X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=fs%2Fbtrfs%2Fdisk-io.c;h=e83be2e4602c256f959a2b9daf0c3da2b8e9bcab;hb=2da8ca26c6bfad685bfddf39728eac1c83906aa9;hp=dae25e78a6b7d33abf19572e792f996ea3ecc330;hpb=2d69a0f88459fae35df3ddef4934a2dad67e2765;p=safe%2Fjmp%2Flinux-2.6

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index dae25e7..e83be2e 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -16,19 +16,18 @@
  * Boston, MA 021110-1307, USA.
  */
 
-#include <linux/version.h>
 #include <linux/fs.h>
 #include <linux/blkdev.h>
 #include <linux/scatterlist.h>
 #include <linux/swap.h>
 #include <linux/radix-tree.h>
 #include <linux/writeback.h>
-#include <linux/buffer_head.h> // for block_sync_page
+#include <linux/buffer_head.h>
 #include <linux/workqueue.h>
 #include <linux/kthread.h>
 #include <linux/freezer.h>
+#include <linux/crc32c.h>
 #include "compat.h"
-#include "crc32c.h"
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
@@ -37,25 +36,14 @@
 #include "print-tree.h"
 #include "async-thread.h"
 #include "locking.h"
-#include "ref-cache.h"
 #include "tree-log.h"
-
-#if 0
-static int check_tree_block(struct btrfs_root *root, struct extent_buffer *buf)
-{
-	if (extent_buffer_blocknr(buf) != btrfs_header_blocknr(buf)) {
-		printk(KERN_CRIT "buf blocknr(buf) is %llu, header is %llu\n",
-		       (unsigned long long)extent_buffer_blocknr(buf),
-		       (unsigned long long)btrfs_header_blocknr(buf));
-		return 1;
-	}
-	return 0;
-}
-#endif
+#include "free-space-cache.h"
 
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
 
+static atomic_t btrfs_bdi_num = ATOMIC_INIT(0);
+
 /*
  * end_io_wq structs are used to do processing in task context when an IO is
  * complete.  This is used during reads to verify checksums, and it is used
@@ -89,6 +77,40 @@ struct async_submit_bio {
 	struct btrfs_work work;
 };
 
+/* These are used to set the lockdep class on the extent buffer locks.
+ * The class is set by the readpage_end_io_hook after the buffer has
+ * passed csum validation but before the pages are unlocked.
+ *
+ * The lockdep class is also set by btrfs_init_new_buffer on freshly
+ * allocated blocks.
+ *
+ * The class is based on the level in the tree block, which allows lockdep
+ * to know that lower nodes nest inside the locks of higher nodes.
+ *
+ * We also add a check to make sure the highest level of the tree is
+ * the same as our lockdep setup here.  If BTRFS_MAX_LEVEL changes, this
+ * code needs update as well.
+ */
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+# if BTRFS_MAX_LEVEL != 8
+#  error
+# endif
+static struct lock_class_key btrfs_eb_class[BTRFS_MAX_LEVEL + 1];
+static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = {
+	/* leaf */
+	"btrfs-extent-00",
+	"btrfs-extent-01",
+	"btrfs-extent-02",
+	"btrfs-extent-03",
+	"btrfs-extent-04",
+	"btrfs-extent-05",
+	"btrfs-extent-06",
+	"btrfs-extent-07",
+	/* highest possible level */
+	"btrfs-extent-08",
+};
+#endif
+
 /*
  * extents on the btree inode are pretty simple, there's one extent
  * that covers the entire device
@@ -128,23 +150,13 @@ static struct extent_map *btree_get_extent(struct inode *inode,
 		u64 failed_start = em->start;
 		u64 failed_len = em->len;
 
-		printk("failed to insert %Lu %Lu -> %Lu into tree\n",
-		       em->start, em->len, em->block_start);
 		free_extent_map(em);
 		em = lookup_extent_mapping(em_tree, start, len);
 		if (em) {
-			printk("after failing, found %Lu %Lu %Lu\n",
-			       em->start, em->len, em->block_start);
 			ret = 0;
 		} else {
 			em = lookup_extent_mapping(em_tree, failed_start,
 						   failed_len);
-			if (em) {
-				printk("double failure lookup gives us "
-				       "%Lu %Lu -> %Lu\n", em->start,
-				       em->len, em->block_start);
-				free_extent_map(em);
-			}
 			ret = -EIO;
 		}
 	} else if (ret) {
@@ -161,7 +173,7 @@ out:
 
 u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len)
 {
-	return btrfs_crc32c(seed, data, len);
+	return crc32c(seed, data, len);
 }
 
 void btrfs_csum_final(u32 crc, char *result)
@@ -191,15 +203,12 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 	unsigned long inline_result;
 
 	len = buf->len - offset;
-	while(len > 0) {
+	while (len > 0) {
 		err = map_private_extent_buffer(buf, offset, 32,
 					&map_token, &kaddr,
 					&map_start, &map_len, KM_USER0);
-		if (err) {
-			printk("failed to map extent buffer! %lu\n",
-			       offset);
+		if (err)
 			return 1;
-		}
 		cur_len = min(len, map_len - (offset - map_start));
 		crc = btrfs_csum_data(root, kaddr + offset - map_start,
 				      crc, cur_len);
@@ -218,17 +227,20 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 	btrfs_csum_final(crc, result);
 
 	if (verify) {
-		/* FIXME, this is not good */
 		if (memcmp_extent_buffer(buf, result, 0, csum_size)) {
 			u32 val;
 			u32 found = 0;
 			memcpy(&found, result, csum_size);
 
 			read_extent_buffer(buf, &val, 0, csum_size);
-			printk("btrfs: %s checksum verify failed on %llu "
-			       "wanted %X found %X level %d\n",
-			       root->fs_info->sb->s_id,
-			       buf->start, val, found, btrfs_header_level(buf));
+			if (printk_ratelimit()) {
+				printk(KERN_INFO "btrfs: %s checksum verify "
+				       "failed on %llu wanted %X found %X "
+				       "level %d\n",
+				       root->fs_info->sb->s_id,
+				       (unsigned long long)buf->start, val, found,
+				       btrfs_header_level(buf));
+			}
 			if (result != (char *)&inline_result)
 				kfree(result);
 			return 1;
@@ -261,10 +273,13 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 		ret = 0;
 		goto out;
 	}
-	printk("parent transid verify failed on %llu wanted %llu found %llu\n",
-	       (unsigned long long)eb->start,
-	       (unsigned long long)parent_transid,
-	       (unsigned long long)btrfs_header_generation(eb));
+	if (printk_ratelimit()) {
+		printk("parent transid verify failed on %llu wanted %llu "
+		       "found %llu\n",
+		       (unsigned long long)eb->start,
+		       (unsigned long long)parent_transid,
+		       (unsigned long long)btrfs_header_generation(eb));
+	}
 	ret = 1;
 	clear_extent_buffer_uptodate(io_tree, eb);
 out:
@@ -293,7 +308,7 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
 		if (!ret &&
 		    !verify_parent_transid(io_tree, eb, parent_transid))
 			return ret;
-printk("read extent buffer pages failed with ret %d mirror no %d\n", ret, mirror_num);
+
 		num_copies = btrfs_num_copies(&root->fs_info->mapping_tree,
 					      eb->start, eb->len);
 		if (num_copies == 1)
@@ -307,9 +322,10 @@ printk("read extent buffer pages failed with ret %d mirror no %d\n", ret, mirror
 }
 
 /*
- * checksum a dirty tree block before IO.  This has extra checks to make
- * sure we only fill in the checksum field in the first page of a multi-page block
+ * checksum a dirty tree block before IO.  This has extra checks to make sure
+ * we only fill in the checksum field in the first page of a multi-page block
  */
+
 static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 {
 	struct extent_io_tree *tree;
@@ -327,28 +343,22 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 	if (!page->private)
 		goto out;
 	len = page->private >> 2;
-	if (len == 0) {
-		WARN_ON(1);
-	}
+	WARN_ON(len == 0);
+
 	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
 	ret = btree_read_extent_buffer_pages(root, eb, start + PAGE_CACHE_SIZE,
 					     btrfs_header_generation(eb));
 	BUG_ON(ret);
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
-		printk("warning: eb start incorrect %Lu buffer %Lu len %lu\n",
-		       start, found_start, len);
 		WARN_ON(1);
 		goto err;
 	}
 	if (eb->first_page != page) {
-		printk("bad first page %lu %lu\n", eb->first_page->index,
-		       page->index);
 		WARN_ON(1);
 		goto err;
 	}
 	if (!PageUptodate(page)) {
-		printk("csum not up to date page %lu\n", page->index);
 		WARN_ON(1);
 		goto err;
 	}
@@ -380,6 +390,15 @@ static int check_tree_block_fsid(struct btrfs_root *root,
 	return ret;
 }
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level)
+{
+	lockdep_set_class_and_name(&eb->lock,
+			   &btrfs_eb_class[level],
+			   btrfs_eb_name[level]);
+}
+#endif
+
 static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 			       struct extent_state *state)
 {
@@ -396,34 +415,42 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		goto out;
 	if (!page->private)
 		goto out;
+
 	len = page->private >> 2;
-	if (len == 0) {
-		WARN_ON(1);
-	}
+	WARN_ON(len == 0);
+
 	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
 
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
-		printk("bad tree block start %llu %llu\n",
-		       (unsigned long long)found_start,
-		       (unsigned long long)eb->start);
+		if (printk_ratelimit()) {
+			printk(KERN_INFO "btrfs bad tree block start "
+			       "%llu %llu\n",
+			       (unsigned long long)found_start,
+			       (unsigned long long)eb->start);
+		}
 		ret = -EIO;
 		goto err;
 	}
 	if (eb->first_page != page) {
-		printk("bad first page %lu %lu\n", eb->first_page->index,
-		       page->index);
+		printk(KERN_INFO "btrfs bad first page %lu %lu\n",
+		       eb->first_page->index, page->index);
 		WARN_ON(1);
 		ret = -EIO;
 		goto err;
 	}
 	if (check_tree_block_fsid(root, eb)) {
-		printk("bad fsid on block %Lu\n", eb->start);
+		if (printk_ratelimit()) {
+			printk(KERN_INFO "btrfs bad fsid on block %llu\n",
+			       (unsigned long long)eb->start);
+		}
 		ret = -EIO;
 		goto err;
 	}
 	found_level = btrfs_header_level(eb);
 
+	btrfs_set_buffer_lockdep_class(eb, found_level);
+
 	ret = csum_tree_block(root, eb, 1);
 	if (ret)
 		ret = -EIO;
@@ -565,20 +592,13 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 	async->bio_flags = bio_flags;
 
 	atomic_inc(&fs_info->nr_async_submits);
+
+	if (rw & (1 << BIO_RW_SYNCIO))
+		btrfs_set_work_high_prio(&async->work);
+
 	btrfs_queue_worker(&fs_info->workers, &async->work);
-#if 0
-	int limit = btrfs_async_submit_limit(fs_info);
-	if (atomic_read(&fs_info->nr_async_submits) > limit) {
-		wait_event_timeout(fs_info->async_submit_wait,
-			   (atomic_read(&fs_info->nr_async_submits) < limit),
-			   HZ/10);
 
-		wait_event_timeout(fs_info->async_submit_wait,
-			   (atomic_read(&fs_info->nr_async_bios) < limit),
-			   HZ/10);
-	}
-#endif
-	while(atomic_read(&fs_info->async_submit_draining) &&
+	while (atomic_read(&fs_info->async_submit_draining) &&
 	      atomic_read(&fs_info->nr_async_submits)) {
 		wait_event(fs_info->async_submit_wait,
 			   (atomic_read(&fs_info->nr_async_submits) == 0));
@@ -594,7 +614,7 @@ static int btree_csum_one_bio(struct bio *bio)
 	struct btrfs_root *root;
 
 	WARN_ON(bio->bi_vcnt <= 0);
-	while(bio_index < bio->bi_vcnt) {
+	while (bio_index < bio->bi_vcnt) {
 		root = BTRFS_I(bvec->bv_page->mapping->host)->root;
 		csum_dirty_buffer(root, bvec->bv_page);
 		bio_index++;
@@ -642,6 +662,7 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 		return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio,
 				     mirror_num, 0);
 	}
+
 	/*
 	 * kthread helpers are used to submit writes so that checksumming
 	 * can happen in parallel across all CPUs
@@ -655,14 +676,31 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 static int btree_writepage(struct page *page, struct writeback_control *wbc)
 {
 	struct extent_io_tree *tree;
+	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+	struct extent_buffer *eb;
+	int was_dirty;
+
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	if (!(current->flags & PF_MEMALLOC)) {
+		return extent_write_full_page(tree, page,
+					      btree_get_extent, wbc);
+	}
 
-	if (current->flags & PF_MEMALLOC) {
-		redirty_page_for_writepage(wbc, page);
-		unlock_page(page);
-		return 0;
+	redirty_page_for_writepage(wbc, page);
+	eb = btrfs_find_tree_block(root, page_offset(page),
+				      PAGE_CACHE_SIZE);
+	WARN_ON(!eb);
+
+	was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
+	if (!was_dirty) {
+		spin_lock(&root->fs_info->delalloc_lock);
+		root->fs_info->dirty_metadata_bytes += PAGE_CACHE_SIZE;
+		spin_unlock(&root->fs_info->delalloc_lock);
 	}
-	return extent_write_full_page(tree, page, btree_get_extent, wbc);
+	free_extent_buffer(eb);
+
+	unlock_page(page);
+	return 0;
 }
 
 static int btree_writepages(struct address_space *mapping,
@@ -671,18 +709,17 @@ static int btree_writepages(struct address_space *mapping,
 	struct extent_io_tree *tree;
 	tree = &BTRFS_I(mapping->host)->io_tree;
 	if (wbc->sync_mode == WB_SYNC_NONE) {
+		struct btrfs_root *root = BTRFS_I(mapping->host)->root;
 		u64 num_dirty;
-		u64 start = 0;
 		unsigned long thresh = 32 * 1024 * 1024;
 
 		if (wbc->for_kupdate)
 			return 0;
 
-		num_dirty = count_range_bits(tree, &start, (u64)-1,
-					     thresh, EXTENT_DIRTY);
-		if (num_dirty < thresh) {
+		/* this is a bit racy, but that's ok */
+		num_dirty = root->fs_info->dirty_metadata_bytes;
+		if (num_dirty < thresh)
 			return 0;
-		}
 	}
 	return extent_writepages(tree, mapping, btree_get_extent, wbc);
 }
@@ -701,15 +738,14 @@ static int btree_releasepage(struct page *page, gfp_t gfp_flags)
 	int ret;
 
 	if (PageWriteback(page) || PageDirty(page))
-	    return 0;
+		return 0;
 
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
 	map = &BTRFS_I(page->mapping->host)->extent_tree;
 
 	ret = try_release_extent_state(map, tree, page, gfp_flags);
-	if (!ret) {
+	if (!ret)
 		return 0;
-	}
 
 	ret = try_release_extent_buffer(tree, page);
 	if (ret == 1) {
@@ -728,35 +764,14 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	extent_invalidatepage(tree, page, offset);
 	btree_releasepage(page, GFP_NOFS);
 	if (PagePrivate(page)) {
-		printk("warning page private not zero on page %Lu\n",
-		       page_offset(page));
+		printk(KERN_WARNING "btrfs warning page private not zero "
+		       "on page %llu\n", (unsigned long long)page_offset(page));
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
 		page_cache_release(page);
 	}
 }
 
-#if 0
-static int btree_writepage(struct page *page, struct writeback_control *wbc)
-{
-	struct buffer_head *bh;
-	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
-	struct buffer_head *head;
-	if (!page_has_buffers(page)) {
-		create_empty_buffers(page, root->fs_info->sb->s_blocksize,
-					(1 << BH_Dirty)|(1 << BH_Uptodate));
-	}
-	head = page_buffers(page);
-	bh = head;
-	do {
-		if (buffer_dirty(bh))
-			csum_tree_block(root, bh, 0);
-		bh = bh->b_this_page;
-	} while (bh != head);
-	return block_write_full_page(page, btree_get_block, wbc);
-}
-#endif
-
 static struct address_space_operations btree_aops = {
 	.readpage	= btree_readpage,
 	.writepage	= btree_writepage,
@@ -813,7 +828,7 @@ int btrfs_write_tree_block(struct extent_buffer *buf)
 int btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
 {
 	return btrfs_wait_on_page_writeback_range(buf->first_page->mapping,
-				  buf->start, buf->start + buf->len -1);
+				  buf->start, buf->start + buf->len - 1);
 }
 
 struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
@@ -832,11 +847,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
 
-	if (ret == 0) {
-		buf->flags |= EXTENT_UPTODATE;
-	} else {
-		WARN_ON(1);
-	}
+	if (ret == 0)
+		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
 	return buf;
 
 }
@@ -847,7 +859,19 @@ int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	if (btrfs_header_generation(buf) ==
 	    root->fs_info->running_transaction->transid) {
-		WARN_ON(!btrfs_tree_locked(buf));
+		btrfs_assert_tree_locked(buf);
+
+		if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)) {
+			spin_lock(&root->fs_info->delalloc_lock);
+			if (root->fs_info->dirty_metadata_bytes >= buf->len)
+				root->fs_info->dirty_metadata_bytes -= buf->len;
+			else
+				WARN_ON(1);
+			spin_unlock(&root->fs_info->delalloc_lock);
+		}
+
+		/* ugh, clear_extent_buffer_dirty needs to lock the page */
+		btrfs_set_lock_blocking(buf);
 		clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
 					  buf);
 	}
@@ -861,7 +885,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 {
 	root->node = NULL;
 	root->commit_root = NULL;
-	root->ref_tree = NULL;
 	root->sectorsize = sectorsize;
 	root->nodesize = nodesize;
 	root->leafsize = leafsize;
@@ -876,20 +899,27 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->last_inode_alloc = 0;
 	root->name = NULL;
 	root->in_sysfs = 0;
+	root->inode_tree.rb_node = NULL;
 
 	INIT_LIST_HEAD(&root->dirty_list);
 	INIT_LIST_HEAD(&root->orphan_list);
-	INIT_LIST_HEAD(&root->dead_list);
+	INIT_LIST_HEAD(&root->root_list);
 	spin_lock_init(&root->node_lock);
 	spin_lock_init(&root->list_lock);
+	spin_lock_init(&root->inode_lock);
 	mutex_init(&root->objectid_mutex);
 	mutex_init(&root->log_mutex);
+	init_waitqueue_head(&root->log_writer_wait);
+	init_waitqueue_head(&root->log_commit_wait[0]);
+	init_waitqueue_head(&root->log_commit_wait[1]);
+	atomic_set(&root->log_commit[0], 0);
+	atomic_set(&root->log_commit[1], 0);
+	atomic_set(&root->log_writers, 0);
+	root->log_batch = 0;
+	root->log_transid = 0;
 	extent_io_tree_init(&root->dirty_log_pages,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
 
-	btrfs_leaf_ref_tree_init(&root->ref_tree_struct);
-	root->ref_tree = &root->ref_tree_struct;
-
 	memset(&root->root_key, 0, sizeof(root->root_key));
 	memset(&root->root_item, 0, sizeof(root->root_item));
 	memset(&root->defrag_progress, 0, sizeof(root->defrag_progress));
@@ -928,6 +958,7 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
 	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     blocksize, generation);
+	root->commit_root = btrfs_root_node(root);
 	BUG_ON(!root->node);
 	return 0;
 }
@@ -944,7 +975,7 @@ int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
 	if (!log_root_tree)
 		return 0;
 
-	while(1) {
+	while (1) {
 		ret = find_first_extent_bit(&log_root_tree->dirty_log_pages,
 				    0, &start, &end, EXTENT_DIRTY);
 		if (ret)
@@ -968,15 +999,16 @@ int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
-int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
-			     struct btrfs_fs_info *fs_info)
+static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
+					 struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root;
 	struct btrfs_root *tree_root = fs_info->tree_root;
+	struct extent_buffer *leaf;
 
 	root = kzalloc(sizeof(*root), GFP_NOFS);
 	if (!root)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	__setup_root(tree_root->nodesize, tree_root->leafsize,
 		     tree_root->sectorsize, tree_root->stripesize,
@@ -985,24 +1017,74 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
 	root->root_key.objectid = BTRFS_TREE_LOG_OBJECTID;
 	root->root_key.type = BTRFS_ROOT_ITEM_KEY;
 	root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
+	/*
+	 * log trees do not get reference counted because they go away
+	 * before a real commit is actually done.  They do store pointers
+	 * to file data extents, and those reference counts still get
+	 * updated (along with back refs to the log tree).
+	 */
 	root->ref_cows = 0;
 
-	root->node = btrfs_alloc_free_block(trans, root, root->leafsize,
-					    0, BTRFS_TREE_LOG_OBJECTID,
-					    trans->transid, 0, 0, 0);
+	leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
+				      BTRFS_TREE_LOG_OBJECTID, NULL, 0, 0, 0);
+	if (IS_ERR(leaf)) {
+		kfree(root);
+		return ERR_CAST(leaf);
+	}
 
-	btrfs_set_header_nritems(root->node, 0);
-	btrfs_set_header_level(root->node, 0);
-	btrfs_set_header_bytenr(root->node, root->node->start);
-	btrfs_set_header_generation(root->node, trans->transid);
-	btrfs_set_header_owner(root->node, BTRFS_TREE_LOG_OBJECTID);
+	memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
+	btrfs_set_header_bytenr(leaf, leaf->start);
+	btrfs_set_header_generation(leaf, trans->transid);
+	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
+	btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
+	root->node = leaf;
 
 	write_extent_buffer(root->node, root->fs_info->fsid,
 			    (unsigned long)btrfs_header_fsid(root->node),
 			    BTRFS_FSID_SIZE);
 	btrfs_mark_buffer_dirty(root->node);
 	btrfs_tree_unlock(root->node);
-	fs_info->log_root_tree = root;
+	return root;
+}
+
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *log_root;
+
+	log_root = alloc_log_tree(trans, fs_info);
+	if (IS_ERR(log_root))
+		return PTR_ERR(log_root);
+	WARN_ON(fs_info->log_root_tree);
+	fs_info->log_root_tree = log_root;
+	return 0;
+}
+
+int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root)
+{
+	struct btrfs_root *log_root;
+	struct btrfs_inode_item *inode_item;
+
+	log_root = alloc_log_tree(trans, root->fs_info);
+	if (IS_ERR(log_root))
+		return PTR_ERR(log_root);
+
+	log_root->last_trans = trans->transid;
+	log_root->root_key.offset = root->root_key.objectid;
+
+	inode_item = &log_root->root_item.inode;
+	inode_item->generation = cpu_to_le64(1);
+	inode_item->size = cpu_to_le64(3);
+	inode_item->nlink = cpu_to_le32(1);
+	inode_item->nbytes = cpu_to_le64(root->leafsize);
+	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+
+	btrfs_set_root_node(&log_root->root_item, log_root->node);
+
+	WARN_ON(root->log_root);
+	root->log_root = log_root;
+	root->log_transid = 0;
 	return 0;
 }
 
@@ -1060,6 +1142,7 @@ out:
 	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     blocksize, generation);
+	root->commit_root = btrfs_root_node(root);
 	BUG_ON(!root->node);
 insert:
 	if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
@@ -1126,7 +1209,7 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 	}
 	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
 		ret = btrfs_find_dead_roots(fs_info->tree_root,
-					    root->root_key.objectid, root);
+					    root->root_key.objectid);
 		BUG_ON(ret);
 		btrfs_orphan_cleanup(root);
 	}
@@ -1165,39 +1248,15 @@ struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
 	root->in_sysfs = 1;
 	return root;
 }
-#if 0
-static int add_hasher(struct btrfs_fs_info *info, char *type) {
-	struct btrfs_hasher *hasher;
-
-	hasher = kmalloc(sizeof(*hasher), GFP_NOFS);
-	if (!hasher)
-		return -ENOMEM;
-	hasher->hash_tfm = crypto_alloc_hash(type, 0, CRYPTO_ALG_ASYNC);
-	if (!hasher->hash_tfm) {
-		kfree(hasher);
-		return -EINVAL;
-	}
-	spin_lock(&info->hash_lock);
-	list_add(&hasher->list, &info->hashers);
-	spin_unlock(&info->hash_lock);
-	return 0;
-}
-#endif
 
 static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 {
 	struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
 	int ret = 0;
-	struct list_head *cur;
 	struct btrfs_device *device;
 	struct backing_dev_info *bdi;
-#if 0
-	if ((bdi_bits & (1 << BDI_write_congested)) &&
-	    btrfs_congested_async(info, 0))
-		return 1;
-#endif
-	list_for_each(cur, &info->fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+
+	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
 		bdi = blk_get_backing_dev_info(device->bdev);
@@ -1215,20 +1274,17 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
  */
 static void __unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
 {
-	struct list_head *cur;
 	struct btrfs_device *device;
 	struct btrfs_fs_info *info;
 
 	info = (struct btrfs_fs_info *)bdi->unplug_io_data;
-	list_for_each(cur, &info->fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
 
 		bdi = blk_get_backing_dev_info(device->bdev);
-		if (bdi->unplug_io_fn) {
+		if (bdi->unplug_io_fn)
 			bdi->unplug_io_fn(bdi, page);
-		}
 	}
 }
 
@@ -1288,12 +1344,25 @@ static void btrfs_unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
 	free_extent_map(em);
 }
 
+/*
+ * If this fails, caller must call bdi_destroy() to get rid of the
+ * bdi again.
+ */
 static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
 {
-	bdi_init(bdi);
+	int err;
+
+	bdi->capabilities = BDI_CAP_MAP_COPY;
+	err = bdi_init(bdi);
+	if (err)
+		return err;
+
+	err = bdi_register(bdi, NULL, "btrfs-%d",
+				atomic_inc_return(&btrfs_bdi_num));
+	if (err)
+		return err;
+
 	bdi->ra_pages	= default_backing_dev_info.ra_pages;
-	bdi->state		= 0;
-	bdi->capabilities	= default_backing_dev_info.capabilities;
 	bdi->unplug_io_fn	= btrfs_unplug_io_fn;
 	bdi->unplug_io_data	= info;
 	bdi->congested_fn	= btrfs_congested_fn;
@@ -1335,8 +1404,6 @@ static int bio_ready_for_csum(struct bio *bio)
 
 	ret = extent_range_uptodate(io_tree, start + length,
 				    start + buf_len - 1);
-	if (ret == 1)
-		return ret;
 	return ret;
 }
 
@@ -1419,11 +1486,6 @@ static int transaction_kthread(void *arg)
 		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
 		mutex_lock(&root->fs_info->transaction_kthread_mutex);
 
-		if (root->fs_info->total_ref_cache_size > 20 * 1024 * 1024) {
-			printk("btrfs: total reference cache size %Lu\n",
-				root->fs_info->total_ref_cache_size);
-		}
-
 		mutex_lock(&root->fs_info->trans_mutex);
 		cur = root->fs_info->running_transaction;
 		if (!cur) {
@@ -1440,6 +1502,7 @@ static int transaction_kthread(void *arg)
 		mutex_unlock(&root->fs_info->trans_mutex);
 		trans = btrfs_start_transaction(root, 1);
 		ret = btrfs_commit_transaction(trans, root);
+
 sleep:
 		wake_up_process(root->fs_info->cleaner_kthread);
 		mutex_unlock(&root->fs_info->transaction_kthread_mutex);
@@ -1499,7 +1562,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->dead_roots);
 	INIT_LIST_HEAD(&fs_info->hashers);
 	INIT_LIST_HEAD(&fs_info->delalloc_inodes);
-	spin_lock_init(&fs_info->hash_lock);
+	INIT_LIST_HEAD(&fs_info->ordered_operations);
 	spin_lock_init(&fs_info->delalloc_lock);
 	spin_lock_init(&fs_info->new_trans_lock);
 	spin_lock_init(&fs_info->ref_cache_lock);
@@ -1518,15 +1581,15 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	atomic_set(&fs_info->async_delalloc_pages, 0);
 	atomic_set(&fs_info->async_submit_draining, 0);
 	atomic_set(&fs_info->nr_async_bios, 0);
-	atomic_set(&fs_info->throttles, 0);
-	atomic_set(&fs_info->throttle_gen, 0);
 	fs_info->sb = sb;
 	fs_info->max_extent = (u64)-1;
 	fs_info->max_inline = 8192 * 1024;
-	setup_bdi(fs_info, &fs_info->bdi);
+	if (setup_bdi(fs_info, &fs_info->bdi))
+		goto fail_bdi;
 	fs_info->btree_inode = new_inode(sb);
 	fs_info->btree_inode->i_ino = 1;
 	fs_info->btree_inode->i_nlink = 1;
+	fs_info->metadata_ratio = 8;
 
 	fs_info->thread_pool_size = min_t(unsigned long,
 					  num_online_cpus() + 2, 8);
@@ -1546,6 +1609,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->btree_inode->i_mapping->a_ops = &btree_aops;
 	fs_info->btree_inode->i_mapping->backing_dev_info = &fs_info->bdi;
 
+	RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node);
 	extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
 			     fs_info->btree_inode->i_mapping,
 			     GFP_NOFS);
@@ -1559,47 +1623,31 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	extent_io_tree_init(&fs_info->pinned_extents,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_io_tree_init(&fs_info->pending_del,
-			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_io_tree_init(&fs_info->extent_ins,
-			     fs_info->btree_inode->i_mapping, GFP_NOFS);
 	fs_info->do_barriers = 1;
 
-	INIT_LIST_HEAD(&fs_info->dead_reloc_roots);
-	btrfs_leaf_ref_tree_init(&fs_info->reloc_ref_tree);
-	btrfs_leaf_ref_tree_init(&fs_info->shared_ref_tree);
-
 	BTRFS_I(fs_info->btree_inode)->root = tree_root;
 	memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
 	       sizeof(struct btrfs_key));
 	insert_inode_hash(fs_info->btree_inode);
 
 	mutex_init(&fs_info->trans_mutex);
+	mutex_init(&fs_info->ordered_operations_mutex);
 	mutex_init(&fs_info->tree_log_mutex);
 	mutex_init(&fs_info->drop_mutex);
-	mutex_init(&fs_info->extent_ins_mutex);
-	mutex_init(&fs_info->pinned_mutex);
 	mutex_init(&fs_info->chunk_mutex);
 	mutex_init(&fs_info->transaction_kthread_mutex);
 	mutex_init(&fs_info->cleaner_mutex);
 	mutex_init(&fs_info->volume_mutex);
 	mutex_init(&fs_info->tree_reloc_mutex);
+	init_rwsem(&fs_info->extent_commit_sem);
+
+	btrfs_init_free_cluster(&fs_info->meta_alloc_cluster);
+	btrfs_init_free_cluster(&fs_info->data_alloc_cluster);
+
 	init_waitqueue_head(&fs_info->transaction_throttle);
 	init_waitqueue_head(&fs_info->transaction_wait);
 	init_waitqueue_head(&fs_info->async_submit_wait);
-	init_waitqueue_head(&fs_info->tree_log_wait);
-	atomic_set(&fs_info->tree_log_commit, 0);
-	atomic_set(&fs_info->tree_log_writers, 0);
-	fs_info->tree_log_transid = 0;
 
-#if 0
-	ret = add_hasher(fs_info, "crc32c");
-	if (ret) {
-		printk("btrfs: failed hash setup, modprobe cryptomgr?\n");
-		err = -ENOMEM;
-		goto fail_iput;
-	}
-#endif
 	__setup_root(4096, 4096, 4096, 4096, tree_root,
 		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
 
@@ -1630,17 +1678,23 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	if (features) {
 		printk(KERN_ERR "BTRFS: couldn't mount because of "
 		       "unsupported optional features (%Lx).\n",
-		       features);
+		       (unsigned long long)features);
 		err = -EINVAL;
 		goto fail_iput;
 	}
 
+	features = btrfs_super_incompat_flags(disk_super);
+	if (!(features & BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF)) {
+		features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF;
+		btrfs_set_super_incompat_flags(disk_super, features);
+	}
+
 	features = btrfs_super_compat_ro_flags(disk_super) &
 		~BTRFS_FEATURE_COMPAT_RO_SUPP;
 	if (!(sb->s_flags & MS_RDONLY) && features) {
 		printk(KERN_ERR "BTRFS: couldn't mount RDWR because of "
 		       "unsupported option features (%Lx).\n",
-		       features);
+		       (unsigned long long)features);
 		err = -EINVAL;
 		goto fail_iput;
 	}
@@ -1687,6 +1741,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	 * low idle thresh
 	 */
 	fs_info->endio_workers.idle_thresh = 4;
+	fs_info->endio_meta_workers.idle_thresh = 4;
+
 	fs_info->endio_write_workers.idle_thresh = 64;
 	fs_info->endio_meta_write_workers.idle_thresh = 64;
 
@@ -1720,7 +1776,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
 		    sizeof(disk_super->magic))) {
-		printk("btrfs: valid FS not found on %s\n", sb->s_id);
+		printk(KERN_INFO "btrfs: valid FS not found on %s\n", sb->s_id);
 		goto fail_sb_buffer;
 	}
 
@@ -1728,9 +1784,9 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	ret = btrfs_read_sys_array(tree_root);
 	mutex_unlock(&fs_info->chunk_mutex);
 	if (ret) {
-		printk("btrfs: failed to read the system array on %s\n",
-		       sb->s_id);
-		goto fail_sys_array;
+		printk(KERN_WARNING "btrfs: failed to read the system "
+		       "array on %s\n", sb->s_id);
+		goto fail_sb_buffer;
 	}
 
 	blocksize = btrfs_level_size(tree_root,
@@ -1744,16 +1800,24 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 					   btrfs_super_chunk_root(disk_super),
 					   blocksize, generation);
 	BUG_ON(!chunk_root->node);
+	if (!test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) {
+		printk(KERN_WARNING "btrfs: failed to read chunk root on %s\n",
+		       sb->s_id);
+		goto fail_chunk_root;
+	}
+	btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
+	chunk_root->commit_root = btrfs_root_node(chunk_root);
 
 	read_extent_buffer(chunk_root->node, fs_info->chunk_tree_uuid,
-	         (unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node),
-		 BTRFS_UUID_SIZE);
+	   (unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node),
+	   BTRFS_UUID_SIZE);
 
 	mutex_lock(&fs_info->chunk_mutex);
 	ret = btrfs_read_chunk_tree(chunk_root);
 	mutex_unlock(&fs_info->chunk_mutex);
 	if (ret) {
-		printk("btrfs: failed to read chunk tree on %s\n", sb->s_id);
+		printk(KERN_WARNING "btrfs: failed to read chunk tree on %s\n",
+		       sb->s_id);
 		goto fail_chunk_root;
 	}
 
@@ -1768,7 +1832,13 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 					  blocksize, generation);
 	if (!tree_root->node)
 		goto fail_chunk_root;
-
+	if (!test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) {
+		printk(KERN_WARNING "btrfs: failed to read tree root on %s\n",
+		       sb->s_id);
+		goto fail_tree_root;
+	}
+	btrfs_set_root_node(&tree_root->root_item, tree_root->node);
+	tree_root->commit_root = btrfs_root_node(tree_root);
 
 	ret = find_and_setup_root(tree_root, fs_info,
 				  BTRFS_EXTENT_TREE_OBJECTID, extent_root);
@@ -1778,15 +1848,14 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	ret = find_and_setup_root(tree_root, fs_info,
 				  BTRFS_DEV_TREE_OBJECTID, dev_root);
-	dev_root->track_dirty = 1;
-
 	if (ret)
 		goto fail_extent_root;
+	dev_root->track_dirty = 1;
 
 	ret = find_and_setup_root(tree_root, fs_info,
 				  BTRFS_CSUM_TREE_OBJECTID, csum_root);
 	if (ret)
-		goto fail_extent_root;
+		goto fail_dev_root;
 
 	csum_root->track_dirty = 1;
 
@@ -1799,20 +1868,29 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->system_alloc_profile = fs_info->metadata_alloc_profile;
 	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
 					       "btrfs-cleaner");
-	if (!fs_info->cleaner_kthread)
+	if (IS_ERR(fs_info->cleaner_kthread))
 		goto fail_csum_root;
 
 	fs_info->transaction_kthread = kthread_run(transaction_kthread,
 						   tree_root,
 						   "btrfs-transaction");
-	if (!fs_info->transaction_kthread)
+	if (IS_ERR(fs_info->transaction_kthread))
 		goto fail_cleaner;
 
+	if (!btrfs_test_opt(tree_root, SSD) &&
+	    !btrfs_test_opt(tree_root, NOSSD) &&
+	    !fs_info->fs_devices->rotating) {
+		printk(KERN_INFO "Btrfs detected SSD devices, enabling SSD "
+		       "mode\n");
+		btrfs_set_opt(fs_info->mount_opt, SSD);
+	}
+
 	if (btrfs_super_log_root(disk_super) != 0) {
 		u64 bytenr = btrfs_super_log_root(disk_super);
 
 		if (fs_devices->rw_devices == 0) {
-			printk("Btrfs log replay required on RO media\n");
+			printk(KERN_WARNING "Btrfs log replay required "
+			       "on RO media\n");
 			err = -EIO;
 			goto fail_trans_kthread;
 		}
@@ -1839,7 +1917,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	}
 
 	if (!(sb->s_flags & MS_RDONLY)) {
-		ret = btrfs_cleanup_reloc_trees(tree_root);
+		ret = btrfs_recover_relocation(tree_root);
 		BUG_ON(ret);
 	}
 
@@ -1850,6 +1928,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->fs_root = btrfs_read_fs_root_no_name(fs_info, &location);
 	if (!fs_info->fs_root)
 		goto fail_trans_kthread;
+
 	return tree_root;
 
 fail_trans_kthread:
@@ -1866,14 +1945,19 @@ fail_cleaner:
 
 fail_csum_root:
 	free_extent_buffer(csum_root->node);
+	free_extent_buffer(csum_root->commit_root);
+fail_dev_root:
+	free_extent_buffer(dev_root->node);
+	free_extent_buffer(dev_root->commit_root);
 fail_extent_root:
 	free_extent_buffer(extent_root->node);
+	free_extent_buffer(extent_root->commit_root);
 fail_tree_root:
 	free_extent_buffer(tree_root->node);
+	free_extent_buffer(tree_root->commit_root);
 fail_chunk_root:
 	free_extent_buffer(chunk_root->node);
-fail_sys_array:
-	free_extent_buffer(dev_root->node);
+	free_extent_buffer(chunk_root->commit_root);
 fail_sb_buffer:
 	btrfs_stop_workers(&fs_info->fixup_workers);
 	btrfs_stop_workers(&fs_info->delalloc_workers);
@@ -1886,13 +1970,14 @@ fail_sb_buffer:
 fail_iput:
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	iput(fs_info->btree_inode);
-fail:
+
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
-
+fail_bdi:
+	bdi_destroy(&fs_info->bdi);
+fail:
 	kfree(extent_root);
 	kfree(tree_root);
-	bdi_destroy(&fs_info->bdi);
 	kfree(fs_info);
 	kfree(chunk_root);
 	kfree(dev_root);
@@ -1962,6 +2047,17 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 	return latest;
 }
 
+/*
+ * this should be called twice, once with wait == 0 and
+ * once with wait == 1.  When wait == 0 is done, all the buffer heads
+ * we write are pinned.
+ *
+ * They are released when wait == 1 is done.
+ * max_mirrors must be the same for both runs, and it indicates how
+ * many supers on this one device should be written.
+ *
+ * max_mirrors == 0 means to write them all.
+ */
 static int write_dev_supers(struct btrfs_device *device,
 			    struct btrfs_super_block *sb,
 			    int do_barriers, int wait, int max_mirrors)
@@ -1997,12 +2093,16 @@ static int write_dev_supers(struct btrfs_device *device,
 			bh = __find_get_block(device->bdev, bytenr / 4096,
 					      BTRFS_SUPER_INFO_SIZE);
 			BUG_ON(!bh);
-			brelse(bh);
 			wait_on_buffer(bh);
-			if (buffer_uptodate(bh)) {
-				brelse(bh);
-				continue;
-			}
+			if (!buffer_uptodate(bh))
+				errors++;
+
+			/* drop our reference */
+			brelse(bh);
+
+			/* drop the reference from the wait == 0 run */
+			brelse(bh);
+			continue;
 		} else {
 			btrfs_set_super_bytenr(sb, bytenr);
 
@@ -2013,12 +2113,18 @@ static int write_dev_supers(struct btrfs_device *device,
 					      BTRFS_CSUM_SIZE);
 			btrfs_csum_final(crc, sb->csum);
 
+			/*
+			 * one reference for us, and we leave it for the
+			 * caller
+			 */
 			bh = __getblk(device->bdev, bytenr / 4096,
 				      BTRFS_SUPER_INFO_SIZE);
 			memcpy(bh->b_data, sb, BTRFS_SUPER_INFO_SIZE);
 
-			set_buffer_uptodate(bh);
+			/* one reference for submit_bh */
 			get_bh(bh);
+
+			set_buffer_uptodate(bh);
 			lock_buffer(bh);
 			bh->b_end_io = btrfs_end_buffer_write_sync;
 		}
@@ -2030,31 +2136,24 @@ static int write_dev_supers(struct btrfs_device *device,
 				       device->name);
 				set_buffer_uptodate(bh);
 				device->barriers = 0;
+				/* one reference for submit_bh */
 				get_bh(bh);
 				lock_buffer(bh);
-				ret = submit_bh(WRITE, bh);
+				ret = submit_bh(WRITE_SYNC, bh);
 			}
 		} else {
-			ret = submit_bh(WRITE, bh);
+			ret = submit_bh(WRITE_SYNC, bh);
 		}
 
-		if (!ret && wait) {
-			wait_on_buffer(bh);
-			if (!buffer_uptodate(bh))
-				errors++;
-		} else if (ret) {
+		if (ret)
 			errors++;
-		}
-		if (wait)
-			brelse(bh);
 	}
 	return errors < i ? 0 : -1;
 }
 
 int write_all_supers(struct btrfs_root *root, int max_mirrors)
 {
-	struct list_head *cur;
-	struct list_head *head = &root->fs_info->fs_devices->devices;
+	struct list_head *head;
 	struct btrfs_device *dev;
 	struct btrfs_super_block *sb;
 	struct btrfs_dev_item *dev_item;
@@ -2069,8 +2168,10 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 
 	sb = &root->fs_info->super_for_commit;
 	dev_item = &sb->dev_item;
-	list_for_each(cur, head) {
-		dev = list_entry(cur, struct btrfs_device, dev_list);
+
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	head = &root->fs_info->fs_devices->devices;
+	list_for_each_entry(dev, head, dev_list) {
 		if (!dev->bdev) {
 			total_errors++;
 			continue;
@@ -2097,13 +2198,13 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 			total_errors++;
 	}
 	if (total_errors > max_errors) {
-		printk("btrfs: %d errors while writing supers\n", total_errors);
+		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
+		       total_errors);
 		BUG();
 	}
 
 	total_errors = 0;
-	list_for_each(cur, head) {
-		dev = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(dev, head, dev_list) {
 		if (!dev->bdev)
 			continue;
 		if (!dev->in_fs_metadata || !dev->writeable)
@@ -2113,8 +2214,10 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 		if (ret)
 			total_errors++;
 	}
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 	if (total_errors > max_errors) {
-		printk("btrfs: %d errors while writing supers\n", total_errors);
+		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
+		       total_errors);
 		BUG();
 	}
 	return 0;
@@ -2131,22 +2234,18 @@ int write_ctree_super(struct btrfs_trans_handle *trans,
 
 int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 {
+	WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
 	radix_tree_delete(&fs_info->fs_roots_radix,
 			  (unsigned long)root->root_key.objectid);
 	if (root->anon_super.s_dev) {
 		down_write(&root->anon_super.s_umount);
 		kill_anon_super(&root->anon_super);
 	}
-#if 0
-	if (root->in_sysfs)
-		btrfs_sysfs_del_root(root);
-#endif
 	if (root->node)
 		free_extent_buffer(root->node);
 	if (root->commit_root)
 		free_extent_buffer(root->commit_root);
-	if (root->name)
-		kfree(root->name);
+	kfree(root->name);
 	kfree(root);
 	return 0;
 }
@@ -2157,7 +2256,7 @@ static int del_fs_roots(struct btrfs_fs_info *fs_info)
 	struct btrfs_root *gang[8];
 	int i;
 
-	while(1) {
+	while (1) {
 		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
 					     (void **)gang, 0,
 					     ARRAY_SIZE(gang));
@@ -2182,10 +2281,12 @@ int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
 					     ARRAY_SIZE(gang));
 		if (!ret)
 			break;
+
+		root_objectid = gang[ret - 1]->root_key.objectid + 1;
 		for (i = 0; i < ret; i++) {
 			root_objectid = gang[i]->root_key.objectid;
 			ret = btrfs_find_dead_roots(fs_info->tree_root,
-						    root_objectid, gang[i]);
+						    root_objectid);
 			BUG_ON(ret);
 			btrfs_orphan_cleanup(gang[i]);
 		}
@@ -2228,36 +2329,35 @@ int close_ctree(struct btrfs_root *root)
 
 	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
 		ret =  btrfs_commit_super(root);
-		if (ret) {
-			printk("btrfs: commit super returns %d\n", ret);
-		}
+		if (ret)
+			printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
 	}
 
+	fs_info->closing = 2;
+	smp_mb();
+
 	if (fs_info->delalloc_bytes) {
-		printk("btrfs: at unmount delalloc count %Lu\n",
-		       fs_info->delalloc_bytes);
+		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
+		       (unsigned long long)fs_info->delalloc_bytes);
 	}
 	if (fs_info->total_ref_cache_size) {
-		printk("btrfs: at umount reference cache size %Lu\n",
-			fs_info->total_ref_cache_size);
+		printk(KERN_INFO "btrfs: at umount reference cache size %llu\n",
+		       (unsigned long long)fs_info->total_ref_cache_size);
 	}
 
-	if (fs_info->extent_root->node)
-		free_extent_buffer(fs_info->extent_root->node);
-
-	if (fs_info->tree_root->node)
-		free_extent_buffer(fs_info->tree_root->node);
-
-	if (root->fs_info->chunk_root->node);
-		free_extent_buffer(root->fs_info->chunk_root->node);
-
-	if (root->fs_info->dev_root->node);
-		free_extent_buffer(root->fs_info->dev_root->node);
-
-	if (root->fs_info->csum_root->node);
-		free_extent_buffer(root->fs_info->csum_root->node);
+	free_extent_buffer(fs_info->extent_root->node);
+	free_extent_buffer(fs_info->extent_root->commit_root);
+	free_extent_buffer(fs_info->tree_root->node);
+	free_extent_buffer(fs_info->tree_root->commit_root);
+	free_extent_buffer(root->fs_info->chunk_root->node);
+	free_extent_buffer(root->fs_info->chunk_root->commit_root);
+	free_extent_buffer(root->fs_info->dev_root->node);
+	free_extent_buffer(root->fs_info->dev_root->commit_root);
+	free_extent_buffer(root->fs_info->csum_root->node);
+	free_extent_buffer(root->fs_info->csum_root->commit_root);
 
 	btrfs_free_block_groups(root->fs_info);
+	btrfs_free_pinned_extents(root->fs_info);
 
 	del_fs_roots(fs_info);
 
@@ -2272,16 +2372,6 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->submit_workers);
 
-#if 0
-	while(!list_empty(&fs_info->hashers)) {
-		struct btrfs_hasher *hasher;
-		hasher = list_entry(fs_info->hashers.next, struct btrfs_hasher,
-				    hashers);
-		list_del(&hasher->hashers);
-		crypto_free_hash(&fs_info->hash_tfm);
-		kfree(hasher);
-	}
-#endif
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
 
@@ -2321,15 +2411,24 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	u64 transid = btrfs_header_generation(buf);
 	struct inode *btree_inode = root->fs_info->btree_inode;
+	int was_dirty;
 
-	WARN_ON(!btrfs_tree_locked(buf));
+	btrfs_assert_tree_locked(buf);
 	if (transid != root->fs_info->generation) {
-		printk(KERN_CRIT "transid mismatch buffer %llu, found %Lu running %Lu\n",
+		printk(KERN_CRIT "btrfs transid mismatch buffer %llu, "
+		       "found %llu running %llu\n",
 			(unsigned long long)buf->start,
-			transid, root->fs_info->generation);
+			(unsigned long long)transid,
+			(unsigned long long)root->fs_info->generation);
 		WARN_ON(1);
 	}
-	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, buf);
+	was_dirty = set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
+					    buf);
+	if (!was_dirty) {
+		spin_lock(&root->fs_info->delalloc_lock);
+		root->fs_info->dirty_metadata_bytes += buf->len;
+		spin_unlock(&root->fs_info->delalloc_lock);
+	}
 }
 
 void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
@@ -2338,17 +2437,14 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
 	 * looks as though older kernels can get into trouble with
 	 * this code, they end up stuck in balance_dirty_pages forever
 	 */
-	struct extent_io_tree *tree;
 	u64 num_dirty;
-	u64 start = 0;
 	unsigned long thresh = 32 * 1024 * 1024;
-	tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
 
-	if (current_is_pdflush() || current->flags & PF_MEMALLOC)
+	if (current->flags & PF_MEMALLOC)
 		return;
 
-	num_dirty = count_range_bits(tree, &start, (u64)-1,
-				     thresh, EXTENT_DIRTY);
+	num_dirty = root->fs_info->dirty_metadata_bytes;
+
 	if (num_dirty > thresh) {
 		balance_dirty_pages_ratelimited_nr(
 				   root->fs_info->btree_inode->i_mapping, 1);
@@ -2361,9 +2457,8 @@ int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	int ret;
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
-	if (ret == 0) {
-		buf->flags |= EXTENT_UPTODATE;
-	}
+	if (ret == 0)
+		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
 	return ret;
 }
 
@@ -2385,9 +2480,17 @@ int btree_lock_page_hook(struct page *page)
 		goto out;
 
 	btrfs_tree_lock(eb);
-	spin_lock(&root->fs_info->hash_lock);
 	btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
-	spin_unlock(&root->fs_info->hash_lock);
+
+	if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
+		spin_lock(&root->fs_info->delalloc_lock);
+		if (root->fs_info->dirty_metadata_bytes >= eb->len)
+			root->fs_info->dirty_metadata_bytes -= eb->len;
+		else
+			WARN_ON(1);
+		spin_unlock(&root->fs_info->delalloc_lock);
+	}
+
 	btrfs_tree_unlock(eb);
 	free_extent_buffer(eb);
 out: