, i_data);
-}
-
 /*
  * Dynamic state flags of NILFS on-memory inode (i_state)
  */
@@ -298,6 +292,7 @@ struct nilfs_super_block **nilfs_prepare_super(struct super_block *sb,
 					       int flip);
 int nilfs_commit_super(struct super_block *sb, int flag);
 int nilfs_cleanup_super(struct super_block *sb);
+int nilfs_resize_fs(struct super_block *sb, __u64 newsize);
 int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt,
 			    struct nilfs_root **root);
 int nilfs_checkpoint_is_mounted(struct super_block *sb, __u64 cno);

fs/nilfs2/page.c

@@ -37,8 +37,7 @@
 
 #define NILFS_BUFFER_INHERENT_BITS  \
 	((1UL << BH_Uptodate) | (1UL << BH_Mapped) | (1UL << BH_NILFS_Node) | \
-	 (1UL << BH_NILFS_Volatile) | (1UL << BH_NILFS_Allocated) | \
-	 (1UL << BH_NILFS_Checked))
+	 (1UL << BH_NILFS_Volatile) | (1UL << BH_NILFS_Checked))
 
 static struct buffer_head *
 __nilfs_get_page_block(struct page *page, unsigned long block, pgoff_t index,
@@ -59,19 +58,6 @@ __nilfs_get_page_block(struct page *page, unsigned long block, pgoff_t index,
 	return bh;
 }
 
-/*
- * Since the page cache of B-tree node pages or data page cache of pseudo
- * inodes does not have a valid mapping->host pointer, calling
- * mark_buffer_dirty() for their buffers causes a NULL pointer dereference;
- * it calls __mark_inode_dirty(NULL) through __set_page_dirty().
- * To avoid this problem, the old style mark_buffer_dirty() is used instead.
- */
-void nilfs_mark_buffer_dirty(struct buffer_head *bh)
-{
-	if (!buffer_dirty(bh) && !test_set_buffer_dirty(bh))
-		__set_page_dirty_nobuffers(bh->b_page);
-}
-
 struct buffer_head *nilfs_grab_buffer(struct inode *inode,
 				      struct address_space *mapping,
 				      unsigned long blkoff,
@@ -183,7 +169,7 @@ int nilfs_page_buffers_clean(struct page *page)
 void nilfs_page_bug(struct page *page)
 {
 	struct address_space *m;
-	unsigned long ino = 0;
+	unsigned long ino;
 
 	if (unlikely(!page)) {
 		printk(KERN_CRIT "NILFS_PAGE_BUG(NULL)\n");
@@ -191,11 +177,8 @@ void nilfs_page_bug(struct page *page)
 	}
 
 	m = page->mapping;
-	if (m) {
-		struct inode *inode = NILFS_AS_I(m);
-		if (inode != NULL)
-			ino = inode->i_ino;
-	}
+	ino = m ? m->host->i_ino : 0;
+
 	printk(KERN_CRIT "NILFS_PAGE_BUG(%p): cnt=%d index#=%llu flags=0x%lx "
 	       "mapping=%p ino=%lu\n",
 	       page, atomic_read(&page->_count),
@@ -216,56 +199,6 @@ void nilfs_page_bug(struct page *page)
 	}
 }
 
-/**
- * nilfs_alloc_private_page - allocate a private page with buffer heads
- *
- * Return Value: On success, a pointer to the allocated page is returned.
- * On error, NULL is returned.
- */
-struct page *nilfs_alloc_private_page(struct block_device *bdev, int size,
-				      unsigned long state)
-{
-	struct buffer_head *bh, *head, *tail;
-	struct page *page;
-
-	page = alloc_page(GFP_NOFS); /* page_count of the returned page is 1 */
-	if (unlikely(!page))
-		return NULL;
-
-	lock_page(page);
-	head = alloc_page_buffers(page, size, 0);
-	if (unlikely(!head)) {
-		unlock_page(page);
-		__free_page(page);
-		return NULL;
-	}
-
-	bh = head;
-	do {
-		bh->b_state = (1UL << BH_NILFS_Allocated) | state;
-		tail = bh;
-		bh->b_bdev = bdev;
-		bh = bh->b_this_page;
-	} while (bh);
-
-	tail->b_this_page = head;
-	attach_page_buffers(page, head);
-
-	return page;
-}
-
-void nilfs_free_private_page(struct page *page)
-{
-	BUG_ON(!PageLocked(page));
-	BUG_ON(page->mapping);
-
-	if (page_has_buffers(page) && !try_to_free_buffers(page))
-		NILFS_PAGE_BUG(page, "failed to free page");
-
-	unlock_page(page);
-	__free_page(page);
-}
-
 /**
  * nilfs_copy_page -- copy the page with buffers
  * @dst: destination page
@@ -492,10 +425,10 @@ unsigned nilfs_page_count_clean_buffers(struct page *page,
 	return nc;
 }
 
-void nilfs_mapping_init(struct address_space *mapping,
+void nilfs_mapping_init(struct address_space *mapping, struct inode *inode,
 			struct backing_dev_info *bdi)
 {
-	mapping->host = NULL;
+	mapping->host = inode;
 	mapping->flags = 0;
 	mapping_set_gfp_mask(mapping, GFP_NOFS);
 	mapping->assoc_mapping = NULL;

fs/nilfs2/page.h

@@ -38,14 +38,12 @@ enum {
 	BH_NILFS_Redirected,
 };
 
-BUFFER_FNS(NILFS_Allocated, nilfs_allocated)	/* nilfs private buffers */
 BUFFER_FNS(NILFS_Node, nilfs_node)		/* nilfs node buffers */
 BUFFER_FNS(NILFS_Volatile, nilfs_volatile)
 BUFFER_FNS(NILFS_Checked, nilfs_checked)	/* buffer is verified */
 BUFFER_FNS(NILFS_Redirected, nilfs_redirected)	/* redirected to a copy */
 
 
-void nilfs_mark_buffer_dirty(struct buffer_head *bh);
 int __nilfs_clear_page_dirty(struct page *);
 
 struct buffer_head *nilfs_grab_buffer(struct inode *, struct address_space *,
@@ -54,14 +52,11 @@ void nilfs_forget_buffer(struct buffer_head *);
 void nilfs_copy_buffer(struct buffer_head *, struct buffer_head *);
 int nilfs_page_buffers_clean(struct page *);
 void nilfs_page_bug(struct page *);
-struct page *nilfs_alloc_private_page(struct block_device *, int,
-				      unsigned long);
-void nilfs_free_private_page(struct page *);
 
 int nilfs_copy_dirty_pages(struct address_space *, struct address_space *);
 void nilfs_copy_back_pages(struct address_space *, struct address_space *);
 void nilfs_clear_dirty_pages(struct address_space *);
-void nilfs_mapping_init(struct address_space *mapping,
+void nilfs_mapping_init(struct address_space *mapping, struct inode *inode,
 			struct backing_dev_info *bdi);
 unsigned nilfs_page_count_clean_buffers(struct page *, unsigned, unsigned);
 unsigned long nilfs_find_uncommitted_extent(struct inode *inode,

fs/nilfs2/recovery.c

@@ -387,9 +387,9 @@ static int nilfs_scan_dsync_log(struct the_nilfs *nilfs, sector_t start_blocknr,
 static void dispose_recovery_list(struct list_head *head)
 {
 	while (!list_empty(head)) {
-		struct nilfs_recovery_block *rb
-			= list_entry(head->next,
-				     struct nilfs_recovery_block, list);
+		struct nilfs_recovery_block *rb;
+
+		rb = list_first_entry(head, struct nilfs_recovery_block, list);
 		list_del(&rb->list);
 		kfree(rb);
 	}
@@ -416,9 +416,9 @@ static int nilfs_segment_list_add(struct list_head *head, __u64 segnum)
 void nilfs_dispose_segment_list(struct list_head *head)
 {
 	while (!list_empty(head)) {
-		struct nilfs_segment_entry *ent
-			= list_entry(head->next,
-				     struct nilfs_segment_entry, list);
+		struct nilfs_segment_entry *ent;
+
+		ent = list_first_entry(head, struct nilfs_segment_entry, list);
 		list_del(&ent->list);
 		kfree(ent);
 	}

fs/nilfs2/segbuf.c

@@ -239,12 +239,15 @@ nilfs_segbuf_fill_in_super_root_crc(struct nilfs_segment_buffer *segbuf,
 				    u32 seed)
 {
 	struct nilfs_super_root *raw_sr;
+	struct the_nilfs *nilfs = segbuf->sb_super->s_fs_info;
+	unsigned srsize;
 	u32 crc;
 
 	raw_sr = (struct nilfs_super_root *)segbuf->sb_super_root->b_data;
+	srsize = NILFS_SR_BYTES(nilfs->ns_inode_size);
 	crc = crc32_le(seed,
 		       (unsigned char *)raw_sr + sizeof(raw_sr->sr_sum),
-		       NILFS_SR_BYTES - sizeof(raw_sr->sr_sum));
+		       srsize - sizeof(raw_sr->sr_sum));
 	raw_sr->sr_sum = cpu_to_le32(crc);
 }
 
@@ -254,18 +257,6 @@ static void nilfs_release_buffers(struct list_head *list)
 
 	list_for_each_entry_safe(bh, n, list, b_assoc_buffers) {
 		list_del_init(&bh->b_assoc_buffers);
-		if (buffer_nilfs_allocated(bh)) {
-			struct page *clone_page = bh->b_page;
-
-			/* remove clone page */
-			brelse(bh);
-			page_cache_release(clone_page); /* for each bh */
-			if (page_count(clone_page) <= 2) {
-				lock_page(clone_page);
-				nilfs_free_private_page(clone_page);
-			}
-			continue;
-		}
 		brelse(bh);
 	}
 }

fs/nilfs2/segment.c

@@ -655,13 +655,10 @@ static size_t nilfs_lookup_dirty_data_buffers(struct inode *inode,
 		if (unlikely(page->index > last))
 			break;
 
-		if (mapping->host) {
-			lock_page(page);
-			if (!page_has_buffers(page))
-				create_empty_buffers(page,
-						     1 << inode->i_blkbits, 0);
-			unlock_page(page);
-		}
+		lock_page(page);
+		if (!page_has_buffers(page))
+			create_empty_buffers(page, 1 << inode->i_blkbits, 0);
+		unlock_page(page);
 
 		bh = head = page_buffers(page);
 		do {
@@ -809,7 +806,7 @@ static int nilfs_segctor_create_checkpoint(struct nilfs_sc_info *sci)
 		/* The following code is duplicated with cpfile.  But, it is
 		   needed to collect the checkpoint even if it was not newly
 		   created */
-		nilfs_mdt_mark_buffer_dirty(bh_cp);
+		mark_buffer_dirty(bh_cp);
 		nilfs_mdt_mark_dirty(nilfs->ns_cpfile);
 		nilfs_cpfile_put_checkpoint(
 			nilfs->ns_cpfile, nilfs->ns_cno, bh_cp);
@@ -889,12 +886,14 @@ static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci,
 {
 	struct buffer_head *bh_sr;
 	struct nilfs_super_root *raw_sr;
-	unsigned isz = nilfs->ns_inode_size;
+	unsigned isz, srsz;
 
 	bh_sr = NILFS_LAST_SEGBUF(&sci->sc_segbufs)->sb_super_root;
 	raw_sr = (struct nilfs_super_root *)bh_sr->b_data;
+	isz = nilfs->ns_inode_size;
+	srsz = NILFS_SR_BYTES(isz);
 
-	raw_sr->sr_bytes = cpu_to_le16(NILFS_SR_BYTES);
+	raw_sr->sr_bytes = cpu_to_le16(srsz);
 	raw_sr->sr_nongc_ctime
 		= cpu_to_le64(nilfs_doing_gc() ?
 			      nilfs->ns_nongc_ctime : sci->sc_seg_ctime);
@@ -906,6 +905,7 @@ static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci,
 				 NILFS_SR_CPFILE_OFFSET(isz), 1);
 	nilfs_write_inode_common(nilfs->ns_sufile, (void *)raw_sr +
 				 NILFS_SR_SUFILE_OFFSET(isz), 1);
+	memset((void *)raw_sr + srsz, 0, nilfs->ns_blocksize - srsz);
 }
 
 static void nilfs_redirty_inodes(struct list_head *head)
@@ -954,8 +954,8 @@ static int nilfs_segctor_apply_buffers(struct nilfs_sc_info *sci,
 
  dispose_buffers:
 	while (!list_empty(listp)) {
-		bh = list_entry(listp->next, struct buffer_head,
-				b_assoc_buffers);
+		bh = list_first_entry(listp, struct buffer_head,
+				      b_assoc_buffers);
 		list_del_init(&bh->b_assoc_buffers);
 		brelse(bh);
 	}
@@ -1500,10 +1500,7 @@ nilfs_segctor_update_payload_blocknr(struct nilfs_sc_info *sci,
 			nblocks = le32_to_cpu(finfo->fi_nblocks);
 			ndatablk = le32_to_cpu(finfo->fi_ndatablk);
 
-			if (buffer_nilfs_node(bh))
-				inode = NILFS_BTNC_I(bh->b_page->mapping);
-			else
-				inode = NILFS_AS_I(bh->b_page->mapping);
+			inode = bh->b_page->mapping->host;
 
 			if (mode == SC_LSEG_DSYNC)
 				sc_op = &nilfs_sc_dsync_ops;
@@ -1556,83 +1553,24 @@ static int nilfs_segctor_assign(struct nilfs_sc_info *sci, int mode)
 	return 0;
 }
 
-static int
-nilfs_copy_replace_page_buffers(struct page *page, struct list_head *out)
-{
-	struct page *clone_page;
-	struct buffer_head *bh, *head, *bh2;
-	void *kaddr;
-
-	bh = head = page_buffers(page);
-
-	clone_page = nilfs_alloc_private_page(bh->b_bdev, bh->b_size, 0);
-	if (unlikely(!clone_page))
-		return -ENOMEM;
-
-	bh2 = page_buffers(clone_page);
-	kaddr = kmap_atomic(page, KM_USER0);
-	do {
-		if (list_empty(&bh->b_assoc_buffers))
-			continue;
-		get_bh(bh2);
-		page_cache_get(clone_page); /* for each bh */
-		memcpy(bh2->b_data, kaddr + bh_offset(bh), bh2->b_size);
-		bh2->b_blocknr = bh->b_blocknr;
-		list_replace(&bh->b_assoc_buffers, &bh2->b_assoc_buffers);
-		list_add_tail(&bh->b_assoc_buffers, out);
-	} while (bh = bh->b_this_page, bh2 = bh2->b_this_page, bh != head);
-	kunmap_atomic(kaddr, KM_USER0);
-
-	if (!TestSetPageWriteback(clone_page))
-		account_page_writeback(clone_page);
-	unlock_page(clone_page);
-
-	return 0;
-}
-
-static int nilfs_test_page_to_be_frozen(struct page *page)
-{
-	struct address_space *mapping = page->mapping;
-
-	if (!mapping || !mapping->host || S_ISDIR(mapping->host->i_mode))
-		return 0;
-
-	if (page_mapped(page)) {
-		ClearPageChecked(page);
-		return 1;
-	}
-	return PageChecked(page);
-}
-
-static int nilfs_begin_page_io(struct page *page, struct list_head *out)
+static void nilfs_begin_page_io(struct page *page)
 {
 	if (!page || PageWriteback(page))
 		/* For split b-tree node pages, this function may be called
 		   twice.  We ignore the 2nd or later calls by this check. */
-		return 0;
+		return;
 
 	lock_page(page);
 	clear_page_dirty_for_io(page);
 	set_page_writeback(page);
 	unlock_page(page);
-
-	if (nilfs_test_page_to_be_frozen(page)) {
-		int err = nilfs_copy_replace_page_buffers(page, out);
-		if (unlikely(err))
-			return err;
-	}
-	return 0;
 }
 
-static int nilfs_segctor_prepare_write(struct nilfs_sc_info *sci,
-				       struct page **failed_page)
+static void nilfs_segctor_prepare_write(struct nilfs_sc_info *sci)
 {
 	struct nilfs_segment_buffer *segbuf;
 	struct page *bd_page = NULL, *fs_page = NULL;
-	struct list_head *list = &sci->sc_copied_buffers;
-	int err;
 
-	*failed_page = NULL;
 	list_for_each_entry(segbuf, &sci->sc_segbufs, sb_list) {
 		struct buffer_head *bh;
 
@@ -1662,11 +1600,7 @@ static int nilfs_segctor_prepare_write(struct nilfs_sc_info *sci,
 				break;
 			}
 			if (bh->b_page != fs_page) {
-				err = nilfs_begin_page_io(fs_page, list);
-				if (unlikely(err)) {
-					*failed_page = fs_page;
-					goto out;
-				}
+				nilfs_begin_page_io(fs_page);
 				fs_page = bh->b_page;
 			}
 		}
@@ -1677,11 +1611,7 @@ static int nilfs_segctor_prepare_write(struct nilfs_sc_info *sci,
 		set_page_writeback(bd_page);
 		unlock_page(bd_page);
 	}
-	err = nilfs_begin_page_io(fs_page, list);
-	if (unlikely(err))
-		*failed_page = fs_page;
- out:
-	return err;
+	nilfs_begin_page_io(fs_page);
 }
 
 static int nilfs_segctor_write(struct nilfs_sc_info *sci,
@@ -1694,24 +1624,6 @@ static int nilfs_segctor_write(struct nilfs_sc_info *sci,
 	return ret;
 }
 
-static void __nilfs_end_page_io(struct page *page, int err)
-{
-	if (!err) {
-		if (!nilfs_page_buffers_clean(page))
-			__set_page_dirty_nobuffers(page);
-		ClearPageError(page);
-	} else {
-		__set_page_dirty_nobuffers(page);
-		SetPageError(page);
-	}
-
-	if (buffer_nilfs_allocated(page_buffers(page))) {
-		if (TestClearPageWriteback(page))
-			dec_zone_page_state(page, NR_WRITEBACK);
-	} else
-		end_page_writeback(page);
-}
-
 static void nilfs_end_page_io(struct page *page, int err)
 {
 	if (!page)
@@ -1738,40 +1650,19 @@ static void nilfs_end_page_io(struct page *page, int err)
 		return;
 	}
 
-	__nilfs_end_page_io(page, err);
-}
-
-static void nilfs_clear_copied_buffers(struct list_head *list, int err)
-{
-	struct buffer_head *bh, *head;
-	struct page *page;
-
-	while (!list_empty(list)) {
-		bh = list_entry(list->next, struct buffer_head,
-				b_assoc_buffers);
-		page = bh->b_page;
-		page_cache_get(page);
-		head = bh = page_buffers(page);
-		do {
-			if (!list_empty(&bh->b_assoc_buffers)) {
-				list_del_init(&bh->b_assoc_buffers);
-				if (!err) {
-					set_buffer_uptodate(bh);
-					clear_buffer_dirty(bh);
-					clear_buffer_delay(bh);
-					clear_buffer_nilfs_volatile(bh);
-				}
-				brelse(bh); /* for b_assoc_buffers */
-			}
-		} while ((bh = bh->b_this_page) != head);
-
-		__nilfs_end_page_io(page, err);
-		page_cache_release(page);
+	if (!err) {
+		if (!nilfs_page_buffers_clean(page))
+			__set_page_dirty_nobuffers(page);
+		ClearPageError(page);
+	} else {
+		__set_page_dirty_nobuffers(page);
+		SetPageError(page);
 	}
+
+	end_page_writeback(page);
 }
 
-static void nilfs_abort_logs(struct list_head *logs, struct page *failed_page,
-			     int err)
+static void nilfs_abort_logs(struct list_head *logs, int err)
 {
 	struct nilfs_segment_buffer *segbuf;
 	struct page *bd_page = NULL, *fs_page = NULL;
@@ -1801,8 +1692,6 @@ static void nilfs_abort_logs(struct list_head *logs, struct page *failed_page,
 			}
 			if (bh->b_page != fs_page) {
 				nilfs_end_page_io(fs_page, err);
-				if (fs_page && fs_page == failed_page)
-					return;
 				fs_page = bh->b_page;
 			}
 		}
@@ -1821,12 +1710,11 @@ static void nilfs_segctor_abort_construction(struct nilfs_sc_info *sci,
 
 	list_splice_tail_init(&sci->sc_write_logs, &logs);
 	ret = nilfs_wait_on_logs(&logs);
-	nilfs_abort_logs(&logs, NULL, ret ? : err);
+	nilfs_abort_logs(&logs, ret ? : err);
 
 	list_splice_tail_init(&sci->sc_segbufs, &logs);
 	nilfs_cancel_segusage(&logs, nilfs->ns_sufile);
 	nilfs_free_incomplete_logs(&logs, nilfs);
-	nilfs_clear_copied_buffers(&sci->sc_copied_buffers, err);
 
 	if (sci->sc_stage.flags & NILFS_CF_SUFREED) {
 		ret = nilfs_sufile_cancel_freev(nilfs->ns_sufile,
@@ -1920,8 +1808,6 @@ static void nilfs_segctor_complete_write(struct nilfs_sc_info *sci)
 
 	nilfs_end_page_io(fs_page, 0);
 
-	nilfs_clear_copied_buffers(&sci->sc_copied_buffers, 0);
-
 	nilfs_drop_collected_inodes(&sci->sc_dirty_files);
 
 	if (nilfs_doing_gc())
@@ -1979,7 +1865,7 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci,
 					      "failed to get inode block.\n");
 				return err;
 			}
-			nilfs_mdt_mark_buffer_dirty(ibh);
+			mark_buffer_dirty(ibh);
 			nilfs_mdt_mark_dirty(ifile);
 			spin_lock(&nilfs->ns_inode_lock);
 			if (likely(!ii->i_bh))
@@ -1991,8 +1877,7 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci,
 
 		clear_bit(NILFS_I_QUEUED, &ii->i_state);
 		set_bit(NILFS_I_BUSY, &ii->i_state);
-		list_del(&ii->i_dirty);
-		list_add_tail(&ii->i_dirty, &sci->sc_dirty_files);
+		list_move_tail(&ii->i_dirty, &sci->sc_dirty_files);
 	}
 	spin_unlock(&nilfs->ns_inode_lock);
 
@@ -2014,8 +1899,7 @@ static void nilfs_segctor_drop_written_files(struct nilfs_sc_info *sci,
 		clear_bit(NILFS_I_BUSY, &ii->i_state);
 		brelse(ii->i_bh);
 		ii->i_bh = NULL;
-		list_del(&ii->i_dirty);
-		list_add_tail(&ii->i_dirty, &ti->ti_garbage);
+		list_move_tail(&ii->i_dirty, &ti->ti_garbage);
 	}
 	spin_unlock(&nilfs->ns_inode_lock);
 }
@@ -2026,7 +1910,6 @@ static void nilfs_segctor_drop_written_files(struct nilfs_sc_info *sci,
 static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 {
 	struct the_nilfs *nilfs = sci->sc_super->s_fs_info;
-	struct page *failed_page;
 	int err;
 
 	sci->sc_stage.scnt = NILFS_ST_INIT;
@@ -2081,11 +1964,7 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode)
 		nilfs_segctor_update_segusage(sci, nilfs->ns_sufile);
 
 		/* Write partial segments */
-		err = nilfs_segctor_prepare_write(sci, &failed_page);
-		if (err) {
-			nilfs_abort_logs(&sci->sc_segbufs, failed_page, err);
-			goto failed_to_write;
-		}
+		nilfs_segctor_prepare_write(sci);
 
 		nilfs_add_checksums_on_logs(&sci->sc_segbufs,
 					    nilfs->ns_crc_seed);
@@ -2687,7 +2566,6 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct super_block *sb,
 	INIT_LIST_HEAD(&sci->sc_segbufs);
 	INIT_LIST_HEAD(&sci->sc_write_logs);
 	INIT_LIST_HEAD(&sci->sc_gc_inodes);
-	INIT_LIST_HEAD(&sci->sc_copied_buffers);
 	init_timer(&sci->sc_timer);
 
 	sci->sc_interval = HZ * NILFS_SC_DEFAULT_TIMEOUT;
@@ -2741,8 +2619,6 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci)
 	if (flag || !nilfs_segctor_confirm(sci))
 		nilfs_segctor_write_out(sci);
 
-	WARN_ON(!list_empty(&sci->sc_copied_buffers));
-
 	if (!list_empty(&sci->sc_dirty_files)) {
 		nilfs_warning(sci->sc_super, __func__,
 			      "dirty file(s) after the final construction\n");

fs/nilfs2/segment.h

@@ -92,7 +92,6 @@ struct nilfs_segsum_pointer {
  * @sc_nblk_inc: Block count of current generation
  * @sc_dirty_files: List of files to be written
  * @sc_gc_inodes: List of GC inodes having blocks to be written
- * @sc_copied_buffers: List of copied buffers (buffer heads) to freeze data
  * @sc_freesegs: array of segment numbers to be freed
  * @sc_nfreesegs: number of segments on @sc_freesegs
  * @sc_dsync_inode: inode whose data pages are written for a sync operation
@@ -136,7 +135,6 @@ struct nilfs_sc_info {
 
 	struct list_head	sc_dirty_files;
 	struct list_head	sc_gc_inodes;
-	struct list_head	sc_copied_buffers;
 
 	__u64		       *sc_freesegs;
 	size_t			sc_nfreesegs;

fs/nilfs2/sufile.c

@@ -33,7 +33,9 @@
 
 struct nilfs_sufile_info {
 	struct nilfs_mdt_info mi;
-	unsigned long ncleansegs;
+	unsigned long ncleansegs;/* number of clean segments */
+	__u64 allocmin;		/* lower limit of allocatable segment range */
+	__u64 allocmax;		/* upper limit of allocatable segment range */
 };
 
 static inline struct nilfs_sufile_info *NILFS_SUI(struct inode *sufile)
@@ -96,6 +98,13 @@ nilfs_sufile_get_segment_usage_block(struct inode *sufile, __u64 segnum,
 				   create, NULL, bhp);
 }
 
+static int nilfs_sufile_delete_segment_usage_block(struct inode *sufile,
+						   __u64 segnum)
+{
+	return nilfs_mdt_delete_block(sufile,
+				      nilfs_sufile_get_blkoff(sufile, segnum));
+}
+
 static void nilfs_sufile_mod_counter(struct buffer_head *header_bh,
 				     u64 ncleanadd, u64 ndirtyadd)
 {
@@ -108,7 +117,7 @@ static void nilfs_sufile_mod_counter(struct buffer_head *header_bh,
 	le64_add_cpu(&header->sh_ndirtysegs, ndirtyadd);
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(header_bh);
+	mark_buffer_dirty(header_bh);
 }
 
 /**
@@ -247,6 +256,35 @@ int nilfs_sufile_update(struct inode *sufile, __u64 segnum, int create,
 	return ret;
 }
 
+/**
+ * nilfs_sufile_set_alloc_range - limit range of segment to be allocated
+ * @sufile: inode of segment usage file
+ * @start: minimum segment number of allocatable region (inclusive)
+ * @end: maximum segment number of allocatable region (inclusive)
+ *
+ * Return Value: On success, 0 is returned.  On error, one of the
+ * following negative error codes is returned.
+ *
+ * %-ERANGE - invalid segment region
+ */
+int nilfs_sufile_set_alloc_range(struct inode *sufile, __u64 start, __u64 end)
+{
+	struct nilfs_sufile_info *sui = NILFS_SUI(sufile);
+	__u64 nsegs;
+	int ret = -ERANGE;
+
+	down_write(&NILFS_MDT(sufile)->mi_sem);
+	nsegs = nilfs_sufile_get_nsegments(sufile);
+
+	if (start <= end && end < nsegs) {
+		sui->allocmin = start;
+		sui->allocmax = end;
+		ret = 0;
+	}
+	up_write(&NILFS_MDT(sufile)->mi_sem);
+	return ret;
+}
+
 /**
  * nilfs_sufile_alloc - allocate a segment
  * @sufile: inode of segment usage file
@@ -269,11 +307,12 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 	struct buffer_head *header_bh, *su_bh;
 	struct nilfs_sufile_header *header;
 	struct nilfs_segment_usage *su;
+	struct nilfs_sufile_info *sui = NILFS_SUI(sufile);
 	size_t susz = NILFS_MDT(sufile)->mi_entry_size;
 	__u64 segnum, maxsegnum, last_alloc;
 	void *kaddr;
-	unsigned long nsegments, ncleansegs, nsus;
-	int ret, i, j;
+	unsigned long nsegments, ncleansegs, nsus, cnt;
+	int ret, j;
 
 	down_write(&NILFS_MDT(sufile)->mi_sem);
 
@@ -287,13 +326,31 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 	kunmap_atomic(kaddr, KM_USER0);
 
 	nsegments = nilfs_sufile_get_nsegments(sufile);
+	maxsegnum = sui->allocmax;
 	segnum = last_alloc + 1;
-	maxsegnum = nsegments - 1;
-	for (i = 0; i < nsegments; i += nsus) {
-		if (segnum >= nsegments) {
-			/* wrap around */
-			segnum = 0;
-			maxsegnum = last_alloc;
+	if (segnum < sui->allocmin || segnum > sui->allocmax)
+		segnum = sui->allocmin;
+
+	for (cnt = 0; cnt < nsegments; cnt += nsus) {
+		if (segnum > maxsegnum) {
+			if (cnt < sui->allocmax - sui->allocmin + 1) {
+				/*
+				 * wrap around in the limited region.
+				 * if allocation started from
+				 * sui->allocmin, this never happens.
+				 */
+				segnum = sui->allocmin;
+				maxsegnum = last_alloc;
+			} else if (segnum > sui->allocmin &&
+				   sui->allocmax + 1 < nsegments) {
+				segnum = sui->allocmax + 1;
+				maxsegnum = nsegments - 1;
+			} else if (sui->allocmin > 0)  {
+				segnum = 0;
+				maxsegnum = sui->allocmin - 1;
+			} else {
+				break; /* never happens */
+			}
 		}
 		ret = nilfs_sufile_get_segment_usage_block(sufile, segnum, 1,
 							   &su_bh);
@@ -319,9 +376,9 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump)
 			header->sh_last_alloc = cpu_to_le64(segnum);
 			kunmap_atomic(kaddr, KM_USER0);
 
-			NILFS_SUI(sufile)->ncleansegs--;
-			nilfs_mdt_mark_buffer_dirty(header_bh);
-			nilfs_mdt_mark_buffer_dirty(su_bh);
+			sui->ncleansegs--;
+			mark_buffer_dirty(header_bh);
+			mark_buffer_dirty(su_bh);
 			nilfs_mdt_mark_dirty(sufile);
 			brelse(su_bh);
 			*segnump = segnum;
@@ -364,7 +421,7 @@ void nilfs_sufile_do_cancel_free(struct inode *sufile, __u64 segnum,
 	nilfs_sufile_mod_counter(header_bh, -1, 1);
 	NILFS_SUI(sufile)->ncleansegs--;
 
-	nilfs_mdt_mark_buffer_dirty(su_bh);
+	mark_buffer_dirty(su_bh);
 	nilfs_mdt_mark_dirty(sufile);
 }
 
@@ -395,7 +452,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum,
 	nilfs_sufile_mod_counter(header_bh, clean ? (u64)-1 : 0, dirty ? 0 : 1);
 	NILFS_SUI(sufile)->ncleansegs -= clean;
 
-	nilfs_mdt_mark_buffer_dirty(su_bh);
+	mark_buffer_dirty(su_bh);
 	nilfs_mdt_mark_dirty(sufile);
 }
 
@@ -421,7 +478,7 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum,
 	sudirty = nilfs_segment_usage_dirty(su);
 	nilfs_segment_usage_set_clean(su);
 	kunmap_atomic(kaddr, KM_USER0);
-	nilfs_mdt_mark_buffer_dirty(su_bh);
+	mark_buffer_dirty(su_bh);
 
 	nilfs_sufile_mod_counter(header_bh, 1, sudirty ? (u64)-1 : 0);
 	NILFS_SUI(sufile)->ncleansegs++;
@@ -441,7 +498,7 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum)
 
 	ret = nilfs_sufile_get_segment_usage_block(sufile, segnum, 0, &bh);
 	if (!ret) {
-		nilfs_mdt_mark_buffer_dirty(bh);
+		mark_buffer_dirty(bh);
 		nilfs_mdt_mark_dirty(sufile);
 		brelse(bh);
 	}
@@ -476,7 +533,7 @@ int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum,
 	su->su_nblocks = cpu_to_le32(nblocks);
 	kunmap_atomic(kaddr, KM_USER0);
 
-	nilfs_mdt_mark_buffer_dirty(bh);
+	mark_buffer_dirty(bh);
 	nilfs_mdt_mark_dirty(sufile);
 	brelse(bh);
 
@@ -505,7 +562,7 @@ int nilfs_sufile_get_stat(struct inode *sufile, struct nilfs_sustat *sustat)
 {
 	struct buffer_head *header_bh;
 	struct nilfs_sufile_header *header;
-	struct the_nilfs *nilfs = NILFS_I_NILFS(sufile);
+	struct the_nilfs *nilfs = sufile->i_sb->s_fs_info;
 	void *kaddr;
 	int ret;
 
@@ -555,10 +612,182 @@ void nilfs_sufile_do_set_error(struct inode *sufile, __u64 segnum,
 		nilfs_sufile_mod_counter(header_bh, -1, 0);
 		NILFS_SUI(sufile)->ncleansegs--;
 	}
-	nilfs_mdt_mark_buffer_dirty(su_bh);
+	mark_buffer_dirty(su_bh);
 	nilfs_mdt_mark_dirty(sufile);
 }
 
+/**
+  * nilfs_sufile_truncate_range - truncate range of segment array
+  * @sufile: inode of segment usage file
+  * @start: start segment number (inclusive)
+  * @end: end segment number (inclusive)
+  *
+  * Return Value: On success, 0 is returned.  On error, one of the
+  * following negative error codes is returned.
+  *
+  * %-EIO - I/O error.
+  *
+  * %-ENOMEM - Insufficient amount of memory available.
+  *
+  * %-EINVAL - Invalid number of segments specified
+  *
+  * %-EBUSY - Dirty or active segments are present in the range
+  */
+static int nilfs_sufile_truncate_range(struct inode *sufile,
+				       __u64 start, __u64 end)
+{
+	struct the_nilfs *nilfs = sufile->i_sb->s_fs_info;
+	struct buffer_head *header_bh;
+	struct buffer_head *su_bh;
+	struct nilfs_segment_usage *su, *su2;
+	size_t susz = NILFS_MDT(sufile)->mi_entry_size;
+	unsigned long segusages_per_block;
+	unsigned long nsegs, ncleaned;
+	__u64 segnum;
+	void *kaddr;
+	ssize_t n, nc;
+	int ret;
+	int j;
+
+	nsegs = nilfs_sufile_get_nsegments(sufile);
+
+	ret = -EINVAL;
+	if (start > end || start >= nsegs)
+		goto out;
+
+	ret = nilfs_sufile_get_header_block(sufile, &header_bh);
+	if (ret < 0)
+		goto out;
+
+	segusages_per_block = nilfs_sufile_segment_usages_per_block(sufile);
+	ncleaned = 0;
+
+	for (segnum = start; segnum <= end; segnum += n) {
+		n = min_t(unsigned long,
+			  segusages_per_block -
+				  nilfs_sufile_get_offset(sufile, segnum),
+			  end - segnum + 1);
+		ret = nilfs_sufile_get_segment_usage_block(sufile, segnum, 0,
+							   &su_bh);
+		if (ret < 0) {
+			if (ret != -ENOENT)
+				goto out_header;
+			/* hole */
+			continue;
+		}
+		kaddr = kmap_atomic(su_bh->b_page, KM_USER0);
+		su = nilfs_sufile_block_get_segment_usage(
+			sufile, segnum, su_bh, kaddr);
+		su2 = su;
+		for (j = 0; j < n; j++, su = (void *)su + susz) {
+			if ((le32_to_cpu(su->su_flags) &
+			     ~(1UL << NILFS_SEGMENT_USAGE_ERROR)) ||
+			    nilfs_segment_is_active(nilfs, segnum + j)) {
+				ret = -EBUSY;
+				kunmap_atomic(kaddr, KM_USER0);
+				brelse(su_bh);
+				goto out_header;
+			}
+		}
+		nc = 0;
+		for (su = su2, j = 0; j < n; j++, su = (void *)su + susz) {
+			if (nilfs_segment_usage_error(su)) {
+				nilfs_segment_usage_set_clean(su);
+				nc++;
+			}
+		}
+		kunmap_atomic(kaddr, KM_USER0);
+		if (nc > 0) {
+			mark_buffer_dirty(su_bh);
+			ncleaned += nc;
+		}
+		brelse(su_bh);
+
+		if (n == segusages_per_block) {
+			/* make hole */
+			nilfs_sufile_delete_segment_usage_block(sufile, segnum);
+		}
+	}
+	ret = 0;
+
+out_header:
+	if (ncleaned > 0) {
+		NILFS_SUI(sufile)->ncleansegs += ncleaned;
+		nilfs_sufile_mod_counter(header_bh, ncleaned, 0);
+		nilfs_mdt_mark_dirty(sufile);
+	}
+	brelse(header_bh);
+out:
+	return ret;
+}
+
+/**
+ * nilfs_sufile_resize - resize segment array
+ * @sufile: inode of segment usage file
+ * @newnsegs: new number of segments
+ *
+ * Return Value: On success, 0 is returned.  On error, one of the
+ * following negative error codes is returned.
+ *
+ * %-EIO - I/O error.
+ *
+ * %-ENOMEM - Insufficient amount of memory available.
+ *
+ * %-ENOSPC - Enough free space is not left for shrinking
+ *
+ * %-EBUSY - Dirty or active segments exist in the region to be truncated
+ */
+int nilfs_sufile_resize(struct inode *sufile, __u64 newnsegs)
+{
+	struct the_nilfs *nilfs = sufile->i_sb->s_fs_info;
+	struct buffer_head *header_bh;
+	struct nilfs_sufile_header *header;
+	struct nilfs_sufile_info *sui = NILFS_SUI(sufile);
+	void *kaddr;
+	unsigned long nsegs, nrsvsegs;
+	int ret = 0;
+
+	down_write(&NILFS_MDT(sufile)->mi_sem);
+
+	nsegs = nilfs_sufile_get_nsegments(sufile);
+	if (nsegs == newnsegs)
+		goto out;
+
+	ret = -ENOSPC;
+	nrsvsegs = nilfs_nrsvsegs(nilfs, newnsegs);
+	if (newnsegs < nsegs && nsegs - newnsegs + nrsvsegs > sui->ncleansegs)
+		goto out;
+
+	ret = nilfs_sufile_get_header_block(sufile, &header_bh);
+	if (ret < 0)
+		goto out;
+
+	if (newnsegs > nsegs) {
+		sui->ncleansegs += newnsegs - nsegs;
+	} else /* newnsegs < nsegs */ {
+		ret = nilfs_sufile_truncate_range(sufile, newnsegs, nsegs - 1);
+		if (ret < 0)
+			goto out_header;
+
+		sui->ncleansegs -= nsegs - newnsegs;
+	}
+
+	kaddr = kmap_atomic(header_bh->b_page, KM_USER0);
+	header = kaddr + bh_offset(header_bh);
+	header->sh_ncleansegs = cpu_to_le64(sui->ncleansegs);
+	kunmap_atomic(kaddr, KM_USER0);
+
+	mark_buffer_dirty(header_bh);
+	nilfs_mdt_mark_dirty(sufile);
+	nilfs_set_nsegments(nilfs, newnsegs);
+
+out_header:
+	brelse(header_bh);
+out:
+	up_write(&NILFS_MDT(sufile)->mi_sem);
+	return ret;
+}
+
 /**
  * nilfs_sufile_get_suinfo -
  * @sufile: inode of segment usage file
@@ -583,7 +812,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf,
 	struct nilfs_segment_usage *su;
 	struct nilfs_suinfo *si = buf;
 	size_t susz = NILFS_MDT(sufile)->mi_entry_size;
-	struct the_nilfs *nilfs = NILFS_I_NILFS(sufile);
+	struct the_nilfs *nilfs = sufile->i_sb->s_fs_info;
 	void *kaddr;
 	unsigned long nsegs, segusages_per_block;
 	ssize_t n;
@@ -679,6 +908,9 @@ int nilfs_sufile_read(struct super_block *sb, size_t susize,
 	kunmap_atomic(kaddr, KM_USER0);
 	brelse(header_bh);
 
+	sui->allocmax = nilfs_sufile_get_nsegments(sufile) - 1;
+	sui->allocmin = 0;
+
 	unlock_new_inode(sufile);
  out:
 	*inodep = sufile;

fs/nilfs2/sufile.h

@@ -31,11 +31,12 @@
 
 static inline unsigned long nilfs_sufile_get_nsegments(struct inode *sufile)
 {
-	return NILFS_I_NILFS(sufile)->ns_nsegments;
+	return ((struct the_nilfs *)sufile->i_sb->s_fs_info)->ns_nsegments;
 }
 
 unsigned long nilfs_sufile_get_ncleansegs(struct inode *sufile);
 
+int nilfs_sufile_set_alloc_range(struct inode *sufile, __u64 start, __u64 end);
 int nilfs_sufile_alloc(struct inode *, __u64 *);
 int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum);
 int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum,
@@ -61,6 +62,7 @@ void nilfs_sufile_do_cancel_free(struct inode *, __u64, struct buffer_head *,
 void nilfs_sufile_do_set_error(struct inode *, __u64, struct buffer_head *,
 			       struct buffer_head *);
 
+int nilfs_sufile_resize(struct inode *sufile, __u64 newnsegs);
 int nilfs_sufile_read(struct super_block *sb, size_t susize,
 		      struct nilfs_inode *raw_inode, struct inode **inodep);
 

fs/nilfs2/super.c

@@ -56,6 +56,7 @@
 #include "btnode.h"
 #include "page.h"
 #include "cpfile.h"
+#include "sufile.h" /* nilfs_sufile_resize(), nilfs_sufile_set_alloc_range() */
 #include "ifile.h"
 #include "dat.h"
 #include "segment.h"
@@ -165,7 +166,7 @@ struct inode *nilfs_alloc_inode(struct super_block *sb)
 	ii->i_state = 0;
 	ii->i_cno = 0;
 	ii->vfs_inode.i_version = 1;
-	nilfs_btnode_cache_init(&ii->i_btnode_cache, sb->s_bdi);
+	nilfs_mapping_init(&ii->i_btnode_cache, &ii->vfs_inode, sb->s_bdi);
 	return &ii->vfs_inode;
 }
 
@@ -347,6 +348,134 @@ int nilfs_cleanup_super(struct super_block *sb)
 	return ret;
 }
 
+/**
+ * nilfs_move_2nd_super - relocate secondary super block
+ * @sb: super block instance
+ * @sb2off: new offset of the secondary super block (in bytes)
+ */
+static int nilfs_move_2nd_super(struct super_block *sb, loff_t sb2off)
+{
+	struct the_nilfs *nilfs = sb->s_fs_info;
+	struct buffer_head *nsbh;
+	struct nilfs_super_block *nsbp;
+	sector_t blocknr, newblocknr;
+	unsigned long offset;
+	int sb2i = -1;  /* array index of the secondary superblock */
+	int ret = 0;
+
+	/* nilfs->ns_sem must be locked by the caller. */
+	if (nilfs->ns_sbh[1] &&
+	    nilfs->ns_sbh[1]->b_blocknr > nilfs->ns_first_data_block) {
+		sb2i = 1;
+		blocknr = nilfs->ns_sbh[1]->b_blocknr;
+	} else if (nilfs->ns_sbh[0]->b_blocknr > nilfs->ns_first_data_block) {
+		sb2i = 0;
+		blocknr = nilfs->ns_sbh[0]->b_blocknr;
+	}
+	if (sb2i >= 0 && (u64)blocknr << nilfs->ns_blocksize_bits == sb2off)
+		goto out;  /* super block location is unchanged */
+
+	/* Get new super block buffer */
+	newblocknr = sb2off >> nilfs->ns_blocksize_bits;
+	offset = sb2off & (nilfs->ns_blocksize - 1);
+	nsbh = sb_getblk(sb, newblocknr);
+	if (!nsbh) {
+		printk(KERN_WARNING
+		       "NILFS warning: unable to move secondary superblock "
+		       "to block %llu\n", (unsigned long long)newblocknr);
+		ret = -EIO;
+		goto out;
+	}
+	nsbp = (void *)nsbh->b_data + offset;
+	memset(nsbp, 0, nilfs->ns_blocksize);
+
+	if (sb2i >= 0) {
+		memcpy(nsbp, nilfs->ns_sbp[sb2i], nilfs->ns_sbsize);
+		brelse(nilfs->ns_sbh[sb2i]);
+		nilfs->ns_sbh[sb2i] = nsbh;
+		nilfs->ns_sbp[sb2i] = nsbp;
+	} else if (nilfs->ns_sbh[0]->b_blocknr < nilfs->ns_first_data_block) {
+		/* secondary super block will be restored to index 1 */
+		nilfs->ns_sbh[1] = nsbh;
+		nilfs->ns_sbp[1] = nsbp;
+	} else {
+		brelse(nsbh);
+	}
+out:
+	return ret;
+}
+
+/**
+ * nilfs_resize_fs - resize the filesystem
+ * @sb: super block instance
+ * @newsize: new size of the filesystem (in bytes)
+ */
+int nilfs_resize_fs(struct super_block *sb, __u64 newsize)
+{
+	struct the_nilfs *nilfs = sb->s_fs_info;
+	struct nilfs_super_block **sbp;
+	__u64 devsize, newnsegs;
+	loff_t sb2off;
+	int ret;
+
+	ret = -ERANGE;
+	devsize = i_size_read(sb->s_bdev->bd_inode);
+	if (newsize > devsize)
+		goto out;
+
+	/*
+	 * Write lock is required to protect some functions depending
+	 * on the number of segments, the number of reserved segments,
+	 * and so forth.
+	 */
+	down_write(&nilfs->ns_segctor_sem);
+
+	sb2off = NILFS_SB2_OFFSET_BYTES(newsize);
+	newnsegs = sb2off >> nilfs->ns_blocksize_bits;
+	do_div(newnsegs, nilfs->ns_blocks_per_segment);
+
+	ret = nilfs_sufile_resize(nilfs->ns_sufile, newnsegs);
+	up_write(&nilfs->ns_segctor_sem);
+	if (ret < 0)
+		goto out;
+
+	ret = nilfs_construct_segment(sb);
+	if (ret < 0)
+		goto out;
+
+	down_write(&nilfs->ns_sem);
+	nilfs_move_2nd_super(sb, sb2off);
+	ret = -EIO;
+	sbp = nilfs_prepare_super(sb, 0);
+	if (likely(sbp)) {
+		nilfs_set_log_cursor(sbp[0], nilfs);
+		/*
+		 * Drop NILFS_RESIZE_FS flag for compatibility with
+		 * mount-time resize which may be implemented in a
+		 * future release.
+		 */
+		sbp[0]->s_state = cpu_to_le16(le16_to_cpu(sbp[0]->s_state) &
+					      ~NILFS_RESIZE_FS);
+		sbp[0]->s_dev_size = cpu_to_le64(newsize);
+		sbp[0]->s_nsegments = cpu_to_le64(nilfs->ns_nsegments);
+		if (sbp[1])
+			memcpy(sbp[1], sbp[0], nilfs->ns_sbsize);
+		ret = nilfs_commit_super(sb, NILFS_SB_COMMIT_ALL);
+	}
+	up_write(&nilfs->ns_sem);
+
+	/*
+	 * Reset the range of allocatable segments last.  This order
+	 * is important in the case of expansion because the secondary
+	 * superblock must be protected from log write until migration
+	 * completes.
+	 */
+	if (!ret)
+		nilfs_sufile_set_alloc_range(nilfs->ns_sufile, 0, newnsegs - 1);
+out:
+	return ret;
+}
+
 static void nilfs_put_super(struct super_block *sb)
 {
 	struct the_nilfs *nilfs = sb->s_fs_info;

fs/nilfs2/the_nilfs.c

@@ -363,6 +363,24 @@ static unsigned long long nilfs_max_size(unsigned int blkbits)
 	return res;
 }
 
+/**
+ * nilfs_nrsvsegs - calculate the number of reserved segments
+ * @nilfs: nilfs object
+ * @nsegs: total number of segments
+ */
+unsigned long nilfs_nrsvsegs(struct the_nilfs *nilfs, unsigned long nsegs)
+{
+	return max_t(unsigned long, NILFS_MIN_NRSVSEGS,
+		     DIV_ROUND_UP(nsegs * nilfs->ns_r_segments_percentage,
+				  100));
+}
+
+void nilfs_set_nsegments(struct the_nilfs *nilfs, unsigned long nsegs)
+{
+	nilfs->ns_nsegments = nsegs;
+	nilfs->ns_nrsvsegs = nilfs_nrsvsegs(nilfs, nsegs);
+}
+
 static int nilfs_store_disk_layout(struct the_nilfs *nilfs,
 				   struct nilfs_super_block *sbp)
 {
@@ -389,13 +407,9 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs,
 	}
 
 	nilfs->ns_first_data_block = le64_to_cpu(sbp->s_first_data_block);
-	nilfs->ns_nsegments = le64_to_cpu(sbp->s_nsegments);
 	nilfs->ns_r_segments_percentage =
 		le32_to_cpu(sbp->s_r_segments_percentage);
-	nilfs->ns_nrsvsegs =
-		max_t(unsigned long, NILFS_MIN_NRSVSEGS,
-		      DIV_ROUND_UP(nilfs->ns_nsegments *
-				   nilfs->ns_r_segments_percentage, 100));
+	nilfs_set_nsegments(nilfs, le64_to_cpu(sbp->s_nsegments));
 	nilfs->ns_crc_seed = le32_to_cpu(sbp->s_crc_seed);
 	return 0;
 }

fs/nilfs2/the_nilfs.h

@@ -268,6 +268,8 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev);
 void destroy_nilfs(struct the_nilfs *nilfs);
 int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data);
 int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb);
+unsigned long nilfs_nrsvsegs(struct the_nilfs *nilfs, unsigned long nsegs);
+void nilfs_set_nsegments(struct the_nilfs *nilfs, unsigned long nsegs);
 int nilfs_discard_segments(struct the_nilfs *, __u64 *, size_t);
 int nilfs_count_free_blocks(struct the_nilfs *, sector_t *);
 struct nilfs_root *nilfs_lookup_root(struct the_nilfs *nilfs, __u64 cno);

include/linux/nilfs2_fs.h

@@ -107,7 +107,7 @@ struct nilfs_super_root {
 #define NILFS_SR_DAT_OFFSET(inode_size)     NILFS_SR_MDT_OFFSET(inode_size, 0)
 #define NILFS_SR_CPFILE_OFFSET(inode_size)  NILFS_SR_MDT_OFFSET(inode_size, 1)
 #define NILFS_SR_SUFILE_OFFSET(inode_size)  NILFS_SR_MDT_OFFSET(inode_size, 2)
-#define NILFS_SR_BYTES                  (sizeof(struct nilfs_super_root))
+#define NILFS_SR_BYTES(inode_size)	    NILFS_SR_MDT_OFFSET(inode_size, 3)
 
 /*
  * Maximal mount counts
@@ -845,5 +845,7 @@ struct nilfs_bdesc {
 	_IOR(NILFS_IOCTL_IDENT, 0x8A, __u64)
 #define NILFS_IOCTL_RESIZE  \
 	_IOW(NILFS_IOCTL_IDENT, 0x8B, __u64)
+#define NILFS_IOCTL_SET_ALLOC_RANGE  \
+	_IOW(NILFS_IOCTL_IDENT, 0x8C, __u64[2])
 
 #endif	/* _LINUX_NILFS_FS_H */