Merge branch 'x86/urgent' into x86/xen

Conflicts:
	arch/frv/include/asm/pgtable.h
	arch/x86/include/asm/required-features.h
	arch/x86/xen/mmu.c

Merge reason: x86/xen was on a .29 base still, move it to a fresher
              branch and pick up Xen fixes as well, plus resolve
              conflicts

Signed-off-by: Ingo Molnar <mingo@elte.hu>

21 files changed, 569 insertions, 466 deletions

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 9adf5e4f7e96..94212844a9bc 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -1,25 +1,10 @@
-ifneq ($(KERNELRELEASE),)
-# kbuild part of makefile
 
 obj-$(CONFIG_BTRFS_FS) := btrfs.o
-btrfs-y := super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
+
+btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   file-item.o inode-item.o inode-map.o disk-io.o \
 	   transaction.o inode.o file.o tree-defrag.o \
 	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   ref-cache.o export.o tree-log.o acl.o free-space-cache.o zlib.o \
 	   compression.o delayed-ref.o
-else
-
-# Normal Makefile
-
-KERNELDIR := /lib/modules/`uname -r`/build
-all:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` CONFIG_BTRFS_FS=m modules
-
-modules_install:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` modules_install
-clean:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` clean
-
-endif
diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
index 7fdd184a528d..cbba000dccbe 100644
--- a/fs/btrfs/acl.c
+++ b/fs/btrfs/acl.c
@@ -60,15 +60,20 @@ static struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
 		return ERR_PTR(-EINVAL);
 	}
 
+	/* Handle the cached NULL acl case without locking */
+	acl = ACCESS_ONCE(*p_acl);
+	if (!acl)
+		return acl;
+
 	spin_lock(&inode->i_lock);
-	if (*p_acl != BTRFS_ACL_NOT_CACHED)
-		acl = posix_acl_dup(*p_acl);
+	acl = *p_acl;
+	if (acl != BTRFS_ACL_NOT_CACHED)
+		acl = posix_acl_dup(acl);
 	spin_unlock(&inode->i_lock);
 
-	if (acl)
+	if (acl != BTRFS_ACL_NOT_CACHED)
 		return acl;
 
-
 	size = __btrfs_getxattr(inode, name, "", 0);
 	if (size > 0) {
 		value = kzalloc(size, GFP_NOFS);
@@ -80,9 +85,12 @@ static struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
 			btrfs_update_cached_acl(inode, p_acl, acl);
 		}
 		kfree(value);
-	} else if (size == -ENOENT) {
+	} else if (size == -ENOENT || size == -ENODATA || size == 0) {
+		/* FIXME, who returns -ENOENT?  I think nobody */
 		acl = NULL;
 		btrfs_update_cached_acl(inode, p_acl, acl);
+	} else {
+		acl = ERR_PTR(-EIO);
 	}
 
 	return acl;
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 51bfdfc8fcda..502c3d61de62 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -25,6 +25,7 @@
 #define WORK_QUEUED_BIT 0
 #define WORK_DONE_BIT 1
 #define WORK_ORDER_DONE_BIT 2
+#define WORK_HIGH_PRIO_BIT 3
 
 /*
  * container for the kthread task pointer and the list of pending work
@@ -36,6 +37,7 @@ struct btrfs_worker_thread {
 
 	/* list of struct btrfs_work that are waiting for service */
 	struct list_head pending;
+	struct list_head prio_pending;
 
 	/* list of worker threads from struct btrfs_workers */
 	struct list_head worker_list;
@@ -103,10 +105,16 @@ static noinline int run_ordered_completions(struct btrfs_workers *workers,
 
 	spin_lock_irqsave(&workers->lock, flags);
 
-	while (!list_empty(&workers->order_list)) {
-		work = list_entry(workers->order_list.next,
-				  struct btrfs_work, order_list);
-
+	while (1) {
+		if (!list_empty(&workers->prio_order_list)) {
+			work = list_entry(workers->prio_order_list.next,
+					  struct btrfs_work, order_list);
+		} else if (!list_empty(&workers->order_list)) {
+			work = list_entry(workers->order_list.next,
+					  struct btrfs_work, order_list);
+		} else {
+			break;
+		}
 		if (!test_bit(WORK_DONE_BIT, &work->flags))
 			break;
 
@@ -143,8 +151,14 @@ static int worker_loop(void *arg)
 	do {
 		spin_lock_irq(&worker->lock);
 again_locked:
-		while (!list_empty(&worker->pending)) {
-			cur = worker->pending.next;
+		while (1) {
+			if (!list_empty(&worker->prio_pending))
+				cur = worker->prio_pending.next;
+			else if (!list_empty(&worker->pending))
+				cur = worker->pending.next;
+			else
+				break;
+
 			work = list_entry(cur, struct btrfs_work, list);
 			list_del(&work->list);
 			clear_bit(WORK_QUEUED_BIT, &work->flags);
@@ -163,7 +177,6 @@ again_locked:
 
 			spin_lock_irq(&worker->lock);
 			check_idle_worker(worker);
-
 		}
 		if (freezing(current)) {
 			worker->working = 0;
@@ -178,7 +191,8 @@ again_locked:
 				 * jump_in?
 				 */
 				smp_mb();
-				if (!list_empty(&worker->pending))
+				if (!list_empty(&worker->pending) ||
+				    !list_empty(&worker->prio_pending))
 					continue;
 
 				/*
@@ -191,7 +205,8 @@ again_locked:
 				 */
 				schedule_timeout(1);
 				smp_mb();
-				if (!list_empty(&worker->pending))
+				if (!list_empty(&worker->pending) ||
+				    !list_empty(&worker->prio_pending))
 					continue;
 
 				if (kthread_should_stop())
@@ -200,7 +215,8 @@ again_locked:
 				/* still no more work?, sleep for real */
 				spin_lock_irq(&worker->lock);
 				set_current_state(TASK_INTERRUPTIBLE);
-				if (!list_empty(&worker->pending))
+				if (!list_empty(&worker->pending) ||
+				    !list_empty(&worker->prio_pending))
 					goto again_locked;
 
 				/*
@@ -248,6 +264,7 @@ void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max)
 	INIT_LIST_HEAD(&workers->worker_list);
 	INIT_LIST_HEAD(&workers->idle_list);
 	INIT_LIST_HEAD(&workers->order_list);
+	INIT_LIST_HEAD(&workers->prio_order_list);
 	spin_lock_init(&workers->lock);
 	workers->max_workers = max;
 	workers->idle_thresh = 32;
@@ -273,6 +290,7 @@ int btrfs_start_workers(struct btrfs_workers *workers, int num_workers)
 		}
 
 		INIT_LIST_HEAD(&worker->pending);
+		INIT_LIST_HEAD(&worker->prio_pending);
 		INIT_LIST_HEAD(&worker->worker_list);
 		spin_lock_init(&worker->lock);
 		atomic_set(&worker->num_pending, 0);
@@ -396,7 +414,10 @@ int btrfs_requeue_work(struct btrfs_work *work)
 		goto out;
 
 	spin_lock_irqsave(&worker->lock, flags);
-	list_add_tail(&work->list, &worker->pending);
+	if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags))
+		list_add_tail(&work->list, &worker->prio_pending);
+	else
+		list_add_tail(&work->list, &worker->pending);
 	atomic_inc(&worker->num_pending);
 
 	/* by definition we're busy, take ourselves off the idle
@@ -422,6 +443,11 @@ out:
 	return 0;
 }
 
+void btrfs_set_work_high_prio(struct btrfs_work *work)
+{
+	set_bit(WORK_HIGH_PRIO_BIT, &work->flags);
+}
+
 /*
  * places a struct btrfs_work into the pending queue of one of the kthreads
  */
@@ -438,7 +464,12 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work)
 	worker = find_worker(workers);
 	if (workers->ordered) {
 		spin_lock_irqsave(&workers->lock, flags);
-		list_add_tail(&work->order_list, &workers->order_list);
+		if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags)) {
+			list_add_tail(&work->order_list,
+				      &workers->prio_order_list);
+		} else {
+			list_add_tail(&work->order_list, &workers->order_list);
+		}
 		spin_unlock_irqrestore(&workers->lock, flags);
 	} else {
 		INIT_LIST_HEAD(&work->order_list);
@@ -446,7 +477,10 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work)
 
 	spin_lock_irqsave(&worker->lock, flags);
 
-	list_add_tail(&work->list, &worker->pending);
+	if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags))
+		list_add_tail(&work->list, &worker->prio_pending);
+	else
+		list_add_tail(&work->list, &worker->pending);
 	atomic_inc(&worker->num_pending);
 	check_busy_worker(worker);
 
diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h
index 31be4ed8b63e..1b511c109db6 100644
--- a/fs/btrfs/async-thread.h
+++ b/fs/btrfs/async-thread.h
@@ -85,6 +85,7 @@ struct btrfs_workers {
 	 * of work items waiting for completion
 	 */
 	struct list_head order_list;
+	struct list_head prio_order_list;
 
 	/* lock for finding the next worker thread to queue on */
 	spinlock_t lock;
@@ -98,4 +99,5 @@ int btrfs_start_workers(struct btrfs_workers *workers, int num_workers);
 int btrfs_stop_workers(struct btrfs_workers *workers);
 void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max);
 int btrfs_requeue_work(struct btrfs_work *work);
+void btrfs_set_work_high_prio(struct btrfs_work *work);
 #endif
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index e5b2533b691a..a99f1c2a710d 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -1325,12 +1325,12 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 	int ret = 0;
 	int blocksize;
 
-	parent = path->nodes[level - 1];
+	parent = path->nodes[level + 1];
 	if (!parent)
 		return 0;
 
 	nritems = btrfs_header_nritems(parent);
-	slot = path->slots[level];
+	slot = path->slots[level + 1];
 	blocksize = btrfs_level_size(root, level);
 
 	if (slot > 0) {
@@ -1341,7 +1341,7 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 			block1 = 0;
 		free_extent_buffer(eb);
 	}
-	if (slot < nritems) {
+	if (slot + 1 < nritems) {
 		block2 = btrfs_node_blockptr(parent, slot + 1);
 		gen = btrfs_node_ptr_generation(parent, slot + 1);
 		eb = btrfs_find_tree_block(root, block2, blocksize);
@@ -1351,7 +1351,11 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 	}
 	if (block1 || block2) {
 		ret = -EAGAIN;
+
+		/* release the whole path */
 		btrfs_release_path(root, path);
+
+		/* read the blocks */
 		if (block1)
 			readahead_tree_block(root, block1, blocksize, 0);
 		if (block2)
@@ -1361,7 +1365,7 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 			eb = read_tree_block(root, block1, blocksize, 0);
 			free_extent_buffer(eb);
 		}
-		if (block1) {
+		if (block2) {
 			eb = read_tree_block(root, block2, blocksize, 0);
 			free_extent_buffer(eb);
 		}
@@ -1481,12 +1485,15 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 	 * of the btree by dropping locks before
 	 * we read.
 	 */
-	btrfs_release_path(NULL, p);
+	btrfs_unlock_up_safe(p, level + 1);
+	btrfs_set_path_blocking(p);
+
 	if (tmp)
 		free_extent_buffer(tmp);
 	if (p->reada)
 		reada_for_search(root, p, level, slot, key->objectid);
 
+	btrfs_release_path(NULL, p);
 	tmp = read_tree_block(root, blocknr, blocksize, gen);
 	if (tmp)
 		free_extent_buffer(tmp);
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index ad96495dedc5..4414a5d9983a 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -881,6 +881,9 @@ struct btrfs_fs_info {
 	u64 metadata_alloc_profile;
 	u64 system_alloc_profile;
 
+	unsigned data_chunk_allocations;
+	unsigned metadata_ratio;
+
 	void *bdev_holder;
 };
 
@@ -2174,7 +2177,8 @@ int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
 extern struct file_operations btrfs_file_operations;
 int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct inode *inode,
-		       u64 start, u64 end, u64 inline_limit, u64 *hint_block);
+		       u64 start, u64 end, u64 locked_end,
+		       u64 inline_limit, u64 *hint_block);
 int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      struct inode *inode, u64 start, u64 end);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 92caa8035f36..0ff16d3331da 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -232,10 +232,14 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 			memcpy(&found, result, csum_size);
 
 			read_extent_buffer(buf, &val, 0, csum_size);
-			printk(KERN_INFO "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;
@@ -268,10 +272,13 @@ 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:
@@ -415,9 +422,12 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
-		printk(KERN_INFO "btrfs 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;
 	}
@@ -429,8 +439,10 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		goto err;
 	}
 	if (check_tree_block_fsid(root, eb)) {
-		printk(KERN_INFO "btrfs bad fsid on block %llu\n",
-		       (unsigned long long)eb->start);
+		if (printk_ratelimit()) {
+			printk(KERN_INFO "btrfs bad fsid on block %llu\n",
+			       (unsigned long long)eb->start);
+		}
 		ret = -EIO;
 		goto err;
 	}
@@ -579,19 +591,12 @@ 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) &&
 	      atomic_read(&fs_info->nr_async_submits)) {
 		wait_event(fs_info->async_submit_wait,
@@ -656,6 +661,7 @@ 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
@@ -765,27 +771,6 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	}
 }
 
-#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,
@@ -1273,11 +1258,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 	int ret = 0;
 	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_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
@@ -1599,6 +1580,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	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);
@@ -1689,7 +1671,7 @@ 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;
 	}
@@ -1699,7 +1681,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	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;
 	}
@@ -2095,10 +2077,10 @@ static int write_dev_supers(struct btrfs_device *device,
 				device->barriers = 0;
 				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) {
@@ -2291,7 +2273,7 @@ int close_ctree(struct btrfs_root *root)
 
 	if (fs_info->delalloc_bytes) {
 		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
-		       fs_info->delalloc_bytes);
+		       (unsigned long long)fs_info->delalloc_bytes);
 	}
 	if (fs_info->total_ref_cache_size) {
 		printk(KERN_INFO "btrfs: at umount reference cache size %llu\n",
@@ -2328,16 +2310,6 @@ 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);
 
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 178df4c67de4..e4966444811b 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -1844,10 +1844,14 @@ again:
 		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
 		       ", %llu bytes_used, %llu bytes_reserved, "
 		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use"
-		       "%llu total\n", bytes, data_sinfo->bytes_delalloc,
-		       data_sinfo->bytes_used, data_sinfo->bytes_reserved,
-		       data_sinfo->bytes_pinned, data_sinfo->bytes_readonly,
-		       data_sinfo->bytes_may_use, data_sinfo->total_bytes);
+		       "%llu total\n", (unsigned long long)bytes,
+		       (unsigned long long)data_sinfo->bytes_delalloc,
+		       (unsigned long long)data_sinfo->bytes_used,
+		       (unsigned long long)data_sinfo->bytes_reserved,
+		       (unsigned long long)data_sinfo->bytes_pinned,
+		       (unsigned long long)data_sinfo->bytes_readonly,
+		       (unsigned long long)data_sinfo->bytes_may_use,
+		       (unsigned long long)data_sinfo->total_bytes);
 		return -ENOSPC;
 	}
 	data_sinfo->bytes_may_use += bytes;
@@ -1918,15 +1922,29 @@ void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
 	spin_unlock(&info->lock);
 }
 
+static void force_metadata_allocation(struct btrfs_fs_info *info)
+{
+	struct list_head *head = &info->space_info;
+	struct btrfs_space_info *found;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(found, head, list) {
+		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
+			found->force_alloc = 1;
+	}
+	rcu_read_unlock();
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
 {
 	struct btrfs_space_info *space_info;
+	struct btrfs_fs_info *fs_info = extent_root->fs_info;
 	u64 thresh;
 	int ret = 0;
 
-	mutex_lock(&extent_root->fs_info->chunk_mutex);
+	mutex_lock(&fs_info->chunk_mutex);
 
 	flags = btrfs_reduce_alloc_profile(extent_root, flags);
 
@@ -1958,6 +1976,18 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 	}
 	spin_unlock(&space_info->lock);
 
+	/*
+	 * if we're doing a data chunk, go ahead and make sure that
+	 * we keep a reasonable number of metadata chunks allocated in the
+	 * FS as well.
+	 */
+	if (flags & BTRFS_BLOCK_GROUP_DATA) {
+		fs_info->data_chunk_allocations++;
+		if (!(fs_info->data_chunk_allocations %
+		      fs_info->metadata_ratio))
+			force_metadata_allocation(fs_info);
+	}
+
 	ret = btrfs_alloc_chunk(trans, extent_root, flags);
 	if (ret)
 		space_info->full = 1;
@@ -2798,9 +2828,12 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
 				    info->bytes_pinned - info->bytes_reserved),
 	       (info->full) ? "" : "not ");
 	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
-	       " may_use=%llu, used=%llu\n", info->total_bytes,
-	       info->bytes_pinned, info->bytes_delalloc, info->bytes_may_use,
-	       info->bytes_used);
+	       " may_use=%llu, used=%llu\n",
+	       (unsigned long long)info->total_bytes,
+	       (unsigned long long)info->bytes_pinned,
+	       (unsigned long long)info->bytes_delalloc,
+	       (unsigned long long)info->bytes_may_use,
+	       (unsigned long long)info->bytes_used);
 
 	down_read(&info->groups_sem);
 	list_for_each_entry(cache, &info->block_groups, list) {
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index eb2bee8b7fbf..fe9eb990e443 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -17,12 +17,6 @@
 #include "ctree.h"
 #include "btrfs_inode.h"
 
-/* temporary define until extent_map moves out of btrfs */
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *, struct kmem_cache *,
-						    unsigned long));
-
 static struct kmem_cache *extent_state_cache;
 static struct kmem_cache *extent_buffer_cache;
 
@@ -50,20 +44,23 @@ struct extent_page_data {
 	/* tells writepage not to lock the state bits for this range
 	 * it still does the unlocking
 	 */
-	int extent_locked;
+	unsigned int extent_locked:1;
+
+	/* tells the submit_bio code to use a WRITE_SYNC */
+	unsigned int sync_io:1;
 };
 
 int __init extent_io_init(void)
 {
-	extent_state_cache = btrfs_cache_create("extent_state",
-					    sizeof(struct extent_state), 0,
-					    NULL);
+	extent_state_cache = kmem_cache_create("extent_state",
+			sizeof(struct extent_state), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_state_cache)
 		return -ENOMEM;
 
-	extent_buffer_cache = btrfs_cache_create("extent_buffers",
-					    sizeof(struct extent_buffer), 0,
-					    NULL);
+	extent_buffer_cache = kmem_cache_create("extent_buffers",
+			sizeof(struct extent_buffer), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_buffer_cache)
 		goto free_state_cache;
 	return 0;
@@ -1404,69 +1401,6 @@ out:
 	return total_bytes;
 }
 
-#if 0
-/*
- * helper function to lock both pages and extents in the tree.
- * pages must be locked first.
- */
-static int lock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	unsigned long index = start >> PAGE_CACHE_SHIFT;
-	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
-	struct page *page;
-	int err;
-
-	while (index <= end_index) {
-		page = grab_cache_page(tree->mapping, index);
-		if (!page) {
-			err = -ENOMEM;
-			goto failed;
-		}
-		if (IS_ERR(page)) {
-			err = PTR_ERR(page);
-			goto failed;
-		}
-		index++;
-	}
-	lock_extent(tree, start, end, GFP_NOFS);
-	return 0;
-
-failed:
-	/*
-	 * we failed above in getting the page at 'index', so we undo here
-	 * up to but not including the page at 'index'
-	 */
-	end_index = index;
-	index = start >> PAGE_CACHE_SHIFT;
-	while (index < end_index) {
-		page = find_get_page(tree->mapping, index);
-		unlock_page(page);
-		page_cache_release(page);
-		index++;
-	}
-	return err;
-}
-
-/*
- * helper function to unlock both pages and extents in the tree.
- */
-static int unlock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	unsigned long index = start >> PAGE_CACHE_SHIFT;
-	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
-	struct page *page;
-
-	while (index <= end_index) {
-		page = find_get_page(tree->mapping, index);
-		unlock_page(page);
-		page_cache_release(page);
-		index++;
-	}
-	unlock_extent(tree, start, end, GFP_NOFS);
-	return 0;
-}
-#endif
-
 /*
  * set the private field for a given byte offset in the tree.  If there isn't
  * an extent_state there already, this does nothing.
@@ -2101,6 +2035,16 @@ int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
 	return ret;
 }
 
+static noinline void update_nr_written(struct page *page,
+				      struct writeback_control *wbc,
+				      unsigned long nr_written)
+{
+	wbc->nr_to_write -= nr_written;
+	if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
+	    wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
+		page->mapping->writeback_index = page->index + nr_written;
+}
+
 /*
  * the writepage semantics are similar to regular writepage.  extent
  * records are inserted to lock ranges in the tree, and as dirty areas
@@ -2136,8 +2080,14 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 	u64 delalloc_end;
 	int page_started;
 	int compressed;
+	int write_flags;
 	unsigned long nr_written = 0;
 
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		write_flags = WRITE_SYNC_PLUG;
+	else
+		write_flags = WRITE;
+
 	WARN_ON(!PageLocked(page));
 	pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
 	if (page->index > end_index ||
@@ -2164,6 +2114,12 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 	delalloc_end = 0;
 	page_started = 0;
 	if (!epd->extent_locked) {
+		/*
+		 * make sure the wbc mapping index is at least updated
+		 * to this page.
+		 */
+		update_nr_written(page, wbc, 0);
+
 		while (delalloc_end < page_end) {
 			nr_delalloc = find_lock_delalloc_range(inode, tree,
 						       page,
@@ -2185,7 +2141,13 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 		 */
 		if (page_started) {
 			ret = 0;
-			goto update_nr_written;
+			/*
+			 * we've unlocked the page, so we can't update
+			 * the mapping's writeback index, just update
+			 * nr_to_write.
+			 */
+			wbc->nr_to_write -= nr_written;
+			goto done_unlocked;
 		}
 	}
 	lock_extent(tree, start, page_end, GFP_NOFS);
@@ -2198,13 +2160,18 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 		if (ret == -EAGAIN) {
 			unlock_extent(tree, start, page_end, GFP_NOFS);
 			redirty_page_for_writepage(wbc, page);
+			update_nr_written(page, wbc, nr_written);
 			unlock_page(page);
 			ret = 0;
-			goto update_nr_written;
+			goto done_unlocked;
 		}
 	}
 
-	nr_written++;
+	/*
+	 * we don't want to touch the inode after unlocking the page,
+	 * so we update the mapping writeback index now
+	 */
+	update_nr_written(page, wbc, nr_written + 1);
 
 	end = page_end;
 	if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0))
@@ -2314,9 +2281,9 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 				       (unsigned long long)end);
 			}
 
-			ret = submit_extent_page(WRITE, tree, page, sector,
-						 iosize, pg_offset, bdev,
-						 &epd->bio, max_nr,
+			ret = submit_extent_page(write_flags, tree, page,
+						 sector, iosize, pg_offset,
+						 bdev, &epd->bio, max_nr,
 						 end_bio_extent_writepage,
 						 0, 0, 0);
 			if (ret)
@@ -2336,11 +2303,8 @@ done:
 		unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
 	unlock_page(page);
 
-update_nr_written:
-	wbc->nr_to_write -= nr_written;
-	if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
-	    wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
-		page->mapping->writeback_index = page->index + nr_written;
+done_unlocked:
+
 	return 0;
 }
 
@@ -2460,15 +2424,23 @@ retry:
 	return ret;
 }
 
-static noinline void flush_write_bio(void *data)
+static void flush_epd_write_bio(struct extent_page_data *epd)
 {
-	struct extent_page_data *epd = data;
 	if (epd->bio) {
-		submit_one_bio(WRITE, epd->bio, 0, 0);
+		if (epd->sync_io)
+			submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
+		else
+			submit_one_bio(WRITE, epd->bio, 0, 0);
 		epd->bio = NULL;
 	}
 }
 
+static noinline void flush_write_bio(void *data)
+{
+	struct extent_page_data *epd = data;
+	flush_epd_write_bio(epd);
+}
+
 int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
 			  get_extent_t *get_extent,
 			  struct writeback_control *wbc)
@@ -2480,23 +2452,22 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
 		.tree = tree,
 		.get_extent = get_extent,
 		.extent_locked = 0,
+		.sync_io = wbc->sync_mode == WB_SYNC_ALL,
 	};
 	struct writeback_control wbc_writepages = {
 		.bdi		= wbc->bdi,
-		.sync_mode	= WB_SYNC_NONE,
+		.sync_mode	= wbc->sync_mode,
 		.older_than_this = NULL,
 		.nr_to_write	= 64,
 		.range_start	= page_offset(page) + PAGE_CACHE_SIZE,
 		.range_end	= (loff_t)-1,
 	};
 
-
 	ret = __extent_writepage(page, wbc, &epd);
 
 	extent_write_cache_pages(tree, mapping, &wbc_writepages,
 				 __extent_writepage, &epd, flush_write_bio);
-	if (epd.bio)
-		submit_one_bio(WRITE, epd.bio, 0, 0);
+	flush_epd_write_bio(&epd);
 	return ret;
 }
 
@@ -2515,6 +2486,7 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
 		.tree = tree,
 		.get_extent = get_extent,
 		.extent_locked = 1,
+		.sync_io = mode == WB_SYNC_ALL,
 	};
 	struct writeback_control wbc_writepages = {
 		.bdi		= inode->i_mapping->backing_dev_info,
@@ -2540,8 +2512,7 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
 		start += PAGE_CACHE_SIZE;
 	}
 
-	if (epd.bio)
-		submit_one_bio(WRITE, epd.bio, 0, 0);
+	flush_epd_write_bio(&epd);
 	return ret;
 }
 
@@ -2556,13 +2527,13 @@ int extent_writepages(struct extent_io_tree *tree,
 		.tree = tree,
 		.get_extent = get_extent,
 		.extent_locked = 0,
+		.sync_io = wbc->sync_mode == WB_SYNC_ALL,
 	};
 
 	ret = extent_write_cache_pages(tree, mapping, wbc,
 				       __extent_writepage, &epd,
 				       flush_write_bio);
-	if (epd.bio)
-		submit_one_bio(WRITE, epd.bio, 0, 0);
+	flush_epd_write_bio(&epd);
 	return ret;
 }
 
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index b187917b36fa..30c9365861e6 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -6,19 +6,14 @@
 #include <linux/hardirq.h>
 #include "extent_map.h"
 
-/* temporary define until extent_map moves out of btrfs */
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *, struct kmem_cache *,
-						    unsigned long));
 
 static struct kmem_cache *extent_map_cache;
 
 int __init extent_map_init(void)
 {
-	extent_map_cache = btrfs_cache_create("extent_map",
-					    sizeof(struct extent_map), 0,
-					    NULL);
+	extent_map_cache = kmem_cache_create("extent_map",
+			sizeof(struct extent_map), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_map_cache)
 		return -ENOMEM;
 	return 0;
@@ -43,7 +38,6 @@ void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask)
 	tree->map.rb_node = NULL;
 	spin_lock_init(&tree->lock);
 }
-EXPORT_SYMBOL(extent_map_tree_init);
 
 /**
  * alloc_extent_map - allocate new extent map structure
@@ -64,7 +58,6 @@ struct extent_map *alloc_extent_map(gfp_t mask)
 	atomic_set(&em->refs, 1);
 	return em;
 }
-EXPORT_SYMBOL(alloc_extent_map);
 
 /**
  * free_extent_map - drop reference count of an extent_map
@@ -83,7 +76,6 @@ void free_extent_map(struct extent_map *em)
 		kmem_cache_free(extent_map_cache, em);
 	}
 }
-EXPORT_SYMBOL(free_extent_map);
 
 static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
 				   struct rb_node *node)
@@ -264,7 +256,6 @@ int add_extent_mapping(struct extent_map_tree *tree,
 out:
 	return ret;
 }
-EXPORT_SYMBOL(add_extent_mapping);
 
 /* simple helper to do math around the end of an extent, handling wrap */
 static u64 range_end(u64 start, u64 len)
@@ -326,7 +317,6 @@ found:
 out:
 	return em;
 }
-EXPORT_SYMBOL(lookup_extent_mapping);
 
 /**
  * remove_extent_mapping - removes an extent_map from the extent tree
@@ -346,4 +336,3 @@ int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
 	em->in_tree = 0;
 	return ret;
 }
-EXPORT_SYMBOL(remove_extent_mapping);
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 9c9fb46ccd08..1d51dc38bb49 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -272,83 +272,6 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 	return 0;
 }
 
-int btrfs_check_file(struct btrfs_root *root, struct inode *inode)
-{
-	return 0;
-#if 0
-	struct btrfs_path *path;
-	struct btrfs_key found_key;
-	struct extent_buffer *leaf;
-	struct btrfs_file_extent_item *extent;
-	u64 last_offset = 0;
-	int nritems;
-	int slot;
-	int found_type;
-	int ret;
-	int err = 0;
-	u64 extent_end = 0;
-
-	path = btrfs_alloc_path();
-	ret = btrfs_lookup_file_extent(NULL, root, path, inode->i_ino,
-				       last_offset, 0);
-	while (1) {
-		nritems = btrfs_header_nritems(path->nodes[0]);
-		if (path->slots[0] >= nritems) {
-			ret = btrfs_next_leaf(root, path);
-			if (ret)
-				goto out;
-			nritems = btrfs_header_nritems(path->nodes[0]);
-		}
-		slot = path->slots[0];
-		leaf = path->nodes[0];
-		btrfs_item_key_to_cpu(leaf, &found_key, slot);
-		if (found_key.objectid != inode->i_ino)
-			break;
-		if (found_key.type != BTRFS_EXTENT_DATA_KEY)
-			goto out;
-
-		if (found_key.offset < last_offset) {
-			WARN_ON(1);
-			btrfs_print_leaf(root, leaf);
-			printk(KERN_ERR "inode %lu found offset %llu "
-			       "expected %llu\n", inode->i_ino,
-			       (unsigned long long)found_key.offset,
-			       (unsigned long long)last_offset);
-			err = 1;
-			goto out;
-		}
-		extent = btrfs_item_ptr(leaf, slot,
-					struct btrfs_file_extent_item);
-		found_type = btrfs_file_extent_type(leaf, extent);
-		if (found_type == BTRFS_FILE_EXTENT_REG) {
-			extent_end = found_key.offset +
-			     btrfs_file_extent_num_bytes(leaf, extent);
-		} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
-			struct btrfs_item *item;
-			item = btrfs_item_nr(leaf, slot);
-			extent_end = found_key.offset +
-			     btrfs_file_extent_inline_len(leaf, extent);
-			extent_end = (extent_end + root->sectorsize - 1) &
-				~((u64)root->sectorsize - 1);
-		}
-		last_offset = extent_end;
-		path->slots[0]++;
-	}
-	if (0 && last_offset < inode->i_size) {
-		WARN_ON(1);
-		btrfs_print_leaf(root, leaf);
-		printk(KERN_ERR "inode %lu found offset %llu size %llu\n",
-		       inode->i_ino, (unsigned long long)last_offset,
-		       (unsigned long long)inode->i_size);
-		err = 1;
-
-	}
-out:
-	btrfs_free_path(path);
-	return err;
-#endif
-}
-
 /*
  * this is very complex, but the basic idea is to drop all extents
  * in the range start - end.  hint_block is filled in with a block number
@@ -363,15 +286,16 @@ out:
  */
 noinline int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct inode *inode,
-		       u64 start, u64 end, u64 inline_limit, u64 *hint_byte)
+		       u64 start, u64 end, u64 locked_end,
+		       u64 inline_limit, u64 *hint_byte)
 {
 	u64 extent_end = 0;
-	u64 locked_end = end;
 	u64 search_start = start;
 	u64 leaf_start;
 	u64 ram_bytes = 0;
 	u64 orig_parent = 0;
 	u64 disk_bytenr = 0;
+	u64 orig_locked_end = locked_end;
 	u8 compression;
 	u8 encryption;
 	u16 other_encoding = 0;
@@ -684,11 +608,10 @@ next_slot:
 	}
 out:
 	btrfs_free_path(path);
-	if (locked_end > end) {
-		unlock_extent(&BTRFS_I(inode)->io_tree, end, locked_end - 1,
-			      GFP_NOFS);
+	if (locked_end > orig_locked_end) {
+		unlock_extent(&BTRFS_I(inode)->io_tree, orig_locked_end,
+			      locked_end - 1, GFP_NOFS);
 	}
-	btrfs_check_file(root, inode);
 	return ret;
 }
 
@@ -830,7 +753,7 @@ again:
 
 		ret = btrfs_del_items(trans, root, path, del_slot, del_nr);
 		BUG_ON(ret);
-		goto done;
+		goto release;
 	} else if (split == start) {
 		if (locked_end < extent_end) {
 			ret = try_lock_extent(&BTRFS_I(inode)->io_tree,
@@ -926,6 +849,8 @@ again:
 	}
 done:
 	btrfs_mark_buffer_dirty(leaf);
+
+release:
 	btrfs_release_path(root, path);
 	if (split_end && split == start) {
 		split = end;
@@ -1131,7 +1056,7 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		if (will_write) {
 			btrfs_fdatawrite_range(inode->i_mapping, pos,
 					       pos + write_bytes - 1,
-					       WB_SYNC_NONE);
+					       WB_SYNC_ALL);
 		} else {
 			balance_dirty_pages_ratelimited_nr(inode->i_mapping,
 							   num_pages);
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 768b9523662d..0bc93657b460 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -332,13 +332,17 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			printk(KERN_ERR "couldn't find space %llu to free\n",
 			       (unsigned long long)offset);
 			printk(KERN_ERR "cached is %d, offset %llu bytes %llu\n",
-			       block_group->cached, block_group->key.objectid,
-			       block_group->key.offset);
+			       block_group->cached,
+			       (unsigned long long)block_group->key.objectid,
+			       (unsigned long long)block_group->key.offset);
 			btrfs_dump_free_space(block_group, bytes);
 		} else if (info) {
 			printk(KERN_ERR "hmm, found offset=%llu bytes=%llu, "
 			       "but wanted offset=%llu bytes=%llu\n",
-			       info->offset, info->bytes, offset, bytes);
+			       (unsigned long long)info->offset,
+			       (unsigned long long)info->bytes,
+			       (unsigned long long)offset,
+			       (unsigned long long)bytes);
 		}
 		WARN_ON(1);
 	}
@@ -357,8 +361,9 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 		info = rb_entry(n, struct btrfs_free_space, offset_index);
 		if (info->bytes >= bytes)
 			count++;
-		printk(KERN_ERR "entry offset %llu, bytes %llu\n", info->offset,
-		       info->bytes);
+		printk(KERN_ERR "entry offset %llu, bytes %llu\n",
+		       (unsigned long long)info->offset,
+		       (unsigned long long)info->bytes);
 	}
 	printk(KERN_INFO "%d blocks of free space at or bigger than bytes is"
 	       "\n", count);
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index cc7334d833c9..9abbced1123d 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -79,7 +79,7 @@ int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
 	}
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
-	search_start = max(search_start, BTRFS_FIRST_FREE_OBJECTID);
+	search_start = max(search_start, (u64)BTRFS_FIRST_FREE_OBJECTID);
 	search_key.objectid = search_start;
 	search_key.type = 0;
 	search_key.offset = 0;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index a0d1dd492a58..90c23eb28829 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -70,7 +70,6 @@ static struct extent_io_ops btrfs_extent_io_ops;
 static struct kmem_cache *btrfs_inode_cachep;
 struct kmem_cache *btrfs_trans_handle_cachep;
 struct kmem_cache *btrfs_transaction_cachep;
-struct kmem_cache *btrfs_bit_radix_cachep;
 struct kmem_cache *btrfs_path_cachep;
 
 #define S_SHIFT 12
@@ -234,7 +233,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 	}
 
 	ret = btrfs_drop_extents(trans, root, inode, start,
-				 aligned_end, start, &hint_byte);
+				 aligned_end, aligned_end, start, &hint_byte);
 	BUG_ON(ret);
 
 	if (isize > actual_end)
@@ -1439,6 +1438,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 				       struct inode *inode, u64 file_pos,
 				       u64 disk_bytenr, u64 disk_num_bytes,
 				       u64 num_bytes, u64 ram_bytes,
+				       u64 locked_end,
 				       u8 compression, u8 encryption,
 				       u16 other_encoding, int extent_type)
 {
@@ -1455,7 +1455,8 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 
 	path->leave_spinning = 1;
 	ret = btrfs_drop_extents(trans, root, inode, file_pos,
-				 file_pos + num_bytes, file_pos, &hint);
+				 file_pos + num_bytes, locked_end,
+				 file_pos, &hint);
 	BUG_ON(ret);
 
 	ins.objectid = inode->i_ino;
@@ -1590,6 +1591,8 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 						ordered_extent->disk_len,
 						ordered_extent->len,
 						ordered_extent->len,
+						ordered_extent->file_offset +
+						ordered_extent->len,
 						compressed, 0, 0,
 						BTRFS_FILE_EXTENT_REG);
 		BUG_ON(ret);
@@ -1819,10 +1822,12 @@ good:
 	return 0;
 
 zeroit:
-	printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
-	       "private %llu\n", page->mapping->host->i_ino,
-	       (unsigned long long)start, csum,
-	       (unsigned long long)private);
+	if (printk_ratelimit()) {
+		printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
+		       "private %llu\n", page->mapping->host->i_ino,
+		       (unsigned long long)start, csum,
+		       (unsigned long long)private);
+	}
 	memset(kaddr + offset, 1, end - start + 1);
 	flush_dcache_page(page);
 	kunmap_atomic(kaddr, KM_USER0);
@@ -2011,6 +2016,57 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 }
 
 /*
+ * very simple check to peek ahead in the leaf looking for xattrs.  If we
+ * don't find any xattrs, we know there can't be any acls.
+ *
+ * slot is the slot the inode is in, objectid is the objectid of the inode
+ */
+static noinline int acls_after_inode_item(struct extent_buffer *leaf,
+					  int slot, u64 objectid)
+{
+	u32 nritems = btrfs_header_nritems(leaf);
+	struct btrfs_key found_key;
+	int scanned = 0;
+
+	slot++;
+	while (slot < nritems) {
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		/* we found a different objectid, there must not be acls */
+		if (found_key.objectid != objectid)
+			return 0;
+
+		/* we found an xattr, assume we've got an acl */
+		if (found_key.type == BTRFS_XATTR_ITEM_KEY)
+			return 1;
+
+		/*
+		 * we found a key greater than an xattr key, there can't
+		 * be any acls later on
+		 */
+		if (found_key.type > BTRFS_XATTR_ITEM_KEY)
+			return 0;
+
+		slot++;
+		scanned++;
+
+		/*
+		 * it goes inode, inode backrefs, xattrs, extents,
+		 * so if there are a ton of hard links to an inode there can
+		 * be a lot of backrefs.  Don't waste time searching too hard,
+		 * this is just an optimization
+		 */
+		if (scanned >= 8)
+			break;
+	}
+	/* we hit the end of the leaf before we found an xattr or
+	 * something larger than an xattr.  We have to assume the inode
+	 * has acls
+	 */
+	return 1;
+}
+
+/*
  * read an inode from the btree into the in-memory inode
  */
 void btrfs_read_locked_inode(struct inode *inode)
@@ -2021,6 +2077,7 @@ void btrfs_read_locked_inode(struct inode *inode)
 	struct btrfs_timespec *tspec;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key location;
+	int maybe_acls;
 	u64 alloc_group_block;
 	u32 rdev;
 	int ret;
@@ -2067,6 +2124,16 @@ void btrfs_read_locked_inode(struct inode *inode)
 
 	alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
 
+	/*
+	 * try to precache a NULL acl entry for files that don't have
+	 * any xattrs or acls
+	 */
+	maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino);
+	if (!maybe_acls) {
+		BTRFS_I(inode)->i_acl = NULL;
+		BTRFS_I(inode)->i_default_acl = NULL;
+	}
+
 	BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
 						alloc_group_block, 0);
 	btrfs_free_path(path);
@@ -2877,6 +2944,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 			err = btrfs_drop_extents(trans, root, inode,
 						 cur_offset,
 						 cur_offset + hole_size,
+						 block_end,
 						 cur_offset, &hint_byte);
 			if (err)
 				break;
@@ -3041,8 +3109,8 @@ static noinline void init_btrfs_i(struct inode *inode)
 {
 	struct btrfs_inode *bi = BTRFS_I(inode);
 
-	bi->i_acl = NULL;
-	bi->i_default_acl = NULL;
+	bi->i_acl = BTRFS_ACL_NOT_CACHED;
+	bi->i_default_acl = BTRFS_ACL_NOT_CACHED;
 
 	bi->generation = 0;
 	bi->sequence = 0;
@@ -4634,47 +4702,36 @@ void btrfs_destroy_cachep(void)
 		kmem_cache_destroy(btrfs_trans_handle_cachep);
 	if (btrfs_transaction_cachep)
 		kmem_cache_destroy(btrfs_transaction_cachep);
-	if (btrfs_bit_radix_cachep)
-		kmem_cache_destroy(btrfs_bit_radix_cachep);
 	if (btrfs_path_cachep)
 		kmem_cache_destroy(btrfs_path_cachep);
 }
 
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *))
-{
-	return kmem_cache_create(name, size, 0, (SLAB_RECLAIM_ACCOUNT |
-				 SLAB_MEM_SPREAD | extra_flags), ctor);
-}
-
 int btrfs_init_cachep(void)
 {
-	btrfs_inode_cachep = btrfs_cache_create("btrfs_inode_cache",
-					  sizeof(struct btrfs_inode),
-					  0, init_once);
+	btrfs_inode_cachep = kmem_cache_create("btrfs_inode_cache",
+			sizeof(struct btrfs_inode), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, init_once);
 	if (!btrfs_inode_cachep)
 		goto fail;
-	btrfs_trans_handle_cachep =
-			btrfs_cache_create("btrfs_trans_handle_cache",
-					   sizeof(struct btrfs_trans_handle),
-					   0, NULL);
+
+	btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle_cache",
+			sizeof(struct btrfs_trans_handle), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_trans_handle_cachep)
 		goto fail;
-	btrfs_transaction_cachep = btrfs_cache_create("btrfs_transaction_cache",
-					     sizeof(struct btrfs_transaction),
-					     0, NULL);
+
+	btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction_cache",
+			sizeof(struct btrfs_transaction), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_transaction_cachep)
 		goto fail;
-	btrfs_path_cachep = btrfs_cache_create("btrfs_path_cache",
-					 sizeof(struct btrfs_path),
-					 0, NULL);
+
+	btrfs_path_cachep = kmem_cache_create("btrfs_path_cache",
+			sizeof(struct btrfs_path), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_path_cachep)
 		goto fail;
-	btrfs_bit_radix_cachep = btrfs_cache_create("btrfs_radix", 256,
-					      SLAB_DESTROY_BY_RCU, NULL);
-	if (!btrfs_bit_radix_cachep)
-		goto fail;
+
 	return 0;
 fail:
 	btrfs_destroy_cachep();
@@ -4970,10 +5027,10 @@ out_fail:
 	return err;
 }
 
-static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
-			       u64 alloc_hint, int mode)
+static int prealloc_file_range(struct btrfs_trans_handle *trans,
+			       struct inode *inode, u64 start, u64 end,
+			       u64 locked_end, u64 alloc_hint, int mode)
 {
-	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key ins;
 	u64 alloc_size;
@@ -4981,10 +5038,6 @@ static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
 	u64 num_bytes = end - start;
 	int ret = 0;
 
-	trans = btrfs_join_transaction(root, 1);
-	BUG_ON(!trans);
-	btrfs_set_trans_block_group(trans, inode);
-
 	while (num_bytes > 0) {
 		alloc_size = min(num_bytes, root->fs_info->max_extent);
 		ret = btrfs_reserve_extent(trans, root, alloc_size,
@@ -4997,7 +5050,8 @@ static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
 		ret = insert_reserved_file_extent(trans, inode,
 						  cur_offset, ins.objectid,
 						  ins.offset, ins.offset,
-						  ins.offset, 0, 0, 0,
+						  ins.offset, locked_end,
+						  0, 0, 0,
 						  BTRFS_FILE_EXTENT_PREALLOC);
 		BUG_ON(ret);
 		num_bytes -= ins.offset;
@@ -5015,7 +5069,6 @@ out:
 		BUG_ON(ret);
 	}
 
-	btrfs_end_transaction(trans, root);
 	return ret;
 }
 
@@ -5027,13 +5080,21 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 	u64 alloc_start;
 	u64 alloc_end;
 	u64 alloc_hint = 0;
+	u64 locked_end;
 	u64 mask = BTRFS_I(inode)->root->sectorsize - 1;
 	struct extent_map *em;
+	struct btrfs_trans_handle *trans;
 	int ret;
 
 	alloc_start = offset & ~mask;
 	alloc_end =  (offset + len + mask) & ~mask;
 
+	/*
+	 * wait for ordered IO before we have any locks.  We'll loop again
+	 * below with the locks held.
+	 */
+	btrfs_wait_ordered_range(inode, alloc_start, alloc_end - alloc_start);
+
 	mutex_lock(&inode->i_mutex);
 	if (alloc_start > inode->i_size) {
 		ret = btrfs_cont_expand(inode, alloc_start);
@@ -5041,10 +5102,21 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 			goto out;
 	}
 
+	locked_end = alloc_end - 1;
 	while (1) {
 		struct btrfs_ordered_extent *ordered;
-		lock_extent(&BTRFS_I(inode)->io_tree, alloc_start,
-			    alloc_end - 1, GFP_NOFS);
+
+		trans = btrfs_start_transaction(BTRFS_I(inode)->root, 1);
+		if (!trans) {
+			ret = -EIO;
+			goto out;
+		}
+
+		/* the extent lock is ordered inside the running
+		 * transaction
+		 */
+		lock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+			    GFP_NOFS);
 		ordered = btrfs_lookup_first_ordered_extent(inode,
 							    alloc_end - 1);
 		if (ordered &&
@@ -5052,7 +5124,13 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		    ordered->file_offset < alloc_end) {
 			btrfs_put_ordered_extent(ordered);
 			unlock_extent(&BTRFS_I(inode)->io_tree,
-				      alloc_start, alloc_end - 1, GFP_NOFS);
+				      alloc_start, locked_end, GFP_NOFS);
+			btrfs_end_transaction(trans, BTRFS_I(inode)->root);
+
+			/*
+			 * we can't wait on the range with the transaction
+			 * running or with the extent lock held
+			 */
 			btrfs_wait_ordered_range(inode, alloc_start,
 						 alloc_end - alloc_start);
 		} else {
@@ -5070,8 +5148,9 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		last_byte = min(extent_map_end(em), alloc_end);
 		last_byte = (last_byte + mask) & ~mask;
 		if (em->block_start == EXTENT_MAP_HOLE) {
-			ret = prealloc_file_range(inode, cur_offset,
-					last_byte, alloc_hint, mode);
+			ret = prealloc_file_range(trans, inode, cur_offset,
+					last_byte, locked_end + 1,
+					alloc_hint, mode);
 			if (ret < 0) {
 				free_extent_map(em);
 				break;
@@ -5087,8 +5166,10 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 			break;
 		}
 	}
-	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, alloc_end - 1,
+	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 		      GFP_NOFS);
+
+	btrfs_end_transaction(trans, BTRFS_I(inode)->root);
 out:
 	mutex_unlock(&inode->i_mutex);
 	return ret;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 7594bec1be10..5e94ea6e1cbe 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -461,15 +461,9 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
-
-	if (!vol_args)
-		return -ENOMEM;
-
-	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
-		ret = -EFAULT;
-		goto out;
-	}
+	vol_args = memdup_user(arg, sizeof(*vol_args));
+	if (IS_ERR(vol_args))
+		return PTR_ERR(vol_args);
 
 	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
 	namelen = strlen(vol_args->name);
@@ -483,11 +477,13 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
 		*devstr = '\0';
 		devstr = vol_args->name;
 		devid = simple_strtoull(devstr, &end, 10);
-		printk(KERN_INFO "resizing devid %llu\n", devid);
+		printk(KERN_INFO "resizing devid %llu\n",
+		       (unsigned long long)devid);
 	}
 	device = btrfs_find_device(root, devid, NULL, NULL);
 	if (!device) {
-		printk(KERN_INFO "resizer unable to find device %llu\n", devid);
+		printk(KERN_INFO "resizer unable to find device %llu\n",
+		       (unsigned long long)devid);
 		ret = -EINVAL;
 		goto out_unlock;
 	}
@@ -545,7 +541,6 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
 
 out_unlock:
 	mutex_unlock(&root->fs_info->volume_mutex);
-out:
 	kfree(vol_args);
 	return ret;
 }
@@ -565,15 +560,9 @@ static noinline int btrfs_ioctl_snap_create(struct file *file,
 	if (root->fs_info->sb->s_flags & MS_RDONLY)
 		return -EROFS;
 
-	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
-
-	if (!vol_args)
-		return -ENOMEM;
-
-	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
-		ret = -EFAULT;
-		goto out;
-	}
+	vol_args = memdup_user(arg, sizeof(*vol_args));
+	if (IS_ERR(vol_args))
+		return PTR_ERR(vol_args);
 
 	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
 	namelen = strlen(vol_args->name);
@@ -675,19 +664,13 @@ static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+	vol_args = memdup_user(arg, sizeof(*vol_args));
+	if (IS_ERR(vol_args))
+		return PTR_ERR(vol_args);
 
-	if (!vol_args)
-		return -ENOMEM;
-
-	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
-		ret = -EFAULT;
-		goto out;
-	}
 	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
 	ret = btrfs_init_new_device(root, vol_args->name);
 
-out:
 	kfree(vol_args);
 	return ret;
 }
@@ -703,19 +686,13 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg)
 	if (root->fs_info->sb->s_flags & MS_RDONLY)
 		return -EROFS;
 
-	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+	vol_args = memdup_user(arg, sizeof(*vol_args));
+	if (IS_ERR(vol_args))
+		return PTR_ERR(vol_args);
 
-	if (!vol_args)
-		return -ENOMEM;
-
-	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
-		ret = -EFAULT;
-		goto out;
-	}
 	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
 	ret = btrfs_rm_device(root, vol_args->name);
 
-out:
 	kfree(vol_args);
 	return ret;
 }
@@ -830,7 +807,8 @@ static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	BUG_ON(!trans);
 
 	/* punch hole in destination first */
-	btrfs_drop_extents(trans, root, inode, off, off+len, 0, &hint_byte);
+	btrfs_drop_extents(trans, root, inode, off, off + len,
+			   off + len, 0, &hint_byte);
 
 	/* clone data */
 	key.objectid = src->i_ino;
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index 53c87b197d70..d6f0806c682f 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -489,7 +489,7 @@ again:
 	/* start IO across the range first to instantiate any delalloc
 	 * extents
 	 */
-	btrfs_fdatawrite_range(inode->i_mapping, start, orig_end, WB_SYNC_NONE);
+	btrfs_fdatawrite_range(inode->i_mapping, start, orig_end, WB_SYNC_ALL);
 
 	/* The compression code will leave pages locked but return from
 	 * writepage without setting the page writeback.  Starting again
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 9744af9d71e9..3536bdb2d7cb 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -68,7 +68,7 @@ enum {
 	Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,
 	Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,
 	Opt_ssd, Opt_thread_pool, Opt_noacl,  Opt_compress, Opt_notreelog,
-	Opt_flushoncommit, Opt_err,
+	Opt_ratio, Opt_flushoncommit, Opt_err,
 };
 
 static match_table_t tokens = {
@@ -87,6 +87,7 @@ static match_table_t tokens = {
 	{Opt_noacl, "noacl"},
 	{Opt_notreelog, "notreelog"},
 	{Opt_flushoncommit, "flushoncommit"},
+	{Opt_ratio, "metadata_ratio=%d"},
 	{Opt_err, NULL},
 };
 
@@ -195,7 +196,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				info->max_extent = max_t(u64,
 					info->max_extent, root->sectorsize);
 				printk(KERN_INFO "btrfs: max_extent at %llu\n",
-				       info->max_extent);
+				       (unsigned long long)info->max_extent);
 			}
 			break;
 		case Opt_max_inline:
@@ -210,7 +211,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 						root->sectorsize);
 				}
 				printk(KERN_INFO "btrfs: max_inline at %llu\n",
-					info->max_inline);
+					(unsigned long long)info->max_inline);
 			}
 			break;
 		case Opt_alloc_start:
@@ -220,7 +221,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				kfree(num);
 				printk(KERN_INFO
 					"btrfs: allocations start at %llu\n",
-					info->alloc_start);
+					(unsigned long long)info->alloc_start);
 			}
 			break;
 		case Opt_noacl:
@@ -234,6 +235,15 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			printk(KERN_INFO "btrfs: turning on flush-on-commit\n");
 			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
 			break;
+		case Opt_ratio:
+			intarg = 0;
+			match_int(&args[0], &intarg);
+			if (intarg) {
+				info->metadata_ratio = intarg;
+				printk(KERN_INFO "btrfs: metadata ratio %d\n",
+				       info->metadata_ratio);
+			}
+			break;
 		default:
 			break;
 		}
@@ -410,11 +420,14 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
 	if (btrfs_test_opt(root, NOBARRIER))
 		seq_puts(seq, ",nobarrier");
 	if (info->max_extent != (u64)-1)
-		seq_printf(seq, ",max_extent=%llu", info->max_extent);
+		seq_printf(seq, ",max_extent=%llu",
+			   (unsigned long long)info->max_extent);
 	if (info->max_inline != 8192 * 1024)
-		seq_printf(seq, ",max_inline=%llu", info->max_inline);
+		seq_printf(seq, ",max_inline=%llu",
+			   (unsigned long long)info->max_inline);
 	if (info->alloc_start != 0)
-		seq_printf(seq, ",alloc_start=%llu", info->alloc_start);
+		seq_printf(seq, ",alloc_start=%llu",
+			   (unsigned long long)info->alloc_start);
 	if (info->thread_pool_size !=  min_t(unsigned long,
 					     num_online_cpus() + 2, 8))
 		seq_printf(seq, ",thread_pool=%d", info->thread_pool_size);
@@ -635,14 +648,9 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	vol = kmalloc(sizeof(*vol), GFP_KERNEL);
-	if (!vol)
-		return -ENOMEM;
-
-	if (copy_from_user(vol, (void __user *)arg, sizeof(*vol))) {
-		ret = -EFAULT;
-		goto out;
-	}
+	vol = memdup_user((void __user *)arg, sizeof(*vol));
+	if (IS_ERR(vol))
+		return PTR_ERR(vol);
 
 	switch (cmd) {
 	case BTRFS_IOC_SCAN_DEV:
@@ -650,7 +658,7 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
 					    &btrfs_fs_type, &fs_devices);
 		break;
 	}
-out:
+
 	kfree(vol);
 	return ret;
 }
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 2869b3361eb6..01b143605ec1 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -687,7 +687,13 @@ static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info)
 		prepare_to_wait(&info->transaction_wait, &wait,
 				TASK_UNINTERRUPTIBLE);
 		mutex_unlock(&info->trans_mutex);
+
+		atomic_dec(&info->throttles);
+		wake_up(&info->transaction_throttle);
+
 		schedule();
+
+		atomic_inc(&info->throttles);
 		mutex_lock(&info->trans_mutex);
 		finish_wait(&info->transaction_wait, &wait);
 	}
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 25f20ea11f27..db5e212e8445 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -536,7 +536,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 	saved_nbytes = inode_get_bytes(inode);
 	/* drop any overlapping extents */
 	ret = btrfs_drop_extents(trans, root, inode,
-			 start, extent_end, start, &alloc_hint);
+			 start, extent_end, extent_end, start, &alloc_hint);
 	BUG_ON(ret);
 
 	if (found_type == BTRFS_FILE_EXTENT_REG ||
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index e0913e469728..5f01dad4b696 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -125,6 +125,20 @@ static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
 	return NULL;
 }
 
+static void requeue_list(struct btrfs_pending_bios *pending_bios,
+			struct bio *head, struct bio *tail)
+{
+
+	struct bio *old_head;
+
+	old_head = pending_bios->head;
+	pending_bios->head = head;
+	if (pending_bios->tail)
+		tail->bi_next = old_head;
+	else
+		pending_bios->tail = tail;
+}
+
 /*
  * we try to collect pending bios for a device so we don't get a large
  * number of procs sending bios down to the same device.  This greatly
@@ -141,10 +155,12 @@ static noinline int run_scheduled_bios(struct btrfs_device *device)
 	struct bio *pending;
 	struct backing_dev_info *bdi;
 	struct btrfs_fs_info *fs_info;
+	struct btrfs_pending_bios *pending_bios;
 	struct bio *tail;
 	struct bio *cur;
 	int again = 0;
-	unsigned long num_run = 0;
+	unsigned long num_run;
+	unsigned long num_sync_run;
 	unsigned long limit;
 	unsigned long last_waited = 0;
 
@@ -153,20 +169,30 @@ static noinline int run_scheduled_bios(struct btrfs_device *device)
 	limit = btrfs_async_submit_limit(fs_info);
 	limit = limit * 2 / 3;
 
+	/* we want to make sure that every time we switch from the sync
+	 * list to the normal list, we unplug
+	 */
+	num_sync_run = 0;
+
 loop:
 	spin_lock(&device->io_lock);
+	num_run = 0;
 
 loop_lock:
+
 	/* take all the bios off the list at once and process them
 	 * later on (without the lock held).  But, remember the
 	 * tail and other pointers so the bios can be properly reinserted
 	 * into the list if we hit congestion
 	 */
-	pending = device->pending_bios;
-	tail = device->pending_bio_tail;
+	if (device->pending_sync_bios.head)
+		pending_bios = &device->pending_sync_bios;
+	else
+		pending_bios = &device->pending_bios;
+
+	pending = pending_bios->head;
+	tail = pending_bios->tail;
 	WARN_ON(pending && !tail);
-	device->pending_bios = NULL;
-	device->pending_bio_tail = NULL;
 
 	/*
 	 * if pending was null this time around, no bios need processing
@@ -176,16 +202,41 @@ loop_lock:
 	 * device->running_pending is used to synchronize with the
 	 * schedule_bio code.
 	 */
-	if (pending) {
-		again = 1;
-		device->running_pending = 1;
-	} else {
+	if (device->pending_sync_bios.head == NULL &&
+	    device->pending_bios.head == NULL) {
 		again = 0;
 		device->running_pending = 0;
+	} else {
+		again = 1;
+		device->running_pending = 1;
 	}
+
+	pending_bios->head = NULL;
+	pending_bios->tail = NULL;
+
 	spin_unlock(&device->io_lock);
 
+	/*
+	 * if we're doing the regular priority list, make sure we unplug
+	 * for any high prio bios we've sent down
+	 */
+	if (pending_bios == &device->pending_bios && num_sync_run > 0) {
+		num_sync_run = 0;
+		blk_run_backing_dev(bdi, NULL);
+	}
+
 	while (pending) {
+
+		rmb();
+		if (pending_bios != &device->pending_sync_bios &&
+		    device->pending_sync_bios.head &&
+		    num_run > 16) {
+			cond_resched();
+			spin_lock(&device->io_lock);
+			requeue_list(pending_bios, pending, tail);
+			goto loop_lock;
+		}
+
 		cur = pending;
 		pending = pending->bi_next;
 		cur->bi_next = NULL;
@@ -196,10 +247,18 @@ loop_lock:
 			wake_up(&fs_info->async_submit_wait);
 
 		BUG_ON(atomic_read(&cur->bi_cnt) == 0);
-		bio_get(cur);
 		submit_bio(cur->bi_rw, cur);
-		bio_put(cur);
 		num_run++;
+		if (bio_sync(cur))
+			num_sync_run++;
+
+		if (need_resched()) {
+			if (num_sync_run) {
+				blk_run_backing_dev(bdi, NULL);
+				num_sync_run = 0;
+			}
+			cond_resched();
+		}
 
 		/*
 		 * we made progress, there is more work to do and the bdi
@@ -208,7 +267,6 @@ loop_lock:
 		 */
 		if (pending && bdi_write_congested(bdi) && num_run > 16 &&
 		    fs_info->fs_devices->open_devices > 1) {
-			struct bio *old_head;
 			struct io_context *ioc;
 
 			ioc = current->io_context;
@@ -233,17 +291,17 @@ loop_lock:
 				 * against it before looping
 				 */
 				last_waited = ioc->last_waited;
+				if (need_resched()) {
+					if (num_sync_run) {
+						blk_run_backing_dev(bdi, NULL);
+						num_sync_run = 0;
+					}
+					cond_resched();
+				}
 				continue;
 			}
 			spin_lock(&device->io_lock);
-
-			old_head = device->pending_bios;
-			device->pending_bios = pending;
-			if (device->pending_bio_tail)
-				tail->bi_next = old_head;
-			else
-				device->pending_bio_tail = tail;
-
+			requeue_list(pending_bios, pending, tail);
 			device->running_pending = 1;
 
 			spin_unlock(&device->io_lock);
@@ -251,11 +309,18 @@ loop_lock:
 			goto done;
 		}
 	}
+
+	if (num_sync_run) {
+		num_sync_run = 0;
+		blk_run_backing_dev(bdi, NULL);
+	}
+
+	cond_resched();
 	if (again)
 		goto loop;
 
 	spin_lock(&device->io_lock);
-	if (device->pending_bios)
+	if (device->pending_bios.head || device->pending_sync_bios.head)
 		goto loop_lock;
 	spin_unlock(&device->io_lock);
 
@@ -1478,7 +1543,7 @@ static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
 	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
 	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
 	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
-	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
+	btrfs_set_device_total_bytes(leaf, dev_item, device->disk_total_bytes);
 	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
 	btrfs_mark_buffer_dirty(leaf);
 
@@ -1875,14 +1940,6 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 	device->total_bytes = new_size;
 	if (device->writeable)
 		device->fs_devices->total_rw_bytes -= diff;
-	ret = btrfs_update_device(trans, device);
-	if (ret) {
-		unlock_chunks(root);
-		btrfs_end_transaction(trans, root);
-		goto done;
-	}
-	WARN_ON(diff > old_total);
-	btrfs_set_super_total_bytes(super_copy, old_total - diff);
 	unlock_chunks(root);
 	btrfs_end_transaction(trans, root);
 
@@ -1914,7 +1971,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 		length = btrfs_dev_extent_length(l, dev_extent);
 
 		if (key.offset + length <= new_size)
-			goto done;
+			break;
 
 		chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
 		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
@@ -1927,6 +1984,26 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 			goto done;
 	}
 
+	/* Shrinking succeeded, else we would be at "done". */
+	trans = btrfs_start_transaction(root, 1);
+	if (!trans) {
+		ret = -ENOMEM;
+		goto done;
+	}
+	lock_chunks(root);
+
+	device->disk_total_bytes = new_size;
+	/* Now btrfs_update_device() will change the on-disk size. */
+	ret = btrfs_update_device(trans, device);
+	if (ret) {
+		unlock_chunks(root);
+		btrfs_end_transaction(trans, root);
+		goto done;
+	}
+	WARN_ON(diff > old_total);
+	btrfs_set_super_total_bytes(super_copy, old_total - diff);
+	unlock_chunks(root);
+	btrfs_end_transaction(trans, root);
 done:
 	btrfs_free_path(path);
 	return ret;
@@ -2497,7 +2574,7 @@ again:
 			max_errors = 1;
 		}
 	}
-	if (multi_ret && rw == WRITE &&
+	if (multi_ret && (rw & (1 << BIO_RW)) &&
 	    stripes_allocated < stripes_required) {
 		stripes_allocated = map->num_stripes;
 		free_extent_map(em);
@@ -2762,6 +2839,7 @@ static noinline int schedule_bio(struct btrfs_root *root,
 				 int rw, struct bio *bio)
 {
 	int should_queue = 1;
+	struct btrfs_pending_bios *pending_bios;
 
 	/* don't bother with additional async steps for reads, right now */
 	if (!(rw & (1 << BIO_RW))) {
@@ -2783,13 +2861,17 @@ static noinline int schedule_bio(struct btrfs_root *root,
 	bio->bi_rw |= rw;
 
 	spin_lock(&device->io_lock);
+	if (bio_sync(bio))
+		pending_bios = &device->pending_sync_bios;
+	else
+		pending_bios = &device->pending_bios;
 
-	if (device->pending_bio_tail)
-		device->pending_bio_tail->bi_next = bio;
+	if (pending_bios->tail)
+		pending_bios->tail->bi_next = bio;
 
-	device->pending_bio_tail = bio;
-	if (!device->pending_bios)
-		device->pending_bios = bio;
+	pending_bios->tail = bio;
+	if (!pending_bios->head)
+		pending_bios->head = bio;
 	if (device->running_pending)
 		should_queue = 0;
 
@@ -3006,7 +3088,8 @@ static int fill_device_from_item(struct extent_buffer *leaf,
 	unsigned long ptr;
 
 	device->devid = btrfs_device_id(leaf, dev_item);
-	device->total_bytes = btrfs_device_total_bytes(leaf, dev_item);
+	device->disk_total_bytes = btrfs_device_total_bytes(leaf, dev_item);
+	device->total_bytes = device->disk_total_bytes;
 	device->bytes_used = btrfs_device_bytes_used(leaf, dev_item);
 	device->type = btrfs_device_type(leaf, dev_item);
 	device->io_align = btrfs_device_io_align(leaf, dev_item);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 2185de72ff7d..5c3ff6d02fd7 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -23,13 +23,22 @@
 #include "async-thread.h"
 
 struct buffer_head;
+struct btrfs_pending_bios {
+	struct bio *head;
+	struct bio *tail;
+};
+
 struct btrfs_device {
 	struct list_head dev_list;
 	struct list_head dev_alloc_list;
 	struct btrfs_fs_devices *fs_devices;
 	struct btrfs_root *dev_root;
-	struct bio *pending_bios;
-	struct bio *pending_bio_tail;
+
+	/* regular prio bios */
+	struct btrfs_pending_bios pending_bios;
+	/* WRITE_SYNC bios */
+	struct btrfs_pending_bios pending_sync_bios;
+
 	int running_pending;
 	u64 generation;
 
@@ -52,6 +61,9 @@ struct btrfs_device {
 	/* size of the device */
 	u64 total_bytes;
 
+	/* size of the disk */
+	u64 disk_total_bytes;
+
 	/* bytes used */
 	u64 bytes_used;