fedora kernel: b920e9b748c595f970bf80ede7832d39f8d567dav3.4.1-2

Signed-off-by: Anton Arapov <anton@redhat.com>

539 files changed, 26539 insertions, 13592 deletions

diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index 1964f98e74b..b85efa77394 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -594,21 +594,21 @@ static int __init init_v9fs(void)
 	int err;
 	pr_info("Installing v9fs 9p2000 file system support\n");
 	/* TODO: Setup list of registered trasnport modules */
-	err = register_filesystem(&v9fs_fs_type);
-	if (err < 0) {
-		pr_err("Failed to register filesystem\n");
-		return err;
-	}
 
 	err = v9fs_cache_register();
 	if (err < 0) {
 		pr_err("Failed to register v9fs for caching\n");
-		goto out_fs_unreg;
+		return err;
 	}
 
 	err = v9fs_sysfs_init();
 	if (err < 0) {
 		pr_err("Failed to register with sysfs\n");
+		goto out_cache;
+	}
+	err = register_filesystem(&v9fs_fs_type);
+	if (err < 0) {
+		pr_err("Failed to register filesystem\n");
 		goto out_sysfs_cleanup;
 	}
 
@@ -617,8 +617,8 @@ static int __init init_v9fs(void)
 out_sysfs_cleanup:
 	v9fs_sysfs_cleanup();
 
-out_fs_unreg:
-	unregister_filesystem(&v9fs_fs_type);
+out_cache:
+	v9fs_cache_unregister();
 
 	return err;
 }
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 7b0cd87b07c..8c92a9ba833 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -155,9 +155,8 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 		goto release_sb;
 	}
 
-	root = d_alloc_root(inode);
+	root = d_make_root(inode);
 	if (!root) {
-		iput(inode);
 		retval = -ENOMEM;
 		goto release_sb;
 	}
@@ -260,7 +259,7 @@ static int v9fs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	if (v9fs_proto_dotl(v9ses)) {
 		res = p9_client_statfs(fid, &rs);
 		if (res == 0) {
-			buf->f_type = V9FS_MAGIC;
+			buf->f_type = rs.type;
 			buf->f_bsize = rs.bsize;
 			buf->f_blocks = rs.blocks;
 			buf->f_bfree = rs.bfree;
diff --git a/fs/Kconfig b/fs/Kconfig
index d621f02a3f9..f95ae3a027f 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -4,6 +4,10 @@
 
 menu "File systems"
 
+# Use unaligned word dcache accesses
+config DCACHE_WORD_ACCESS
+       bool
+
 if BLOCK
 
 source "fs/ext2/Kconfig"
@@ -210,6 +214,7 @@ source "fs/minix/Kconfig"
 source "fs/omfs/Kconfig"
 source "fs/hpfs/Kconfig"
 source "fs/qnx4/Kconfig"
+source "fs/qnx6/Kconfig"
 source "fs/romfs/Kconfig"
 source "fs/pstore/Kconfig"
 source "fs/sysv/Kconfig"
diff --git a/fs/Makefile b/fs/Makefile
index 93804d4d66e..2fb97793467 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -102,6 +102,7 @@ obj-$(CONFIG_UBIFS_FS)		+= ubifs/
 obj-$(CONFIG_AFFS_FS)		+= affs/
 obj-$(CONFIG_ROMFS_FS)		+= romfs/
 obj-$(CONFIG_QNX4FS_FS)		+= qnx4/
+obj-$(CONFIG_QNX6FS_FS)		+= qnx6/
 obj-$(CONFIG_AUTOFS4_FS)	+= autofs4/
 obj-$(CONFIG_ADFS_FS)		+= adfs/
 obj-$(CONFIG_FUSE_FS)		+= fuse/
diff --git a/fs/adfs/super.c b/fs/adfs/super.c
index 8e3b36ace30..06fdcc9382c 100644
--- a/fs/adfs/super.c
+++ b/fs/adfs/super.c
@@ -483,10 +483,9 @@ static int adfs_fill_super(struct super_block *sb, void *data, int silent)
 
 	sb->s_d_op = &adfs_dentry_operations;
 	root = adfs_iget(sb, &root_obj);
-	sb->s_root = d_alloc_root(root);
+	sb->s_root = d_make_root(root);
 	if (!sb->s_root) {
 		int i;
-		iput(root);
 		for (i = 0; i < asb->s_map_size; i++)
 			brelse(asb->s_map[i].dm_bh);
 		kfree(asb->s_map);
diff --git a/fs/affs/super.c b/fs/affs/super.c
index 8ba73fed796..0782653a05a 100644
--- a/fs/affs/super.c
+++ b/fs/affs/super.c
@@ -473,7 +473,7 @@ got_root:
 	root_inode = affs_iget(sb, root_block);
 	if (IS_ERR(root_inode)) {
 		ret = PTR_ERR(root_inode);
-		goto out_error_noinode;
+		goto out_error;
 	}
 
 	if (AFFS_SB(sb)->s_flags & SF_INTL)
@@ -481,7 +481,7 @@ got_root:
 	else
 		sb->s_d_op = &affs_dentry_operations;
 
-	sb->s_root = d_alloc_root(root_inode);
+	sb->s_root = d_make_root(root_inode);
 	if (!sb->s_root) {
 		printk(KERN_ERR "AFFS: Get root inode failed\n");
 		goto out_error;
@@ -494,9 +494,6 @@ got_root:
 	 * Begin the cascaded cleanup ...
 	 */
 out_error:
-	if (root_inode)
-		iput(root_inode);
-out_error_noinode:
 	kfree(sbi->s_bitmap);
 	affs_brelse(root_bh);
 	kfree(sbi->s_prefix);
diff --git a/fs/afs/file.c b/fs/afs/file.c
index 14d89fa58fe..8f6e9234d56 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -251,7 +251,7 @@ static int afs_readpages(struct file *file, struct address_space *mapping,
 	ASSERT(key != NULL);
 
 	vnode = AFS_FS_I(mapping->host);
-	if (vnode->flags & AFS_VNODE_DELETED) {
+	if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) {
 		_leave(" = -ESTALE");
 		return -ESTALE;
 	}
diff --git a/fs/afs/fsclient.c b/fs/afs/fsclient.c
index 2f213d109c2..b960ff05ea0 100644
--- a/fs/afs/fsclient.c
+++ b/fs/afs/fsclient.c
@@ -365,10 +365,10 @@ static int afs_deliver_fs_fetch_data(struct afs_call *call,
 		_debug("extract data");
 		if (call->count > 0) {
 			page = call->reply3;
-			buffer = kmap_atomic(page, KM_USER0);
+			buffer = kmap_atomic(page);
 			ret = afs_extract_data(call, skb, last, buffer,
 					       call->count);
-			kunmap_atomic(buffer, KM_USER0);
+			kunmap_atomic(buffer);
 			switch (ret) {
 			case 0:		break;
 			case -EAGAIN:	return 0;
@@ -411,9 +411,9 @@ static int afs_deliver_fs_fetch_data(struct afs_call *call,
 	if (call->count < PAGE_SIZE) {
 		_debug("clear");
 		page = call->reply3;
-		buffer = kmap_atomic(page, KM_USER0);
+		buffer = kmap_atomic(page);
 		memset(buffer + call->count, 0, PAGE_SIZE - call->count);
-		kunmap_atomic(buffer, KM_USER0);
+		kunmap_atomic(buffer);
 	}
 
 	_leave(" = 0 [done]");
diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c
index 8f4ce2658b7..298cf8919ec 100644
--- a/fs/afs/mntpt.c
+++ b/fs/afs/mntpt.c
@@ -200,9 +200,9 @@ static struct vfsmount *afs_mntpt_do_automount(struct dentry *mntpt)
 		if (PageError(page))
 			goto error;
 
-		buf = kmap_atomic(page, KM_USER0);
+		buf = kmap_atomic(page);
 		memcpy(devname, buf, size);
-		kunmap_atomic(buf, KM_USER0);
+		kunmap_atomic(buf);
 		page_cache_release(page);
 		page = NULL;
 	}
diff --git a/fs/afs/super.c b/fs/afs/super.c
index 983ec59fc80..f02b31e7e64 100644
--- a/fs/afs/super.c
+++ b/fs/afs/super.c
@@ -301,7 +301,6 @@ static int afs_fill_super(struct super_block *sb,
 {
 	struct afs_super_info *as = sb->s_fs_info;
 	struct afs_fid fid;
-	struct dentry *root = NULL;
 	struct inode *inode = NULL;
 	int ret;
 
@@ -327,18 +326,16 @@ static int afs_fill_super(struct super_block *sb,
 		set_bit(AFS_VNODE_AUTOCELL, &AFS_FS_I(inode)->flags);
 
 	ret = -ENOMEM;
-	root = d_alloc_root(inode);
-	if (!root)
+	sb->s_root = d_make_root(inode);
+	if (!sb->s_root)
 		goto error;
 
 	sb->s_d_op = &afs_fs_dentry_operations;
-	sb->s_root = root;
 
 	_leave(" = 0");
 	return 0;
 
 error:
-	iput(inode);
 	_leave(" = %d", ret);
 	return ret;
 }
diff --git a/fs/aio.c b/fs/aio.c
index b9d64d89a04..e7f2fad7b4c 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -13,7 +13,7 @@
 #include <linux/errno.h>
 #include <linux/time.h>
 #include <linux/aio_abi.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/syscalls.h>
 #include <linux/backing-dev.h>
 #include <linux/uio.h>
@@ -93,9 +93,8 @@ static void aio_free_ring(struct kioctx *ctx)
 		put_page(info->ring_pages[i]);
 
 	if (info->mmap_size) {
-		down_write(&ctx->mm->mmap_sem);
-		do_munmap(ctx->mm, info->mmap_base, info->mmap_size);
-		up_write(&ctx->mm->mmap_sem);
+		BUG_ON(ctx->mm != current->mm);
+		vm_munmap(info->mmap_base, info->mmap_size);
 	}
 
 	if (info->ring_pages && info->ring_pages != info->internal_pages)
@@ -160,7 +159,7 @@ static int aio_setup_ring(struct kioctx *ctx)
 
 	info->nr = nr_events;		/* trusted copy */
 
-	ring = kmap_atomic(info->ring_pages[0], KM_USER0);
+	ring = kmap_atomic(info->ring_pages[0]);
 	ring->nr = nr_events;	/* user copy */
 	ring->id = ctx->user_id;
 	ring->head = ring->tail = 0;
@@ -168,47 +167,38 @@ static int aio_setup_ring(struct kioctx *ctx)
 	ring->compat_features = AIO_RING_COMPAT_FEATURES;
 	ring->incompat_features = AIO_RING_INCOMPAT_FEATURES;
 	ring->header_length = sizeof(struct aio_ring);
-	kunmap_atomic(ring, KM_USER0);
+	kunmap_atomic(ring);
 
 	return 0;
 }
 
 
 /* aio_ring_event: returns a pointer to the event at the given index from
- * kmap_atomic(, km).  Release the pointer with put_aio_ring_event();
+ * kmap_atomic().  Release the pointer with put_aio_ring_event();
  */
 #define AIO_EVENTS_PER_PAGE	(PAGE_SIZE / sizeof(struct io_event))
 #define AIO_EVENTS_FIRST_PAGE	((PAGE_SIZE - sizeof(struct aio_ring)) / sizeof(struct io_event))
 #define AIO_EVENTS_OFFSET	(AIO_EVENTS_PER_PAGE - AIO_EVENTS_FIRST_PAGE)
 
-#define aio_ring_event(info, nr, km) ({					\
+#define aio_ring_event(info, nr) ({					\
 	unsigned pos = (nr) + AIO_EVENTS_OFFSET;			\
 	struct io_event *__event;					\
 	__event = kmap_atomic(						\
-			(info)->ring_pages[pos / AIO_EVENTS_PER_PAGE], km); \
+			(info)->ring_pages[pos / AIO_EVENTS_PER_PAGE]); \
 	__event += pos % AIO_EVENTS_PER_PAGE;				\
 	__event;							\
 })
 
-#define put_aio_ring_event(event, km) do {	\
+#define put_aio_ring_event(event) do {		\
 	struct io_event *__event = (event);	\
 	(void)__event;				\
-	kunmap_atomic((void *)((unsigned long)__event & PAGE_MASK), km); \
+	kunmap_atomic((void *)((unsigned long)__event & PAGE_MASK)); \
 } while(0)
 
 static void ctx_rcu_free(struct rcu_head *head)
 {
 	struct kioctx *ctx = container_of(head, struct kioctx, rcu_head);
-	unsigned nr_events = ctx->max_reqs;
-
 	kmem_cache_free(kioctx_cachep, ctx);
-
-	if (nr_events) {
-		spin_lock(&aio_nr_lock);
-		BUG_ON(aio_nr - nr_events > aio_nr);
-		aio_nr -= nr_events;
-		spin_unlock(&aio_nr_lock);
-	}
 }
 
 /* __put_ioctx
@@ -217,13 +207,19 @@ static void ctx_rcu_free(struct rcu_head *head)
  */
 static void __put_ioctx(struct kioctx *ctx)
 {
+	unsigned nr_events = ctx->max_reqs;
 	BUG_ON(ctx->reqs_active);
 
-	cancel_delayed_work(&ctx->wq);
-	cancel_work_sync(&ctx->wq.work);
+	cancel_delayed_work_sync(&ctx->wq);
 	aio_free_ring(ctx);
 	mmdrop(ctx->mm);
 	ctx->mm = NULL;
+	if (nr_events) {
+		spin_lock(&aio_nr_lock);
+		BUG_ON(aio_nr - nr_events > aio_nr);
+		aio_nr -= nr_events;
+		spin_unlock(&aio_nr_lock);
+	}
 	pr_debug("__put_ioctx: freeing %p\n", ctx);
 	call_rcu(&ctx->rcu_head, ctx_rcu_free);
 }
@@ -247,7 +243,7 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 {
 	struct mm_struct *mm;
 	struct kioctx *ctx;
-	int did_sync = 0;
+	int err = -ENOMEM;
 
 	/* Prevent overflows */
 	if ((nr_events > (0x10000000U / sizeof(struct io_event))) ||
@@ -256,7 +252,7 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 		return ERR_PTR(-EINVAL);
 	}
 
-	if ((unsigned long)nr_events > aio_max_nr)
+	if (!nr_events || (unsigned long)nr_events > aio_max_nr)
 		return ERR_PTR(-EAGAIN);
 
 	ctx = kmem_cache_zalloc(kioctx_cachep, GFP_KERNEL);
@@ -280,25 +276,14 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 		goto out_freectx;
 
 	/* limit the number of system wide aios */
-	do {
-		spin_lock_bh(&aio_nr_lock);
-		if (aio_nr + nr_events > aio_max_nr ||
-		    aio_nr + nr_events < aio_nr)
-			ctx->max_reqs = 0;
-		else
-			aio_nr += ctx->max_reqs;
-		spin_unlock_bh(&aio_nr_lock);
-		if (ctx->max_reqs || did_sync)
-			break;
-
-		/* wait for rcu callbacks to have completed before giving up */
-		synchronize_rcu();
-		did_sync = 1;
-		ctx->max_reqs = nr_events;
-	} while (1);
-
-	if (ctx->max_reqs == 0)
+	spin_lock(&aio_nr_lock);
+	if (aio_nr + nr_events > aio_max_nr ||
+	    aio_nr + nr_events < aio_nr) {
+		spin_unlock(&aio_nr_lock);
 		goto out_cleanup;
+	}
+	aio_nr += ctx->max_reqs;
+	spin_unlock(&aio_nr_lock);
 
 	/* now link into global list. */
 	spin_lock(&mm->ioctx_lock);
@@ -310,27 +295,27 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 	return ctx;
 
 out_cleanup:
-	__put_ioctx(ctx);
-	return ERR_PTR(-EAGAIN);
-
+	err = -EAGAIN;
+	aio_free_ring(ctx);
 out_freectx:
 	mmdrop(mm);
 	kmem_cache_free(kioctx_cachep, ctx);
-	ctx = ERR_PTR(-ENOMEM);
-
-	dprintk("aio: error allocating ioctx %p\n", ctx);
-	return ctx;
+	dprintk("aio: error allocating ioctx %d\n", err);
+	return ERR_PTR(err);
 }
 
-/* aio_cancel_all
+/* kill_ctx
  *	Cancels all outstanding aio requests on an aio context.  Used 
  *	when the processes owning a context have all exited to encourage 
  *	the rapid destruction of the kioctx.
  */
-static void aio_cancel_all(struct kioctx *ctx)
+static void kill_ctx(struct kioctx *ctx)
 {
 	int (*cancel)(struct kiocb *, struct io_event *);
+	struct task_struct *tsk = current;
+	DECLARE_WAITQUEUE(wait, tsk);
 	struct io_event res;
+
 	spin_lock_irq(&ctx->ctx_lock);
 	ctx->dead = 1;
 	while (!list_empty(&ctx->active_reqs)) {
@@ -346,15 +331,7 @@ static void aio_cancel_all(struct kioctx *ctx)
 			spin_lock_irq(&ctx->ctx_lock);
 		}
 	}
-	spin_unlock_irq(&ctx->ctx_lock);
-}
-
-static void wait_for_all_aios(struct kioctx *ctx)
-{
-	struct task_struct *tsk = current;
-	DECLARE_WAITQUEUE(wait, tsk);
 
-	spin_lock_irq(&ctx->ctx_lock);
 	if (!ctx->reqs_active)
 		goto out;
 
@@ -404,19 +381,24 @@ void exit_aio(struct mm_struct *mm)
 		ctx = hlist_entry(mm->ioctx_list.first, struct kioctx, list);
 		hlist_del_rcu(&ctx->list);
 
-		aio_cancel_all(ctx);
-
-		wait_for_all_aios(ctx);
-		/*
-		 * Ensure we don't leave the ctx on the aio_wq
-		 */
-		cancel_work_sync(&ctx->wq.work);
+		kill_ctx(ctx);
 
 		if (1 != atomic_read(&ctx->users))
 			printk(KERN_DEBUG
 				"exit_aio:ioctx still alive: %d %d %d\n",
 				atomic_read(&ctx->users), ctx->dead,
 				ctx->reqs_active);
+		/*
+		 * We don't need to bother with munmap() here -
+		 * exit_mmap(mm) is coming and it'll unmap everything.
+		 * Since aio_free_ring() uses non-zero ->mmap_size
+		 * as indicator that it needs to unmap the area,
+		 * just set it to 0; aio_free_ring() is the only
+		 * place that uses ->mmap_size, so it's safe.
+		 * That way we get all munmap done to current->mm -
+		 * all other callers have ctx->mm == current->mm.
+		 */
+		ctx->ring_info.mmap_size = 0;
 		put_ioctx(ctx);
 	}
 }
@@ -920,7 +902,7 @@ static void aio_kick_handler(struct work_struct *work)
  	unuse_mm(mm);
 	set_fs(oldfs);
 	/*
-	 * we're in a worker thread already, don't use queue_delayed_work,
+	 * we're in a worker thread already; no point using non-zero delay
 	 */
 	if (requeue)
 		queue_delayed_work(aio_wq, &ctx->wq, 0);
@@ -1019,10 +1001,10 @@ int aio_complete(struct kiocb *iocb, long res, long res2)
 	if (kiocbIsCancelled(iocb))
 		goto put_rq;
 
-	ring = kmap_atomic(info->ring_pages[0], KM_IRQ1);
+	ring = kmap_atomic(info->ring_pages[0]);
 
 	tail = info->tail;
-	event = aio_ring_event(info, tail, KM_IRQ0);
+	event = aio_ring_event(info, tail);
 	if (++tail >= info->nr)
 		tail = 0;
 
@@ -1043,8 +1025,8 @@ int aio_complete(struct kiocb *iocb, long res, long res2)
 	info->tail = tail;
 	ring->tail = tail;
 
-	put_aio_ring_event(event, KM_IRQ0);
-	kunmap_atomic(ring, KM_IRQ1);
+	put_aio_ring_event(event);
+	kunmap_atomic(ring);
 
 	pr_debug("added to ring %p at [%lu]\n", iocb, tail);
 
@@ -1089,7 +1071,7 @@ static int aio_read_evt(struct kioctx *ioctx, struct io_event *ent)
 	unsigned long head;
 	int ret = 0;
 
-	ring = kmap_atomic(info->ring_pages[0], KM_USER0);
+	ring = kmap_atomic(info->ring_pages[0]);
 	dprintk("in aio_read_evt h%lu t%lu m%lu\n",
 		 (unsigned long)ring->head, (unsigned long)ring->tail,
 		 (unsigned long)ring->nr);
@@ -1101,18 +1083,18 @@ static int aio_read_evt(struct kioctx *ioctx, struct io_event *ent)
 
 	head = ring->head % info->nr;
 	if (head != ring->tail) {
-		struct io_event *evp = aio_ring_event(info, head, KM_USER1);
+		struct io_event *evp = aio_ring_event(info, head);
 		*ent = *evp;
 		head = (head + 1) % info->nr;
 		smp_mb(); /* finish reading the event before updatng the head */
 		ring->head = head;
 		ret = 1;
-		put_aio_ring_event(evp, KM_USER1);
+		put_aio_ring_event(evp);
 	}
 	spin_unlock(&info->ring_lock);
 
 out:
-	kunmap_atomic(ring, KM_USER0);
+	kunmap_atomic(ring);
 	dprintk("leaving aio_read_evt: %d  h%lu t%lu\n", ret,
 		 (unsigned long)ring->head, (unsigned long)ring->tail);
 	return ret;
@@ -1290,8 +1272,7 @@ static void io_destroy(struct kioctx *ioctx)
 	if (likely(!was_dead))
 		put_ioctx(ioctx);	/* twice for the list */
 
-	aio_cancel_all(ioctx);
-	wait_for_all_aios(ioctx);
+	kill_ctx(ioctx);
 
 	/*
 	 * Wake up any waiters.  The setting of ctx->dead must be seen
@@ -1299,7 +1280,6 @@ static void io_destroy(struct kioctx *ioctx)
 	 * locking done by the above calls to ensure this consistency.
 	 */
 	wake_up_all(&ioctx->wait);
-	put_ioctx(ioctx);	/* once for the lookup */
 }
 
 /* sys_io_setup:
@@ -1336,11 +1316,9 @@ SYSCALL_DEFINE2(io_setup, unsigned, nr_events, aio_context_t __user *, ctxp)
 	ret = PTR_ERR(ioctx);
 	if (!IS_ERR(ioctx)) {
 		ret = put_user(ioctx->user_id, ctxp);
-		if (!ret) {
-			put_ioctx(ioctx);
-			return 0;
-		}
-		io_destroy(ioctx);
+		if (ret)
+			io_destroy(ioctx);
+		put_ioctx(ioctx);
 	}
 
 out:
@@ -1358,6 +1336,7 @@ SYSCALL_DEFINE1(io_destroy, aio_context_t, ctx)
 	struct kioctx *ioctx = lookup_ioctx(ctx);
 	if (likely(NULL != ioctx)) {
 		io_destroy(ioctx);
+		put_ioctx(ioctx);
 		return 0;
 	}
 	pr_debug("EINVAL: io_destroy: invalid context id\n");
@@ -1477,6 +1456,10 @@ static ssize_t aio_setup_vectored_rw(int type, struct kiocb *kiocb, bool compat)
 	if (ret < 0)
 		goto out;
 
+	ret = rw_verify_area(type, kiocb->ki_filp, &kiocb->ki_pos, ret);
+	if (ret < 0)
+		goto out;
+
 	kiocb->ki_nr_segs = kiocb->ki_nbytes;
 	kiocb->ki_cur_seg = 0;
 	/* ki_nbytes/left now reflect bytes instead of segs */
@@ -1488,11 +1471,17 @@ out:
 	return ret;
 }
 
-static ssize_t aio_setup_single_vector(struct kiocb *kiocb)
+static ssize_t aio_setup_single_vector(int type, struct file * file, struct kiocb *kiocb)
 {
+	int bytes;
+
+	bytes = rw_verify_area(type, file, &kiocb->ki_pos, kiocb->ki_left);
+	if (bytes < 0)
+		return bytes;
+
 	kiocb->ki_iovec = &kiocb->ki_inline_vec;
 	kiocb->ki_iovec->iov_base = kiocb->ki_buf;
-	kiocb->ki_iovec->iov_len = kiocb->ki_left;
+	kiocb->ki_iovec->iov_len = bytes;
 	kiocb->ki_nr_segs = 1;
 	kiocb->ki_cur_seg = 0;
 	return 0;
@@ -1517,10 +1506,7 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
 		if (unlikely(!access_ok(VERIFY_WRITE, kiocb->ki_buf,
 			kiocb->ki_left)))
 			break;
-		ret = security_file_permission(file, MAY_READ);
-		if (unlikely(ret))
-			break;
-		ret = aio_setup_single_vector(kiocb);
+		ret = aio_setup_single_vector(READ, file, kiocb);
 		if (ret)
 			break;
 		ret = -EINVAL;
@@ -1535,10 +1521,7 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
 		if (unlikely(!access_ok(VERIFY_READ, kiocb->ki_buf,
 			kiocb->ki_left)))
 			break;
-		ret = security_file_permission(file, MAY_WRITE);
-		if (unlikely(ret))
-			break;
-		ret = aio_setup_single_vector(kiocb);
+		ret = aio_setup_single_vector(WRITE, file, kiocb);
 		if (ret)
 			break;
 		ret = -EINVAL;
@@ -1549,9 +1532,6 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
 		ret = -EBADF;
 		if (unlikely(!(file->f_mode & FMODE_READ)))
 			break;
-		ret = security_file_permission(file, MAY_READ);
-		if (unlikely(ret))
-			break;
 		ret = aio_setup_vectored_rw(READ, kiocb, compat);
 		if (ret)
 			break;
@@ -1563,9 +1543,6 @@ static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
 		ret = -EBADF;
 		if (unlikely(!(file->f_mode & FMODE_WRITE)))
 			break;
-		ret = security_file_permission(file, MAY_WRITE);
-		if (unlikely(ret))
-			break;
 		ret = aio_setup_vectored_rw(WRITE, kiocb, compat);
 		if (ret)
 			break;
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index f11e43ed907..28d39fb84ae 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c
@@ -39,19 +39,6 @@ static const struct dentry_operations anon_inodefs_dentry_operations = {
 	.d_dname	= anon_inodefs_dname,
 };
 
-static struct dentry *anon_inodefs_mount(struct file_system_type *fs_type,
-				int flags, const char *dev_name, void *data)
-{
-	return mount_pseudo(fs_type, "anon_inode:", NULL,
-			&anon_inodefs_dentry_operations, ANON_INODE_FS_MAGIC);
-}
-
-static struct file_system_type anon_inode_fs_type = {
-	.name		= "anon_inodefs",
-	.mount		= anon_inodefs_mount,
-	.kill_sb	= kill_anon_super,
-};
-
 /*
  * nop .set_page_dirty method so that people can use .page_mkwrite on
  * anon inodes.
@@ -65,6 +52,62 @@ static const struct address_space_operations anon_aops = {
 	.set_page_dirty = anon_set_page_dirty,
 };
 
+/*
+ * A single inode exists for all anon_inode files. Contrary to pipes,
+ * anon_inode inodes have no associated per-instance data, so we need
+ * only allocate one of them.
+ */
+static struct inode *anon_inode_mkinode(struct super_block *s)
+{
+	struct inode *inode = new_inode_pseudo(s);
+
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+
+	inode->i_ino = get_next_ino();
+	inode->i_fop = &anon_inode_fops;
+
+	inode->i_mapping->a_ops = &anon_aops;
+
+	/*
+	 * Mark the inode dirty from the very beginning,
+	 * that way it will never be moved to the dirty
+	 * list because mark_inode_dirty() will think
+	 * that it already _is_ on the dirty list.
+	 */
+	inode->i_state = I_DIRTY;
+	inode->i_mode = S_IRUSR | S_IWUSR;
+	inode->i_uid = current_fsuid();
+	inode->i_gid = current_fsgid();
+	inode->i_flags |= S_PRIVATE;
+	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+	return inode;
+}
+
+static struct dentry *anon_inodefs_mount(struct file_system_type *fs_type,
+				int flags, const char *dev_name, void *data)
+{
+	struct dentry *root;
+	root = mount_pseudo(fs_type, "anon_inode:", NULL,
+			&anon_inodefs_dentry_operations, ANON_INODE_FS_MAGIC);
+	if (!IS_ERR(root)) {
+		struct super_block *s = root->d_sb;
+		anon_inode_inode = anon_inode_mkinode(s);
+		if (IS_ERR(anon_inode_inode)) {
+			dput(root);
+			deactivate_locked_super(s);
+			root = ERR_CAST(anon_inode_inode);
+		}
+	}
+	return root;
+}
+
+static struct file_system_type anon_inode_fs_type = {
+	.name		= "anon_inodefs",
+	.mount		= anon_inodefs_mount,
+	.kill_sb	= kill_anon_super,
+};
+
 /**
  * anon_inode_getfile - creates a new file instance by hooking it up to an
  *                      anonymous inode, and a dentry that describe the "class"
@@ -180,38 +223,6 @@ err_put_unused_fd:
 }
 EXPORT_SYMBOL_GPL(anon_inode_getfd);
 
-/*
- * A single inode exists for all anon_inode files. Contrary to pipes,
- * anon_inode inodes have no associated per-instance data, so we need
- * only allocate one of them.
- */
-static struct inode *anon_inode_mkinode(void)
-{
-	struct inode *inode = new_inode_pseudo(anon_inode_mnt->mnt_sb);
-
-	if (!inode)
-		return ERR_PTR(-ENOMEM);
-
-	inode->i_ino = get_next_ino();
-	inode->i_fop = &anon_inode_fops;
-
-	inode->i_mapping->a_ops = &anon_aops;
-
-	/*
-	 * Mark the inode dirty from the very beginning,
-	 * that way it will never be moved to the dirty
-	 * list because mark_inode_dirty() will think
-	 * that it already _is_ on the dirty list.
-	 */
-	inode->i_state = I_DIRTY;
-	inode->i_mode = S_IRUSR | S_IWUSR;
-	inode->i_uid = current_fsuid();
-	inode->i_gid = current_fsgid();
-	inode->i_flags |= S_PRIVATE;
-	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-	return inode;
-}
-
 static int __init anon_inode_init(void)
 {
 	int error;
@@ -224,16 +235,8 @@ static int __init anon_inode_init(void)
 		error = PTR_ERR(anon_inode_mnt);
 		goto err_unregister_filesystem;
 	}
-	anon_inode_inode = anon_inode_mkinode();
-	if (IS_ERR(anon_inode_inode)) {
-		error = PTR_ERR(anon_inode_inode);
-		goto err_mntput;
-	}
-
 	return 0;
 
-err_mntput:
-	kern_unmount(anon_inode_mnt);
 err_unregister_filesystem:
 	unregister_filesystem(&anon_inode_fs_type);
 err_exit:
diff --git a/fs/attr.c b/fs/attr.c
index 95053ad8abc..73f69a6ce9e 100644
--- a/fs/attr.c
+++ b/fs/attr.c
@@ -5,7 +5,7 @@
  *  changes by Thomas Schoebel-Theuer
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/time.h>
 #include <linux/mm.h>
 #include <linux/string.h>
diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c
index d06d95a5c63..aa9103f8f01 100644
--- a/fs/autofs4/dev-ioctl.c
+++ b/fs/autofs4/dev-ioctl.c
@@ -230,7 +230,7 @@ static void autofs_dev_ioctl_fd_install(unsigned int fd, struct file *file)
 	fdt = files_fdtable(files);
 	BUG_ON(fdt->fd[fd] != NULL);
 	rcu_assign_pointer(fdt->fd[fd], file);
-	FD_SET(fd, fdt->close_on_exec);
+	__set_close_on_exec(fd, fdt);
 	spin_unlock(&files->file_lock);
 }
 
diff --git a/fs/autofs4/init.c b/fs/autofs4/init.c
index c038727b405..cddc74b9cdb 100644
--- a/fs/autofs4/init.c
+++ b/fs/autofs4/init.c
@@ -31,11 +31,11 @@ static int __init init_autofs4_fs(void)
 {
 	int err;
 
+	autofs_dev_ioctl_init();
+
 	err = register_filesystem(&autofs_fs_type);
 	if (err)
-		return err;
-
-	autofs_dev_ioctl_init();
+		autofs_dev_ioctl_exit();
 
 	return err;
 }
diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c
index 9ef53a68ea0..6e488ebe778 100644
--- a/fs/autofs4/inode.c
+++ b/fs/autofs4/inode.c
@@ -245,12 +245,9 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	if (!ino)
 		goto fail_free;
 	root_inode = autofs4_get_inode(s, S_IFDIR | 0755);
-	if (!root_inode)
-		goto fail_ino;
-
-	root = d_alloc_root(root_inode);
+	root = d_make_root(root_inode);
 	if (!root)
-		goto fail_iput;
+		goto fail_ino;
 	pipe = NULL;
 
 	root->d_fsdata = ino;
@@ -315,9 +312,6 @@ fail_fput:
 fail_dput:
 	dput(root);
 	goto fail_free;
-fail_iput:
-	printk("autofs: get root dentry failed\n");
-	iput(root_inode);
 fail_ino:
 	kfree(ino);
 fail_free:
diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c
index f624cd0ff84..da8876d38a7 100644
--- a/fs/autofs4/waitq.c
+++ b/fs/autofs4/waitq.c
@@ -91,7 +91,7 @@ static int autofs4_write(struct autofs_sb_info *sbi,
 
 	return (bytes > 0);
 }
-
+	
 static void autofs4_notify_daemon(struct autofs_sb_info *sbi,
 				 struct autofs_wait_queue *wq,
 				 int type)
diff --git a/fs/bad_inode.c b/fs/bad_inode.c
index 22e9a78872f..37268c5bb98 100644
--- a/fs/bad_inode.c
+++ b/fs/bad_inode.c
@@ -9,7 +9,7 @@
  */
 
 #include <linux/fs.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/stat.h>
 #include <linux/time.h>
 #include <linux/namei.h>
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index 6e6d536767f..e18da23d42b 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -852,9 +852,8 @@ befs_fill_super(struct super_block *sb, void *data, int silent)
 		ret = PTR_ERR(root);
 		goto unacquire_priv_sbp;
 	}
-	sb->s_root = d_alloc_root(root);
+	sb->s_root = d_make_root(root);
 	if (!sb->s_root) {
-		iput(root);
 		befs_error(sb, "get root inode failed");
 		goto unacquire_priv_sbp;
 	}
diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c
index b0391bc402b..e23dc7c8b88 100644
--- a/fs/bfs/inode.c
+++ b/fs/bfs/inode.c
@@ -367,9 +367,8 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent)
 		ret = PTR_ERR(inode);
 		goto out2;
 	}
-	s->s_root = d_alloc_root(inode);
+	s->s_root = d_make_root(inode);
 	if (!s->s_root) {
-		iput(inode);
 		ret = -ENOMEM;
 		goto out2;
 	}
diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c
index 1ff94054d35..d146e181d10 100644
--- a/fs/binfmt_aout.c
+++ b/fs/binfmt_aout.c
@@ -26,7 +26,6 @@
 #include <linux/coredump.h>
 #include <linux/slab.h>
 
-#include <asm/system.h>
 #include <asm/uaccess.h>
 #include <asm/cacheflush.h>
 #include <asm/a.out-core.h>
@@ -51,9 +50,7 @@ static int set_brk(unsigned long start, unsigned long end)
 	end = PAGE_ALIGN(end);
 	if (end > start) {
 		unsigned long addr;
-		down_write(&current->mm->mmap_sem);
-		addr = do_brk(start, end - start);
-		up_write(&current->mm->mmap_sem);
+		addr = vm_brk(start, end - start);
 		if (BAD_ADDR(addr))
 			return addr;
 	}
@@ -267,7 +264,6 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 	}
 
 	install_exec_creds(bprm);
- 	current->flags &= ~PF_FORKNOEXEC;
 
 	if (N_MAGIC(ex) == OMAGIC) {
 		unsigned long text_addr, map_size;
@@ -282,9 +278,7 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 		pos = 32;
 		map_size = ex.a_text+ex.a_data;
 #endif
-		down_write(&current->mm->mmap_sem);
-		error = do_brk(text_addr & PAGE_MASK, map_size);
-		up_write(&current->mm->mmap_sem);
+		error = vm_brk(text_addr & PAGE_MASK, map_size);
 		if (error != (text_addr & PAGE_MASK)) {
 			send_sig(SIGKILL, current, 0);
 			return error;
@@ -315,9 +309,7 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 
 		if (!bprm->file->f_op->mmap||((fd_offset & ~PAGE_MASK) != 0)) {
 			loff_t pos = fd_offset;
-			down_write(&current->mm->mmap_sem);
-			do_brk(N_TXTADDR(ex), ex.a_text+ex.a_data);
-			up_write(&current->mm->mmap_sem);
+			vm_brk(N_TXTADDR(ex), ex.a_text+ex.a_data);
 			bprm->file->f_op->read(bprm->file,
 					(char __user *)N_TXTADDR(ex),
 					ex.a_text+ex.a_data, &pos);
@@ -327,24 +319,20 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 			goto beyond_if;
 		}
 
-		down_write(&current->mm->mmap_sem);
-		error = do_mmap(bprm->file, N_TXTADDR(ex), ex.a_text,
+		error = vm_mmap(bprm->file, N_TXTADDR(ex), ex.a_text,
 			PROT_READ | PROT_EXEC,
 			MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE,
 			fd_offset);
-		up_write(&current->mm->mmap_sem);
 
 		if (error != N_TXTADDR(ex)) {
 			send_sig(SIGKILL, current, 0);
 			return error;
 		}
 
-		down_write(&current->mm->mmap_sem);
- 		error = do_mmap(bprm->file, N_DATADDR(ex), ex.a_data,
+		error = vm_mmap(bprm->file, N_DATADDR(ex), ex.a_data,
 				PROT_READ | PROT_WRITE | PROT_EXEC,
 				MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE,
 				fd_offset + ex.a_text);
-		up_write(&current->mm->mmap_sem);
 		if (error != N_DATADDR(ex)) {
 			send_sig(SIGKILL, current, 0);
 			return error;
@@ -414,9 +402,7 @@ static int load_aout_library(struct file *file)
 			       "N_TXTOFF is not page aligned. Please convert library: %s\n",
 			       file->f_path.dentry->d_name.name);
 		}
-		down_write(&current->mm->mmap_sem);
-		do_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss);
-		up_write(&current->mm->mmap_sem);
+		vm_brk(start_addr, ex.a_text + ex.a_data + ex.a_bss);
 		
 		file->f_op->read(file, (char __user *)start_addr,
 			ex.a_text + ex.a_data, &pos);
@@ -427,12 +413,10 @@ static int load_aout_library(struct file *file)
 		goto out;
 	}
 	/* Now use mmap to map the library into memory. */
-	down_write(&current->mm->mmap_sem);
-	error = do_mmap(file, start_addr, ex.a_text + ex.a_data,
+	error = vm_mmap(file, start_addr, ex.a_text + ex.a_data,
 			PROT_READ | PROT_WRITE | PROT_EXEC,
 			MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE,
 			N_TXTOFF(ex));
-	up_write(&current->mm->mmap_sem);
 	retval = error;
 	if (error != start_addr)
 		goto out;
@@ -440,9 +424,7 @@ static int load_aout_library(struct file *file)
 	len = PAGE_ALIGN(ex.a_text + ex.a_data);
 	bss = ex.a_text + ex.a_data + ex.a_bss;
 	if (bss > len) {
-		down_write(&current->mm->mmap_sem);
-		error = do_brk(start_addr + len, bss - len);
-		up_write(&current->mm->mmap_sem);
+		error = vm_brk(start_addr + len, bss - len);
 		retval = error;
 		if (error != start_addr + len)
 			goto out;
@@ -454,7 +436,8 @@ out:
 
 static int __init init_aout_binfmt(void)
 {
-	return register_binfmt(&aout_format);
+	register_binfmt(&aout_format);
+	return 0;
 }
 
 static void __exit exit_aout_binfmt(void)
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index fa7483fc4a4..16f73541707 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -35,6 +35,7 @@
 #include <asm/uaccess.h>
 #include <asm/param.h>
 #include <asm/page.h>
+#include <asm/exec.h>
 
 static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs);
 static int load_elf_library(struct file *);
@@ -81,9 +82,7 @@ static int set_brk(unsigned long start, unsigned long end)
 	end = ELF_PAGEALIGN(end);
 	if (end > start) {
 		unsigned long addr;
-		down_write(&current->mm->mmap_sem);
-		addr = do_brk(start, end - start);
-		up_write(&current->mm->mmap_sem);
+		addr = vm_brk(start, end - start);
 		if (BAD_ADDR(addr))
 			return addr;
 	}
@@ -513,9 +512,7 @@ static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex,
 		elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1);
 
 		/* Map the last of the bss segment */
-		down_write(&current->mm->mmap_sem);
-		error = do_brk(elf_bss, last_bss - elf_bss);
-		up_write(&current->mm->mmap_sem);
+		error = vm_brk(elf_bss, last_bss - elf_bss);
 		if (BAD_ADDR(error))
 			goto out_close;
 	}
@@ -711,17 +708,7 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs)
 	if (retval)
 		goto out_free_dentry;
 
-#ifdef CONFIG_X86_32
-	/*
-	 * Turn off the CS limit completely if exec-shield disabled or
-	 * NX active:
-	 */
-	if (disable_nx || executable_stack != EXSTACK_DISABLE_X || (__supported_pte_mask & _PAGE_NX))
-		arch_add_exec_range(current->mm, -1);
-#endif
-
 	/* OK, This is the point of no return */
-	current->flags &= ~PF_FORKNOEXEC;
 	current->mm->def_flags = def_flags;
 
 	/* Do this immediately, since STACK_TOP as used in setup_arg_pages
@@ -943,7 +930,6 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs)
 #endif /* ARCH_HAS_SETUP_ADDITIONAL_PAGES */
 
 	install_exec_creds(bprm);
-	current->flags &= ~PF_FORKNOEXEC;
 	retval = create_elf_tables(bprm, &loc->elf_ex,
 			  load_addr, interp_load_addr);
 	if (retval < 0) {
@@ -972,10 +958,8 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs)
 		   and some applications "depend" upon this behavior.
 		   Since we do not have the power to recompile these, we
 		   emulate the SVr4 behavior. Sigh. */
-		down_write(&current->mm->mmap_sem);
-		error = do_mmap(NULL, 0, PAGE_SIZE, PROT_READ | PROT_EXEC,
+		error = vm_mmap(NULL, 0, PAGE_SIZE, PROT_READ | PROT_EXEC,
 				MAP_FIXED | MAP_PRIVATE, 0);
-		up_write(&current->mm->mmap_sem);
 	}
 
 #ifdef ELF_PLAT_INIT
@@ -1060,8 +1044,7 @@ static int load_elf_library(struct file *file)
 		eppnt++;
 
 	/* Now use mmap to map the library into memory. */
-	down_write(&current->mm->mmap_sem);
-	error = do_mmap(file,
+	error = vm_mmap(file,
 			ELF_PAGESTART(eppnt->p_vaddr),
 			(eppnt->p_filesz +
 			 ELF_PAGEOFFSET(eppnt->p_vaddr)),
@@ -1069,7 +1052,6 @@ static int load_elf_library(struct file *file)
 			MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE,
 			(eppnt->p_offset -
 			 ELF_PAGEOFFSET(eppnt->p_vaddr)));
-	up_write(&current->mm->mmap_sem);
 	if (error != ELF_PAGESTART(eppnt->p_vaddr))
 		goto out_free_ph;
 
@@ -1082,11 +1064,8 @@ static int load_elf_library(struct file *file)
 	len = ELF_PAGESTART(eppnt->p_filesz + eppnt->p_vaddr +
 			    ELF_MIN_ALIGN - 1);
 	bss = eppnt->p_memsz + eppnt->p_vaddr;
-	if (bss > len) {
-		down_write(&current->mm->mmap_sem);
-		do_brk(len, bss - len);
-		up_write(&current->mm->mmap_sem);
-	}
+	if (bss > len)
+		vm_brk(len, bss - len);
 	error = 0;
 
 out_free_ph:
@@ -1104,6 +1083,29 @@ out:
  */
 
 /*
+ * The purpose of always_dump_vma() is to make sure that special kernel mappings
+ * that are useful for post-mortem analysis are included in every core dump.
+ * In that way we ensure that the core dump is fully interpretable later
+ * without matching up the same kernel and hardware config to see what PC values
+ * meant. These special mappings include - vDSO, vsyscall, and other
+ * architecture specific mappings
+ */
+static bool always_dump_vma(struct vm_area_struct *vma)
+{
+	/* Any vsyscall mappings? */
+	if (vma == get_gate_vma(vma->vm_mm))
+		return true;
+	/*
+	 * arch_vma_name() returns non-NULL for special architecture mappings,
+	 * such as vDSO sections.
+	 */
+	if (arch_vma_name(vma))
+		return true;
+
+	return false;
+}
+
+/*
  * Decide what to dump of a segment, part, all or none.
  */
 static unsigned long vma_dump_size(struct vm_area_struct *vma,
@@ -1111,10 +1113,13 @@ static unsigned long vma_dump_size(struct vm_area_struct *vma,
 {
 #define FILTER(type)	(mm_flags & (1UL << MMF_DUMP_##type))
 
-	/* The vma can be set up to tell us the answer directly.  */
-	if (vma->vm_flags & VM_ALWAYSDUMP)
+	/* always dump the vdso and vsyscall sections */
+	if (always_dump_vma(vma))
 		goto whole;
 
+	if (vma->vm_flags & VM_NODUMP)
+		return 0;
+
 	/* Hugetlb memory check */
 	if (vma->vm_flags & VM_HUGETLB) {
 		if ((vma->vm_flags & VM_SHARED) && FILTER(HUGETLB_SHARED))
@@ -1399,6 +1404,22 @@ static void do_thread_regset_writeback(struct task_struct *task,
 		regset->writeback(task, regset, 1);
 }
 
+#ifndef PR_REG_SIZE
+#define PR_REG_SIZE(S) sizeof(S)
+#endif
+
+#ifndef PRSTATUS_SIZE
+#define PRSTATUS_SIZE(S) sizeof(S)
+#endif
+
+#ifndef PR_REG_PTR
+#define PR_REG_PTR(S) (&((S)->pr_reg))
+#endif
+
+#ifndef SET_PR_FPVALID
+#define SET_PR_FPVALID(S, V) ((S)->pr_fpvalid = (V))
+#endif
+
 static int fill_thread_core_info(struct elf_thread_core_info *t,
 				 const struct user_regset_view *view,
 				 long signr, size_t *total)
@@ -1413,11 +1434,11 @@ static int fill_thread_core_info(struct elf_thread_core_info *t,
 	 */
 	fill_prstatus(&t->prstatus, t->task, signr);
 	(void) view->regsets[0].get(t->task, &view->regsets[0],
-				    0, sizeof(t->prstatus.pr_reg),
-				    &t->prstatus.pr_reg, NULL);
+				    0, PR_REG_SIZE(t->prstatus.pr_reg),
+				    PR_REG_PTR(&t->prstatus), NULL);
 
 	fill_note(&t->notes[0], "CORE", NT_PRSTATUS,
-		  sizeof(t->prstatus), &t->prstatus);
+		  PRSTATUS_SIZE(t->prstatus), &t->prstatus);
 	*total += notesize(&t->notes[0]);
 
 	do_thread_regset_writeback(t->task, &view->regsets[0]);
@@ -1447,7 +1468,7 @@ static int fill_thread_core_info(struct elf_thread_core_info *t,
 						  regset->core_note_type,
 						  size, data);
 				else {
-					t->prstatus.pr_fpvalid = 1;
+					SET_PR_FPVALID(&t->prstatus, 1);
 					fill_note(&t->notes[i], "CORE",
 						  NT_PRFPREG, size, data);
 				}
@@ -2086,7 +2107,8 @@ out:
 
 static int __init init_elf_binfmt(void)
 {
-	return register_binfmt(&elf_format);
+	register_binfmt(&elf_format);
+	return 0;
 }
 
 static void __exit exit_elf_binfmt(void)
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 30745f459fa..d390a0fffc6 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -39,6 +39,7 @@
 #include <asm/uaccess.h>
 #include <asm/param.h>
 #include <asm/pgalloc.h>
+#include <asm/exec.h>
 
 typedef char *elf_caddr_t;
 
@@ -91,7 +92,8 @@ static struct linux_binfmt elf_fdpic_format = {
 
 static int __init init_elf_fdpic_binfmt(void)
 {
-	return register_binfmt(&elf_fdpic_format);
+	register_binfmt(&elf_fdpic_format);
+	return 0;
 }
 
 static void __exit exit_elf_fdpic_binfmt(void)
@@ -334,8 +336,6 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 	current->mm->context.exec_fdpic_loadmap = 0;
 	current->mm->context.interp_fdpic_loadmap = 0;
 
-	current->flags &= ~PF_FORKNOEXEC;
-
 #ifdef CONFIG_MMU
 	elf_fdpic_arch_lay_out_mm(&exec_params,
 				  &interp_params,
@@ -390,21 +390,17 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 	    (executable_stack == EXSTACK_DEFAULT && VM_STACK_FLAGS & VM_EXEC))
 		stack_prot |= PROT_EXEC;
 
-	down_write(&current->mm->mmap_sem);
-	current->mm->start_brk = do_mmap(NULL, 0, stack_size, stack_prot,
+	current->mm->start_brk = vm_mmap(NULL, 0, stack_size, stack_prot,
 					 MAP_PRIVATE | MAP_ANONYMOUS |
 					 MAP_UNINITIALIZED | MAP_GROWSDOWN,
 					 0);
 
 	if (IS_ERR_VALUE(current->mm->start_brk)) {
-		up_write(&current->mm->mmap_sem);
 		retval = current->mm->start_brk;
 		current->mm->start_brk = 0;
 		goto error_kill;
 	}
 
-	up_write(&current->mm->mmap_sem);
-
 	current->mm->brk = current->mm->start_brk;
 	current->mm->context.end_brk = current->mm->start_brk;
 	current->mm->context.end_brk +=
@@ -413,7 +409,6 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 #endif
 
 	install_exec_creds(bprm);
-	current->flags &= ~PF_FORKNOEXEC;
 	if (create_elf_fdpic_tables(bprm, current->mm,
 				    &exec_params, &interp_params) < 0)
 		goto error_kill;
@@ -956,10 +951,8 @@ static int elf_fdpic_map_file_constdisp_on_uclinux(
 	if (params->flags & ELF_FDPIC_FLAG_EXECUTABLE)
 		mflags |= MAP_EXECUTABLE;
 
-	down_write(&mm->mmap_sem);
-	maddr = do_mmap(NULL, load_addr, top - base,
+	maddr = vm_mmap(NULL, load_addr, top - base,
 			PROT_READ | PROT_WRITE | PROT_EXEC, mflags, 0);
-	up_write(&mm->mmap_sem);
 	if (IS_ERR_VALUE(maddr))
 		return (int) maddr;
 
@@ -1097,10 +1090,8 @@ static int elf_fdpic_map_file_by_direct_mmap(struct elf_fdpic_params *params,
 
 		/* create the mapping */
 		disp = phdr->p_vaddr & ~PAGE_MASK;
-		down_write(&mm->mmap_sem);
-		maddr = do_mmap(file, maddr, phdr->p_memsz + disp, prot, flags,
+		maddr = vm_mmap(file, maddr, phdr->p_memsz + disp, prot, flags,
 				phdr->p_offset - disp);
-		up_write(&mm->mmap_sem);
 
 		kdebug("mmap[%d] <file> sz=%lx pr=%x fl=%x of=%lx --> %08lx",
 		       loop, phdr->p_memsz + disp, prot, flags,
@@ -1144,10 +1135,8 @@ static int elf_fdpic_map_file_by_direct_mmap(struct elf_fdpic_params *params,
 			unsigned long xmaddr;
 
 			flags |= MAP_FIXED | MAP_ANONYMOUS;
-			down_write(&mm->mmap_sem);
-			xmaddr = do_mmap(NULL, xaddr, excess - excess1,
+			xmaddr = vm_mmap(NULL, xaddr, excess - excess1,
 					 prot, flags, 0);
-			up_write(&mm->mmap_sem);
 
 			kdebug("mmap[%d] <anon>"
 			       " ad=%lx sz=%lx pr=%x fl=%x of=0 --> %08lx",
diff --git a/fs/binfmt_em86.c b/fs/binfmt_em86.c
index b8e8b0acf9b..2790c7e1912 100644
--- a/fs/binfmt_em86.c
+++ b/fs/binfmt_em86.c
@@ -100,7 +100,8 @@ static struct linux_binfmt em86_format = {
 
 static int __init init_em86_binfmt(void)
 {
-	return register_binfmt(&em86_format);
+	register_binfmt(&em86_format);
+	return 0;
 }
 
 static void __exit exit_em86_binfmt(void)
diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c
index 1bffbe0ed77..6b2daf99fab 100644
--- a/fs/binfmt_flat.c
+++ b/fs/binfmt_flat.c
@@ -15,7 +15,7 @@
  *	JAN/99 -- coded full program relocation (gerg@snapgear.com)
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/kernel.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
@@ -37,7 +37,6 @@
 #include <linux/syscalls.h>
 
 #include <asm/byteorder.h>
-#include <asm/system.h>
 #include <asm/uaccess.h>
 #include <asm/unaligned.h>
 #include <asm/cacheflush.h>
@@ -543,10 +542,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 		 */
 		DBG_FLT("BINFMT_FLAT: ROM mapping of file (we hope)\n");
 
-		down_write(&current->mm->mmap_sem);
-		textpos = do_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC,
+		textpos = vm_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC,
 				  MAP_PRIVATE|MAP_EXECUTABLE, 0);
-		up_write(&current->mm->mmap_sem);
 		if (!textpos || IS_ERR_VALUE(textpos)) {
 			if (!textpos)
 				textpos = (unsigned long) -ENOMEM;
@@ -557,10 +554,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 
 		len = data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long);
 		len = PAGE_ALIGN(len);
-		down_write(&current->mm->mmap_sem);
-		realdatastart = do_mmap(0, 0, len,
+		realdatastart = vm_mmap(0, 0, len,
 			PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, 0);
-		up_write(&current->mm->mmap_sem);
 
 		if (realdatastart == 0 || IS_ERR_VALUE(realdatastart)) {
 			if (!realdatastart)
@@ -604,10 +599,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 
 		len = text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long);
 		len = PAGE_ALIGN(len);
-		down_write(&current->mm->mmap_sem);
-		textpos = do_mmap(0, 0, len,
+		textpos = vm_mmap(0, 0, len,
 			PROT_READ | PROT_EXEC | PROT_WRITE, MAP_PRIVATE, 0);
-		up_write(&current->mm->mmap_sem);
 
 		if (!textpos || IS_ERR_VALUE(textpos)) {
 			if (!textpos)
@@ -902,7 +895,6 @@ static int load_flat_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 						libinfo.lib_list[j].start_data:UNLOADED_LIB;
 
 	install_exec_creds(bprm);
- 	current->flags &= ~PF_FORKNOEXEC;
 
 	set_binfmt(&flat_format);
 
@@ -950,7 +942,8 @@ static int load_flat_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 
 static int __init init_flat_binfmt(void)
 {
-	return register_binfmt(&flat_format);
+	register_binfmt(&flat_format);
+	return 0;
 }
 
 /****************************************************************************/
diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c
index a9198dfd5f8..613aa061823 100644
--- a/fs/binfmt_misc.c
+++ b/fs/binfmt_misc.c
@@ -19,6 +19,7 @@
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/sched.h>
+#include <linux/magic.h>
 #include <linux/binfmts.h>
 #include <linux/slab.h>
 #include <linux/ctype.h>
@@ -699,7 +700,7 @@ static int bm_fill_super(struct super_block * sb, void * data, int silent)
 		[3] = {"register", &bm_register_operations, S_IWUSR},
 		/* last one */ {""}
 	};
-	int err = simple_fill_super(sb, 0x42494e4d, bm_files);
+	int err = simple_fill_super(sb, BINFMTFS_MAGIC, bm_files);
 	if (!err)
 		sb->s_op = &s_ops;
 	return err;
@@ -726,11 +727,8 @@ static struct file_system_type bm_fs_type = {
 static int __init init_misc_binfmt(void)
 {
 	int err = register_filesystem(&bm_fs_type);
-	if (!err) {
-		err = insert_binfmt(&misc_format);
-		if (err)
-			unregister_filesystem(&bm_fs_type);
-	}
+	if (!err)
+		insert_binfmt(&misc_format);
 	return err;
 }
 
diff --git a/fs/binfmt_script.c b/fs/binfmt_script.c
index 396a9884591..d3b8c1f6315 100644
--- a/fs/binfmt_script.c
+++ b/fs/binfmt_script.c
@@ -105,7 +105,8 @@ static struct linux_binfmt script_format = {
 
 static int __init init_script_binfmt(void)
 {
-	return register_binfmt(&script_format);
+	register_binfmt(&script_format);
+	return 0;
 }
 
 static void __exit exit_script_binfmt(void)
diff --git a/fs/binfmt_som.c b/fs/binfmt_som.c
index cc8560f6c9b..4517aaff61b 100644
--- a/fs/binfmt_som.c
+++ b/fs/binfmt_som.c
@@ -147,10 +147,8 @@ static int map_som_binary(struct file *file,
 	code_size = SOM_PAGEALIGN(hpuxhdr->exec_tsize);
 	current->mm->start_code = code_start;
 	current->mm->end_code = code_start + code_size;
-	down_write(&current->mm->mmap_sem);
-	retval = do_mmap(file, code_start, code_size, prot,
+	retval = vm_mmap(file, code_start, code_size, prot,
 			flags, SOM_PAGESTART(hpuxhdr->exec_tfile));
-	up_write(&current->mm->mmap_sem);
 	if (retval < 0 && retval > -1024)
 		goto out;
 
@@ -158,20 +156,16 @@ static int map_som_binary(struct file *file,
 	data_size = SOM_PAGEALIGN(hpuxhdr->exec_dsize);
 	current->mm->start_data = data_start;
 	current->mm->end_data = bss_start = data_start + data_size;
-	down_write(&current->mm->mmap_sem);
-	retval = do_mmap(file, data_start, data_size,
+	retval = vm_mmap(file, data_start, data_size,
 			prot | PROT_WRITE, flags,
 			SOM_PAGESTART(hpuxhdr->exec_dfile));
-	up_write(&current->mm->mmap_sem);
 	if (retval < 0 && retval > -1024)
 		goto out;
 
 	som_brk = bss_start + SOM_PAGEALIGN(hpuxhdr->exec_bsize);
 	current->mm->start_brk = current->mm->brk = som_brk;
-	down_write(&current->mm->mmap_sem);
-	retval = do_mmap(NULL, bss_start, som_brk - bss_start,
+	retval = vm_mmap(NULL, bss_start, som_brk - bss_start,
 			prot | PROT_WRITE, MAP_FIXED | MAP_PRIVATE, 0);
-	up_write(&current->mm->mmap_sem);
 	if (retval > 0 || retval < -1024)
 		retval = 0;
 out:
@@ -225,7 +219,6 @@ load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 		goto out_free;
 
 	/* OK, This is the point of no return */
-	current->flags &= ~PF_FORKNOEXEC;
 	current->personality = PER_HPUX;
 	setup_new_exec(bprm);
 
@@ -289,7 +282,8 @@ static int load_som_library(struct file *f)
 
 static int __init init_som_binfmt(void)
 {
-	return register_binfmt(&som_format);
+	register_binfmt(&som_format);
+	return 0;
 }
 
 static void __exit exit_som_binfmt(void)
diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c
index c2183f3917c..e85c04b9f61 100644
--- a/fs/bio-integrity.c
+++ b/fs/bio-integrity.c
@@ -357,7 +357,7 @@ static void bio_integrity_generate(struct bio *bio)
 	bix.sector_size = bi->sector_size;
 
 	bio_for_each_segment(bv, bio, i) {
-		void *kaddr = kmap_atomic(bv->bv_page, KM_USER0);
+		void *kaddr = kmap_atomic(bv->bv_page);
 		bix.data_buf = kaddr + bv->bv_offset;
 		bix.data_size = bv->bv_len;
 		bix.prot_buf = prot_buf;
@@ -371,7 +371,7 @@ static void bio_integrity_generate(struct bio *bio)
 		total += sectors * bi->tuple_size;
 		BUG_ON(total > bio->bi_integrity->bip_size);
 
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 	}
 }
 
@@ -498,7 +498,7 @@ static int bio_integrity_verify(struct bio *bio)
 	bix.sector_size = bi->sector_size;
 
 	bio_for_each_segment(bv, bio, i) {
-		void *kaddr = kmap_atomic(bv->bv_page, KM_USER0);
+		void *kaddr = kmap_atomic(bv->bv_page);
 		bix.data_buf = kaddr + bv->bv_offset;
 		bix.data_size = bv->bv_len;
 		bix.prot_buf = prot_buf;
@@ -507,7 +507,7 @@ static int bio_integrity_verify(struct bio *bio)
 		ret = bi->verify_fn(&bix);
 
 		if (ret) {
-			kunmap_atomic(kaddr, KM_USER0);
+			kunmap_atomic(kaddr);
 			return ret;
 		}
 
@@ -517,7 +517,7 @@ static int bio_integrity_verify(struct bio *bio)
 		total += sectors * bi->tuple_size;
 		BUG_ON(total > bio->bi_integrity->bip_size);
 
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 	}
 
 	return ret;
diff --git a/fs/bio.c b/fs/bio.c
index b980ecde026..84da8853904 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -22,7 +22,7 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
 #include <scsi/sg.h>		/* for struct sg_iovec */
@@ -505,9 +505,14 @@ EXPORT_SYMBOL(bio_clone);
 int bio_get_nr_vecs(struct block_device *bdev)
 {
 	struct request_queue *q = bdev_get_queue(bdev);
-	return min_t(unsigned,
+	int nr_pages;
+
+	nr_pages = min_t(unsigned,
 		     queue_max_segments(q),
 		     queue_max_sectors(q) / (PAGE_SIZE >> 9) + 1);
+
+	return min_t(unsigned, nr_pages, BIO_MAX_PAGES);
+
 }
 EXPORT_SYMBOL(bio_get_nr_vecs);
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 5e9f198f771..ba11c30f302 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -16,6 +16,7 @@
 #include <linux/blkdev.h>
 #include <linux/module.h>
 #include <linux/blkpg.h>
+#include <linux/magic.h>
 #include <linux/buffer_head.h>
 #include <linux/swap.h>
 #include <linux/pagevec.h>
@@ -69,7 +70,7 @@ static void bdev_inode_switch_bdi(struct inode *inode,
 	spin_unlock(&dst->wb.list_lock);
 }
 
-static sector_t max_block(struct block_device *bdev)
+sector_t blkdev_max_block(struct block_device *bdev)
 {
 	sector_t retval = ~((sector_t)0);
 	loff_t sz = i_size_read(bdev->bd_inode);
@@ -109,7 +110,7 @@ void invalidate_bdev(struct block_device *bdev)
 	/* 99% of the time, we don't need to flush the cleancache on the bdev.
 	 * But, for the strange corners, lets be cautious
 	 */
-	cleancache_flush_inode(mapping);
+	cleancache_invalidate_inode(mapping);
 }
 EXPORT_SYMBOL(invalidate_bdev);
 
@@ -162,7 +163,7 @@ static int
 blkdev_get_block(struct inode *inode, sector_t iblock,
 		struct buffer_head *bh, int create)
 {
-	if (iblock >= max_block(I_BDEV(inode))) {
+	if (iblock >= blkdev_max_block(I_BDEV(inode))) {
 		if (create)
 			return -EIO;
 
@@ -184,7 +185,7 @@ static int
 blkdev_get_blocks(struct inode *inode, sector_t iblock,
 		struct buffer_head *bh, int create)
 {
-	sector_t end_block = max_block(I_BDEV(inode));
+	sector_t end_block = blkdev_max_block(I_BDEV(inode));
 	unsigned long max_blocks = bh->b_size >> inode->i_blkbits;
 
 	if ((iblock + max_blocks) > end_block) {
@@ -506,7 +507,7 @@ static const struct super_operations bdev_sops = {
 static struct dentry *bd_mount(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *data)
 {
-	return mount_pseudo(fs_type, "bdev:", &bdev_sops, NULL, 0x62646576);
+	return mount_pseudo(fs_type, "bdev:", &bdev_sops, NULL, BDEVFS_MAGIC);
 }
 
 static struct file_system_type bd_type = {
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 0cc20b35c1c..42704149b72 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -171,11 +171,11 @@ out:
 	spin_unlock_irqrestore(&workers->lock, flags);
 }
 
-static noinline int run_ordered_completions(struct btrfs_workers *workers,
+static noinline void run_ordered_completions(struct btrfs_workers *workers,
 					    struct btrfs_work *work)
 {
 	if (!workers->ordered)
-		return 0;
+		return;
 
 	set_bit(WORK_DONE_BIT, &work->flags);
 
@@ -213,7 +213,6 @@ static noinline int run_ordered_completions(struct btrfs_workers *workers,
 	}
 
 	spin_unlock(&workers->order_lock);
-	return 0;
 }
 
 static void put_worker(struct btrfs_worker_thread *worker)
@@ -399,7 +398,7 @@ again:
 /*
  * this will wait for all the worker threads to shutdown
  */
-int btrfs_stop_workers(struct btrfs_workers *workers)
+void btrfs_stop_workers(struct btrfs_workers *workers)
 {
 	struct list_head *cur;
 	struct btrfs_worker_thread *worker;
@@ -427,7 +426,6 @@ int btrfs_stop_workers(struct btrfs_workers *workers)
 		put_worker(worker);
 	}
 	spin_unlock_irq(&workers->lock);
-	return 0;
 }
 
 /*
@@ -615,14 +613,14 @@ found:
  * it was taken from.  It is intended for use with long running work functions
  * that make some progress and want to give the cpu up for others.
  */
-int btrfs_requeue_work(struct btrfs_work *work)
+void btrfs_requeue_work(struct btrfs_work *work)
 {
 	struct btrfs_worker_thread *worker = work->worker;
 	unsigned long flags;
 	int wake = 0;
 
 	if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags))
-		goto out;
+		return;
 
 	spin_lock_irqsave(&worker->lock, flags);
 	if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags))
@@ -649,9 +647,6 @@ int btrfs_requeue_work(struct btrfs_work *work)
 	if (wake)
 		wake_up_process(worker->task);
 	spin_unlock_irqrestore(&worker->lock, flags);
-out:
-
-	return 0;
 }
 
 void btrfs_set_work_high_prio(struct btrfs_work *work)
diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h
index f34cc31fa3c..063698b90ce 100644
--- a/fs/btrfs/async-thread.h
+++ b/fs/btrfs/async-thread.h
@@ -111,9 +111,9 @@ struct btrfs_workers {
 
 void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work);
 int btrfs_start_workers(struct btrfs_workers *workers);
-int btrfs_stop_workers(struct btrfs_workers *workers);
+void btrfs_stop_workers(struct btrfs_workers *workers);
 void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max,
 			struct btrfs_workers *async_starter);
-int btrfs_requeue_work(struct btrfs_work *work);
+void btrfs_requeue_work(struct btrfs_work *work);
 void btrfs_set_work_high_prio(struct btrfs_work *work);
 #endif
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 56136d9046f..bcec0675023 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -22,6 +22,7 @@
 #include "ulist.h"
 #include "transaction.h"
 #include "delayed-ref.h"
+#include "locking.h"
 
 /*
  * this structure records all encountered refs on the way up to the root
@@ -893,18 +894,22 @@ static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 	s64 bytes_left = size - 1;
 	struct extent_buffer *eb = eb_in;
 	struct btrfs_key found_key;
+	int leave_spinning = path->leave_spinning;
 
 	if (bytes_left >= 0)
 		dest[bytes_left] = '\0';
 
+	path->leave_spinning = 1;
 	while (1) {
 		len = btrfs_inode_ref_name_len(eb, iref);
 		bytes_left -= len;
 		if (bytes_left >= 0)
 			read_extent_buffer(eb, dest + bytes_left,
 						(unsigned long)(iref + 1), len);
-		if (eb != eb_in)
+		if (eb != eb_in) {
+			btrfs_tree_read_unlock_blocking(eb);
 			free_extent_buffer(eb);
+		}
 		ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
 		if (ret > 0)
 			ret = -ENOENT;
@@ -919,8 +924,11 @@ static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 		slot = path->slots[0];
 		eb = path->nodes[0];
 		/* make sure we can use eb after releasing the path */
-		if (eb != eb_in)
+		if (eb != eb_in) {
 			atomic_inc(&eb->refs);
+			btrfs_tree_read_lock(eb);
+			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
+		}
 		btrfs_release_path(path);
 
 		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
@@ -931,6 +939,7 @@ static char *iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 	}
 
 	btrfs_release_path(path);
+	path->leave_spinning = leave_spinning;
 
 	if (ret)
 		return ERR_PTR(ret);
@@ -1247,7 +1256,7 @@ static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
 				struct btrfs_path *path,
 				iterate_irefs_t *iterate, void *ctx)
 {
-	int ret;
+	int ret = 0;
 	int slot;
 	u32 cur;
 	u32 len;
@@ -1259,7 +1268,8 @@ static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
 	struct btrfs_inode_ref *iref;
 	struct btrfs_key found_key;
 
-	while (1) {
+	while (!ret) {
+		path->leave_spinning = 1;
 		ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
 					&found_key);
 		if (ret < 0)
@@ -1275,6 +1285,8 @@ static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
 		eb = path->nodes[0];
 		/* make sure we can use eb after releasing the path */
 		atomic_inc(&eb->refs);
+		btrfs_tree_read_lock(eb);
+		btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
 		btrfs_release_path(path);
 
 		item = btrfs_item_nr(eb, slot);
@@ -1288,13 +1300,12 @@ static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
 				 (unsigned long long)found_key.objectid,
 				 (unsigned long long)fs_root->objectid);
 			ret = iterate(parent, iref, eb, ctx);
-			if (ret) {
-				free_extent_buffer(eb);
+			if (ret)
 				break;
-			}
 			len = sizeof(*iref) + name_len;
 			iref = (struct btrfs_inode_ref *)((char *)iref + len);
 		}
+		btrfs_tree_read_unlock_blocking(eb);
 		free_extent_buffer(eb);
 	}
 
@@ -1359,12 +1370,6 @@ int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
 				inode_to_path, ipath);
 }
 
-/*
- * allocates space to return multiple file system paths for an inode.
- * total_bytes to allocate are passed, note that space usable for actual path
- * information will be total_bytes - sizeof(struct inode_fs_paths).
- * the returned pointer must be freed with free_ipath() in the end.
- */
 struct btrfs_data_container *init_data_container(u32 total_bytes)
 {
 	struct btrfs_data_container *data;
@@ -1420,5 +1425,8 @@ struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
 
 void free_ipath(struct inode_fs_paths *ipath)
 {
+	if (!ipath)
+		return;
+	kfree(ipath->fspath);
 	kfree(ipath);
 }
diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index d986824bb2b..c053e90f200 100644
--- a/fs/btrfs/check-integrity.c
+++ b/fs/btrfs/check-integrity.c
@@ -89,7 +89,6 @@
 #include "disk-io.h"
 #include "transaction.h"
 #include "extent_io.h"
-#include "disk-io.h"
 #include "volumes.h"
 #include "print-tree.h"
 #include "locking.h"
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index d02c27cd14c..86eff48dab7 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -120,10 +120,10 @@ static int check_compressed_csum(struct inode *inode,
 		page = cb->compressed_pages[i];
 		csum = ~(u32)0;
 
-		kaddr = kmap_atomic(page, KM_USER0);
+		kaddr = kmap_atomic(page);
 		csum = btrfs_csum_data(root, kaddr, csum, PAGE_CACHE_SIZE);
 		btrfs_csum_final(csum, (char *)&csum);
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 
 		if (csum != *cb_sum) {
 			printk(KERN_INFO "btrfs csum failed ino %llu "
@@ -226,8 +226,8 @@ out:
  * Clear the writeback bits on all of the file
  * pages for a compressed write
  */
-static noinline int end_compressed_writeback(struct inode *inode, u64 start,
-					     unsigned long ram_size)
+static noinline void end_compressed_writeback(struct inode *inode, u64 start,
+					      unsigned long ram_size)
 {
 	unsigned long index = start >> PAGE_CACHE_SHIFT;
 	unsigned long end_index = (start + ram_size - 1) >> PAGE_CACHE_SHIFT;
@@ -253,7 +253,6 @@ static noinline int end_compressed_writeback(struct inode *inode, u64 start,
 		index += ret;
 	}
 	/* the inode may be gone now */
-	return 0;
 }
 
 /*
@@ -392,20 +391,21 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start,
 			 */
 			atomic_inc(&cb->pending_bios);
 			ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			if (!skip_sum) {
 				ret = btrfs_csum_one_bio(root, inode, bio,
 							 start, 1);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 			}
 
 			ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			bio_put(bio);
 
 			bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS);
+			BUG_ON(!bio);
 			bio->bi_private = cb;
 			bio->bi_end_io = end_compressed_bio_write;
 			bio_add_page(bio, page, PAGE_CACHE_SIZE, 0);
@@ -421,15 +421,15 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start,
 	bio_get(bio);
 
 	ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 
 	if (!skip_sum) {
 		ret = btrfs_csum_one_bio(root, inode, bio, start, 1);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 
 	ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 
 	bio_put(bio);
 	return 0;
@@ -497,7 +497,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 		 * sure they map to this compressed extent on disk.
 		 */
 		set_page_extent_mapped(page);
-		lock_extent(tree, last_offset, end, GFP_NOFS);
+		lock_extent(tree, last_offset, end);
 		read_lock(&em_tree->lock);
 		em = lookup_extent_mapping(em_tree, last_offset,
 					   PAGE_CACHE_SIZE);
@@ -507,7 +507,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 		    (last_offset + PAGE_CACHE_SIZE > extent_map_end(em)) ||
 		    (em->block_start >> 9) != cb->orig_bio->bi_sector) {
 			free_extent_map(em);
-			unlock_extent(tree, last_offset, end, GFP_NOFS);
+			unlock_extent(tree, last_offset, end);
 			unlock_page(page);
 			page_cache_release(page);
 			break;
@@ -521,10 +521,10 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			if (zero_offset) {
 				int zeros;
 				zeros = PAGE_CACHE_SIZE - zero_offset;
-				userpage = kmap_atomic(page, KM_USER0);
+				userpage = kmap_atomic(page);
 				memset(userpage + zero_offset, 0, zeros);
 				flush_dcache_page(page);
-				kunmap_atomic(userpage, KM_USER0);
+				kunmap_atomic(userpage);
 			}
 		}
 
@@ -535,7 +535,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			nr_pages++;
 			page_cache_release(page);
 		} else {
-			unlock_extent(tree, last_offset, end, GFP_NOFS);
+			unlock_extent(tree, last_offset, end);
 			unlock_page(page);
 			page_cache_release(page);
 			break;
@@ -662,7 +662,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 			bio_get(comp_bio);
 
 			ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			/*
 			 * inc the count before we submit the bio so
@@ -675,19 +675,20 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 			if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) {
 				ret = btrfs_lookup_bio_sums(root, inode,
 							comp_bio, sums);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 			}
 			sums += (comp_bio->bi_size + root->sectorsize - 1) /
 				root->sectorsize;
 
 			ret = btrfs_map_bio(root, READ, comp_bio,
 					    mirror_num, 0);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			bio_put(comp_bio);
 
 			comp_bio = compressed_bio_alloc(bdev, cur_disk_byte,
 							GFP_NOFS);
+			BUG_ON(!comp_bio);
 			comp_bio->bi_private = cb;
 			comp_bio->bi_end_io = end_compressed_bio_read;
 
@@ -698,15 +699,15 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	bio_get(comp_bio);
 
 	ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 
 	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) {
 		ret = btrfs_lookup_bio_sums(root, inode, comp_bio, sums);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 
 	ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 
 	bio_put(comp_bio);
 	return 0;
@@ -734,7 +735,7 @@ struct btrfs_compress_op *btrfs_compress_op[] = {
 	&btrfs_lzo_compress,
 };
 
-int __init btrfs_init_compress(void)
+void __init btrfs_init_compress(void)
 {
 	int i;
 
@@ -744,7 +745,6 @@ int __init btrfs_init_compress(void)
 		atomic_set(&comp_alloc_workspace[i], 0);
 		init_waitqueue_head(&comp_workspace_wait[i]);
 	}
-	return 0;
 }
 
 /*
@@ -993,9 +993,9 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start,
 		bytes = min(PAGE_CACHE_SIZE - *pg_offset,
 			    PAGE_CACHE_SIZE - buf_offset);
 		bytes = min(bytes, working_bytes);
-		kaddr = kmap_atomic(page_out, KM_USER0);
+		kaddr = kmap_atomic(page_out);
 		memcpy(kaddr + *pg_offset, buf + buf_offset, bytes);
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 		flush_dcache_page(page_out);
 
 		*pg_offset += bytes;
diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
index a12059f4f0f..9afb0a62ae8 100644
--- a/fs/btrfs/compression.h
+++ b/fs/btrfs/compression.h
@@ -19,7 +19,7 @@
 #ifndef __BTRFS_COMPRESSION_
 #define __BTRFS_COMPRESSION_
 
-int btrfs_init_compress(void);
+void btrfs_init_compress(void);
 void btrfs_exit_compress(void);
 
 int btrfs_compress_pages(int type, struct address_space *mapping,
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 0639a555e16..4106264fbc6 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -36,7 +36,7 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      struct extent_buffer *dst_buf,
 			      struct extent_buffer *src_buf);
-static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		   struct btrfs_path *path, int level, int slot);
 
 struct btrfs_path *btrfs_alloc_path(void)
@@ -156,10 +156,23 @@ struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
 {
 	struct extent_buffer *eb;
 
-	rcu_read_lock();
-	eb = rcu_dereference(root->node);
-	extent_buffer_get(eb);
-	rcu_read_unlock();
+	while (1) {
+		rcu_read_lock();
+		eb = rcu_dereference(root->node);
+
+		/*
+		 * RCU really hurts here, we could free up the root node because
+		 * it was cow'ed but we may not get the new root node yet so do
+		 * the inc_not_zero dance and if it doesn't work then
+		 * synchronize_rcu and try again.
+		 */
+		if (atomic_inc_not_zero(&eb->refs)) {
+			rcu_read_unlock();
+			break;
+		}
+		rcu_read_unlock();
+		synchronize_rcu();
+	}
 	return eb;
 }
 
@@ -207,10 +220,12 @@ struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root)
  */
 static void add_root_to_dirty_list(struct btrfs_root *root)
 {
+	spin_lock(&root->fs_info->trans_lock);
 	if (root->track_dirty && list_empty(&root->dirty_list)) {
 		list_add(&root->dirty_list,
 			 &root->fs_info->dirty_cowonly_roots);
 	}
+	spin_unlock(&root->fs_info->trans_lock);
 }
 
 /*
@@ -331,8 +346,13 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 	if (btrfs_block_can_be_shared(root, buf)) {
 		ret = btrfs_lookup_extent_info(trans, root, buf->start,
 					       buf->len, &refs, &flags);
-		BUG_ON(ret);
-		BUG_ON(refs == 0);
+		if (ret)
+			return ret;
+		if (refs == 0) {
+			ret = -EROFS;
+			btrfs_std_error(root->fs_info, ret);
+			return ret;
+		}
 	} else {
 		refs = 1;
 		if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
@@ -351,14 +371,14 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 		     root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) &&
 		    !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) {
 			ret = btrfs_inc_ref(trans, root, buf, 1, 1);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			if (root->root_key.objectid ==
 			    BTRFS_TREE_RELOC_OBJECTID) {
 				ret = btrfs_dec_ref(trans, root, buf, 0, 1);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 				ret = btrfs_inc_ref(trans, root, cow, 1, 1);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 			}
 			new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
 		} else {
@@ -368,14 +388,15 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 				ret = btrfs_inc_ref(trans, root, cow, 1, 1);
 			else
 				ret = btrfs_inc_ref(trans, root, cow, 0, 1);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		}
 		if (new_flags != 0) {
 			ret = btrfs_set_disk_extent_flags(trans, root,
 							  buf->start,
 							  buf->len,
 							  new_flags, 0);
-			BUG_ON(ret);
+			if (ret)
+				return ret;
 		}
 	} else {
 		if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
@@ -384,9 +405,9 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 				ret = btrfs_inc_ref(trans, root, cow, 1, 1);
 			else
 				ret = btrfs_inc_ref(trans, root, cow, 0, 1);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 			ret = btrfs_dec_ref(trans, root, buf, 1, 1);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		}
 		clean_tree_block(trans, root, buf);
 		*last_ref = 1;
@@ -415,7 +436,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_disk_key disk_key;
 	struct extent_buffer *cow;
-	int level;
+	int level, ret;
 	int last_ref = 0;
 	int unlock_orig = 0;
 	u64 parent_start;
@@ -467,7 +488,11 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 			    (unsigned long)btrfs_header_fsid(cow),
 			    BTRFS_FSID_SIZE);
 
-	update_ref_for_cow(trans, root, buf, cow, &last_ref);
+	ret = update_ref_for_cow(trans, root, buf, cow, &last_ref);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		return ret;
+	}
 
 	if (root->ref_cows)
 		btrfs_reloc_cow_block(trans, root, buf, cow);
@@ -504,7 +529,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 	}
 	if (unlock_orig)
 		btrfs_tree_unlock(buf);
-	free_extent_buffer(buf);
+	free_extent_buffer_stale(buf);
 	btrfs_mark_buffer_dirty(cow);
 	*cow_ret = cow;
 	return 0;
@@ -700,7 +725,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 
 		cur = btrfs_find_tree_block(root, blocknr, blocksize);
 		if (cur)
-			uptodate = btrfs_buffer_uptodate(cur, gen);
+			uptodate = btrfs_buffer_uptodate(cur, gen, 0);
 		else
 			uptodate = 0;
 		if (!cur || !uptodate) {
@@ -934,7 +959,12 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 
 		/* promote the child to a root */
 		child = read_node_slot(root, mid, 0);
-		BUG_ON(!child);
+		if (!child) {
+			ret = -EROFS;
+			btrfs_std_error(root->fs_info, ret);
+			goto enospc;
+		}
+
 		btrfs_tree_lock(child);
 		btrfs_set_lock_blocking(child);
 		ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
@@ -959,7 +989,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		root_sub_used(root, mid->len);
 		btrfs_free_tree_block(trans, root, mid, 0, 1, 0);
 		/* once for the root ptr */
-		free_extent_buffer(mid);
+		free_extent_buffer_stale(mid);
 		return 0;
 	}
 	if (btrfs_header_nritems(mid) >
@@ -1010,13 +1040,10 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		if (btrfs_header_nritems(right) == 0) {
 			clean_tree_block(trans, root, right);
 			btrfs_tree_unlock(right);
-			wret = del_ptr(trans, root, path, level + 1, pslot +
-				       1);
-			if (wret)
-				ret = wret;
+			del_ptr(trans, root, path, level + 1, pslot + 1);
 			root_sub_used(root, right->len);
 			btrfs_free_tree_block(trans, root, right, 0, 1, 0);
-			free_extent_buffer(right);
+			free_extent_buffer_stale(right);
 			right = NULL;
 		} else {
 			struct btrfs_disk_key right_key;
@@ -1035,7 +1062,11 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		 * otherwise we would have pulled some pointers from the
 		 * right
 		 */
-		BUG_ON(!left);
+		if (!left) {
+			ret = -EROFS;
+			btrfs_std_error(root->fs_info, ret);
+			goto enospc;
+		}
 		wret = balance_node_right(trans, root, mid, left);
 		if (wret < 0) {
 			ret = wret;
@@ -1051,12 +1082,10 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	if (btrfs_header_nritems(mid) == 0) {
 		clean_tree_block(trans, root, mid);
 		btrfs_tree_unlock(mid);
-		wret = del_ptr(trans, root, path, level + 1, pslot);
-		if (wret)
-			ret = wret;
+		del_ptr(trans, root, path, level + 1, pslot);
 		root_sub_used(root, mid->len);
 		btrfs_free_tree_block(trans, root, mid, 0, 1, 0);
-		free_extent_buffer(mid);
+		free_extent_buffer_stale(mid);
 		mid = NULL;
 	} else {
 		/* update the parent key to reflect our changes */
@@ -1331,7 +1360,12 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 		block1 = btrfs_node_blockptr(parent, slot - 1);
 		gen = btrfs_node_ptr_generation(parent, slot - 1);
 		eb = btrfs_find_tree_block(root, block1, blocksize);
-		if (eb && btrfs_buffer_uptodate(eb, gen))
+		/*
+		 * if we get -eagain from btrfs_buffer_uptodate, we
+		 * don't want to return eagain here.  That will loop
+		 * forever
+		 */
+		if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0)
 			block1 = 0;
 		free_extent_buffer(eb);
 	}
@@ -1339,7 +1373,7 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 		block2 = btrfs_node_blockptr(parent, slot + 1);
 		gen = btrfs_node_ptr_generation(parent, slot + 1);
 		eb = btrfs_find_tree_block(root, block2, blocksize);
-		if (eb && btrfs_buffer_uptodate(eb, gen))
+		if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0)
 			block2 = 0;
 		free_extent_buffer(eb);
 	}
@@ -1382,7 +1416,8 @@ static noinline int reada_for_balance(struct btrfs_root *root,
  * if lowest_unlock is 1, level 0 won't be unlocked
  */
 static noinline void unlock_up(struct btrfs_path *path, int level,
-			       int lowest_unlock)
+			       int lowest_unlock, int min_write_lock_level,
+			       int *write_lock_level)
 {
 	int i;
 	int skip_level = level;
@@ -1414,6 +1449,11 @@ static noinline void unlock_up(struct btrfs_path *path, int level,
 		if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
 			btrfs_tree_unlock_rw(t, path->locks[i]);
 			path->locks[i] = 0;
+			if (write_lock_level &&
+			    i > min_write_lock_level &&
+			    i <= *write_lock_level) {
+				*write_lock_level = i - 1;
+			}
 		}
 	}
 }
@@ -1471,8 +1511,9 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 
 	tmp = btrfs_find_tree_block(root, blocknr, blocksize);
 	if (tmp) {
-		if (btrfs_buffer_uptodate(tmp, 0)) {
-			if (btrfs_buffer_uptodate(tmp, gen)) {
+		/* first we do an atomic uptodate check */
+		if (btrfs_buffer_uptodate(tmp, 0, 1) > 0) {
+			if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
 				/*
 				 * we found an up to date block without
 				 * sleeping, return
@@ -1490,8 +1531,9 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 			free_extent_buffer(tmp);
 			btrfs_set_path_blocking(p);
 
+			/* now we're allowed to do a blocking uptodate check */
 			tmp = read_tree_block(root, blocknr, blocksize, gen);
-			if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+			if (tmp && btrfs_buffer_uptodate(tmp, gen, 0) > 0) {
 				*eb_ret = tmp;
 				return 0;
 			}
@@ -1526,7 +1568,7 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 		 * and give up so that our caller doesn't loop forever
 		 * on our EAGAINs.
 		 */
-		if (!btrfs_buffer_uptodate(tmp, 0))
+		if (!btrfs_buffer_uptodate(tmp, 0, 0))
 			ret = -EIO;
 		free_extent_buffer(tmp);
 	}
@@ -1637,6 +1679,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	/* everything at write_lock_level or lower must be write locked */
 	int write_lock_level = 0;
 	u8 lowest_level = 0;
+	int min_write_lock_level;
 
 	lowest_level = p->lowest_level;
 	WARN_ON(lowest_level && ins_len > 0);
@@ -1664,6 +1707,8 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (cow && (p->keep_locks || p->lowest_level))
 		write_lock_level = BTRFS_MAX_LEVEL;
 
+	min_write_lock_level = write_lock_level;
+
 again:
 	/*
 	 * we try very hard to do read locks on the root
@@ -1795,7 +1840,8 @@ cow_done:
 				goto again;
 			}
 
-			unlock_up(p, level, lowest_unlock);
+			unlock_up(p, level, lowest_unlock,
+				  min_write_lock_level, &write_lock_level);
 
 			if (level == lowest_level) {
 				if (dec)
@@ -1857,7 +1903,8 @@ cow_done:
 				}
 			}
 			if (!p->search_for_split)
-				unlock_up(p, level, lowest_unlock);
+				unlock_up(p, level, lowest_unlock,
+					  min_write_lock_level, &write_lock_level);
 			goto done;
 		}
 	}
@@ -1881,15 +1928,12 @@ done:
  * fixing up pointers when a given leaf/node is not in slot 0 of the
  * higher levels
  *
- * If this fails to write a tree block, it returns -1, but continues
- * fixing up the blocks in ram so the tree is consistent.
  */
-static int fixup_low_keys(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root, struct btrfs_path *path,
-			  struct btrfs_disk_key *key, int level)
+static void fixup_low_keys(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, struct btrfs_path *path,
+			   struct btrfs_disk_key *key, int level)
 {
 	int i;
-	int ret = 0;
 	struct extent_buffer *t;
 
 	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
@@ -1902,7 +1946,6 @@ static int fixup_low_keys(struct btrfs_trans_handle *trans,
 		if (tslot != 0)
 			break;
 	}
-	return ret;
 }
 
 /*
@@ -1911,9 +1954,9 @@ static int fixup_low_keys(struct btrfs_trans_handle *trans,
  * This function isn't completely safe. It's the caller's responsibility
  * that the new key won't break the order
  */
-int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root, struct btrfs_path *path,
-			    struct btrfs_key *new_key)
+void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, struct btrfs_path *path,
+			     struct btrfs_key *new_key)
 {
 	struct btrfs_disk_key disk_key;
 	struct extent_buffer *eb;
@@ -1923,13 +1966,11 @@ int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
 	slot = path->slots[0];
 	if (slot > 0) {
 		btrfs_item_key(eb, &disk_key, slot - 1);
-		if (comp_keys(&disk_key, new_key) >= 0)
-			return -1;
+		BUG_ON(comp_keys(&disk_key, new_key) >= 0);
 	}
 	if (slot < btrfs_header_nritems(eb) - 1) {
 		btrfs_item_key(eb, &disk_key, slot + 1);
-		if (comp_keys(&disk_key, new_key) <= 0)
-			return -1;
+		BUG_ON(comp_keys(&disk_key, new_key) <= 0);
 	}
 
 	btrfs_cpu_key_to_disk(&disk_key, new_key);
@@ -1937,7 +1978,6 @@ int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(eb);
 	if (slot == 0)
 		fixup_low_keys(trans, root, path, &disk_key, 1);
-	return 0;
 }
 
 /*
@@ -2140,12 +2180,11 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans,
  *
  * slot and level indicate where you want the key to go, and
  * blocknr is the block the key points to.
- *
- * returns zero on success and < 0 on any error
  */
-static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
-		      *root, struct btrfs_path *path, struct btrfs_disk_key
-		      *key, u64 bytenr, int slot, int level)
+static void insert_ptr(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct btrfs_path *path,
+		       struct btrfs_disk_key *key, u64 bytenr,
+		       int slot, int level)
 {
 	struct extent_buffer *lower;
 	int nritems;
@@ -2155,8 +2194,7 @@ static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
 	lower = path->nodes[level];
 	nritems = btrfs_header_nritems(lower);
 	BUG_ON(slot > nritems);
-	if (nritems == BTRFS_NODEPTRS_PER_BLOCK(root))
-		BUG();
+	BUG_ON(nritems == BTRFS_NODEPTRS_PER_BLOCK(root));
 	if (slot != nritems) {
 		memmove_extent_buffer(lower,
 			      btrfs_node_key_ptr_offset(slot + 1),
@@ -2169,7 +2207,6 @@ static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
 	btrfs_set_node_ptr_generation(lower, slot, trans->transid);
 	btrfs_set_header_nritems(lower, nritems + 1);
 	btrfs_mark_buffer_dirty(lower);
-	return 0;
 }
 
 /*
@@ -2190,7 +2227,6 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	struct btrfs_disk_key disk_key;
 	int mid;
 	int ret;
-	int wret;
 	u32 c_nritems;
 
 	c = path->nodes[level];
@@ -2247,11 +2283,8 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(c);
 	btrfs_mark_buffer_dirty(split);
 
-	wret = insert_ptr(trans, root, path, &disk_key, split->start,
-			  path->slots[level + 1] + 1,
-			  level + 1);
-	if (wret)
-		ret = wret;
+	insert_ptr(trans, root, path, &disk_key, split->start,
+		   path->slots[level + 1] + 1, level + 1);
 
 	if (path->slots[level] >= mid) {
 		path->slots[level] -= mid;
@@ -2320,6 +2353,7 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
 {
 	struct extent_buffer *left = path->nodes[0];
 	struct extent_buffer *upper = path->nodes[1];
+	struct btrfs_map_token token;
 	struct btrfs_disk_key disk_key;
 	int slot;
 	u32 i;
@@ -2331,6 +2365,8 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
 	u32 data_end;
 	u32 this_item_size;
 
+	btrfs_init_map_token(&token);
+
 	if (empty)
 		nr = 0;
 	else
@@ -2408,8 +2444,8 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
 	push_space = BTRFS_LEAF_DATA_SIZE(root);
 	for (i = 0; i < right_nritems; i++) {
 		item = btrfs_item_nr(right, i);
-		push_space -= btrfs_item_size(right, item);
-		btrfs_set_item_offset(right, item, push_space);
+		push_space -= btrfs_token_item_size(right, item, &token);
+		btrfs_set_token_item_offset(right, item, push_space, &token);
 	}
 
 	left_nritems -= push_items;
@@ -2537,9 +2573,11 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 	u32 old_left_nritems;
 	u32 nr;
 	int ret = 0;
-	int wret;
 	u32 this_item_size;
 	u32 old_left_item_size;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	if (empty)
 		nr = min(right_nritems, max_slot);
@@ -2600,9 +2638,10 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 
 		item = btrfs_item_nr(left, i);
 
-		ioff = btrfs_item_offset(left, item);
-		btrfs_set_item_offset(left, item,
-		      ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size));
+		ioff = btrfs_token_item_offset(left, item, &token);
+		btrfs_set_token_item_offset(left, item,
+		      ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size),
+		      &token);
 	}
 	btrfs_set_header_nritems(left, old_left_nritems + push_items);
 
@@ -2632,8 +2671,9 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 	for (i = 0; i < right_nritems; i++) {
 		item = btrfs_item_nr(right, i);
 
-		push_space = push_space - btrfs_item_size(right, item);
-		btrfs_set_item_offset(right, item, push_space);
+		push_space = push_space - btrfs_token_item_size(right,
+								item, &token);
+		btrfs_set_token_item_offset(right, item, push_space, &token);
 	}
 
 	btrfs_mark_buffer_dirty(left);
@@ -2643,9 +2683,7 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 		clean_tree_block(trans, root, right);
 
 	btrfs_item_key(right, &disk_key, 0);
-	wret = fixup_low_keys(trans, root, path, &disk_key, 1);
-	if (wret)
-		ret = wret;
+	fixup_low_keys(trans, root, path, &disk_key, 1);
 
 	/* then fixup the leaf pointer in the path */
 	if (path->slots[0] < push_items) {
@@ -2716,7 +2754,8 @@ static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
 			      path->nodes[1], slot - 1, &left);
 	if (ret) {
 		/* we hit -ENOSPC, but it isn't fatal here */
-		ret = 1;
+		if (ret == -ENOSPC)
+			ret = 1;
 		goto out;
 	}
 
@@ -2738,22 +2777,21 @@ out:
 /*
  * split the path's leaf in two, making sure there is at least data_size
  * available for the resulting leaf level of the path.
- *
- * returns 0 if all went well and < 0 on failure.
  */
-static noinline int copy_for_split(struct btrfs_trans_handle *trans,
-			       struct btrfs_root *root,
-			       struct btrfs_path *path,
-			       struct extent_buffer *l,
-			       struct extent_buffer *right,
-			       int slot, int mid, int nritems)
+static noinline void copy_for_split(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root,
+				    struct btrfs_path *path,
+				    struct extent_buffer *l,
+				    struct extent_buffer *right,
+				    int slot, int mid, int nritems)
 {
 	int data_copy_size;
 	int rt_data_off;
 	int i;
-	int ret = 0;
-	int wret;
 	struct btrfs_disk_key disk_key;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	nritems = nritems - mid;
 	btrfs_set_header_nritems(right, nritems);
@@ -2775,17 +2813,15 @@ static noinline int copy_for_split(struct btrfs_trans_handle *trans,
 		struct btrfs_item *item = btrfs_item_nr(right, i);
 		u32 ioff;
 
-		ioff = btrfs_item_offset(right, item);
-		btrfs_set_item_offset(right, item, ioff + rt_data_off);
+		ioff = btrfs_token_item_offset(right, item, &token);
+		btrfs_set_token_item_offset(right, item,
+					    ioff + rt_data_off, &token);
 	}
 
 	btrfs_set_header_nritems(l, mid);
-	ret = 0;
 	btrfs_item_key(right, &disk_key, 0);
-	wret = insert_ptr(trans, root, path, &disk_key, right->start,
-			  path->slots[1] + 1, 1);
-	if (wret)
-		ret = wret;
+	insert_ptr(trans, root, path, &disk_key, right->start,
+		   path->slots[1] + 1, 1);
 
 	btrfs_mark_buffer_dirty(right);
 	btrfs_mark_buffer_dirty(l);
@@ -2803,8 +2839,6 @@ static noinline int copy_for_split(struct btrfs_trans_handle *trans,
 	}
 
 	BUG_ON(path->slots[0] < 0);
-
-	return ret;
 }
 
 /*
@@ -2993,12 +3027,8 @@ again:
 	if (split == 0) {
 		if (mid <= slot) {
 			btrfs_set_header_nritems(right, 0);
-			wret = insert_ptr(trans, root, path,
-					  &disk_key, right->start,
-					  path->slots[1] + 1, 1);
-			if (wret)
-				ret = wret;
-
+			insert_ptr(trans, root, path, &disk_key, right->start,
+				   path->slots[1] + 1, 1);
 			btrfs_tree_unlock(path->nodes[0]);
 			free_extent_buffer(path->nodes[0]);
 			path->nodes[0] = right;
@@ -3006,29 +3036,21 @@ again:
 			path->slots[1] += 1;
 		} else {
 			btrfs_set_header_nritems(right, 0);
-			wret = insert_ptr(trans, root, path,
-					  &disk_key,
-					  right->start,
+			insert_ptr(trans, root, path, &disk_key, right->start,
 					  path->slots[1], 1);
-			if (wret)
-				ret = wret;
 			btrfs_tree_unlock(path->nodes[0]);
 			free_extent_buffer(path->nodes[0]);
 			path->nodes[0] = right;
 			path->slots[0] = 0;
-			if (path->slots[1] == 0) {
-				wret = fixup_low_keys(trans, root,
-						path, &disk_key, 1);
-				if (wret)
-					ret = wret;
-			}
+			if (path->slots[1] == 0)
+				fixup_low_keys(trans, root, path,
+					       &disk_key, 1);
 		}
 		btrfs_mark_buffer_dirty(right);
 		return ret;
 	}
 
-	ret = copy_for_split(trans, root, path, l, right, slot, mid, nritems);
-	BUG_ON(ret);
+	copy_for_split(trans, root, path, l, right, slot, mid, nritems);
 
 	if (split == 2) {
 		BUG_ON(num_doubles != 0);
@@ -3036,7 +3058,7 @@ again:
 		goto again;
 	}
 
-	return ret;
+	return 0;
 
 push_for_double:
 	push_for_double_split(trans, root, path, data_size);
@@ -3238,11 +3260,9 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
 		return ret;
 
 	path->slots[0]++;
-	ret = setup_items_for_insert(trans, root, path, new_key, &item_size,
-				     item_size, item_size +
-				     sizeof(struct btrfs_item), 1);
-	BUG_ON(ret);
-
+	setup_items_for_insert(trans, root, path, new_key, &item_size,
+			       item_size, item_size +
+			       sizeof(struct btrfs_item), 1);
 	leaf = path->nodes[0];
 	memcpy_extent_buffer(leaf,
 			     btrfs_item_ptr_offset(leaf, path->slots[0]),
@@ -3257,10 +3277,10 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
  * off the end of the item or if we shift the item to chop bytes off
  * the front.
  */
-int btrfs_truncate_item(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root,
-			struct btrfs_path *path,
-			u32 new_size, int from_end)
+void btrfs_truncate_item(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 struct btrfs_path *path,
+			 u32 new_size, int from_end)
 {
 	int slot;
 	struct extent_buffer *leaf;
@@ -3271,13 +3291,16 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans,
 	unsigned int old_size;
 	unsigned int size_diff;
 	int i;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	leaf = path->nodes[0];
 	slot = path->slots[0];
 
 	old_size = btrfs_item_size_nr(leaf, slot);
 	if (old_size == new_size)
-		return 0;
+		return;
 
 	nritems = btrfs_header_nritems(leaf);
 	data_end = leaf_data_end(root, leaf);
@@ -3297,8 +3320,9 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans,
 		u32 ioff;
 		item = btrfs_item_nr(leaf, i);
 
-		ioff = btrfs_item_offset(leaf, item);
-		btrfs_set_item_offset(leaf, item, ioff + size_diff);
+		ioff = btrfs_token_item_offset(leaf, item, &token);
+		btrfs_set_token_item_offset(leaf, item,
+					    ioff + size_diff, &token);
 	}
 
 	/* shift the data */
@@ -3350,15 +3374,14 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans,
 		btrfs_print_leaf(root, leaf);
 		BUG();
 	}
-	return 0;
 }
 
 /*
  * make the item pointed to by the path bigger, data_size is the new size.
  */
-int btrfs_extend_item(struct btrfs_trans_handle *trans,
-		      struct btrfs_root *root, struct btrfs_path *path,
-		      u32 data_size)
+void btrfs_extend_item(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct btrfs_path *path,
+		       u32 data_size)
 {
 	int slot;
 	struct extent_buffer *leaf;
@@ -3368,6 +3391,9 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans,
 	unsigned int old_data;
 	unsigned int old_size;
 	int i;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	leaf = path->nodes[0];
 
@@ -3397,8 +3423,9 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans,
 		u32 ioff;
 		item = btrfs_item_nr(leaf, i);
 
-		ioff = btrfs_item_offset(leaf, item);
-		btrfs_set_item_offset(leaf, item, ioff - data_size);
+		ioff = btrfs_token_item_offset(leaf, item, &token);
+		btrfs_set_token_item_offset(leaf, item,
+					    ioff - data_size, &token);
 	}
 
 	/* shift the data */
@@ -3416,7 +3443,6 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans,
 		btrfs_print_leaf(root, leaf);
 		BUG();
 	}
-	return 0;
 }
 
 /*
@@ -3441,6 +3467,9 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
 	unsigned int data_end;
 	struct btrfs_disk_key disk_key;
 	struct btrfs_key found_key;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	for (i = 0; i < nr; i++) {
 		if (total_size + data_size[i] + sizeof(struct btrfs_item) >
@@ -3506,8 +3535,9 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
 			u32 ioff;
 
 			item = btrfs_item_nr(leaf, i);
-			ioff = btrfs_item_offset(leaf, item);
-			btrfs_set_item_offset(leaf, item, ioff - total_data);
+			ioff = btrfs_token_item_offset(leaf, item, &token);
+			btrfs_set_token_item_offset(leaf, item,
+						    ioff - total_data, &token);
 		}
 		/* shift the items */
 		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
@@ -3534,9 +3564,10 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
 		btrfs_set_item_key(leaf, &disk_key, slot + i);
 		item = btrfs_item_nr(leaf, slot + i);
-		btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
+		btrfs_set_token_item_offset(leaf, item,
+					    data_end - data_size[i], &token);
 		data_end -= data_size[i];
-		btrfs_set_item_size(leaf, item, data_size[i]);
+		btrfs_set_token_item_size(leaf, item, data_size[i], &token);
 	}
 	btrfs_set_header_nritems(leaf, nritems + nr);
 	btrfs_mark_buffer_dirty(leaf);
@@ -3544,7 +3575,7 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
 	ret = 0;
 	if (slot == 0) {
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
-		ret = fixup_low_keys(trans, root, path, &disk_key, 1);
+		fixup_low_keys(trans, root, path, &disk_key, 1);
 	}
 
 	if (btrfs_leaf_free_space(root, leaf) < 0) {
@@ -3562,19 +3593,21 @@ out:
  * to save stack depth by doing the bulk of the work in a function
  * that doesn't call btrfs_search_slot
  */
-int setup_items_for_insert(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *root, struct btrfs_path *path,
-			   struct btrfs_key *cpu_key, u32 *data_size,
-			   u32 total_data, u32 total_size, int nr)
+void setup_items_for_insert(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, struct btrfs_path *path,
+			    struct btrfs_key *cpu_key, u32 *data_size,
+			    u32 total_data, u32 total_size, int nr)
 {
 	struct btrfs_item *item;
 	int i;
 	u32 nritems;
 	unsigned int data_end;
 	struct btrfs_disk_key disk_key;
-	int ret;
 	struct extent_buffer *leaf;
 	int slot;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	leaf = path->nodes[0];
 	slot = path->slots[0];
@@ -3606,8 +3639,9 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans,
 			u32 ioff;
 
 			item = btrfs_item_nr(leaf, i);
-			ioff = btrfs_item_offset(leaf, item);
-			btrfs_set_item_offset(leaf, item, ioff - total_data);
+			ioff = btrfs_token_item_offset(leaf, item, &token);
+			btrfs_set_token_item_offset(leaf, item,
+						    ioff - total_data, &token);
 		}
 		/* shift the items */
 		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
@@ -3626,17 +3660,17 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans,
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
 		btrfs_set_item_key(leaf, &disk_key, slot + i);
 		item = btrfs_item_nr(leaf, slot + i);
-		btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
+		btrfs_set_token_item_offset(leaf, item,
+					    data_end - data_size[i], &token);
 		data_end -= data_size[i];
-		btrfs_set_item_size(leaf, item, data_size[i]);
+		btrfs_set_token_item_size(leaf, item, data_size[i], &token);
 	}
 
 	btrfs_set_header_nritems(leaf, nritems + nr);
 
-	ret = 0;
 	if (slot == 0) {
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
-		ret = fixup_low_keys(trans, root, path, &disk_key, 1);
+		fixup_low_keys(trans, root, path, &disk_key, 1);
 	}
 	btrfs_unlock_up_safe(path, 1);
 	btrfs_mark_buffer_dirty(leaf);
@@ -3645,7 +3679,6 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans,
 		btrfs_print_leaf(root, leaf);
 		BUG();
 	}
-	return ret;
 }
 
 /*
@@ -3672,16 +3705,14 @@ int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
 	if (ret == 0)
 		return -EEXIST;
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	slot = path->slots[0];
 	BUG_ON(slot < 0);
 
-	ret = setup_items_for_insert(trans, root, path, cpu_key, data_size,
+	setup_items_for_insert(trans, root, path, cpu_key, data_size,
 			       total_data, total_size, nr);
-
-out:
-	return ret;
+	return 0;
 }
 
 /*
@@ -3717,13 +3748,11 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
  * the tree should have been previously balanced so the deletion does not
  * empty a node.
  */
-static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		   struct btrfs_path *path, int level, int slot)
+static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		    struct btrfs_path *path, int level, int slot)
 {
 	struct extent_buffer *parent = path->nodes[level];
 	u32 nritems;
-	int ret = 0;
-	int wret;
 
 	nritems = btrfs_header_nritems(parent);
 	if (slot != nritems - 1) {
@@ -3743,12 +3772,9 @@ static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		struct btrfs_disk_key disk_key;
 
 		btrfs_node_key(parent, &disk_key, 0);
-		wret = fixup_low_keys(trans, root, path, &disk_key, level + 1);
-		if (wret)
-			ret = wret;
+		fixup_low_keys(trans, root, path, &disk_key, level + 1);
 	}
 	btrfs_mark_buffer_dirty(parent);
-	return ret;
 }
 
 /*
@@ -3761,17 +3787,13 @@ static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
  * The path must have already been setup for deleting the leaf, including
  * all the proper balancing.  path->nodes[1] must be locked.
  */
-static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
-				   struct btrfs_root *root,
-				   struct btrfs_path *path,
-				   struct extent_buffer *leaf)
+static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root,
+				    struct btrfs_path *path,
+				    struct extent_buffer *leaf)
 {
-	int ret;
-
 	WARN_ON(btrfs_header_generation(leaf) != trans->transid);
-	ret = del_ptr(trans, root, path, 1, path->slots[1]);
-	if (ret)
-		return ret;
+	del_ptr(trans, root, path, 1, path->slots[1]);
 
 	/*
 	 * btrfs_free_extent is expensive, we want to make sure we
@@ -3781,8 +3803,9 @@ static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
 
 	root_sub_used(root, leaf->len);
 
+	extent_buffer_get(leaf);
 	btrfs_free_tree_block(trans, root, leaf, 0, 1, 0);
-	return 0;
+	free_extent_buffer_stale(leaf);
 }
 /*
  * delete the item at the leaf level in path.  If that empties
@@ -3799,6 +3822,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	int wret;
 	int i;
 	u32 nritems;
+	struct btrfs_map_token token;
+
+	btrfs_init_map_token(&token);
 
 	leaf = path->nodes[0];
 	last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
@@ -3820,8 +3846,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 			u32 ioff;
 
 			item = btrfs_item_nr(leaf, i);
-			ioff = btrfs_item_offset(leaf, item);
-			btrfs_set_item_offset(leaf, item, ioff + dsize);
+			ioff = btrfs_token_item_offset(leaf, item, &token);
+			btrfs_set_token_item_offset(leaf, item,
+						    ioff + dsize, &token);
 		}
 
 		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
@@ -3839,8 +3866,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		} else {
 			btrfs_set_path_blocking(path);
 			clean_tree_block(trans, root, leaf);
-			ret = btrfs_del_leaf(trans, root, path, leaf);
-			BUG_ON(ret);
+			btrfs_del_leaf(trans, root, path, leaf);
 		}
 	} else {
 		int used = leaf_space_used(leaf, 0, nritems);
@@ -3848,10 +3874,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 			struct btrfs_disk_key disk_key;
 
 			btrfs_item_key(leaf, &disk_key, 0);
-			wret = fixup_low_keys(trans, root, path,
-					      &disk_key, 1);
-			if (wret)
-				ret = wret;
+			fixup_low_keys(trans, root, path, &disk_key, 1);
 		}
 
 		/* delete the leaf if it is mostly empty */
@@ -3879,9 +3902,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 
 			if (btrfs_header_nritems(leaf) == 0) {
 				path->slots[1] = slot;
-				ret = btrfs_del_leaf(trans, root, path, leaf);
-				BUG_ON(ret);
+				btrfs_del_leaf(trans, root, path, leaf);
 				free_extent_buffer(leaf);
+				ret = 0;
 			} else {
 				/* if we're still in the path, make sure
 				 * we're dirty.  Otherwise, one of the
@@ -4029,7 +4052,7 @@ again:
 			tmp = btrfs_find_tree_block(root, blockptr,
 					    btrfs_level_size(root, level - 1));
 
-			if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+			if (tmp && btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
 				free_extent_buffer(tmp);
 				break;
 			}
@@ -4059,18 +4082,18 @@ find_next_key:
 		path->slots[level] = slot;
 		if (level == path->lowest_level) {
 			ret = 0;
-			unlock_up(path, level, 1);
+			unlock_up(path, level, 1, 0, NULL);
 			goto out;
 		}
 		btrfs_set_path_blocking(path);
 		cur = read_node_slot(root, cur, slot);
-		BUG_ON(!cur);
+		BUG_ON(!cur); /* -ENOMEM */
 
 		btrfs_tree_read_lock(cur);
 
 		path->locks[level - 1] = BTRFS_READ_LOCK;
 		path->nodes[level - 1] = cur;
-		unlock_up(path, level, 1);
+		unlock_up(path, level, 1, 0, NULL);
 		btrfs_clear_path_blocking(path, NULL, 0);
 	}
 out:
@@ -4152,7 +4175,8 @@ next:
 				struct extent_buffer *cur;
 				cur = btrfs_find_tree_block(root, blockptr,
 					    btrfs_level_size(root, level - 1));
-				if (!cur || !btrfs_buffer_uptodate(cur, gen)) {
+				if (!cur ||
+				    btrfs_buffer_uptodate(cur, gen, 1) <= 0) {
 					slot++;
 					if (cur)
 						free_extent_buffer(cur);
@@ -4306,7 +4330,7 @@ again:
 	}
 	ret = 0;
 done:
-	unlock_up(path, 0, 1);
+	unlock_up(path, 0, 1, 0, NULL);
 	path->leave_spinning = old_spinning;
 	if (!old_spinning)
 		btrfs_set_path_blocking(path);
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 91128d7e9d3..8fd72331d60 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -48,6 +48,8 @@ struct btrfs_ordered_sum;
 
 #define BTRFS_MAGIC "_BHRfS_M"
 
+#define BTRFS_MAX_MIRRORS 2
+
 #define BTRFS_MAX_LEVEL 8
 
 #define BTRFS_COMPAT_EXTENT_TREE_V0
@@ -138,6 +140,12 @@ struct btrfs_ordered_sum;
 #define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2
 
 /*
+ * the max metadata block size.  This limit is somewhat artificial,
+ * but the memmove costs go through the roof for larger blocks.
+ */
+#define BTRFS_MAX_METADATA_BLOCKSIZE 65536
+
+/*
  * we can actually store much bigger names, but lets not confuse the rest
  * of linux
  */
@@ -461,6 +469,19 @@ struct btrfs_super_block {
 #define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL	(1ULL << 1)
 #define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS	(1ULL << 2)
 #define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO	(1ULL << 3)
+/*
+ * some patches floated around with a second compression method
+ * lets save that incompat here for when they do get in
+ * Note we don't actually support it, we're just reserving the
+ * number
+ */
+#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZOv2	(1ULL << 4)
+
+/*
+ * older kernels tried to do bigger metadata blocks, but the
+ * code was pretty buggy.  Lets not let them try anymore.
+ */
+#define BTRFS_FEATURE_INCOMPAT_BIG_METADATA	(1ULL << 5)
 
 #define BTRFS_FEATURE_COMPAT_SUPP		0ULL
 #define BTRFS_FEATURE_COMPAT_RO_SUPP		0ULL
@@ -468,6 +489,7 @@ struct btrfs_super_block {
 	(BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF |		\
 	 BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL |	\
 	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS |		\
+	 BTRFS_FEATURE_INCOMPAT_BIG_METADATA |		\
 	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO)
 
 /*
@@ -829,6 +851,21 @@ struct btrfs_csum_item {
  */
 #define BTRFS_AVAIL_ALLOC_BIT_SINGLE	(1ULL << 48)
 
+#define BTRFS_EXTENDED_PROFILE_MASK	(BTRFS_BLOCK_GROUP_PROFILE_MASK | \
+					 BTRFS_AVAIL_ALLOC_BIT_SINGLE)
+
+static inline u64 chunk_to_extended(u64 flags)
+{
+	if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0)
+		flags |= BTRFS_AVAIL_ALLOC_BIT_SINGLE;
+
+	return flags;
+}
+static inline u64 extended_to_chunk(u64 flags)
+{
+	return flags & ~BTRFS_AVAIL_ALLOC_BIT_SINGLE;
+}
+
 struct btrfs_block_group_item {
 	__le64 used;
 	__le64 chunk_objectid;
@@ -1041,7 +1078,7 @@ struct btrfs_fs_info {
 	 * is required instead of the faster short fsync log commits
 	 */
 	u64 last_trans_log_full_commit;
-	unsigned long mount_opt:21;
+	unsigned long mount_opt;
 	unsigned long compress_type:4;
 	u64 max_inline;
 	u64 alloc_start;
@@ -1503,6 +1540,7 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_SKIP_BALANCE	(1 << 19)
 #define BTRFS_MOUNT_CHECK_INTEGRITY	(1 << 20)
 #define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21)
+#define BTRFS_MOUNT_PANIC_ON_FATAL_ERROR	(1 << 22)
 
 #define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
@@ -1526,6 +1564,17 @@ struct btrfs_ioctl_defrag_range_args {
 
 #define BTRFS_INODE_ROOT_ITEM_INIT	(1 << 31)
 
+struct btrfs_map_token {
+	struct extent_buffer *eb;
+	char *kaddr;
+	unsigned long offset;
+};
+
+static inline void btrfs_init_map_token (struct btrfs_map_token *token)
+{
+	memset(token, 0, sizeof(*token));
+}
+
 /* some macros to generate set/get funcs for the struct fields.  This
  * assumes there is a lefoo_to_cpu for every type, so lets make a simple
  * one for u8:
@@ -1549,20 +1598,22 @@ struct btrfs_ioctl_defrag_range_args {
 #ifndef BTRFS_SETGET_FUNCS
 #define BTRFS_SETGET_FUNCS(name, type, member, bits)			\
 u##bits btrfs_##name(struct extent_buffer *eb, type *s);		\
+u##bits btrfs_token_##name(struct extent_buffer *eb, type *s, struct btrfs_map_token *token);		\
+void btrfs_set_token_##name(struct extent_buffer *eb, type *s, u##bits val, struct btrfs_map_token *token);\
 void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val);
 #endif
 
 #define BTRFS_SETGET_HEADER_FUNCS(name, type, member, bits)		\
 static inline u##bits btrfs_##name(struct extent_buffer *eb)		\
 {									\
-	type *p = page_address(eb->first_page);				\
+	type *p = page_address(eb->pages[0]);				\
 	u##bits res = le##bits##_to_cpu(p->member);			\
 	return res;							\
 }									\
 static inline void btrfs_set_##name(struct extent_buffer *eb,		\
 				    u##bits val)			\
 {									\
-	type *p = page_address(eb->first_page);				\
+	type *p = page_address(eb->pages[0]);				\
 	p->member = cpu_to_le##bits(val);				\
 }
 
@@ -2466,8 +2517,7 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 				  struct btrfs_root *root,
 				  u64 num_bytes, u64 min_alloc_size,
 				  u64 empty_size, u64 hint_byte,
-				  u64 search_end, struct btrfs_key *ins,
-				  u64 data);
+				  struct btrfs_key *ins, u64 data);
 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		  struct extent_buffer *buf, int full_backref, int for_cow);
 int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
@@ -2484,8 +2534,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans,
 int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len);
 int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root,
 				       u64 start, u64 len);
-int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
-				struct btrfs_root *root);
+void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root);
 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root);
 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
@@ -2548,8 +2598,8 @@ void btrfs_block_rsv_release(struct btrfs_root *root,
 			     u64 num_bytes);
 int btrfs_set_block_group_ro(struct btrfs_root *root,
 			     struct btrfs_block_group_cache *cache);
-int btrfs_set_block_group_rw(struct btrfs_root *root,
-			     struct btrfs_block_group_cache *cache);
+void btrfs_set_block_group_rw(struct btrfs_root *root,
+			      struct btrfs_block_group_cache *cache);
 void btrfs_put_block_group_cache(struct btrfs_fs_info *info);
 u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo);
 int btrfs_error_unpin_extent_range(struct btrfs_root *root,
@@ -2568,9 +2618,9 @@ int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2);
 int btrfs_previous_item(struct btrfs_root *root,
 			struct btrfs_path *path, u64 min_objectid,
 			int type);
-int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root, struct btrfs_path *path,
-			    struct btrfs_key *new_key);
+void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, struct btrfs_path *path,
+			     struct btrfs_key *new_key);
 struct extent_buffer *btrfs_root_node(struct btrfs_root *root);
 struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root);
 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
@@ -2590,12 +2640,13 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans,
 		      struct extent_buffer **cow_ret, u64 new_root_objectid);
 int btrfs_block_can_be_shared(struct btrfs_root *root,
 			      struct extent_buffer *buf);
-int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root
-		      *root, struct btrfs_path *path, u32 data_size);
-int btrfs_truncate_item(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root,
-			struct btrfs_path *path,
-			u32 new_size, int from_end);
+void btrfs_extend_item(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct btrfs_path *path,
+		       u32 data_size);
+void btrfs_truncate_item(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 struct btrfs_path *path,
+			 u32 new_size, int from_end);
 int btrfs_split_item(struct btrfs_trans_handle *trans,
 		     struct btrfs_root *root,
 		     struct btrfs_path *path,
@@ -2629,10 +2680,10 @@ static inline int btrfs_del_item(struct btrfs_trans_handle *trans,
 	return btrfs_del_items(trans, root, path, path->slots[0], 1);
 }
 
-int setup_items_for_insert(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *root, struct btrfs_path *path,
-			   struct btrfs_key *cpu_key, u32 *data_size,
-			   u32 total_data, u32 total_size, int nr);
+void setup_items_for_insert(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, struct btrfs_path *path,
+			    struct btrfs_key *cpu_key, u32 *data_size,
+			    u32 total_data, u32 total_size, int nr);
 int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
 		      *root, struct btrfs_key *key, void *data, u32 data_size);
 int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
@@ -2659,9 +2710,9 @@ static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p)
 }
 int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path);
 int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf);
-void btrfs_drop_snapshot(struct btrfs_root *root,
-			 struct btrfs_block_rsv *block_rsv, int update_ref,
-			 int for_reloc);
+int __must_check btrfs_drop_snapshot(struct btrfs_root *root,
+				     struct btrfs_block_rsv *block_rsv,
+				     int update_ref, int for_reloc);
 int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 			struct btrfs_root *root,
 			struct extent_buffer *node,
@@ -2687,24 +2738,6 @@ static inline void free_fs_info(struct btrfs_fs_info *fs_info)
 	kfree(fs_info->super_for_commit);
 	kfree(fs_info);
 }
-/**
- * profile_is_valid - tests whether a given profile is valid and reduced
- * @flags: profile to validate
- * @extended: if true @flags is treated as an extended profile
- */
-static inline int profile_is_valid(u64 flags, int extended)
-{
-	u64 mask = ~BTRFS_BLOCK_GROUP_PROFILE_MASK;
-
-	flags &= ~BTRFS_BLOCK_GROUP_TYPE_MASK;
-	if (extended)
-		mask &= ~BTRFS_AVAIL_ALLOC_BIT_SINGLE;
-
-	if (flags & mask)
-		return 0;
-	/* true if zero or exactly one bit set */
-	return (flags & (~flags + 1)) == flags;
-}
 
 /* root-item.c */
 int btrfs_find_root_ref(struct btrfs_root *tree_root,
@@ -2723,9 +2756,10 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
 		      *root, struct btrfs_key *key, struct btrfs_root_item
 		      *item);
-int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
-		      *root, struct btrfs_key *key, struct btrfs_root_item
-		      *item);
+int __must_check btrfs_update_root(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct btrfs_key *key,
+				   struct btrfs_root_item *item);
 int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct
 			 btrfs_root_item *item, struct btrfs_key *key);
 int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid);
@@ -2909,7 +2943,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root);
 void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root);
 int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size);
-int btrfs_invalidate_inodes(struct btrfs_root *root);
+void btrfs_invalidate_inodes(struct btrfs_root *root);
 void btrfs_add_delayed_iput(struct inode *inode);
 void btrfs_run_delayed_iputs(struct btrfs_root *root);
 int btrfs_prealloc_file_range(struct inode *inode, int mode,
@@ -2961,13 +2995,41 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size);
 /* super.c */
 int btrfs_parse_options(struct btrfs_root *root, char *options);
 int btrfs_sync_fs(struct super_block *sb, int wait);
+void btrfs_printk(struct btrfs_fs_info *fs_info, const char *fmt, ...);
 void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
-		     unsigned int line, int errno);
+		     unsigned int line, int errno, const char *fmt, ...);
+
+void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root, const char *function,
+			       unsigned int line, int errno);
+
+#define btrfs_abort_transaction(trans, root, errno)		\
+do {								\
+	__btrfs_abort_transaction(trans, root, __func__,	\
+				  __LINE__, errno);		\
+} while (0)
 
 #define btrfs_std_error(fs_info, errno)				\
 do {								\
 	if ((errno))						\
-		__btrfs_std_error((fs_info), __func__, __LINE__, (errno));\
+		__btrfs_std_error((fs_info), __func__,		\
+				   __LINE__, (errno), NULL);	\
+} while (0)
+
+#define btrfs_error(fs_info, errno, fmt, args...)		\
+do {								\
+	__btrfs_std_error((fs_info), __func__, __LINE__,	\
+			  (errno), fmt, ##args);		\
+} while (0)
+
+void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function,
+		   unsigned int line, int errno, const char *fmt, ...);
+
+#define btrfs_panic(fs_info, errno, fmt, args...)			\
+do {									\
+	struct btrfs_fs_info *_i = (fs_info);				\
+	__btrfs_panic(_i, __func__, __LINE__, errno, fmt, ##args);	\
+	BUG_ON(!(_i->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR));	\
 } while (0)
 
 /* acl.c */
@@ -3003,16 +3065,17 @@ void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
 void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
 			      struct btrfs_pending_snapshot *pending,
 			      u64 *bytes_to_reserve);
-void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
+int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 			      struct btrfs_pending_snapshot *pending);
 
 /* scrub.c */
 int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
 		    struct btrfs_scrub_progress *progress, int readonly);
-int btrfs_scrub_pause(struct btrfs_root *root);
-int btrfs_scrub_pause_super(struct btrfs_root *root);
-int btrfs_scrub_continue(struct btrfs_root *root);
-int btrfs_scrub_continue_super(struct btrfs_root *root);
+void btrfs_scrub_pause(struct btrfs_root *root);
+void btrfs_scrub_pause_super(struct btrfs_root *root);
+void btrfs_scrub_continue(struct btrfs_root *root);
+void btrfs_scrub_continue_super(struct btrfs_root *root);
+int __btrfs_scrub_cancel(struct btrfs_fs_info *info);
 int btrfs_scrub_cancel(struct btrfs_root *root);
 int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev);
 int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid);
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index fe4cd0f1cef..03e3748d84d 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -115,6 +115,7 @@ static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode)
 	return NULL;
 }
 
+/* Will return either the node or PTR_ERR(-ENOMEM) */
 static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node(
 							struct inode *inode)
 {
@@ -836,10 +837,8 @@ static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans,
 	btrfs_clear_path_blocking(path, NULL, 0);
 
 	/* insert the keys of the items */
-	ret = setup_items_for_insert(trans, root, path, keys, data_size,
-				     total_data_size, total_size, nitems);
-	if (ret)
-		goto error;
+	setup_items_for_insert(trans, root, path, keys, data_size,
+			       total_data_size, total_size, nitems);
 
 	/* insert the dir index items */
 	slot = path->slots[0];
@@ -1108,16 +1107,25 @@ static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
-/* Called when committing the transaction. */
+/*
+ * Called when committing the transaction.
+ * Returns 0 on success.
+ * Returns < 0 on error and returns with an aborted transaction with any
+ * outstanding delayed items cleaned up.
+ */
 int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
 			    struct btrfs_root *root)
 {
+	struct btrfs_root *curr_root = root;
 	struct btrfs_delayed_root *delayed_root;
 	struct btrfs_delayed_node *curr_node, *prev_node;
 	struct btrfs_path *path;
 	struct btrfs_block_rsv *block_rsv;
 	int ret = 0;
 
+	if (trans->aborted)
+		return -EIO;
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -1130,17 +1138,18 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
 
 	curr_node = btrfs_first_delayed_node(delayed_root);
 	while (curr_node) {
-		root = curr_node->root;
-		ret = btrfs_insert_delayed_items(trans, path, root,
+		curr_root = curr_node->root;
+		ret = btrfs_insert_delayed_items(trans, path, curr_root,
 						 curr_node);
 		if (!ret)
-			ret = btrfs_delete_delayed_items(trans, path, root,
-							 curr_node);
+			ret = btrfs_delete_delayed_items(trans, path,
+						curr_root, curr_node);
 		if (!ret)
-			ret = btrfs_update_delayed_inode(trans, root, path,
-							 curr_node);
+			ret = btrfs_update_delayed_inode(trans, curr_root,
+						path, curr_node);
 		if (ret) {
 			btrfs_release_delayed_node(curr_node);
+			btrfs_abort_transaction(trans, root, ret);
 			break;
 		}
 
@@ -1151,6 +1160,7 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
 
 	btrfs_free_path(path);
 	trans->block_rsv = block_rsv;
+
 	return ret;
 }
 
@@ -1371,6 +1381,7 @@ void btrfs_balance_delayed_items(struct btrfs_root *root)
 	btrfs_wq_run_delayed_node(delayed_root, root, 0);
 }
 
+/* Will return 0 or -ENOMEM */
 int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root, const char *name,
 				   int name_len, struct inode *dir,
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 66e4f29505a..69f22e3ab3b 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -420,7 +420,7 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing,
  * this does all the dirty work in terms of maintaining the correct
  * overall modification count.
  */
-static noinline int add_delayed_ref_head(struct btrfs_fs_info *fs_info,
+static noinline void add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 					struct btrfs_trans_handle *trans,
 					struct btrfs_delayed_ref_node *ref,
 					u64 bytenr, u64 num_bytes,
@@ -487,20 +487,19 @@ static noinline int add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(ref);
+		kfree(head_ref);
 	} else {
 		delayed_refs->num_heads++;
 		delayed_refs->num_heads_ready++;
 		delayed_refs->num_entries++;
 		trans->delayed_ref_updates++;
 	}
-	return 0;
 }
 
 /*
  * helper to insert a delayed tree ref into the rbtree.
  */
-static noinline int add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
+static noinline void add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 					 struct btrfs_trans_handle *trans,
 					 struct btrfs_delayed_ref_node *ref,
 					 u64 bytenr, u64 num_bytes, u64 parent,
@@ -549,18 +548,17 @@ static noinline int add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(ref);
+		kfree(full_ref);
 	} else {
 		delayed_refs->num_entries++;
 		trans->delayed_ref_updates++;
 	}
-	return 0;
 }
 
 /*
  * helper to insert a delayed data ref into the rbtree.
  */
-static noinline int add_delayed_data_ref(struct btrfs_fs_info *fs_info,
+static noinline void add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 					 struct btrfs_trans_handle *trans,
 					 struct btrfs_delayed_ref_node *ref,
 					 u64 bytenr, u64 num_bytes, u64 parent,
@@ -611,12 +609,11 @@ static noinline int add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(ref);
+		kfree(full_ref);
 	} else {
 		delayed_refs->num_entries++;
 		trans->delayed_ref_updates++;
 	}
-	return 0;
 }
 
 /*
@@ -634,7 +631,6 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 	struct btrfs_delayed_tree_ref *ref;
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-	int ret;
 
 	BUG_ON(extent_op && extent_op->is_data);
 	ref = kmalloc(sizeof(*ref), GFP_NOFS);
@@ -656,14 +652,12 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 	 * insert both the head node and the new ref without dropping
 	 * the spin lock
 	 */
-	ret = add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
+	add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
 				   num_bytes, action, 0);
-	BUG_ON(ret);
 
-	ret = add_delayed_tree_ref(fs_info, trans, &ref->node, bytenr,
+	add_delayed_tree_ref(fs_info, trans, &ref->node, bytenr,
 				   num_bytes, parent, ref_root, level, action,
 				   for_cow);
-	BUG_ON(ret);
 	if (!need_ref_seq(for_cow, ref_root) &&
 	    waitqueue_active(&delayed_refs->seq_wait))
 		wake_up(&delayed_refs->seq_wait);
@@ -685,7 +679,6 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	struct btrfs_delayed_data_ref *ref;
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-	int ret;
 
 	BUG_ON(extent_op && !extent_op->is_data);
 	ref = kmalloc(sizeof(*ref), GFP_NOFS);
@@ -707,14 +700,12 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	 * insert both the head node and the new ref without dropping
 	 * the spin lock
 	 */
-	ret = add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
+	add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
 				   num_bytes, action, 1);
-	BUG_ON(ret);
 
-	ret = add_delayed_data_ref(fs_info, trans, &ref->node, bytenr,
+	add_delayed_data_ref(fs_info, trans, &ref->node, bytenr,
 				   num_bytes, parent, ref_root, owner, offset,
 				   action, for_cow);
-	BUG_ON(ret);
 	if (!need_ref_seq(for_cow, ref_root) &&
 	    waitqueue_active(&delayed_refs->seq_wait))
 		wake_up(&delayed_refs->seq_wait);
@@ -729,7 +720,6 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-	int ret;
 
 	head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
 	if (!head_ref)
@@ -740,10 +730,9 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 	delayed_refs = &trans->transaction->delayed_refs;
 	spin_lock(&delayed_refs->lock);
 
-	ret = add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
+	add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
 				   num_bytes, BTRFS_UPDATE_DELAYED_HEAD,
 				   extent_op->is_data);
-	BUG_ON(ret);
 
 	if (waitqueue_active(&delayed_refs->seq_wait))
 		wake_up(&delayed_refs->seq_wait);
diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c
index 31d84e78129..c1a074d0696 100644
--- a/fs/btrfs/dir-item.c
+++ b/fs/btrfs/dir-item.c
@@ -49,9 +49,8 @@ static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle
 		di = btrfs_match_dir_item_name(root, path, name, name_len);
 		if (di)
 			return ERR_PTR(-EEXIST);
-		ret = btrfs_extend_item(trans, root, path, data_size);
-	}
-	if (ret < 0)
+		btrfs_extend_item(trans, root, path, data_size);
+	} else if (ret < 0)
 		return ERR_PTR(ret);
 	WARN_ON(ret > 0);
 	leaf = path->nodes[0];
@@ -116,6 +115,7 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,
  * 'location' is the key to stuff into the directory item, 'type' is the
  * type of the inode we're pointing to, and 'index' is the sequence number
  * to use for the second index (if one is created).
+ * Will return 0 or -ENOMEM
  */
 int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root
 			  *root, const char *name, int name_len,
@@ -383,8 +383,8 @@ int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans,
 		start = btrfs_item_ptr_offset(leaf, path->slots[0]);
 		memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
 			item_len - (ptr + sub_item_len - start));
-		ret = btrfs_truncate_item(trans, root, path,
-					  item_len - sub_item_len, 1);
+		btrfs_truncate_item(trans, root, path,
+				    item_len - sub_item_len, 1);
 	}
 	return ret;
 }
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 534266fe505..a7ffc88a7db 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -48,20 +48,19 @@
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
 static void free_fs_root(struct btrfs_root *root);
-static void btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
+static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
 				    int read_only);
-static int btrfs_destroy_ordered_operations(struct btrfs_root *root);
-static int btrfs_destroy_ordered_extents(struct btrfs_root *root);
+static void btrfs_destroy_ordered_operations(struct btrfs_root *root);
+static void btrfs_destroy_ordered_extents(struct btrfs_root *root);
 static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 				      struct btrfs_root *root);
-static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t);
-static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root);
+static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t);
+static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root);
 static int btrfs_destroy_marked_extents(struct btrfs_root *root,
 					struct extent_io_tree *dirty_pages,
 					int mark);
 static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
 				       struct extent_io_tree *pinned_extents);
-static int btrfs_cleanup_transaction(struct btrfs_root *root);
 
 /*
  * end_io_wq structs are used to do processing in task context when an IO is
@@ -99,6 +98,7 @@ struct async_submit_bio {
 	 */
 	u64 bio_offset;
 	struct btrfs_work work;
+	int error;
 };
 
 /*
@@ -323,7 +323,8 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
  * in the wrong place.
  */
 static int verify_parent_transid(struct extent_io_tree *io_tree,
-				 struct extent_buffer *eb, u64 parent_transid)
+				 struct extent_buffer *eb, u64 parent_transid,
+				 int atomic)
 {
 	struct extent_state *cached_state = NULL;
 	int ret;
@@ -331,9 +332,12 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 	if (!parent_transid || btrfs_header_generation(eb) == parent_transid)
 		return 0;
 
+	if (atomic)
+		return -EAGAIN;
+
 	lock_extent_bits(io_tree, eb->start, eb->start + eb->len - 1,
-			 0, &cached_state, GFP_NOFS);
-	if (extent_buffer_uptodate(io_tree, eb, cached_state) &&
+			 0, &cached_state);
+	if (extent_buffer_uptodate(eb) &&
 	    btrfs_header_generation(eb) == parent_transid) {
 		ret = 0;
 		goto out;
@@ -344,7 +348,7 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 		       (unsigned long long)parent_transid,
 		       (unsigned long long)btrfs_header_generation(eb));
 	ret = 1;
-	clear_extent_buffer_uptodate(io_tree, eb, &cached_state);
+	clear_extent_buffer_uptodate(eb);
 out:
 	unlock_extent_cached(io_tree, eb->start, eb->start + eb->len - 1,
 			     &cached_state, GFP_NOFS);
@@ -360,9 +364,11 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
 					  u64 start, u64 parent_transid)
 {
 	struct extent_io_tree *io_tree;
+	int failed = 0;
 	int ret;
 	int num_copies = 0;
 	int mirror_num = 0;
+	int failed_mirror = 0;
 
 	clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
 	io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
@@ -370,9 +376,9 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
 		ret = read_extent_buffer_pages(io_tree, eb, start,
 					       WAIT_COMPLETE,
 					       btree_get_extent, mirror_num);
-		if (!ret &&
-		    !verify_parent_transid(io_tree, eb, parent_transid))
-			return ret;
+		if (!ret && !verify_parent_transid(io_tree, eb,
+						   parent_transid, 0))
+			break;
 
 		/*
 		 * This buffer's crc is fine, but its contents are corrupted, so
@@ -380,18 +386,30 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
 		 * any less wrong.
 		 */
 		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
-			return ret;
+			break;
 
 		num_copies = btrfs_num_copies(&root->fs_info->mapping_tree,
 					      eb->start, eb->len);
 		if (num_copies == 1)
-			return ret;
+			break;
+
+		if (!failed_mirror) {
+			failed = 1;
+			failed_mirror = eb->read_mirror;
+		}
 
 		mirror_num++;
+		if (mirror_num == failed_mirror)
+			mirror_num++;
+
 		if (mirror_num > num_copies)
-			return ret;
+			break;
 	}
-	return -EIO;
+
+	if (failed && !ret)
+		repair_eb_io_failure(root, eb, failed_mirror);
+
+	return ret;
 }
 
 /*
@@ -404,50 +422,27 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 	struct extent_io_tree *tree;
 	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
 	u64 found_start;
-	unsigned long len;
 	struct extent_buffer *eb;
-	int ret;
 
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
 
-	if (page->private == EXTENT_PAGE_PRIVATE) {
-		WARN_ON(1);
-		goto out;
-	}
-	if (!page->private) {
-		WARN_ON(1);
-		goto out;
-	}
-	len = page->private >> 2;
-	WARN_ON(len == 0);
-
-	eb = alloc_extent_buffer(tree, start, len, page);
-	if (eb == NULL) {
-		WARN_ON(1);
-		goto out;
-	}
-	ret = btree_read_extent_buffer_pages(root, eb, start + PAGE_CACHE_SIZE,
-					     btrfs_header_generation(eb));
-	BUG_ON(ret);
-	WARN_ON(!btrfs_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN));
-
+	eb = (struct extent_buffer *)page->private;
+	if (page != eb->pages[0])
+		return 0;
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
 		WARN_ON(1);
-		goto err;
+		return 0;
 	}
-	if (eb->first_page != page) {
+	if (eb->pages[0] != page) {
 		WARN_ON(1);
-		goto err;
+		return 0;
 	}
 	if (!PageUptodate(page)) {
 		WARN_ON(1);
-		goto err;
+		return 0;
 	}
 	csum_tree_block(root, eb, 0);
-err:
-	free_extent_buffer(eb);
-out:
 	return 0;
 }
 
@@ -537,34 +532,75 @@ static noinline int check_leaf(struct btrfs_root *root,
 	return 0;
 }
 
+struct extent_buffer *find_eb_for_page(struct extent_io_tree *tree,
+				       struct page *page, int max_walk)
+{
+	struct extent_buffer *eb;
+	u64 start = page_offset(page);
+	u64 target = start;
+	u64 min_start;
+
+	if (start < max_walk)
+		min_start = 0;
+	else
+		min_start = start - max_walk;
+
+	while (start >= min_start) {
+		eb = find_extent_buffer(tree, start, 0);
+		if (eb) {
+			/*
+			 * we found an extent buffer and it contains our page
+			 * horray!
+			 */
+			if (eb->start <= target &&
+			    eb->start + eb->len > target)
+				return eb;
+
+			/* we found an extent buffer that wasn't for us */
+			free_extent_buffer(eb);
+			return NULL;
+		}
+		if (start == 0)
+			break;
+		start -= PAGE_CACHE_SIZE;
+	}
+	return NULL;
+}
+
 static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
-			       struct extent_state *state)
+			       struct extent_state *state, int mirror)
 {
 	struct extent_io_tree *tree;
 	u64 found_start;
 	int found_level;
-	unsigned long len;
 	struct extent_buffer *eb;
 	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
 	int ret = 0;
+	int reads_done;
 
-	tree = &BTRFS_I(page->mapping->host)->io_tree;
-	if (page->private == EXTENT_PAGE_PRIVATE)
-		goto out;
 	if (!page->private)
 		goto out;
 
-	len = page->private >> 2;
-	WARN_ON(len == 0);
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	eb = (struct extent_buffer *)page->private;
+
+	/* the pending IO might have been the only thing that kept this buffer
+	 * in memory.  Make sure we have a ref for all this other checks
+	 */
+	extent_buffer_get(eb);
+
+	reads_done = atomic_dec_and_test(&eb->io_pages);
+	if (!reads_done)
+		goto err;
 
-	eb = alloc_extent_buffer(tree, start, len, page);
-	if (eb == NULL) {
+	eb->read_mirror = mirror;
+	if (test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) {
 		ret = -EIO;
-		goto out;
+		goto err;
 	}
 
 	found_start = btrfs_header_bytenr(eb);
-	if (found_start != start) {
+	if (found_start != eb->start) {
 		printk_ratelimited(KERN_INFO "btrfs bad tree block start "
 			       "%llu %llu\n",
 			       (unsigned long long)found_start,
@@ -572,13 +608,6 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		ret = -EIO;
 		goto err;
 	}
-	if (eb->first_page != page) {
-		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_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n",
 			       (unsigned long long)eb->start);
@@ -606,48 +635,31 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		ret = -EIO;
 	}
 
-	end = min_t(u64, eb->len, PAGE_CACHE_SIZE);
-	end = eb->start + end - 1;
+	if (!ret)
+		set_extent_buffer_uptodate(eb);
 err:
 	if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) {
 		clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags);
 		btree_readahead_hook(root, eb, eb->start, ret);
 	}
 
+	if (ret)
+		clear_extent_buffer_uptodate(eb);
 	free_extent_buffer(eb);
 out:
 	return ret;
 }
 
-static int btree_io_failed_hook(struct bio *failed_bio,
-			 struct page *page, u64 start, u64 end,
-			 int mirror_num, struct extent_state *state)
+static int btree_io_failed_hook(struct page *page, int failed_mirror)
 {
-	struct extent_io_tree *tree;
-	unsigned long len;
 	struct extent_buffer *eb;
 	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
 
-	tree = &BTRFS_I(page->mapping->host)->io_tree;
-	if (page->private == EXTENT_PAGE_PRIVATE)
-		goto out;
-	if (!page->private)
-		goto out;
-
-	len = page->private >> 2;
-	WARN_ON(len == 0);
-
-	eb = alloc_extent_buffer(tree, start, len, page);
-	if (eb == NULL)
-		goto out;
-
-	if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) {
-		clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags);
+	eb = (struct extent_buffer *)page->private;
+	set_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
+	eb->read_mirror = failed_mirror;
+	if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags))
 		btree_readahead_hook(root, eb, eb->start, -EIO);
-	}
-	free_extent_buffer(eb);
-
-out:
 	return -EIO;	/* we fixed nothing */
 }
 
@@ -719,11 +731,14 @@ unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info)
 static void run_one_async_start(struct btrfs_work *work)
 {
 	struct async_submit_bio *async;
+	int ret;
 
 	async = container_of(work, struct  async_submit_bio, work);
-	async->submit_bio_start(async->inode, async->rw, async->bio,
-			       async->mirror_num, async->bio_flags,
-			       async->bio_offset);
+	ret = async->submit_bio_start(async->inode, async->rw, async->bio,
+				      async->mirror_num, async->bio_flags,
+				      async->bio_offset);
+	if (ret)
+		async->error = ret;
 }
 
 static void run_one_async_done(struct btrfs_work *work)
@@ -744,6 +759,12 @@ static void run_one_async_done(struct btrfs_work *work)
 	    waitqueue_active(&fs_info->async_submit_wait))
 		wake_up(&fs_info->async_submit_wait);
 
+	/* If an error occured we just want to clean up the bio and move on */
+	if (async->error) {
+		bio_endio(async->bio, async->error);
+		return;
+	}
+
 	async->submit_bio_done(async->inode, async->rw, async->bio,
 			       async->mirror_num, async->bio_flags,
 			       async->bio_offset);
@@ -785,6 +806,8 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 	async->bio_flags = bio_flags;
 	async->bio_offset = bio_offset;
 
+	async->error = 0;
+
 	atomic_inc(&fs_info->nr_async_submits);
 
 	if (rw & REQ_SYNC)
@@ -806,15 +829,18 @@ static int btree_csum_one_bio(struct bio *bio)
 	struct bio_vec *bvec = bio->bi_io_vec;
 	int bio_index = 0;
 	struct btrfs_root *root;
+	int ret = 0;
 
 	WARN_ON(bio->bi_vcnt <= 0);
 	while (bio_index < bio->bi_vcnt) {
 		root = BTRFS_I(bvec->bv_page->mapping->host)->root;
-		csum_dirty_buffer(root, bvec->bv_page);
+		ret = csum_dirty_buffer(root, bvec->bv_page);
+		if (ret)
+			break;
 		bio_index++;
 		bvec++;
 	}
-	return 0;
+	return ret;
 }
 
 static int __btree_submit_bio_start(struct inode *inode, int rw,
@@ -826,8 +852,7 @@ static int __btree_submit_bio_start(struct inode *inode, int rw,
 	 * when we're called for a write, we're already in the async
 	 * submission context.  Just jump into btrfs_map_bio
 	 */
-	btree_csum_one_bio(bio);
-	return 0;
+	return btree_csum_one_bio(bio);
 }
 
 static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
@@ -847,15 +872,16 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 {
 	int ret;
 
-	ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info,
-					  bio, 1);
-	BUG_ON(ret);
-
 	if (!(rw & REQ_WRITE)) {
+
 		/*
 		 * called for a read, do the setup so that checksum validation
 		 * can happen in the async kernel threads
 		 */
+		ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info,
+					  bio, 1);
+		if (ret)
+			return ret;
 		return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio,
 				     mirror_num, 0);
 	}
@@ -893,34 +919,6 @@ static int btree_migratepage(struct address_space *mapping,
 }
 #endif
 
-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);
-	}
-
-	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);
-	}
-	free_extent_buffer(eb);
-
-	unlock_page(page);
-	return 0;
-}
 
 static int btree_writepages(struct address_space *mapping,
 			    struct writeback_control *wbc)
@@ -940,7 +938,7 @@ static int btree_writepages(struct address_space *mapping,
 		if (num_dirty < thresh)
 			return 0;
 	}
-	return extent_writepages(tree, mapping, btree_get_extent, wbc);
+	return btree_write_cache_pages(mapping, wbc);
 }
 
 static int btree_readpage(struct file *file, struct page *page)
@@ -952,16 +950,8 @@ static int btree_readpage(struct file *file, struct page *page)
 
 static int btree_releasepage(struct page *page, gfp_t gfp_flags)
 {
-	struct extent_io_tree *tree;
-	struct extent_map_tree *map;
-	int ret;
-
 	if (PageWriteback(page) || PageDirty(page))
 		return 0;
-
-	tree = &BTRFS_I(page->mapping->host)->io_tree;
-	map = &BTRFS_I(page->mapping->host)->extent_tree;
-
 	/*
 	 * We need to mask out eg. __GFP_HIGHMEM and __GFP_DMA32 as we're doing
 	 * slab allocation from alloc_extent_state down the callchain where
@@ -969,18 +959,7 @@ static int btree_releasepage(struct page *page, gfp_t gfp_flags)
 	 */
 	gfp_flags &= ~GFP_SLAB_BUG_MASK;
 
-	ret = try_release_extent_state(map, tree, page, gfp_flags);
-	if (!ret)
-		return 0;
-
-	ret = try_release_extent_buffer(tree, page);
-	if (ret == 1) {
-		ClearPagePrivate(page);
-		set_page_private(page, 0);
-		page_cache_release(page);
-	}
-
-	return ret;
+	return try_release_extent_buffer(page, gfp_flags);
 }
 
 static void btree_invalidatepage(struct page *page, unsigned long offset)
@@ -998,15 +977,28 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	}
 }
 
+static int btree_set_page_dirty(struct page *page)
+{
+	struct extent_buffer *eb;
+
+	BUG_ON(!PagePrivate(page));
+	eb = (struct extent_buffer *)page->private;
+	BUG_ON(!eb);
+	BUG_ON(!test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
+	BUG_ON(!atomic_read(&eb->refs));
+	btrfs_assert_tree_locked(eb);
+	return __set_page_dirty_nobuffers(page);
+}
+
 static const struct address_space_operations btree_aops = {
 	.readpage	= btree_readpage,
-	.writepage	= btree_writepage,
 	.writepages	= btree_writepages,
 	.releasepage	= btree_releasepage,
 	.invalidatepage = btree_invalidatepage,
 #ifdef CONFIG_MIGRATION
 	.migratepage	= btree_migratepage,
 #endif
+	.set_page_dirty = btree_set_page_dirty,
 };
 
 int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize,
@@ -1049,7 +1041,7 @@ int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize,
 	if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) {
 		free_extent_buffer(buf);
 		return -EIO;
-	} else if (extent_buffer_uptodate(io_tree, buf, NULL)) {
+	} else if (extent_buffer_uptodate(buf)) {
 		*eb = buf;
 	} else {
 		free_extent_buffer(buf);
@@ -1074,20 +1066,20 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
 	struct extent_buffer *eb;
 
 	eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
-				 bytenr, blocksize, NULL);
+				 bytenr, blocksize);
 	return eb;
 }
 
 
 int btrfs_write_tree_block(struct extent_buffer *buf)
 {
-	return filemap_fdatawrite_range(buf->first_page->mapping, buf->start,
+	return filemap_fdatawrite_range(buf->pages[0]->mapping, buf->start,
 					buf->start + buf->len - 1);
 }
 
 int btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
 {
-	return filemap_fdatawait_range(buf->first_page->mapping,
+	return filemap_fdatawait_range(buf->pages[0]->mapping,
 				       buf->start, buf->start + buf->len - 1);
 }
 
@@ -1102,17 +1094,13 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 		return NULL;
 
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
-
-	if (ret == 0)
-		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
 	return buf;
 
 }
 
-int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		     struct extent_buffer *buf)
+void clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		      struct extent_buffer *buf)
 {
-	struct inode *btree_inode = root->fs_info->btree_inode;
 	if (btrfs_header_generation(buf) ==
 	    root->fs_info->running_transaction->transid) {
 		btrfs_assert_tree_locked(buf);
@@ -1121,23 +1109,27 @@ int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 			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);
+			else {
+				spin_unlock(&root->fs_info->delalloc_lock);
+				btrfs_panic(root->fs_info, -EOVERFLOW,
+					  "Can't clear %lu bytes from "
+					  " dirty_mdatadata_bytes (%lu)",
+					  buf->len,
+					  root->fs_info->dirty_metadata_bytes);
+			}
 			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);
+		clear_extent_buffer_dirty(buf);
 	}
-	return 0;
 }
 
-static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
-			u32 stripesize, struct btrfs_root *root,
-			struct btrfs_fs_info *fs_info,
-			u64 objectid)
+static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
+			 u32 stripesize, struct btrfs_root *root,
+			 struct btrfs_fs_info *fs_info,
+			 u64 objectid)
 {
 	root->node = NULL;
 	root->commit_root = NULL;
@@ -1189,13 +1181,12 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->defrag_running = 0;
 	root->root_key.objectid = objectid;
 	root->anon_dev = 0;
-	return 0;
 }
 
-static int find_and_setup_root(struct btrfs_root *tree_root,
-			       struct btrfs_fs_info *fs_info,
-			       u64 objectid,
-			       struct btrfs_root *root)
+static int __must_check find_and_setup_root(struct btrfs_root *tree_root,
+					    struct btrfs_fs_info *fs_info,
+					    u64 objectid,
+					    struct btrfs_root *root)
 {
 	int ret;
 	u32 blocksize;
@@ -1208,14 +1199,15 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
 				   &root->root_item, &root->root_key);
 	if (ret > 0)
 		return -ENOENT;
-	BUG_ON(ret);
+	else if (ret < 0)
+		return ret;
 
 	generation = btrfs_root_generation(&root->root_item);
 	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
 	root->commit_root = NULL;
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     blocksize, generation);
-	if (!root->node || !btrfs_buffer_uptodate(root->node, generation)) {
+	if (!root->node || !btrfs_buffer_uptodate(root->node, generation, 0)) {
 		free_extent_buffer(root->node);
 		root->node = NULL;
 		return -EIO;
@@ -1377,7 +1369,7 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
 	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);
+	BUG_ON(!root->node); /* -ENOMEM */
 out:
 	if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
 		root->ref_cows = 1;
@@ -1513,41 +1505,6 @@ static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
 	return 0;
 }
 
-static int bio_ready_for_csum(struct bio *bio)
-{
-	u64 length = 0;
-	u64 buf_len = 0;
-	u64 start = 0;
-	struct page *page;
-	struct extent_io_tree *io_tree = NULL;
-	struct bio_vec *bvec;
-	int i;
-	int ret;
-
-	bio_for_each_segment(bvec, bio, i) {
-		page = bvec->bv_page;
-		if (page->private == EXTENT_PAGE_PRIVATE) {
-			length += bvec->bv_len;
-			continue;
-		}
-		if (!page->private) {
-			length += bvec->bv_len;
-			continue;
-		}
-		length = bvec->bv_len;
-		buf_len = page->private >> 2;
-		start = page_offset(page) + bvec->bv_offset;
-		io_tree = &BTRFS_I(page->mapping->host)->io_tree;
-	}
-	/* are we fully contained in this bio? */
-	if (buf_len <= length)
-		return 1;
-
-	ret = extent_range_uptodate(io_tree, start + length,
-				    start + buf_len - 1);
-	return ret;
-}
-
 /*
  * called by the kthread helper functions to finally call the bio end_io
  * functions.  This is where read checksum verification actually happens
@@ -1563,17 +1520,6 @@ static void end_workqueue_fn(struct btrfs_work *work)
 	bio = end_io_wq->bio;
 	fs_info = end_io_wq->info;
 
-	/* metadata bio reads are special because the whole tree block must
-	 * be checksummed at once.  This makes sure the entire block is in
-	 * ram and up to date before trying to verify things.  For
-	 * blocksize <= pagesize, it is basically a noop
-	 */
-	if (!(bio->bi_rw & REQ_WRITE) && end_io_wq->metadata &&
-	    !bio_ready_for_csum(bio)) {
-		btrfs_queue_worker(&fs_info->endio_meta_workers,
-				   &end_io_wq->work);
-		return;
-	}
 	error = end_io_wq->error;
 	bio->bi_private = end_io_wq->private;
 	bio->bi_end_io = end_io_wq->end_io;
@@ -1614,9 +1560,10 @@ static int transaction_kthread(void *arg)
 	u64 transid;
 	unsigned long now;
 	unsigned long delay;
-	int ret;
+	bool cannot_commit;
 
 	do {
+		cannot_commit = false;
 		delay = HZ * 30;
 		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
 		mutex_lock(&root->fs_info->transaction_kthread_mutex);
@@ -1638,11 +1585,14 @@ static int transaction_kthread(void *arg)
 		transid = cur->transid;
 		spin_unlock(&root->fs_info->trans_lock);
 
+		/* If the file system is aborted, this will always fail. */
 		trans = btrfs_join_transaction(root);
-		BUG_ON(IS_ERR(trans));
+		if (IS_ERR(trans)) {
+			cannot_commit = true;
+			goto sleep;
+		}
 		if (transid == trans->transid) {
-			ret = btrfs_commit_transaction(trans, root);
-			BUG_ON(ret);
+			btrfs_commit_transaction(trans, root);
 		} else {
 			btrfs_end_transaction(trans, root);
 		}
@@ -1653,7 +1603,8 @@ sleep:
 		if (!try_to_freeze()) {
 			set_current_state(TASK_INTERRUPTIBLE);
 			if (!kthread_should_stop() &&
-			    !btrfs_transaction_blocked(root->fs_info))
+			    (!btrfs_transaction_blocked(root->fs_info) ||
+			     cannot_commit))
 				schedule_timeout(delay);
 			__set_current_state(TASK_RUNNING);
 		}
@@ -2042,6 +1993,7 @@ int open_ctree(struct super_block *sb,
 	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);
+	BTRFS_I(fs_info->btree_inode)->io_tree.track_uptodate = 0;
 	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree);
 
 	BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;
@@ -2084,6 +2036,7 @@ int open_ctree(struct super_block *sb,
 	__setup_root(4096, 4096, 4096, 4096, tree_root,
 		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
 
+	invalidate_bdev(fs_devices->latest_bdev);
 	bh = btrfs_read_dev_super(fs_devices->latest_bdev);
 	if (!bh) {
 		err = -EINVAL;
@@ -2104,7 +2057,12 @@ int open_ctree(struct super_block *sb,
 	/* check FS state, whether FS is broken. */
 	fs_info->fs_state |= btrfs_super_flags(disk_super);
 
-	btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY);
+	ret = btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY);
+	if (ret) {
+		printk(KERN_ERR "btrfs: superblock contains fatal errors\n");
+		err = ret;
+		goto fail_alloc;
+	}
 
 	/*
 	 * run through our array of backup supers and setup
@@ -2135,10 +2093,55 @@ int open_ctree(struct super_block *sb,
 		goto fail_alloc;
 	}
 
+	if (btrfs_super_leafsize(disk_super) !=
+	    btrfs_super_nodesize(disk_super)) {
+		printk(KERN_ERR "BTRFS: couldn't mount because metadata "
+		       "blocksizes don't match.  node %d leaf %d\n",
+		       btrfs_super_nodesize(disk_super),
+		       btrfs_super_leafsize(disk_super));
+		err = -EINVAL;
+		goto fail_alloc;
+	}
+	if (btrfs_super_leafsize(disk_super) > BTRFS_MAX_METADATA_BLOCKSIZE) {
+		printk(KERN_ERR "BTRFS: couldn't mount because metadata "
+		       "blocksize (%d) was too large\n",
+		       btrfs_super_leafsize(disk_super));
+		err = -EINVAL;
+		goto fail_alloc;
+	}
+
 	features = btrfs_super_incompat_flags(disk_super);
 	features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF;
 	if (tree_root->fs_info->compress_type & BTRFS_COMPRESS_LZO)
 		features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO;
+
+	/*
+	 * flag our filesystem as having big metadata blocks if
+	 * they are bigger than the page size
+	 */
+	if (btrfs_super_leafsize(disk_super) > PAGE_CACHE_SIZE) {
+		if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA))
+			printk(KERN_INFO "btrfs flagging fs with big metadata feature\n");
+		features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA;
+	}
+
+	nodesize = btrfs_super_nodesize(disk_super);
+	leafsize = btrfs_super_leafsize(disk_super);
+	sectorsize = btrfs_super_sectorsize(disk_super);
+	stripesize = btrfs_super_stripesize(disk_super);
+
+	/*
+	 * mixed block groups end up with duplicate but slightly offset
+	 * extent buffers for the same range.  It leads to corruptions
+	 */
+	if ((features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) &&
+	    (sectorsize != leafsize)) {
+		printk(KERN_WARNING "btrfs: unequal leaf/node/sector sizes "
+				"are not allowed for mixed block groups on %s\n",
+				sb->s_id);
+		goto fail_alloc;
+	}
+
 	btrfs_set_super_incompat_flags(disk_super, features);
 
 	features = btrfs_super_compat_ro_flags(disk_super) &
@@ -2242,10 +2245,6 @@ int open_ctree(struct super_block *sb,
 	fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
 				    4 * 1024 * 1024 / PAGE_CACHE_SIZE);
 
-	nodesize = btrfs_super_nodesize(disk_super);
-	leafsize = btrfs_super_leafsize(disk_super);
-	sectorsize = btrfs_super_sectorsize(disk_super);
-	stripesize = btrfs_super_stripesize(disk_super);
 	tree_root->nodesize = nodesize;
 	tree_root->leafsize = leafsize;
 	tree_root->sectorsize = sectorsize;
@@ -2260,9 +2259,9 @@ int open_ctree(struct super_block *sb,
 		goto fail_sb_buffer;
 	}
 
-	if (sectorsize < PAGE_SIZE) {
-		printk(KERN_WARNING "btrfs: Incompatible sector size "
-		       "found on %s\n", sb->s_id);
+	if (sectorsize != PAGE_SIZE) {
+		printk(KERN_WARNING "btrfs: Incompatible sector size(%lu) "
+		       "found on %s\n", (unsigned long)sectorsize, sb->s_id);
 		goto fail_sb_buffer;
 	}
 
@@ -2285,7 +2284,7 @@ int open_ctree(struct super_block *sb,
 	chunk_root->node = read_tree_block(chunk_root,
 					   btrfs_super_chunk_root(disk_super),
 					   blocksize, generation);
-	BUG_ON(!chunk_root->node);
+	BUG_ON(!chunk_root->node); /* -ENOMEM */
 	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);
@@ -2425,21 +2424,31 @@ retry_root_backup:
 		log_tree_root->node = read_tree_block(tree_root, bytenr,
 						      blocksize,
 						      generation + 1);
+		/* returns with log_tree_root freed on success */
 		ret = btrfs_recover_log_trees(log_tree_root);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_error(tree_root->fs_info, ret,
+				    "Failed to recover log tree");
+			free_extent_buffer(log_tree_root->node);
+			kfree(log_tree_root);
+			goto fail_trans_kthread;
+		}
 
 		if (sb->s_flags & MS_RDONLY) {
-			ret =  btrfs_commit_super(tree_root);
-			BUG_ON(ret);
+			ret = btrfs_commit_super(tree_root);
+			if (ret)
+				goto fail_trans_kthread;
 		}
 	}
 
 	ret = btrfs_find_orphan_roots(tree_root);
-	BUG_ON(ret);
+	if (ret)
+		goto fail_trans_kthread;
 
 	if (!(sb->s_flags & MS_RDONLY)) {
 		ret = btrfs_cleanup_fs_roots(fs_info);
-		BUG_ON(ret);
+		if (ret) {
+			}
 
 		ret = btrfs_recover_relocation(tree_root);
 		if (ret < 0) {
@@ -2859,6 +2868,8 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	if (total_errors > max_errors) {
 		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
 		       total_errors);
+
+		/* This shouldn't happen. FUA is masked off if unsupported */
 		BUG();
 	}
 
@@ -2875,9 +2886,9 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	}
 	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 	if (total_errors > max_errors) {
-		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
-		       total_errors);
-		BUG();
+		btrfs_error(root->fs_info, -EIO,
+			    "%d errors while writing supers", total_errors);
+		return -EIO;
 	}
 	return 0;
 }
@@ -2891,7 +2902,20 @@ int write_ctree_super(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
+/* Kill all outstanding I/O */
+void btrfs_abort_devices(struct btrfs_root *root)
+{
+	struct list_head *head;
+	struct btrfs_device *dev;
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	head = &root->fs_info->fs_devices->devices;
+	list_for_each_entry_rcu(dev, head, dev_list) {
+		blk_abort_queue(dev->bdev->bd_disk->queue);
+	}
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+}
+
+void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 {
 	spin_lock(&fs_info->fs_roots_radix_lock);
 	radix_tree_delete(&fs_info->fs_roots_radix,
@@ -2904,7 +2928,6 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 	__btrfs_remove_free_space_cache(root->free_ino_pinned);
 	__btrfs_remove_free_space_cache(root->free_ino_ctl);
 	free_fs_root(root);
-	return 0;
 }
 
 static void free_fs_root(struct btrfs_root *root)
@@ -2921,7 +2944,7 @@ static void free_fs_root(struct btrfs_root *root)
 	kfree(root);
 }
 
-static int del_fs_roots(struct btrfs_fs_info *fs_info)
+static void del_fs_roots(struct btrfs_fs_info *fs_info)
 {
 	int ret;
 	struct btrfs_root *gang[8];
@@ -2950,7 +2973,6 @@ static int del_fs_roots(struct btrfs_fs_info *fs_info)
 		for (i = 0; i < ret; i++)
 			btrfs_free_fs_root(fs_info, gang[i]);
 	}
-	return 0;
 }
 
 int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
@@ -2999,14 +3021,21 @@ int btrfs_commit_super(struct btrfs_root *root)
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
 	ret = btrfs_commit_transaction(trans, root);
-	BUG_ON(ret);
+	if (ret)
+		return ret;
 	/* run commit again to drop the original snapshot */
 	trans = btrfs_join_transaction(root);
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
-	btrfs_commit_transaction(trans, root);
+	ret = btrfs_commit_transaction(trans, root);
+	if (ret)
+		return ret;
 	ret = btrfs_write_and_wait_transaction(NULL, root);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_error(root->fs_info, ret,
+			    "Failed to sync btree inode to disk.");
+		return ret;
+	}
 
 	ret = write_ctree_super(NULL, root, 0);
 	return ret;
@@ -3119,33 +3148,32 @@ int close_ctree(struct btrfs_root *root)
 	return 0;
 }
 
-int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid)
+int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
+			  int atomic)
 {
 	int ret;
-	struct inode *btree_inode = buf->first_page->mapping->host;
+	struct inode *btree_inode = buf->pages[0]->mapping->host;
 
-	ret = extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, buf,
-				     NULL);
+	ret = extent_buffer_uptodate(buf);
 	if (!ret)
 		return ret;
 
 	ret = verify_parent_transid(&BTRFS_I(btree_inode)->io_tree, buf,
-				    parent_transid);
+				    parent_transid, atomic);
+	if (ret == -EAGAIN)
+		return ret;
 	return !ret;
 }
 
 int btrfs_set_buffer_uptodate(struct extent_buffer *buf)
 {
-	struct inode *btree_inode = buf->first_page->mapping->host;
-	return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree,
-					  buf);
+	return set_extent_buffer_uptodate(buf);
 }
 
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 {
-	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
 	u64 transid = btrfs_header_generation(buf);
-	struct inode *btree_inode = root->fs_info->btree_inode;
 	int was_dirty;
 
 	btrfs_assert_tree_locked(buf);
@@ -3157,8 +3185,7 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 			(unsigned long long)root->fs_info->generation);
 		WARN_ON(1);
 	}
-	was_dirty = set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
-					    buf);
+	was_dirty = set_extent_buffer_dirty(buf);
 	if (!was_dirty) {
 		spin_lock(&root->fs_info->delalloc_lock);
 		root->fs_info->dirty_metadata_bytes += buf->len;
@@ -3212,12 +3239,8 @@ void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
 
 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)
-		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
-	return ret;
+	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
+	return btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
 }
 
 static int btree_lock_page_hook(struct page *page, void *data,
@@ -3225,17 +3248,21 @@ static int btree_lock_page_hook(struct page *page, void *data,
 {
 	struct inode *inode = page->mapping->host;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
 	struct extent_buffer *eb;
-	unsigned long len;
-	u64 bytenr = page_offset(page);
 
-	if (page->private == EXTENT_PAGE_PRIVATE)
+	/*
+	 * We culled this eb but the page is still hanging out on the mapping,
+	 * carry on.
+	 */
+	if (!PagePrivate(page))
 		goto out;
 
-	len = page->private >> 2;
-	eb = find_extent_buffer(io_tree, bytenr, len);
-	if (!eb)
+	eb = (struct extent_buffer *)page->private;
+	if (!eb) {
+		WARN_ON(1);
+		goto out;
+	}
+	if (page != eb->pages[0])
 		goto out;
 
 	if (!btrfs_try_tree_write_lock(eb)) {
@@ -3254,7 +3281,6 @@ static int btree_lock_page_hook(struct page *page, void *data,
 	}
 
 	btrfs_tree_unlock(eb);
-	free_extent_buffer(eb);
 out:
 	if (!trylock_page(page)) {
 		flush_fn(data);
@@ -3263,15 +3289,23 @@ out:
 	return 0;
 }
 
-static void btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
+static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
 			      int read_only)
 {
+	if (btrfs_super_csum_type(fs_info->super_copy) >= ARRAY_SIZE(btrfs_csum_sizes)) {
+		printk(KERN_ERR "btrfs: unsupported checksum algorithm\n");
+		return -EINVAL;
+	}
+
 	if (read_only)
-		return;
+		return 0;
 
-	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
+	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
 		printk(KERN_WARNING "warning: mount fs with errors, "
 		       "running btrfsck is recommended\n");
+	}
+
+	return 0;
 }
 
 int btrfs_error_commit_super(struct btrfs_root *root)
@@ -3293,7 +3327,7 @@ int btrfs_error_commit_super(struct btrfs_root *root)
 	return ret;
 }
 
-static int btrfs_destroy_ordered_operations(struct btrfs_root *root)
+static void btrfs_destroy_ordered_operations(struct btrfs_root *root)
 {
 	struct btrfs_inode *btrfs_inode;
 	struct list_head splice;
@@ -3315,11 +3349,9 @@ static int btrfs_destroy_ordered_operations(struct btrfs_root *root)
 
 	spin_unlock(&root->fs_info->ordered_extent_lock);
 	mutex_unlock(&root->fs_info->ordered_operations_mutex);
-
-	return 0;
 }
 
-static int btrfs_destroy_ordered_extents(struct btrfs_root *root)
+static void btrfs_destroy_ordered_extents(struct btrfs_root *root)
 {
 	struct list_head splice;
 	struct btrfs_ordered_extent *ordered;
@@ -3351,12 +3383,10 @@ static int btrfs_destroy_ordered_extents(struct btrfs_root *root)
 	}
 
 	spin_unlock(&root->fs_info->ordered_extent_lock);
-
-	return 0;
 }
 
-static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
-				      struct btrfs_root *root)
+int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
+			       struct btrfs_root *root)
 {
 	struct rb_node *node;
 	struct btrfs_delayed_ref_root *delayed_refs;
@@ -3365,6 +3395,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 
 	delayed_refs = &trans->delayed_refs;
 
+again:
 	spin_lock(&delayed_refs->lock);
 	if (delayed_refs->num_entries == 0) {
 		spin_unlock(&delayed_refs->lock);
@@ -3386,6 +3417,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 			struct btrfs_delayed_ref_head *head;
 
 			head = btrfs_delayed_node_to_head(ref);
+			spin_unlock(&delayed_refs->lock);
 			mutex_lock(&head->mutex);
 			kfree(head->extent_op);
 			delayed_refs->num_heads--;
@@ -3393,8 +3425,9 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 				delayed_refs->num_heads_ready--;
 			list_del_init(&head->cluster);
 			mutex_unlock(&head->mutex);
+			btrfs_put_delayed_ref(ref);
+			goto again;
 		}
-
 		spin_unlock(&delayed_refs->lock);
 		btrfs_put_delayed_ref(ref);
 
@@ -3407,7 +3440,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 	return ret;
 }
 
-static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t)
+static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t)
 {
 	struct btrfs_pending_snapshot *snapshot;
 	struct list_head splice;
@@ -3425,11 +3458,9 @@ static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t)
 
 		kfree(snapshot);
 	}
-
-	return 0;
 }
 
-static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
+static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 {
 	struct btrfs_inode *btrfs_inode;
 	struct list_head splice;
@@ -3449,8 +3480,6 @@ static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 	}
 
 	spin_unlock(&root->fs_info->delalloc_lock);
-
-	return 0;
 }
 
 static int btrfs_destroy_marked_extents(struct btrfs_root *root,
@@ -3541,13 +3570,43 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
 	return 0;
 }
 
-static int btrfs_cleanup_transaction(struct btrfs_root *root)
+void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
+				   struct btrfs_root *root)
+{
+	btrfs_destroy_delayed_refs(cur_trans, root);
+	btrfs_block_rsv_release(root, &root->fs_info->trans_block_rsv,
+				cur_trans->dirty_pages.dirty_bytes);
+
+	/* FIXME: cleanup wait for commit */
+	cur_trans->in_commit = 1;
+	cur_trans->blocked = 1;
+	if (waitqueue_active(&root->fs_info->transaction_blocked_wait))
+		wake_up(&root->fs_info->transaction_blocked_wait);
+
+	cur_trans->blocked = 0;
+	if (waitqueue_active(&root->fs_info->transaction_wait))
+		wake_up(&root->fs_info->transaction_wait);
+
+	cur_trans->commit_done = 1;
+	if (waitqueue_active(&cur_trans->commit_wait))
+		wake_up(&cur_trans->commit_wait);
+
+	btrfs_destroy_pending_snapshots(cur_trans);
+
+	btrfs_destroy_marked_extents(root, &cur_trans->dirty_pages,
+				     EXTENT_DIRTY);
+
+	/*
+	memset(cur_trans, 0, sizeof(*cur_trans));
+	kmem_cache_free(btrfs_transaction_cachep, cur_trans);
+	*/
+}
+
+int btrfs_cleanup_transaction(struct btrfs_root *root)
 {
 	struct btrfs_transaction *t;
 	LIST_HEAD(list);
 
-	WARN_ON(1);
-
 	mutex_lock(&root->fs_info->transaction_kthread_mutex);
 
 	spin_lock(&root->fs_info->trans_lock);
@@ -3612,6 +3671,17 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 	return 0;
 }
 
+static int btree_writepage_io_failed_hook(struct bio *bio, struct page *page,
+					  u64 start, u64 end,
+					  struct extent_state *state)
+{
+	struct super_block *sb = page->mapping->host->i_sb;
+	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
+	btrfs_error(fs_info, -EIO,
+		    "Error occured while writing out btree at %llu", start);
+	return -EIO;
+}
+
 static struct extent_io_ops btree_extent_io_ops = {
 	.write_cache_pages_lock_hook = btree_lock_page_hook,
 	.readpage_end_io_hook = btree_readpage_end_io_hook,
@@ -3619,4 +3689,5 @@ static struct extent_io_ops btree_extent_io_ops = {
 	.submit_bio_hook = btree_submit_bio_hook,
 	/* note we're sharing with inode.c for the merge bio hook */
 	.merge_bio_hook = btrfs_merge_bio_hook,
+	.writepage_io_failed_hook = btree_writepage_io_failed_hook,
 };
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index e4bc4741319..ab1830aaf0e 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -44,8 +44,8 @@ int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize,
 			 int mirror_num, struct extent_buffer **eb);
 struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
 						   u64 bytenr, u32 blocksize);
-int clean_tree_block(struct btrfs_trans_handle *trans,
-		     struct btrfs_root *root, struct extent_buffer *buf);
+void clean_tree_block(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root, struct extent_buffer *buf);
 int open_ctree(struct super_block *sb,
 	       struct btrfs_fs_devices *fs_devices,
 	       char *options);
@@ -64,9 +64,10 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info);
 void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr);
 void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr);
-int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
+void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
-int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid);
+int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
+			  int atomic);
 int btrfs_set_buffer_uptodate(struct extent_buffer *buf);
 int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
 u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len);
@@ -85,6 +86,10 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
 			     struct btrfs_fs_info *fs_info);
 int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root);
+int btrfs_cleanup_transaction(struct btrfs_root *root);
+void btrfs_cleanup_one_transaction(struct btrfs_transaction *trans,
+				  struct btrfs_root *root);
+void btrfs_abort_devices(struct btrfs_root *root);
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 void btrfs_init_lockdep(void);
diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c
index 5f77166fd01..e887ee62b6d 100644
--- a/fs/btrfs/export.c
+++ b/fs/btrfs/export.c
@@ -193,7 +193,7 @@ static struct dentry *btrfs_get_parent(struct dentry *child)
 	if (ret < 0)
 		goto fail;
 
-	BUG_ON(ret == 0);
+	BUG_ON(ret == 0); /* Key with offset of -1 found */
 	if (path->slots[0] == 0) {
 		ret = -ENOENT;
 		goto fail;
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index f6f584fd549..49fd7b66d57 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -245,7 +245,7 @@ static int exclude_super_stripes(struct btrfs_root *root,
 		cache->bytes_super += stripe_len;
 		ret = add_excluded_extent(root, cache->key.objectid,
 					  stripe_len);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 
 	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
@@ -253,13 +253,13 @@ static int exclude_super_stripes(struct btrfs_root *root,
 		ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
 				       cache->key.objectid, bytenr,
 				       0, &logical, &nr, &stripe_len);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 
 		while (nr--) {
 			cache->bytes_super += stripe_len;
 			ret = add_excluded_extent(root, logical[nr],
 						  stripe_len);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		}
 
 		kfree(logical);
@@ -321,7 +321,7 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group,
 			total_added += size;
 			ret = btrfs_add_free_space(block_group, start,
 						   size);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM or logic error */
 			start = extent_end + 1;
 		} else {
 			break;
@@ -332,7 +332,7 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group,
 		size = end - start;
 		total_added += size;
 		ret = btrfs_add_free_space(block_group, start, size);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM or logic error */
 	}
 
 	return total_added;
@@ -474,7 +474,8 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
 	int ret = 0;
 
 	caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS);
-	BUG_ON(!caching_ctl);
+	if (!caching_ctl)
+		return -ENOMEM;
 
 	INIT_LIST_HEAD(&caching_ctl->list);
 	mutex_init(&caching_ctl->mutex);
@@ -528,9 +529,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
 	 * allocate blocks for the tree root we can't do the fast caching since
 	 * we likely hold important locks.
 	 */
-	if (trans && (!trans->transaction->in_commit) &&
-	    (root && root != root->fs_info->tree_root) &&
-	    btrfs_test_opt(root, SPACE_CACHE)) {
+	if (fs_info->mount_opt & BTRFS_MOUNT_SPACE_CACHE) {
 		ret = load_free_space_cache(fs_info, cache);
 
 		spin_lock(&cache->lock);
@@ -982,7 +981,7 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans,
 				ret = btrfs_next_leaf(root, path);
 				if (ret < 0)
 					return ret;
-				BUG_ON(ret > 0);
+				BUG_ON(ret > 0); /* Corruption */
 				leaf = path->nodes[0];
 			}
 			btrfs_item_key_to_cpu(leaf, &found_key,
@@ -1008,9 +1007,9 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans,
 				new_size + extra_size, 1);
 	if (ret < 0)
 		return ret;
-	BUG_ON(ret);
+	BUG_ON(ret); /* Corruption */
 
-	ret = btrfs_extend_item(trans, root, path, new_size);
+	btrfs_extend_item(trans, root, path, new_size);
 
 	leaf = path->nodes[0];
 	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
@@ -1478,7 +1477,11 @@ int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
 		err = ret;
 		goto out;
 	}
-	BUG_ON(ret);
+	if (ret && !insert) {
+		err = -ENOENT;
+		goto out;
+	}
+	BUG_ON(ret); /* Corruption */
 
 	leaf = path->nodes[0];
 	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
@@ -1592,13 +1595,13 @@ out:
  * helper to add new inline back ref
  */
 static noinline_for_stack
-int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
-				struct btrfs_root *root,
-				struct btrfs_path *path,
-				struct btrfs_extent_inline_ref *iref,
-				u64 parent, u64 root_objectid,
-				u64 owner, u64 offset, int refs_to_add,
-				struct btrfs_delayed_extent_op *extent_op)
+void setup_inline_extent_backref(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct btrfs_path *path,
+				 struct btrfs_extent_inline_ref *iref,
+				 u64 parent, u64 root_objectid,
+				 u64 owner, u64 offset, int refs_to_add,
+				 struct btrfs_delayed_extent_op *extent_op)
 {
 	struct extent_buffer *leaf;
 	struct btrfs_extent_item *ei;
@@ -1608,7 +1611,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
 	u64 refs;
 	int size;
 	int type;
-	int ret;
 
 	leaf = path->nodes[0];
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
@@ -1617,7 +1619,7 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
 	type = extent_ref_type(parent, owner);
 	size = btrfs_extent_inline_ref_size(type);
 
-	ret = btrfs_extend_item(trans, root, path, size);
+	btrfs_extend_item(trans, root, path, size);
 
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	refs = btrfs_extent_refs(leaf, ei);
@@ -1652,7 +1654,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
 		btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
 	}
 	btrfs_mark_buffer_dirty(leaf);
-	return 0;
 }
 
 static int lookup_extent_backref(struct btrfs_trans_handle *trans,
@@ -1687,12 +1688,12 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans,
  * helper to update/remove inline back ref
  */
 static noinline_for_stack
-int update_inline_extent_backref(struct btrfs_trans_handle *trans,
-				 struct btrfs_root *root,
-				 struct btrfs_path *path,
-				 struct btrfs_extent_inline_ref *iref,
-				 int refs_to_mod,
-				 struct btrfs_delayed_extent_op *extent_op)
+void update_inline_extent_backref(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  struct btrfs_path *path,
+				  struct btrfs_extent_inline_ref *iref,
+				  int refs_to_mod,
+				  struct btrfs_delayed_extent_op *extent_op)
 {
 	struct extent_buffer *leaf;
 	struct btrfs_extent_item *ei;
@@ -1703,7 +1704,6 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans,
 	u32 item_size;
 	int size;
 	int type;
-	int ret;
 	u64 refs;
 
 	leaf = path->nodes[0];
@@ -1745,10 +1745,9 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans,
 			memmove_extent_buffer(leaf, ptr, ptr + size,
 					      end - ptr - size);
 		item_size -= size;
-		ret = btrfs_truncate_item(trans, root, path, item_size, 1);
+		btrfs_truncate_item(trans, root, path, item_size, 1);
 	}
 	btrfs_mark_buffer_dirty(leaf);
-	return 0;
 }
 
 static noinline_for_stack
@@ -1768,13 +1767,13 @@ int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
 					   root_objectid, owner, offset, 1);
 	if (ret == 0) {
 		BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
-		ret = update_inline_extent_backref(trans, root, path, iref,
-						   refs_to_add, extent_op);
+		update_inline_extent_backref(trans, root, path, iref,
+					     refs_to_add, extent_op);
 	} else if (ret == -ENOENT) {
-		ret = setup_inline_extent_backref(trans, root, path, iref,
-						  parent, root_objectid,
-						  owner, offset, refs_to_add,
-						  extent_op);
+		setup_inline_extent_backref(trans, root, path, iref, parent,
+					    root_objectid, owner, offset,
+					    refs_to_add, extent_op);
+		ret = 0;
 	}
 	return ret;
 }
@@ -1804,12 +1803,12 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans,
 				 struct btrfs_extent_inline_ref *iref,
 				 int refs_to_drop, int is_data)
 {
-	int ret;
+	int ret = 0;
 
 	BUG_ON(!is_data && refs_to_drop != 1);
 	if (iref) {
-		ret = update_inline_extent_backref(trans, root, path, iref,
-						   -refs_to_drop, NULL);
+		update_inline_extent_backref(trans, root, path, iref,
+					     -refs_to_drop, NULL);
 	} else if (is_data) {
 		ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
 	} else {
@@ -1835,6 +1834,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 	/* Tell the block device(s) that the sectors can be discarded */
 	ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD,
 			      bytenr, &num_bytes, &bbio, 0);
+	/* Error condition is -ENOMEM */
 	if (!ret) {
 		struct btrfs_bio_stripe *stripe = bbio->stripes;
 		int i;
@@ -1850,7 +1850,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 			if (!ret)
 				discarded_bytes += stripe->length;
 			else if (ret != -EOPNOTSUPP)
-				break;
+				break; /* Logic errors or -ENOMEM, or -EIO but I don't know how that could happen JDM */
 
 			/*
 			 * Just in case we get back EOPNOTSUPP for some reason,
@@ -1869,6 +1869,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 	return ret;
 }
 
+/* Can return -ENOMEM */
 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root,
 			 u64 bytenr, u64 num_bytes, u64 parent,
@@ -1944,7 +1945,8 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	ret = insert_extent_backref(trans, root->fs_info->extent_root,
 				    path, bytenr, parent, root_objectid,
 				    owner, offset, refs_to_add);
-	BUG_ON(ret);
+	if (ret)
+		btrfs_abort_transaction(trans, root, ret);
 out:
 	btrfs_free_path(path);
 	return err;
@@ -2031,6 +2033,9 @@ static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
 	int ret;
 	int err = 0;
 
+	if (trans->aborted)
+		return 0;
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -2128,7 +2133,11 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 			       struct btrfs_delayed_extent_op *extent_op,
 			       int insert_reserved)
 {
-	int ret;
+	int ret = 0;
+
+	if (trans->aborted)
+		return 0;
+
 	if (btrfs_delayed_ref_is_head(node)) {
 		struct btrfs_delayed_ref_head *head;
 		/*
@@ -2146,11 +2155,10 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 				ret = btrfs_del_csums(trans, root,
 						      node->bytenr,
 						      node->num_bytes);
-				BUG_ON(ret);
 			}
 		}
 		mutex_unlock(&head->mutex);
-		return 0;
+		return ret;
 	}
 
 	if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
@@ -2197,6 +2205,10 @@ again:
 	return NULL;
 }
 
+/*
+ * Returns 0 on success or if called with an already aborted transaction.
+ * Returns -ENOMEM or -EIO on failure and will abort the transaction.
+ */
 static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
 				       struct btrfs_root *root,
 				       struct list_head *cluster)
@@ -2285,9 +2297,14 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
 
 				ret = run_delayed_extent_op(trans, root,
 							    ref, extent_op);
-				BUG_ON(ret);
 				kfree(extent_op);
 
+				if (ret) {
+					printk(KERN_DEBUG "btrfs: run_delayed_extent_op returned %d\n", ret);
+					spin_lock(&delayed_refs->lock);
+					return ret;
+				}
+
 				goto next;
 			}
 
@@ -2308,11 +2325,17 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
 
 		ret = run_one_delayed_ref(trans, root, ref, extent_op,
 					  must_insert_reserved);
-		BUG_ON(ret);
 
 		btrfs_put_delayed_ref(ref);
 		kfree(extent_op);
 		count++;
+
+		if (ret) {
+			printk(KERN_DEBUG "btrfs: run_one_delayed_ref returned %d\n", ret);
+			spin_lock(&delayed_refs->lock);
+			return ret;
+		}
+
 next:
 		do_chunk_alloc(trans, root->fs_info->extent_root,
 			       2 * 1024 * 1024,
@@ -2347,6 +2370,9 @@ static void wait_for_more_refs(struct btrfs_delayed_ref_root *delayed_refs,
  * 0, which means to process everything in the tree at the start
  * of the run (but not newly added entries), or it can be some target
  * number you'd like to process.
+ *
+ * Returns 0 on success or if called with an aborted transaction
+ * Returns <0 on error and aborts the transaction
  */
 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root, unsigned long count)
@@ -2362,6 +2388,10 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 	unsigned long num_refs = 0;
 	int consider_waiting;
 
+	/* We'll clean this up in btrfs_cleanup_transaction */
+	if (trans->aborted)
+		return 0;
+
 	if (root == root->fs_info->extent_root)
 		root = root->fs_info->tree_root;
 
@@ -2419,7 +2449,11 @@ again:
 		}
 
 		ret = run_clustered_refs(trans, root, &cluster);
-		BUG_ON(ret < 0);
+		if (ret < 0) {
+			spin_unlock(&delayed_refs->lock);
+			btrfs_abort_transaction(trans, root, ret);
+			return ret;
+		}
 
 		count -= min_t(unsigned long, ret, count);
 
@@ -2584,7 +2618,7 @@ static noinline int check_committed_ref(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret == 0);
+	BUG_ON(ret == 0); /* Corruption */
 
 	ret = -ENOENT;
 	if (path->slots[0] == 0)
@@ -2738,7 +2772,6 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 	}
 	return 0;
 fail:
-	BUG();
 	return ret;
 }
 
@@ -2767,7 +2800,7 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
 	if (ret < 0)
 		goto fail;
-	BUG_ON(ret);
+	BUG_ON(ret); /* Corruption */
 
 	leaf = path->nodes[0];
 	bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
@@ -2775,8 +2808,10 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_release_path(path);
 fail:
-	if (ret)
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
 		return ret;
+	}
 	return 0;
 
 }
@@ -2949,7 +2984,8 @@ again:
 		if (last == 0) {
 			err = btrfs_run_delayed_refs(trans, root,
 						     (unsigned long)-1);
-			BUG_ON(err);
+			if (err) /* File system offline */
+				goto out;
 		}
 
 		cache = btrfs_lookup_first_block_group(root->fs_info, last);
@@ -2976,7 +3012,9 @@ again:
 		last = cache->key.objectid + cache->key.offset;
 
 		err = write_one_cache_group(trans, root, path, cache);
-		BUG_ON(err);
+		if (err) /* File system offline */
+			goto out;
+
 		btrfs_put_block_group(cache);
 	}
 
@@ -2989,7 +3027,8 @@ again:
 		if (last == 0) {
 			err = btrfs_run_delayed_refs(trans, root,
 						     (unsigned long)-1);
-			BUG_ON(err);
+			if (err) /* File system offline */
+				goto out;
 		}
 
 		cache = btrfs_lookup_first_block_group(root->fs_info, last);
@@ -3014,20 +3053,21 @@ again:
 			continue;
 		}
 
-		btrfs_write_out_cache(root, trans, cache, path);
+		err = btrfs_write_out_cache(root, trans, cache, path);
 
 		/*
 		 * If we didn't have an error then the cache state is still
 		 * NEED_WRITE, so we can set it to WRITTEN.
 		 */
-		if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE)
+		if (!err && cache->disk_cache_state == BTRFS_DC_NEED_WRITE)
 			cache->disk_cache_state = BTRFS_DC_WRITTEN;
 		last = cache->key.objectid + cache->key.offset;
 		btrfs_put_block_group(cache);
 	}
+out:
 
 	btrfs_free_path(path);
-	return 0;
+	return err;
 }
 
 int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
@@ -3098,11 +3138,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 
 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
 {
-	u64 extra_flags = flags & BTRFS_BLOCK_GROUP_PROFILE_MASK;
-
-	/* chunk -> extended profile */
-	if (extra_flags == 0)
-		extra_flags = BTRFS_AVAIL_ALLOC_BIT_SINGLE;
+	u64 extra_flags = chunk_to_extended(flags) &
+				BTRFS_EXTENDED_PROFILE_MASK;
 
 	if (flags & BTRFS_BLOCK_GROUP_DATA)
 		fs_info->avail_data_alloc_bits |= extra_flags;
@@ -3113,6 +3150,34 @@ static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
 }
 
 /*
+ * returns target flags in extended format or 0 if restripe for this
+ * chunk_type is not in progress
+ *
+ * should be called with either volume_mutex or balance_lock held
+ */
+static u64 get_restripe_target(struct btrfs_fs_info *fs_info, u64 flags)
+{
+	struct btrfs_balance_control *bctl = fs_info->balance_ctl;
+	u64 target = 0;
+
+	if (!bctl)
+		return 0;
+
+	if (flags & BTRFS_BLOCK_GROUP_DATA &&
+	    bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) {
+		target = BTRFS_BLOCK_GROUP_DATA | bctl->data.target;
+	} else if (flags & BTRFS_BLOCK_GROUP_SYSTEM &&
+		   bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) {
+		target = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target;
+	} else if (flags & BTRFS_BLOCK_GROUP_METADATA &&
+		   bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) {
+		target = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target;
+	}
+
+	return target;
+}
+
+/*
  * @flags: available profiles in extended format (see ctree.h)
  *
  * Returns reduced profile in chunk format.  If profile changing is in
@@ -3128,31 +3193,19 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 	 */
 	u64 num_devices = root->fs_info->fs_devices->rw_devices +
 		root->fs_info->fs_devices->missing_devices;
+	u64 target;
 
-	/* pick restriper's target profile if it's available */
+	/*
+	 * see if restripe for this chunk_type is in progress, if so
+	 * try to reduce to the target profile
+	 */
 	spin_lock(&root->fs_info->balance_lock);
-	if (root->fs_info->balance_ctl) {
-		struct btrfs_balance_control *bctl = root->fs_info->balance_ctl;
-		u64 tgt = 0;
-
-		if ((flags & BTRFS_BLOCK_GROUP_DATA) &&
-		    (bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-		    (flags & bctl->data.target)) {
-			tgt = BTRFS_BLOCK_GROUP_DATA | bctl->data.target;
-		} else if ((flags & BTRFS_BLOCK_GROUP_SYSTEM) &&
-			   (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-			   (flags & bctl->sys.target)) {
-			tgt = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target;
-		} else if ((flags & BTRFS_BLOCK_GROUP_METADATA) &&
-			   (bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-			   (flags & bctl->meta.target)) {
-			tgt = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target;
-		}
-
-		if (tgt) {
+	target = get_restripe_target(root->fs_info, flags);
+	if (target) {
+		/* pick target profile only if it's already available */
+		if ((flags & target) & BTRFS_EXTENDED_PROFILE_MASK) {
 			spin_unlock(&root->fs_info->balance_lock);
-			flags = tgt;
-			goto out;
+			return extended_to_chunk(target);
 		}
 	}
 	spin_unlock(&root->fs_info->balance_lock);
@@ -3180,10 +3233,7 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 		flags &= ~BTRFS_BLOCK_GROUP_RAID0;
 	}
 
-out:
-	/* extended -> chunk profile */
-	flags &= ~BTRFS_AVAIL_ALLOC_BIT_SINGLE;
-	return flags;
+	return extended_to_chunk(flags);
 }
 
 static u64 get_alloc_profile(struct btrfs_root *root, u64 flags)
@@ -3312,8 +3362,7 @@ commit_trans:
 	}
 	data_sinfo->bytes_may_use += bytes;
 	trace_btrfs_space_reservation(root->fs_info, "space_info",
-				      (u64)(unsigned long)data_sinfo,
-				      bytes, 1);
+				      data_sinfo->flags, bytes, 1);
 	spin_unlock(&data_sinfo->lock);
 
 	return 0;
@@ -3334,8 +3383,7 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 	spin_lock(&data_sinfo->lock);
 	data_sinfo->bytes_may_use -= bytes;
 	trace_btrfs_space_reservation(root->fs_info, "space_info",
-				      (u64)(unsigned long)data_sinfo,
-				      bytes, 0);
+				      data_sinfo->flags, bytes, 0);
 	spin_unlock(&data_sinfo->lock);
 }
 
@@ -3396,6 +3444,50 @@ static int should_alloc_chunk(struct btrfs_root *root,
 	return 1;
 }
 
+static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type)
+{
+	u64 num_dev;
+
+	if (type & BTRFS_BLOCK_GROUP_RAID10 ||
+	    type & BTRFS_BLOCK_GROUP_RAID0)
+		num_dev = root->fs_info->fs_devices->rw_devices;
+	else if (type & BTRFS_BLOCK_GROUP_RAID1)
+		num_dev = 2;
+	else
+		num_dev = 1;	/* DUP or single */
+
+	/* metadata for updaing devices and chunk tree */
+	return btrfs_calc_trans_metadata_size(root, num_dev + 1);
+}
+
+static void check_system_chunk(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root, u64 type)
+{
+	struct btrfs_space_info *info;
+	u64 left;
+	u64 thresh;
+
+	info = __find_space_info(root->fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
+	spin_lock(&info->lock);
+	left = info->total_bytes - info->bytes_used - info->bytes_pinned -
+		info->bytes_reserved - info->bytes_readonly;
+	spin_unlock(&info->lock);
+
+	thresh = get_system_chunk_thresh(root, type);
+	if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) {
+		printk(KERN_INFO "left=%llu, need=%llu, flags=%llu\n",
+		       left, thresh, type);
+		dump_space_info(info, 0, 0);
+	}
+
+	if (left < thresh) {
+		u64 flags;
+
+		flags = btrfs_get_alloc_profile(root->fs_info->chunk_root, 0);
+		btrfs_alloc_chunk(trans, root, flags);
+	}
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
@@ -3405,15 +3497,13 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 	int wait_for_alloc = 0;
 	int ret = 0;
 
-	BUG_ON(!profile_is_valid(flags, 0));
-
 	space_info = __find_space_info(extent_root->fs_info, flags);
 	if (!space_info) {
 		ret = update_space_info(extent_root->fs_info, flags,
 					0, 0, &space_info);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
-	BUG_ON(!space_info);
+	BUG_ON(!space_info); /* Logic error */
 
 again:
 	spin_lock(&space_info->lock);
@@ -3468,6 +3558,12 @@ again:
 			force_metadata_allocation(fs_info);
 	}
 
+	/*
+	 * Check if we have enough space in SYSTEM chunk because we may need
+	 * to update devices.
+	 */
+	check_system_chunk(trans, extent_root, flags);
+
 	ret = btrfs_alloc_chunk(trans, extent_root, flags);
 	if (ret < 0 && ret != -ENOSPC)
 		goto out;
@@ -3675,9 +3771,8 @@ again:
 		 */
 		if (current->journal_info)
 			return -EAGAIN;
-		ret = wait_event_interruptible(space_info->wait,
-					       !space_info->flush);
-		/* Must have been interrupted, return */
+		ret = wait_event_killable(space_info->wait, !space_info->flush);
+		/* Must have been killed, return */
 		if (ret)
 			return -EINTR;
 
@@ -3700,9 +3795,7 @@ again:
 		if (used + orig_bytes <= space_info->total_bytes) {
 			space_info->bytes_may_use += orig_bytes;
 			trace_btrfs_space_reservation(root->fs_info,
-					      "space_info",
-					      (u64)(unsigned long)space_info,
-					      orig_bytes, 1);
+				"space_info", space_info->flags, orig_bytes, 1);
 			ret = 0;
 		} else {
 			/*
@@ -3771,9 +3864,7 @@ again:
 		if (used + num_bytes < space_info->total_bytes + avail) {
 			space_info->bytes_may_use += orig_bytes;
 			trace_btrfs_space_reservation(root->fs_info,
-					      "space_info",
-					      (u64)(unsigned long)space_info,
-					      orig_bytes, 1);
+				"space_info", space_info->flags, orig_bytes, 1);
 			ret = 0;
 		} else {
 			wait_ordered = true;
@@ -3836,8 +3927,9 @@ out:
 	return ret;
 }
 
-static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans,
-					     struct btrfs_root *root)
+static struct btrfs_block_rsv *get_block_rsv(
+					const struct btrfs_trans_handle *trans,
+					const struct btrfs_root *root)
 {
 	struct btrfs_block_rsv *block_rsv = NULL;
 
@@ -3918,8 +4010,7 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info,
 			spin_lock(&space_info->lock);
 			space_info->bytes_may_use -= num_bytes;
 			trace_btrfs_space_reservation(fs_info, "space_info",
-					      (u64)(unsigned long)space_info,
-					      num_bytes, 0);
+					space_info->flags, num_bytes, 0);
 			space_info->reservation_progress++;
 			spin_unlock(&space_info->lock);
 		}
@@ -4123,8 +4214,8 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
 
 	num_bytes = calc_global_metadata_size(fs_info);
 
-	spin_lock(&block_rsv->lock);
 	spin_lock(&sinfo->lock);
+	spin_lock(&block_rsv->lock);
 
 	block_rsv->size = num_bytes;
 
@@ -4137,21 +4228,21 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
 		block_rsv->reserved += num_bytes;
 		sinfo->bytes_may_use += num_bytes;
 		trace_btrfs_space_reservation(fs_info, "space_info",
-				      (u64)(unsigned long)sinfo, num_bytes, 1);
+				      sinfo->flags, num_bytes, 1);
 	}
 
 	if (block_rsv->reserved >= block_rsv->size) {
 		num_bytes = block_rsv->reserved - block_rsv->size;
 		sinfo->bytes_may_use -= num_bytes;
 		trace_btrfs_space_reservation(fs_info, "space_info",
-				      (u64)(unsigned long)sinfo, num_bytes, 0);
+				      sinfo->flags, num_bytes, 0);
 		sinfo->reservation_progress++;
 		block_rsv->reserved = block_rsv->size;
 		block_rsv->full = 1;
 	}
 
-	spin_unlock(&sinfo->lock);
 	spin_unlock(&block_rsv->lock);
+	spin_unlock(&sinfo->lock);
 }
 
 static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
@@ -4198,12 +4289,12 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
 		return;
 
 	trace_btrfs_space_reservation(root->fs_info, "transaction",
-				      (u64)(unsigned long)trans,
-				      trans->bytes_reserved, 0);
+				      trans->transid, trans->bytes_reserved, 0);
 	btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved);
 	trans->bytes_reserved = 0;
 }
 
+/* Can only return 0 or -ENOSPC */
 int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
 				  struct inode *inode)
 {
@@ -4540,7 +4631,7 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 	while (total) {
 		cache = btrfs_lookup_block_group(info, bytenr);
 		if (!cache)
-			return -1;
+			return -ENOENT;
 		if (cache->flags & (BTRFS_BLOCK_GROUP_DUP |
 				    BTRFS_BLOCK_GROUP_RAID1 |
 				    BTRFS_BLOCK_GROUP_RAID10))
@@ -4643,7 +4734,7 @@ int btrfs_pin_extent(struct btrfs_root *root,
 	struct btrfs_block_group_cache *cache;
 
 	cache = btrfs_lookup_block_group(root->fs_info, bytenr);
-	BUG_ON(!cache);
+	BUG_ON(!cache); /* Logic error */
 
 	pin_down_extent(root, cache, bytenr, num_bytes, reserved);
 
@@ -4661,7 +4752,7 @@ int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *cache;
 
 	cache = btrfs_lookup_block_group(root->fs_info, bytenr);
-	BUG_ON(!cache);
+	BUG_ON(!cache); /* Logic error */
 
 	/*
 	 * pull in the free space cache (if any) so that our pin
@@ -4706,6 +4797,7 @@ static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
 {
 	struct btrfs_space_info *space_info = cache->space_info;
 	int ret = 0;
+
 	spin_lock(&space_info->lock);
 	spin_lock(&cache->lock);
 	if (reserve != RESERVE_FREE) {
@@ -4716,9 +4808,8 @@ static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
 			space_info->bytes_reserved += num_bytes;
 			if (reserve == RESERVE_ALLOC) {
 				trace_btrfs_space_reservation(cache->fs_info,
-					      "space_info",
-					      (u64)(unsigned long)space_info,
-					      num_bytes, 0);
+						"space_info", space_info->flags,
+						num_bytes, 0);
 				space_info->bytes_may_use -= num_bytes;
 			}
 		}
@@ -4734,7 +4825,7 @@ static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
 	return ret;
 }
 
-int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
+void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -4764,7 +4855,6 @@ int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 	up_write(&fs_info->extent_commit_sem);
 
 	update_global_block_rsv(fs_info);
-	return 0;
 }
 
 static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
@@ -4779,7 +4869,7 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
 			if (cache)
 				btrfs_put_block_group(cache);
 			cache = btrfs_lookup_block_group(fs_info, start);
-			BUG_ON(!cache);
+			BUG_ON(!cache); /* Logic error */
 		}
 
 		len = cache->key.objectid + cache->key.offset - start;
@@ -4816,6 +4906,9 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 	u64 end;
 	int ret;
 
+	if (trans->aborted)
+		return 0;
+
 	if (fs_info->pinned_extents == &fs_info->freed_extents[0])
 		unpin = &fs_info->freed_extents[1];
 	else
@@ -4901,7 +4994,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			ret = remove_extent_backref(trans, extent_root, path,
 						    NULL, refs_to_drop,
 						    is_data);
-			BUG_ON(ret);
+			if (ret)
+				goto abort;
 			btrfs_release_path(path);
 			path->leave_spinning = 1;
 
@@ -4919,10 +5013,11 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 					btrfs_print_leaf(extent_root,
 							 path->nodes[0]);
 			}
-			BUG_ON(ret);
+			if (ret < 0)
+				goto abort;
 			extent_slot = path->slots[0];
 		}
-	} else {
+	} else if (ret == -ENOENT) {
 		btrfs_print_leaf(extent_root, path->nodes[0]);
 		WARN_ON(1);
 		printk(KERN_ERR "btrfs unable to find ref byte nr %llu "
@@ -4932,6 +5027,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 		       (unsigned long long)root_objectid,
 		       (unsigned long long)owner_objectid,
 		       (unsigned long long)owner_offset);
+	} else {
+		goto abort;
 	}
 
 	leaf = path->nodes[0];
@@ -4941,7 +5038,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 		BUG_ON(found_extent || extent_slot != path->slots[0]);
 		ret = convert_extent_item_v0(trans, extent_root, path,
 					     owner_objectid, 0);
-		BUG_ON(ret < 0);
+		if (ret < 0)
+			goto abort;
 
 		btrfs_release_path(path);
 		path->leave_spinning = 1;
@@ -4958,7 +5056,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			       (unsigned long long)bytenr);
 			btrfs_print_leaf(extent_root, path->nodes[0]);
 		}
-		BUG_ON(ret);
+		if (ret < 0)
+			goto abort;
 		extent_slot = path->slots[0];
 		leaf = path->nodes[0];
 		item_size = btrfs_item_size_nr(leaf, extent_slot);
@@ -4995,7 +5094,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			ret = remove_extent_backref(trans, extent_root, path,
 						    iref, refs_to_drop,
 						    is_data);
-			BUG_ON(ret);
+			if (ret)
+				goto abort;
 		}
 	} else {
 		if (found_extent) {
@@ -5012,23 +5112,27 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 
 		ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
 				      num_to_del);
-		BUG_ON(ret);
+		if (ret)
+			goto abort;
 		btrfs_release_path(path);
 
 		if (is_data) {
 			ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
-			BUG_ON(ret);
-		} else {
-			invalidate_mapping_pages(info->btree_inode->i_mapping,
-			     bytenr >> PAGE_CACHE_SHIFT,
-			     (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT);
+			if (ret)
+				goto abort;
 		}
 
 		ret = update_block_group(trans, root, bytenr, num_bytes, 0);
-		BUG_ON(ret);
+		if (ret)
+			goto abort;
 	}
+out:
 	btrfs_free_path(path);
 	return ret;
+
+abort:
+	btrfs_abort_transaction(trans, extent_root, ret);
+	goto out;
 }
 
 /*
@@ -5124,7 +5228,7 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 					parent, root->root_key.objectid,
 					btrfs_header_level(buf),
 					BTRFS_DROP_DELAYED_REF, NULL, for_cow);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 
 	if (!last_ref)
@@ -5158,6 +5262,7 @@ out:
 	btrfs_put_block_group(cache);
 }
 
+/* Can return -ENOMEM */
 int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		      u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid,
 		      u64 owner, u64 offset, int for_cow)
@@ -5179,14 +5284,12 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 					num_bytes,
 					parent, root_objectid, (int)owner,
 					BTRFS_DROP_DELAYED_REF, NULL, for_cow);
-		BUG_ON(ret);
 	} else {
 		ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr,
 						num_bytes,
 						parent, root_objectid, owner,
 						offset, BTRFS_DROP_DELAYED_REF,
 						NULL, for_cow);
-		BUG_ON(ret);
 	}
 	return ret;
 }
@@ -5243,28 +5346,34 @@ wait_block_group_cache_done(struct btrfs_block_group_cache *cache)
 	return 0;
 }
 
-static int get_block_group_index(struct btrfs_block_group_cache *cache)
+static int __get_block_group_index(u64 flags)
 {
 	int index;
-	if (cache->flags & BTRFS_BLOCK_GROUP_RAID10)
+
+	if (flags & BTRFS_BLOCK_GROUP_RAID10)
 		index = 0;
-	else if (cache->flags & BTRFS_BLOCK_GROUP_RAID1)
+	else if (flags & BTRFS_BLOCK_GROUP_RAID1)
 		index = 1;
-	else if (cache->flags & BTRFS_BLOCK_GROUP_DUP)
+	else if (flags & BTRFS_BLOCK_GROUP_DUP)
 		index = 2;
-	else if (cache->flags & BTRFS_BLOCK_GROUP_RAID0)
+	else if (flags & BTRFS_BLOCK_GROUP_RAID0)
 		index = 3;
 	else
 		index = 4;
+
 	return index;
 }
 
+static int get_block_group_index(struct btrfs_block_group_cache *cache)
+{
+	return __get_block_group_index(cache->flags);
+}
+
 enum btrfs_loop_type {
-	LOOP_FIND_IDEAL = 0,
-	LOOP_CACHING_NOWAIT = 1,
-	LOOP_CACHING_WAIT = 2,
-	LOOP_ALLOC_CHUNK = 3,
-	LOOP_NO_EMPTY_SIZE = 4,
+	LOOP_CACHING_NOWAIT = 0,
+	LOOP_CACHING_WAIT = 1,
+	LOOP_ALLOC_CHUNK = 2,
+	LOOP_NO_EMPTY_SIZE = 3,
 };
 
 /*
@@ -5278,7 +5387,6 @@ enum btrfs_loop_type {
 static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *orig_root,
 				     u64 num_bytes, u64 empty_size,
-				     u64 search_start, u64 search_end,
 				     u64 hint_byte, struct btrfs_key *ins,
 				     u64 data)
 {
@@ -5287,6 +5395,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_free_cluster *last_ptr = NULL;
 	struct btrfs_block_group_cache *block_group = NULL;
 	struct btrfs_block_group_cache *used_block_group;
+	u64 search_start = 0;
 	int empty_cluster = 2 * 1024 * 1024;
 	int allowed_chunk_alloc = 0;
 	int done_chunk_alloc = 0;
@@ -5300,8 +5409,6 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	bool failed_alloc = false;
 	bool use_cluster = true;
 	bool have_caching_bg = false;
-	u64 ideal_cache_percent = 0;
-	u64 ideal_cache_offset = 0;
 
 	WARN_ON(num_bytes < root->sectorsize);
 	btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
@@ -5351,7 +5458,6 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 		empty_cluster = 0;
 
 	if (search_start == hint_byte) {
-ideal_cache:
 		block_group = btrfs_lookup_block_group(root->fs_info,
 						       search_start);
 		used_block_group = block_group;
@@ -5363,8 +5469,7 @@ ideal_cache:
 		 * picked out then we don't care that the block group is cached.
 		 */
 		if (block_group && block_group_bits(block_group, data) &&
-		    (block_group->cached != BTRFS_CACHE_NO ||
-		     search_start == ideal_cache_offset)) {
+		    block_group->cached != BTRFS_CACHE_NO) {
 			down_read(&space_info->groups_sem);
 			if (list_empty(&block_group->list) ||
 			    block_group->ro) {
@@ -5418,44 +5523,13 @@ search:
 have_block_group:
 		cached = block_group_cache_done(block_group);
 		if (unlikely(!cached)) {
-			u64 free_percent;
-
 			found_uncached_bg = true;
 			ret = cache_block_group(block_group, trans,
-						orig_root, 1);
-			if (block_group->cached == BTRFS_CACHE_FINISHED)
-				goto alloc;
-
-			free_percent = btrfs_block_group_used(&block_group->item);
-			free_percent *= 100;
-			free_percent = div64_u64(free_percent,
-						 block_group->key.offset);
-			free_percent = 100 - free_percent;
-			if (free_percent > ideal_cache_percent &&
-			    likely(!block_group->ro)) {
-				ideal_cache_offset = block_group->key.objectid;
-				ideal_cache_percent = free_percent;
-			}
-
-			/*
-			 * The caching workers are limited to 2 threads, so we
-			 * can queue as much work as we care to.
-			 */
-			if (loop > LOOP_FIND_IDEAL) {
-				ret = cache_block_group(block_group, trans,
-							orig_root, 0);
-				BUG_ON(ret);
-			}
-
-			/*
-			 * If loop is set for cached only, try the next block
-			 * group.
-			 */
-			if (loop == LOOP_FIND_IDEAL)
-				goto loop;
+						orig_root, 0);
+			BUG_ON(ret < 0);
+			ret = 0;
 		}
 
-alloc:
 		if (unlikely(block_group->ro))
 			goto loop;
 
@@ -5606,11 +5680,6 @@ unclustered_alloc:
 		}
 checks:
 		search_start = stripe_align(root, offset);
-		/* move on to the next group */
-		if (search_start + num_bytes >= search_end) {
-			btrfs_add_free_space(used_block_group, offset, num_bytes);
-			goto loop;
-		}
 
 		/* move on to the next group */
 		if (search_start + num_bytes >
@@ -5661,9 +5730,7 @@ loop:
 	if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES)
 		goto search;
 
-	/* LOOP_FIND_IDEAL, only search caching/cached bg's, and don't wait for
-	 *			for them to make caching progress.  Also
-	 *			determine the best possible bg to cache
+	/*
 	 * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking
 	 *			caching kthreads as we move along
 	 * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
@@ -5673,50 +5740,17 @@ loop:
 	 */
 	if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) {
 		index = 0;
-		if (loop == LOOP_FIND_IDEAL && found_uncached_bg) {
-			found_uncached_bg = false;
-			loop++;
-			if (!ideal_cache_percent)
-				goto search;
-
-			/*
-			 * 1 of the following 2 things have happened so far
-			 *
-			 * 1) We found an ideal block group for caching that
-			 * is mostly full and will cache quickly, so we might
-			 * as well wait for it.
-			 *
-			 * 2) We searched for cached only and we didn't find
-			 * anything, and we didn't start any caching kthreads
-			 * either, so chances are we will loop through and
-			 * start a couple caching kthreads, and then come back
-			 * around and just wait for them.  This will be slower
-			 * because we will have 2 caching kthreads reading at
-			 * the same time when we could have just started one
-			 * and waited for it to get far enough to give us an
-			 * allocation, so go ahead and go to the wait caching
-			 * loop.
-			 */
-			loop = LOOP_CACHING_WAIT;
-			search_start = ideal_cache_offset;
-			ideal_cache_percent = 0;
-			goto ideal_cache;
-		} else if (loop == LOOP_FIND_IDEAL) {
-			/*
-			 * Didn't find a uncached bg, wait on anything we find
-			 * next.
-			 */
-			loop = LOOP_CACHING_WAIT;
-			goto search;
-		}
-
 		loop++;
-
 		if (loop == LOOP_ALLOC_CHUNK) {
 		       if (allowed_chunk_alloc) {
 				ret = do_chunk_alloc(trans, root, num_bytes +
 						     2 * 1024 * 1024, data,
 						     CHUNK_ALLOC_LIMITED);
+				if (ret < 0) {
+					btrfs_abort_transaction(trans,
+								root, ret);
+					goto out;
+				}
 				allowed_chunk_alloc = 0;
 				if (ret == 1)
 					done_chunk_alloc = 1;
@@ -5745,6 +5779,7 @@ loop:
 	} else if (ins->objectid) {
 		ret = 0;
 	}
+out:
 
 	return ret;
 }
@@ -5798,12 +5833,10 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root,
 			 u64 num_bytes, u64 min_alloc_size,
 			 u64 empty_size, u64 hint_byte,
-			 u64 search_end, struct btrfs_key *ins,
-			 u64 data)
+			 struct btrfs_key *ins, u64 data)
 {
 	bool final_tried = false;
 	int ret;
-	u64 search_start = 0;
 
 	data = btrfs_get_alloc_profile(root, data);
 again:
@@ -5811,23 +5844,31 @@ again:
 	 * the only place that sets empty_size is btrfs_realloc_node, which
 	 * is not called recursively on allocations
 	 */
-	if (empty_size || root->ref_cows)
+	if (empty_size || root->ref_cows) {
 		ret = do_chunk_alloc(trans, root->fs_info->extent_root,
 				     num_bytes + 2 * 1024 * 1024, data,
 				     CHUNK_ALLOC_NO_FORCE);
+		if (ret < 0 && ret != -ENOSPC) {
+			btrfs_abort_transaction(trans, root, ret);
+			return ret;
+		}
+	}
 
 	WARN_ON(num_bytes < root->sectorsize);
 	ret = find_free_extent(trans, root, num_bytes, empty_size,
-			       search_start, search_end, hint_byte,
-			       ins, data);
+			       hint_byte, ins, data);
 
 	if (ret == -ENOSPC) {
 		if (!final_tried) {
 			num_bytes = num_bytes >> 1;
 			num_bytes = num_bytes & ~(root->sectorsize - 1);
 			num_bytes = max(num_bytes, min_alloc_size);
-			do_chunk_alloc(trans, root->fs_info->extent_root,
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
 				       num_bytes, data, CHUNK_ALLOC_FORCE);
+			if (ret < 0 && ret != -ENOSPC) {
+				btrfs_abort_transaction(trans, root, ret);
+				return ret;
+			}
 			if (num_bytes == min_alloc_size)
 				final_tried = true;
 			goto again;
@@ -5838,7 +5879,8 @@ again:
 			printk(KERN_ERR "btrfs allocation failed flags %llu, "
 			       "wanted %llu\n", (unsigned long long)data,
 			       (unsigned long long)num_bytes);
-			dump_space_info(sinfo, num_bytes, 1);
+			if (sinfo)
+				dump_space_info(sinfo, num_bytes, 1);
 		}
 	}
 
@@ -5917,7 +5959,10 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	path->leave_spinning = 1;
 	ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
 				      ins, size);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_free_path(path);
+		return ret;
+	}
 
 	leaf = path->nodes[0];
 	extent_item = btrfs_item_ptr(leaf, path->slots[0],
@@ -5947,7 +5992,7 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	btrfs_free_path(path);
 
 	ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
-	if (ret) {
+	if (ret) { /* -ENOENT, logic error */
 		printk(KERN_ERR "btrfs update block group failed for %llu "
 		       "%llu\n", (unsigned long long)ins->objectid,
 		       (unsigned long long)ins->offset);
@@ -5978,7 +6023,10 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	path->leave_spinning = 1;
 	ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
 				      ins, size);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_free_path(path);
+		return ret;
+	}
 
 	leaf = path->nodes[0];
 	extent_item = btrfs_item_ptr(leaf, path->slots[0],
@@ -6008,7 +6056,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	btrfs_free_path(path);
 
 	ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
-	if (ret) {
+	if (ret) { /* -ENOENT, logic error */
 		printk(KERN_ERR "btrfs update block group failed for %llu "
 		       "%llu\n", (unsigned long long)ins->objectid,
 		       (unsigned long long)ins->offset);
@@ -6056,28 +6104,28 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 	if (!caching_ctl) {
 		BUG_ON(!block_group_cache_done(block_group));
 		ret = btrfs_remove_free_space(block_group, start, num_bytes);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	} else {
 		mutex_lock(&caching_ctl->mutex);
 
 		if (start >= caching_ctl->progress) {
 			ret = add_excluded_extent(root, start, num_bytes);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		} else if (start + num_bytes <= caching_ctl->progress) {
 			ret = btrfs_remove_free_space(block_group,
 						      start, num_bytes);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		} else {
 			num_bytes = caching_ctl->progress - start;
 			ret = btrfs_remove_free_space(block_group,
 						      start, num_bytes);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 
 			start = caching_ctl->progress;
 			num_bytes = ins->objectid + ins->offset -
 				    caching_ctl->progress;
 			ret = add_excluded_extent(root, start, num_bytes);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		}
 
 		mutex_unlock(&caching_ctl->mutex);
@@ -6086,7 +6134,7 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_update_reserved_bytes(block_group, ins->offset,
 					  RESERVE_ALLOC_NO_ACCOUNT);
-	BUG_ON(ret);
+	BUG_ON(ret); /* logic error */
 	btrfs_put_block_group(block_group);
 	ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
 					 0, owner, offset, ins, 1);
@@ -6107,6 +6155,7 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 	btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level);
 	btrfs_tree_lock(buf);
 	clean_tree_block(trans, root, buf);
+	clear_bit(EXTENT_BUFFER_STALE, &buf->bflags);
 
 	btrfs_set_lock_blocking(buf);
 	btrfs_set_buffer_uptodate(buf);
@@ -6214,7 +6263,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 		return ERR_CAST(block_rsv);
 
 	ret = btrfs_reserve_extent(trans, root, blocksize, blocksize,
-				   empty_size, hint, (u64)-1, &ins, 0);
+				   empty_size, hint, &ins, 0);
 	if (ret) {
 		unuse_block_rsv(root->fs_info, block_rsv, blocksize);
 		return ERR_PTR(ret);
@@ -6222,7 +6271,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 
 	buf = btrfs_init_new_buffer(trans, root, ins.objectid,
 				    blocksize, level);
-	BUG_ON(IS_ERR(buf));
+	BUG_ON(IS_ERR(buf)); /* -ENOMEM */
 
 	if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
 		if (parent == 0)
@@ -6234,7 +6283,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
 		struct btrfs_delayed_extent_op *extent_op;
 		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
-		BUG_ON(!extent_op);
+		BUG_ON(!extent_op); /* -ENOMEM */
 		if (key)
 			memcpy(&extent_op->key, key, sizeof(extent_op->key));
 		else
@@ -6249,7 +6298,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 					ins.offset, parent, root_objectid,
 					level, BTRFS_ADD_DELAYED_EXTENT,
 					extent_op, for_cow);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 	return buf;
 }
@@ -6319,7 +6368,9 @@ static noinline void reada_walk_down(struct btrfs_trans_handle *trans,
 		/* We don't lock the tree block, it's OK to be racy here */
 		ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize,
 					       &refs, &flags);
-		BUG_ON(ret);
+		/* We don't care about errors in readahead. */
+		if (ret < 0)
+			continue;
 		BUG_ON(refs == 0);
 
 		if (wc->stage == DROP_REFERENCE) {
@@ -6386,7 +6437,9 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
 					       eb->start, eb->len,
 					       &wc->refs[level],
 					       &wc->flags[level]);
-		BUG_ON(ret);
+		BUG_ON(ret == -ENOMEM);
+		if (ret)
+			return ret;
 		BUG_ON(wc->refs[level] == 0);
 	}
 
@@ -6405,12 +6458,12 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
 	if (!(wc->flags[level] & flag)) {
 		BUG_ON(!path->locks[level]);
 		ret = btrfs_inc_ref(trans, root, eb, 1, wc->for_reloc);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 		ret = btrfs_dec_ref(trans, root, eb, 0, wc->for_reloc);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 		ret = btrfs_set_disk_extent_flags(trans, root, eb->start,
 						  eb->len, flag, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 		wc->flags[level] |= flag;
 	}
 
@@ -6482,7 +6535,11 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize,
 				       &wc->refs[level - 1],
 				       &wc->flags[level - 1]);
-	BUG_ON(ret);
+	if (ret < 0) {
+		btrfs_tree_unlock(next);
+		return ret;
+	}
+
 	BUG_ON(wc->refs[level - 1] == 0);
 	*lookup_info = 0;
 
@@ -6511,7 +6568,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 			goto skip;
 	}
 
-	if (!btrfs_buffer_uptodate(next, generation)) {
+	if (!btrfs_buffer_uptodate(next, generation, 0)) {
 		btrfs_tree_unlock(next);
 		free_extent_buffer(next);
 		next = NULL;
@@ -6551,7 +6608,7 @@ skip:
 
 		ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent,
 				root->root_key.objectid, level - 1, 0, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 	btrfs_tree_unlock(next);
 	free_extent_buffer(next);
@@ -6609,7 +6666,10 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 						       eb->start, eb->len,
 						       &wc->refs[level],
 						       &wc->flags[level]);
-			BUG_ON(ret);
+			if (ret < 0) {
+				btrfs_tree_unlock_rw(eb, path->locks[level]);
+				return ret;
+			}
 			BUG_ON(wc->refs[level] == 0);
 			if (wc->refs[level] == 1) {
 				btrfs_tree_unlock_rw(eb, path->locks[level]);
@@ -6629,7 +6689,7 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 			else
 				ret = btrfs_dec_ref(trans, root, eb, 0,
 						    wc->for_reloc);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM */
 		}
 		/* make block locked assertion in clean_tree_block happy */
 		if (!path->locks[level] &&
@@ -6738,7 +6798,7 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
  * also make sure backrefs for the shared block and all lower level
  * blocks are properly updated.
  */
-void btrfs_drop_snapshot(struct btrfs_root *root,
+int btrfs_drop_snapshot(struct btrfs_root *root,
 			 struct btrfs_block_rsv *block_rsv, int update_ref,
 			 int for_reloc)
 {
@@ -6766,7 +6826,10 @@ void btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 
 	trans = btrfs_start_transaction(tree_root, 0);
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans)) {
+		err = PTR_ERR(trans);
+		goto out_free;
+	}
 
 	if (block_rsv)
 		trans->block_rsv = block_rsv;
@@ -6791,7 +6854,7 @@ void btrfs_drop_snapshot(struct btrfs_root *root,
 		path->lowest_level = 0;
 		if (ret < 0) {
 			err = ret;
-			goto out_free;
+			goto out_end_trans;
 		}
 		WARN_ON(ret > 0);
 
@@ -6811,7 +6874,10 @@ void btrfs_drop_snapshot(struct btrfs_root *root,
 						path->nodes[level]->len,
 						&wc->refs[level],
 						&wc->flags[level]);
-			BUG_ON(ret);
+			if (ret < 0) {
+				err = ret;
+				goto out_end_trans;
+			}
 			BUG_ON(wc->refs[level] == 0);
 
 			if (level == root_item->drop_level)
@@ -6862,26 +6928,40 @@ void btrfs_drop_snapshot(struct btrfs_root *root,
 			ret = btrfs_update_root(trans, tree_root,
 						&root->root_key,
 						root_item);
-			BUG_ON(ret);
+			if (ret) {
+				btrfs_abort_transaction(trans, tree_root, ret);
+				err = ret;
+				goto out_end_trans;
+			}
 
 			btrfs_end_transaction_throttle(trans, tree_root);
 			trans = btrfs_start_transaction(tree_root, 0);
-			BUG_ON(IS_ERR(trans));
+			if (IS_ERR(trans)) {
+				err = PTR_ERR(trans);
+				goto out_free;
+			}
 			if (block_rsv)
 				trans->block_rsv = block_rsv;
 		}
 	}
 	btrfs_release_path(path);
-	BUG_ON(err);
+	if (err)
+		goto out_end_trans;
 
 	ret = btrfs_del_root(trans, tree_root, &root->root_key);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, tree_root, ret);
+		goto out_end_trans;
+	}
 
 	if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
 		ret = btrfs_find_last_root(tree_root, root->root_key.objectid,
 					   NULL, NULL);
-		BUG_ON(ret < 0);
-		if (ret > 0) {
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, tree_root, ret);
+			err = ret;
+			goto out_end_trans;
+		} else if (ret > 0) {
 			/* if we fail to delete the orphan item this time
 			 * around, it'll get picked up the next time.
 			 *
@@ -6899,14 +6979,15 @@ void btrfs_drop_snapshot(struct btrfs_root *root,
 		free_extent_buffer(root->commit_root);
 		kfree(root);
 	}
-out_free:
+out_end_trans:
 	btrfs_end_transaction_throttle(trans, tree_root);
+out_free:
 	kfree(wc);
 	btrfs_free_path(path);
 out:
 	if (err)
 		btrfs_std_error(root->fs_info, err);
-	return;
+	return err;
 }
 
 /*
@@ -6983,31 +7064,15 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 {
 	u64 num_devices;
-	u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
-		BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
-
-	if (root->fs_info->balance_ctl) {
-		struct btrfs_balance_control *bctl = root->fs_info->balance_ctl;
-		u64 tgt = 0;
-
-		/* pick restriper's target profile and return */
-		if (flags & BTRFS_BLOCK_GROUP_DATA &&
-		    bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) {
-			tgt = BTRFS_BLOCK_GROUP_DATA | bctl->data.target;
-		} else if (flags & BTRFS_BLOCK_GROUP_SYSTEM &&
-			   bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) {
-			tgt = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target;
-		} else if (flags & BTRFS_BLOCK_GROUP_METADATA &&
-			   bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) {
-			tgt = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target;
-		}
+	u64 stripped;
 
-		if (tgt) {
-			/* extended -> chunk profile */
-			tgt &= ~BTRFS_AVAIL_ALLOC_BIT_SINGLE;
-			return tgt;
-		}
-	}
+	/*
+	 * if restripe for this chunk_type is on pick target profile and
+	 * return, otherwise do the usual balance
+	 */
+	stripped = get_restripe_target(root->fs_info, flags);
+	if (stripped)
+		return extended_to_chunk(stripped);
 
 	/*
 	 * we add in the count of missing devices because we want
@@ -7017,6 +7082,9 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 	num_devices = root->fs_info->fs_devices->rw_devices +
 		root->fs_info->fs_devices->missing_devices;
 
+	stripped = BTRFS_BLOCK_GROUP_RAID0 |
+		BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
+
 	if (num_devices == 1) {
 		stripped |= BTRFS_BLOCK_GROUP_DUP;
 		stripped = flags & ~stripped;
@@ -7029,7 +7097,6 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 		if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
 			     BTRFS_BLOCK_GROUP_RAID10))
 			return stripped | BTRFS_BLOCK_GROUP_DUP;
-		return flags;
 	} else {
 		/* they already had raid on here, just return */
 		if (flags & stripped)
@@ -7042,9 +7109,9 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 		if (flags & BTRFS_BLOCK_GROUP_DUP)
 			return stripped | BTRFS_BLOCK_GROUP_RAID1;
 
-		/* turn single device chunks into raid0 */
-		return stripped | BTRFS_BLOCK_GROUP_RAID0;
+		/* this is drive concat, leave it alone */
 	}
+
 	return flags;
 }
 
@@ -7103,12 +7170,16 @@ int btrfs_set_block_group_ro(struct btrfs_root *root,
 	BUG_ON(cache->ro);
 
 	trans = btrfs_join_transaction(root);
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
 	alloc_flags = update_block_group_flags(root, cache->flags);
-	if (alloc_flags != cache->flags)
-		do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
-			       CHUNK_ALLOC_FORCE);
+	if (alloc_flags != cache->flags) {
+		ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
+				     CHUNK_ALLOC_FORCE);
+		if (ret < 0)
+			goto out;
+	}
 
 	ret = set_block_group_ro(cache, 0);
 	if (!ret)
@@ -7188,7 +7259,7 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo)
 	return free_bytes;
 }
 
-int btrfs_set_block_group_rw(struct btrfs_root *root,
+void btrfs_set_block_group_rw(struct btrfs_root *root,
 			      struct btrfs_block_group_cache *cache)
 {
 	struct btrfs_space_info *sinfo = cache->space_info;
@@ -7204,7 +7275,6 @@ int btrfs_set_block_group_rw(struct btrfs_root *root,
 	cache->ro = 0;
 	spin_unlock(&cache->lock);
 	spin_unlock(&sinfo->lock);
-	return 0;
 }
 
 /*
@@ -7222,6 +7292,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 	u64 min_free;
 	u64 dev_min = 1;
 	u64 dev_nr = 0;
+	u64 target;
 	int index;
 	int full = 0;
 	int ret = 0;
@@ -7262,13 +7333,11 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 	/*
 	 * ok we don't have enough space, but maybe we have free space on our
 	 * devices to allocate new chunks for relocation, so loop through our
-	 * alloc devices and guess if we have enough space.  However, if we
-	 * were marked as full, then we know there aren't enough chunks, and we
-	 * can just return.
+	 * alloc devices and guess if we have enough space.  if this block
+	 * group is going to be restriped, run checks against the target
+	 * profile instead of the current one.
 	 */
 	ret = -1;
-	if (full)
-		goto out;
 
 	/*
 	 * index:
@@ -7278,7 +7347,20 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 	 *      3: raid0
 	 *      4: single
 	 */
-	index = get_block_group_index(block_group);
+	target = get_restripe_target(root->fs_info, block_group->flags);
+	if (target) {
+		index = __get_block_group_index(extended_to_chunk(target));
+	} else {
+		/*
+		 * this is just a balance, so if we were marked as full
+		 * we know there is no space for a new chunk
+		 */
+		if (full)
+			goto out;
+
+		index = get_block_group_index(block_group);
+	}
+
 	if (index == 0) {
 		dev_min = 4;
 		/* Divide by 2 */
@@ -7572,7 +7654,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		ret = update_space_info(info, cache->flags, found_key.offset,
 					btrfs_block_group_used(&cache->item),
 					&space_info);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 		cache->space_info = space_info;
 		spin_lock(&cache->space_info->lock);
 		cache->space_info->bytes_readonly += cache->bytes_super;
@@ -7581,7 +7663,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		__link_block_group(space_info, cache);
 
 		ret = btrfs_add_block_group_cache(root->fs_info, cache);
-		BUG_ON(ret);
+		BUG_ON(ret); /* Logic error */
 
 		set_avail_alloc_bits(root->fs_info, cache->flags);
 		if (btrfs_chunk_readonly(root, cache->key.objectid))
@@ -7663,7 +7745,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 
 	ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
 				&cache->space_info);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 	update_global_block_rsv(root->fs_info);
 
 	spin_lock(&cache->space_info->lock);
@@ -7673,11 +7755,14 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	__link_block_group(cache->space_info, cache);
 
 	ret = btrfs_add_block_group_cache(root->fs_info, cache);
-	BUG_ON(ret);
+	BUG_ON(ret); /* Logic error */
 
 	ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
 				sizeof(cache->item));
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, extent_root, ret);
+		return ret;
+	}
 
 	set_avail_alloc_bits(extent_root->fs_info, type);
 
@@ -7686,11 +7771,8 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 
 static void clear_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
 {
-	u64 extra_flags = flags & BTRFS_BLOCK_GROUP_PROFILE_MASK;
-
-	/* chunk -> extended profile */
-	if (extra_flags == 0)
-		extra_flags = BTRFS_AVAIL_ALLOC_BIT_SINGLE;
+	u64 extra_flags = chunk_to_extended(flags) &
+				BTRFS_EXTENDED_PROFILE_MASK;
 
 	if (flags & BTRFS_BLOCK_GROUP_DATA)
 		fs_info->avail_data_alloc_bits &= ~extra_flags;
@@ -7758,7 +7840,10 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	inode = lookup_free_space_inode(tree_root, block_group, path);
 	if (!IS_ERR(inode)) {
 		ret = btrfs_orphan_add(trans, inode);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_add_delayed_iput(inode);
+			goto out;
+		}
 		clear_nlink(inode);
 		/* One for the block groups ref */
 		spin_lock(&block_group->lock);
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index a55fbe6252d..c9018a05036 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -19,6 +19,7 @@
 #include "btrfs_inode.h"
 #include "volumes.h"
 #include "check-integrity.h"
+#include "locking.h"
 
 static struct kmem_cache *extent_state_cache;
 static struct kmem_cache *extent_buffer_cache;
@@ -53,6 +54,13 @@ struct extent_page_data {
 	unsigned int sync_io:1;
 };
 
+static noinline void flush_write_bio(void *data);
+static inline struct btrfs_fs_info *
+tree_fs_info(struct extent_io_tree *tree)
+{
+	return btrfs_sb(tree->mapping->host->i_sb);
+}
+
 int __init extent_io_init(void)
 {
 	extent_state_cache = kmem_cache_create("extent_state",
@@ -136,6 +144,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask)
 #endif
 	atomic_set(&state->refs, 1);
 	init_waitqueue_head(&state->wq);
+	trace_alloc_extent_state(state, mask, _RET_IP_);
 	return state;
 }
 
@@ -153,6 +162,7 @@ void free_extent_state(struct extent_state *state)
 		list_del(&state->leak_list);
 		spin_unlock_irqrestore(&leak_lock, flags);
 #endif
+		trace_free_extent_state(state, _RET_IP_);
 		kmem_cache_free(extent_state_cache, state);
 	}
 }
@@ -392,20 +402,28 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
 	return 0;
 }
 
+static struct extent_state *next_state(struct extent_state *state)
+{
+	struct rb_node *next = rb_next(&state->rb_node);
+	if (next)
+		return rb_entry(next, struct extent_state, rb_node);
+	else
+		return NULL;
+}
+
 /*
  * utility function to clear some bits in an extent state struct.
- * it will optionally wake up any one waiting on this state (wake == 1), or
- * forcibly remove the state from the tree (delete == 1).
+ * it will optionally wake up any one waiting on this state (wake == 1)
  *
  * If no bits are set on the state struct after clearing things, the
  * struct is freed and removed from the tree
  */
-static int clear_state_bit(struct extent_io_tree *tree,
-			    struct extent_state *state,
-			    int *bits, int wake)
+static struct extent_state *clear_state_bit(struct extent_io_tree *tree,
+					    struct extent_state *state,
+					    int *bits, int wake)
 {
+	struct extent_state *next;
 	int bits_to_clear = *bits & ~EXTENT_CTLBITS;
-	int ret = state->state & bits_to_clear;
 
 	if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
 		u64 range = state->end - state->start + 1;
@@ -417,6 +435,7 @@ static int clear_state_bit(struct extent_io_tree *tree,
 	if (wake)
 		wake_up(&state->wq);
 	if (state->state == 0) {
+		next = next_state(state);
 		if (state->tree) {
 			rb_erase(&state->rb_node, &tree->state);
 			state->tree = NULL;
@@ -426,8 +445,9 @@ static int clear_state_bit(struct extent_io_tree *tree,
 		}
 	} else {
 		merge_state(tree, state);
+		next = next_state(state);
 	}
-	return ret;
+	return next;
 }
 
 static struct extent_state *
@@ -439,6 +459,13 @@ alloc_extent_state_atomic(struct extent_state *prealloc)
 	return prealloc;
 }
 
+void extent_io_tree_panic(struct extent_io_tree *tree, int err)
+{
+	btrfs_panic(tree_fs_info(tree), err, "Locking error: "
+		    "Extent tree was modified by another "
+		    "thread while locked.");
+}
+
 /*
  * clear some bits on a range in the tree.  This may require splitting
  * or inserting elements in the tree, so the gfp mask is used to
@@ -449,8 +476,7 @@ alloc_extent_state_atomic(struct extent_state *prealloc)
  *
  * the range [start, end] is inclusive.
  *
- * This takes the tree lock, and returns < 0 on error, > 0 if any of the
- * bits were already set, or zero if none of the bits were already set.
+ * This takes the tree lock, and returns 0 on success and < 0 on error.
  */
 int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		     int bits, int wake, int delete,
@@ -460,11 +486,9 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	struct extent_state *state;
 	struct extent_state *cached;
 	struct extent_state *prealloc = NULL;
-	struct rb_node *next_node;
 	struct rb_node *node;
 	u64 last_end;
 	int err;
-	int set = 0;
 	int clear = 0;
 
 	if (delete)
@@ -513,14 +537,11 @@ hit_next:
 	WARN_ON(state->end < start);
 	last_end = state->end;
 
-	if (state->end < end && !need_resched())
-		next_node = rb_next(&state->rb_node);
-	else
-		next_node = NULL;
-
 	/* the state doesn't have the wanted bits, go ahead */
-	if (!(state->state & bits))
+	if (!(state->state & bits)) {
+		state = next_state(state);
 		goto next;
+	}
 
 	/*
 	 *     | ---- desired range ---- |
@@ -542,12 +563,14 @@ hit_next:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, start);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
+
 		prealloc = NULL;
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			set |= clear_state_bit(tree, state, &bits, wake);
+			clear_state_bit(tree, state, &bits, wake);
 			if (last_end == (u64)-1)
 				goto out;
 			start = last_end + 1;
@@ -564,26 +587,25 @@ hit_next:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, end + 1);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
+
 		if (wake)
 			wake_up(&state->wq);
 
-		set |= clear_state_bit(tree, prealloc, &bits, wake);
+		clear_state_bit(tree, prealloc, &bits, wake);
 
 		prealloc = NULL;
 		goto out;
 	}
 
-	set |= clear_state_bit(tree, state, &bits, wake);
+	state = clear_state_bit(tree, state, &bits, wake);
 next:
 	if (last_end == (u64)-1)
 		goto out;
 	start = last_end + 1;
-	if (start <= end && next_node) {
-		state = rb_entry(next_node, struct extent_state,
-				 rb_node);
+	if (start <= end && state && !need_resched())
 		goto hit_next;
-	}
 	goto search_again;
 
 out:
@@ -591,7 +613,7 @@ out:
 	if (prealloc)
 		free_extent_state(prealloc);
 
-	return set;
+	return 0;
 
 search_again:
 	if (start > end)
@@ -602,8 +624,8 @@ search_again:
 	goto again;
 }
 
-static int wait_on_state(struct extent_io_tree *tree,
-			 struct extent_state *state)
+static void wait_on_state(struct extent_io_tree *tree,
+			  struct extent_state *state)
 		__releases(tree->lock)
 		__acquires(tree->lock)
 {
@@ -613,7 +635,6 @@ static int wait_on_state(struct extent_io_tree *tree,
 	schedule();
 	spin_lock(&tree->lock);
 	finish_wait(&state->wq, &wait);
-	return 0;
 }
 
 /*
@@ -621,7 +642,7 @@ static int wait_on_state(struct extent_io_tree *tree,
  * The range [start, end] is inclusive.
  * The tree lock is taken by this function
  */
-int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits)
+void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits)
 {
 	struct extent_state *state;
 	struct rb_node *node;
@@ -658,7 +679,6 @@ again:
 	}
 out:
 	spin_unlock(&tree->lock);
-	return 0;
 }
 
 static void set_state_bits(struct extent_io_tree *tree,
@@ -706,9 +726,10 @@ static void uncache_state(struct extent_state **cached_ptr)
  * [start, end] is inclusive This takes the tree lock.
  */
 
-int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
-		   int bits, int exclusive_bits, u64 *failed_start,
-		   struct extent_state **cached_state, gfp_t mask)
+static int __must_check
+__set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		 int bits, int exclusive_bits, u64 *failed_start,
+		 struct extent_state **cached_state, gfp_t mask)
 {
 	struct extent_state *state;
 	struct extent_state *prealloc = NULL;
@@ -742,8 +763,10 @@ again:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = insert_state(tree, prealloc, start, end, &bits);
+		if (err)
+			extent_io_tree_panic(tree, err);
+
 		prealloc = NULL;
-		BUG_ON(err == -EEXIST);
 		goto out;
 	}
 	state = rb_entry(node, struct extent_state, rb_node);
@@ -809,7 +832,9 @@ hit_next:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, start);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
+
 		prealloc = NULL;
 		if (err)
 			goto out;
@@ -846,12 +871,9 @@ hit_next:
 		 */
 		err = insert_state(tree, prealloc, start, this_end,
 				   &bits);
-		BUG_ON(err == -EEXIST);
-		if (err) {
-			free_extent_state(prealloc);
-			prealloc = NULL;
-			goto out;
-		}
+		if (err)
+			extent_io_tree_panic(tree, err);
+
 		cache_state(prealloc, cached_state);
 		prealloc = NULL;
 		start = this_end + 1;
@@ -873,7 +895,8 @@ hit_next:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, end + 1);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
 
 		set_state_bits(tree, prealloc, &bits);
 		cache_state(prealloc, cached_state);
@@ -900,6 +923,15 @@ search_again:
 	goto again;
 }
 
+int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits,
+		   u64 *failed_start, struct extent_state **cached_state,
+		   gfp_t mask)
+{
+	return __set_extent_bit(tree, start, end, bits, 0, failed_start,
+				cached_state, mask);
+}
+
+
 /**
  * convert_extent - convert all bits in a given range from one bit to another
  * @tree:	the io tree to search
@@ -946,7 +978,8 @@ again:
 		}
 		err = insert_state(tree, prealloc, start, end, &bits);
 		prealloc = NULL;
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
 		goto out;
 	}
 	state = rb_entry(node, struct extent_state, rb_node);
@@ -1002,7 +1035,8 @@ hit_next:
 			goto out;
 		}
 		err = split_state(tree, state, prealloc, start);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
 		prealloc = NULL;
 		if (err)
 			goto out;
@@ -1041,12 +1075,8 @@ hit_next:
 		 */
 		err = insert_state(tree, prealloc, start, this_end,
 				   &bits);
-		BUG_ON(err == -EEXIST);
-		if (err) {
-			free_extent_state(prealloc);
-			prealloc = NULL;
-			goto out;
-		}
+		if (err)
+			extent_io_tree_panic(tree, err);
 		prealloc = NULL;
 		start = this_end + 1;
 		goto search_again;
@@ -1065,7 +1095,8 @@ hit_next:
 		}
 
 		err = split_state(tree, state, prealloc, end + 1);
-		BUG_ON(err == -EEXIST);
+		if (err)
+			extent_io_tree_panic(tree, err);
 
 		set_state_bits(tree, prealloc, &bits);
 		clear_state_bit(tree, prealloc, &clear_bits, 0);
@@ -1095,14 +1126,14 @@ search_again:
 int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 		     gfp_t mask)
 {
-	return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
+	return set_extent_bit(tree, start, end, EXTENT_DIRTY, NULL,
 			      NULL, mask);
 }
 
 int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		    int bits, gfp_t mask)
 {
-	return set_extent_bit(tree, start, end, bits, 0, NULL,
+	return set_extent_bit(tree, start, end, bits, NULL,
 			      NULL, mask);
 }
 
@@ -1117,7 +1148,7 @@ int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
 {
 	return set_extent_bit(tree, start, end,
 			      EXTENT_DELALLOC | EXTENT_UPTODATE,
-			      0, NULL, cached_state, mask);
+			      NULL, cached_state, mask);
 }
 
 int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
@@ -1131,7 +1162,7 @@ int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 		     gfp_t mask)
 {
-	return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
+	return set_extent_bit(tree, start, end, EXTENT_NEW, NULL,
 			      NULL, mask);
 }
 
@@ -1139,7 +1170,7 @@ int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
 			struct extent_state **cached_state, gfp_t mask)
 {
 	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0,
-			      NULL, cached_state, mask);
+			      cached_state, mask);
 }
 
 static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
@@ -1155,42 +1186,40 @@ static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
  * us if waiting is desired.
  */
 int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
-		     int bits, struct extent_state **cached_state, gfp_t mask)
+		     int bits, struct extent_state **cached_state)
 {
 	int err;
 	u64 failed_start;
 	while (1) {
-		err = set_extent_bit(tree, start, end, EXTENT_LOCKED | bits,
-				     EXTENT_LOCKED, &failed_start,
-				     cached_state, mask);
-		if (err == -EEXIST && (mask & __GFP_WAIT)) {
+		err = __set_extent_bit(tree, start, end, EXTENT_LOCKED | bits,
+				       EXTENT_LOCKED, &failed_start,
+				       cached_state, GFP_NOFS);
+		if (err == -EEXIST) {
 			wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
 			start = failed_start;
-		} else {
+		} else
 			break;
-		}
 		WARN_ON(start > end);
 	}
 	return err;
 }
 
-int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
+int lock_extent(struct extent_io_tree *tree, u64 start, u64 end)
 {
-	return lock_extent_bits(tree, start, end, 0, NULL, mask);
+	return lock_extent_bits(tree, start, end, 0, NULL);
 }
 
-int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
-		    gfp_t mask)
+int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end)
 {
 	int err;
 	u64 failed_start;
 
-	err = set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED,
-			     &failed_start, NULL, mask);
+	err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED,
+			       &failed_start, NULL, GFP_NOFS);
 	if (err == -EEXIST) {
 		if (failed_start > start)
 			clear_extent_bit(tree, start, failed_start - 1,
-					 EXTENT_LOCKED, 1, 0, NULL, mask);
+					 EXTENT_LOCKED, 1, 0, NULL, GFP_NOFS);
 		return 0;
 	}
 	return 1;
@@ -1203,10 +1232,10 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
 				mask);
 }
 
-int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end)
 {
 	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL,
-				mask);
+				GFP_NOFS);
 }
 
 /*
@@ -1220,7 +1249,7 @@ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end)
 
 	while (index <= end_index) {
 		page = find_get_page(tree->mapping, index);
-		BUG_ON(!page);
+		BUG_ON(!page); /* Pages should be in the extent_io_tree */
 		set_page_writeback(page);
 		page_cache_release(page);
 		index++;
@@ -1343,9 +1372,9 @@ out:
 	return found;
 }
 
-static noinline int __unlock_for_delalloc(struct inode *inode,
-					  struct page *locked_page,
-					  u64 start, u64 end)
+static noinline void __unlock_for_delalloc(struct inode *inode,
+					   struct page *locked_page,
+					   u64 start, u64 end)
 {
 	int ret;
 	struct page *pages[16];
@@ -1355,7 +1384,7 @@ static noinline int __unlock_for_delalloc(struct inode *inode,
 	int i;
 
 	if (index == locked_page->index && end_index == index)
-		return 0;
+		return;
 
 	while (nr_pages > 0) {
 		ret = find_get_pages_contig(inode->i_mapping, index,
@@ -1370,7 +1399,6 @@ static noinline int __unlock_for_delalloc(struct inode *inode,
 		index += ret;
 		cond_resched();
 	}
-	return 0;
 }
 
 static noinline int lock_delalloc_pages(struct inode *inode,
@@ -1500,11 +1528,10 @@ again:
 			goto out_failed;
 		}
 	}
-	BUG_ON(ret);
+	BUG_ON(ret); /* Only valid values are 0 and -EAGAIN */
 
 	/* step three, lock the state bits for the whole range */
-	lock_extent_bits(tree, delalloc_start, delalloc_end,
-			 0, &cached_state, GFP_NOFS);
+	lock_extent_bits(tree, delalloc_start, delalloc_end, 0, &cached_state);
 
 	/* then test to make sure it is all still delalloc */
 	ret = test_range_bit(tree, delalloc_start, delalloc_end,
@@ -1761,39 +1788,34 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
  * helper function to set a given page up to date if all the
  * extents in the tree for that page are up to date
  */
-static int check_page_uptodate(struct extent_io_tree *tree,
-			       struct page *page)
+static void check_page_uptodate(struct extent_io_tree *tree, struct page *page)
 {
 	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL))
 		SetPageUptodate(page);
-	return 0;
 }
 
 /*
  * helper function to unlock a page if all the extents in the tree
  * for that page are unlocked
  */
-static int check_page_locked(struct extent_io_tree *tree,
-			     struct page *page)
+static void check_page_locked(struct extent_io_tree *tree, struct page *page)
 {
 	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0, NULL))
 		unlock_page(page);
-	return 0;
 }
 
 /*
  * helper function to end page writeback if all the extents
  * in the tree for that page are done with writeback
  */
-static int check_page_writeback(struct extent_io_tree *tree,
-			     struct page *page)
+static void check_page_writeback(struct extent_io_tree *tree,
+				 struct page *page)
 {
 	end_page_writeback(page);
-	return 0;
 }
 
 /*
@@ -1912,6 +1934,26 @@ int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start,
 	return 0;
 }
 
+int repair_eb_io_failure(struct btrfs_root *root, struct extent_buffer *eb,
+			 int mirror_num)
+{
+	struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
+	u64 start = eb->start;
+	unsigned long i, num_pages = num_extent_pages(eb->start, eb->len);
+	int ret = 0;
+
+	for (i = 0; i < num_pages; i++) {
+		struct page *p = extent_buffer_page(eb, i);
+		ret = repair_io_failure(map_tree, start, PAGE_CACHE_SIZE,
+					start, p, mirror_num);
+		if (ret)
+			break;
+		start += PAGE_CACHE_SIZE;
+	}
+
+	return ret;
+}
+
 /*
  * each time an IO finishes, we do a fast check in the IO failure tree
  * to see if we need to process or clean up an io_failure_record
@@ -2141,6 +2183,10 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page,
 	}
 
 	bio = bio_alloc(GFP_NOFS, 1);
+	if (!bio) {
+		free_io_failure(inode, failrec, 0);
+		return -EIO;
+	}
 	bio->bi_private = state;
 	bio->bi_end_io = failed_bio->bi_end_io;
 	bio->bi_sector = failrec->logical >> 9;
@@ -2258,6 +2304,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 	u64 start;
 	u64 end;
 	int whole_page;
+	int mirror;
 	int ret;
 
 	if (err)
@@ -2296,17 +2343,22 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 		}
 		spin_unlock(&tree->lock);
 
+		mirror = (int)(unsigned long)bio->bi_bdev;
 		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
 			ret = tree->ops->readpage_end_io_hook(page, start, end,
-							      state);
+							      state, mirror);
 			if (ret)
 				uptodate = 0;
 			else
 				clean_io_failure(start, page);
 		}
-		if (!uptodate) {
-			int failed_mirror;
-			failed_mirror = (int)(unsigned long)bio->bi_bdev;
+
+		if (!uptodate && tree->ops && tree->ops->readpage_io_failed_hook) {
+			ret = tree->ops->readpage_io_failed_hook(page, mirror);
+			if (!ret && !err &&
+			    test_bit(BIO_UPTODATE, &bio->bi_flags))
+				uptodate = 1;
+		} else if (!uptodate) {
 			/*
 			 * The generic bio_readpage_error handles errors the
 			 * following way: If possible, new read requests are
@@ -2317,10 +2369,8 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 			 * can't handle the error it will return -EIO and we
 			 * remain responsible for that page.
 			 */
-			ret = bio_readpage_error(bio, page, start, end,
-							failed_mirror, NULL);
+			ret = bio_readpage_error(bio, page, start, end, mirror, NULL);
 			if (ret == 0) {
-error_handled:
 				uptodate =
 					test_bit(BIO_UPTODATE, &bio->bi_flags);
 				if (err)
@@ -2328,16 +2378,9 @@ error_handled:
 				uncache_state(&cached);
 				continue;
 			}
-			if (tree->ops && tree->ops->readpage_io_failed_hook) {
-				ret = tree->ops->readpage_io_failed_hook(
-							bio, page, start, end,
-							failed_mirror, state);
-				if (ret == 0)
-					goto error_handled;
-			}
 		}
 
-		if (uptodate) {
+		if (uptodate && tree->track_uptodate) {
 			set_extent_uptodate(tree, start, end, &cached,
 					    GFP_ATOMIC);
 		}
@@ -2386,8 +2429,12 @@ btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
 	return bio;
 }
 
-static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
-			  unsigned long bio_flags)
+/*
+ * Since writes are async, they will only return -ENOMEM.
+ * Reads can return the full range of I/O error conditions.
+ */
+static int __must_check submit_one_bio(int rw, struct bio *bio,
+				       int mirror_num, unsigned long bio_flags)
 {
 	int ret = 0;
 	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
@@ -2413,6 +2460,19 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
 	return ret;
 }
 
+static int merge_bio(struct extent_io_tree *tree, struct page *page,
+		     unsigned long offset, size_t size, struct bio *bio,
+		     unsigned long bio_flags)
+{
+	int ret = 0;
+	if (tree->ops && tree->ops->merge_bio_hook)
+		ret = tree->ops->merge_bio_hook(page, offset, size, bio,
+						bio_flags);
+	BUG_ON(ret < 0);
+	return ret;
+
+}
+
 static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			      struct page *page, sector_t sector,
 			      size_t size, unsigned long offset,
@@ -2441,12 +2501,12 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 				sector;
 
 		if (prev_bio_flags != bio_flags || !contig ||
-		    (tree->ops && tree->ops->merge_bio_hook &&
-		     tree->ops->merge_bio_hook(page, offset, page_size, bio,
-					       bio_flags)) ||
+		    merge_bio(tree, page, offset, page_size, bio, bio_flags) ||
 		    bio_add_page(bio, page, page_size, offset) < page_size) {
 			ret = submit_one_bio(rw, bio, mirror_num,
 					     prev_bio_flags);
+			if (ret < 0)
+				return ret;
 			bio = NULL;
 		} else {
 			return 0;
@@ -2473,25 +2533,31 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 	return ret;
 }
 
-void set_page_extent_mapped(struct page *page)
+void attach_extent_buffer_page(struct extent_buffer *eb, struct page *page)
 {
 	if (!PagePrivate(page)) {
 		SetPagePrivate(page);
 		page_cache_get(page);
-		set_page_private(page, EXTENT_PAGE_PRIVATE);
+		set_page_private(page, (unsigned long)eb);
+	} else {
+		WARN_ON(page->private != (unsigned long)eb);
 	}
 }
 
-static void set_page_extent_head(struct page *page, unsigned long len)
+void set_page_extent_mapped(struct page *page)
 {
-	WARN_ON(!PagePrivate(page));
-	set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2);
+	if (!PagePrivate(page)) {
+		SetPagePrivate(page);
+		page_cache_get(page);
+		set_page_private(page, EXTENT_PAGE_PRIVATE);
+	}
 }
 
 /*
  * basic readpage implementation.  Locked extent state structs are inserted
  * into the tree that are removed when the IO is done (by the end_io
  * handlers)
+ * XXX JDM: This needs looking at to ensure proper page locking
  */
 static int __extent_read_full_page(struct extent_io_tree *tree,
 				   struct page *page,
@@ -2531,11 +2597,11 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 
 	end = page_end;
 	while (1) {
-		lock_extent(tree, start, end, GFP_NOFS);
+		lock_extent(tree, start, end);
 		ordered = btrfs_lookup_ordered_extent(inode, start);
 		if (!ordered)
 			break;
-		unlock_extent(tree, start, end, GFP_NOFS);
+		unlock_extent(tree, start, end);
 		btrfs_start_ordered_extent(inode, ordered, 1);
 		btrfs_put_ordered_extent(ordered);
 	}
@@ -2546,10 +2612,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 
 		if (zero_offset) {
 			iosize = PAGE_CACHE_SIZE - zero_offset;
-			userpage = kmap_atomic(page, KM_USER0);
+			userpage = kmap_atomic(page);
 			memset(userpage + zero_offset, 0, iosize);
 			flush_dcache_page(page);
-			kunmap_atomic(userpage, KM_USER0);
+			kunmap_atomic(userpage);
 		}
 	}
 	while (cur <= end) {
@@ -2558,10 +2624,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			struct extent_state *cached = NULL;
 
 			iosize = PAGE_CACHE_SIZE - pg_offset;
-			userpage = kmap_atomic(page, KM_USER0);
+			userpage = kmap_atomic(page);
 			memset(userpage + pg_offset, 0, iosize);
 			flush_dcache_page(page);
-			kunmap_atomic(userpage, KM_USER0);
+			kunmap_atomic(userpage);
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
 					    &cached, GFP_NOFS);
 			unlock_extent_cached(tree, cur, cur + iosize - 1,
@@ -2572,7 +2638,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 				end - cur + 1, 0);
 		if (IS_ERR_OR_NULL(em)) {
 			SetPageError(page);
-			unlock_extent(tree, cur, end, GFP_NOFS);
+			unlock_extent(tree, cur, end);
 			break;
 		}
 		extent_offset = cur - em->start;
@@ -2607,10 +2673,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			char *userpage;
 			struct extent_state *cached = NULL;
 
-			userpage = kmap_atomic(page, KM_USER0);
+			userpage = kmap_atomic(page);
 			memset(userpage + pg_offset, 0, iosize);
 			flush_dcache_page(page);
-			kunmap_atomic(userpage, KM_USER0);
+			kunmap_atomic(userpage);
 
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
 					    &cached, GFP_NOFS);
@@ -2624,7 +2690,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		if (test_range_bit(tree, cur, cur_end,
 				   EXTENT_UPTODATE, 1, NULL)) {
 			check_page_uptodate(tree, page);
-			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			unlock_extent(tree, cur, cur + iosize - 1);
 			cur = cur + iosize;
 			pg_offset += iosize;
 			continue;
@@ -2634,7 +2700,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		 */
 		if (block_start == EXTENT_MAP_INLINE) {
 			SetPageError(page);
-			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			unlock_extent(tree, cur, cur + iosize - 1);
 			cur = cur + iosize;
 			pg_offset += iosize;
 			continue;
@@ -2654,6 +2720,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 					 end_bio_extent_readpage, mirror_num,
 					 *bio_flags,
 					 this_bio_flag);
+			BUG_ON(ret == -ENOMEM);
 			nr++;
 			*bio_flags = this_bio_flag;
 		}
@@ -2756,10 +2823,10 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 	if (page->index == end_index) {
 		char *userpage;
 
-		userpage = kmap_atomic(page, KM_USER0);
+		userpage = kmap_atomic(page);
 		memset(userpage + pg_offset, 0,
 		       PAGE_CACHE_SIZE - pg_offset);
-		kunmap_atomic(userpage, KM_USER0);
+		kunmap_atomic(userpage);
 		flush_dcache_page(page);
 	}
 	pg_offset = 0;
@@ -2795,7 +2862,11 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 						       delalloc_end,
 						       &page_started,
 						       &nr_written);
-			BUG_ON(ret);
+			/* File system has been set read-only */
+			if (ret) {
+				SetPageError(page);
+				goto done;
+			}
 			/*
 			 * delalloc_end is already one less than the total
 			 * length, so we don't subtract one from
@@ -2968,6 +3039,275 @@ done_unlocked:
 	return 0;
 }
 
+static int eb_wait(void *word)
+{
+	io_schedule();
+	return 0;
+}
+
+static void wait_on_extent_buffer_writeback(struct extent_buffer *eb)
+{
+	wait_on_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK, eb_wait,
+		    TASK_UNINTERRUPTIBLE);
+}
+
+static int lock_extent_buffer_for_io(struct extent_buffer *eb,
+				     struct btrfs_fs_info *fs_info,
+				     struct extent_page_data *epd)
+{
+	unsigned long i, num_pages;
+	int flush = 0;
+	int ret = 0;
+
+	if (!btrfs_try_tree_write_lock(eb)) {
+		flush = 1;
+		flush_write_bio(epd);
+		btrfs_tree_lock(eb);
+	}
+
+	if (test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags)) {
+		btrfs_tree_unlock(eb);
+		if (!epd->sync_io)
+			return 0;
+		if (!flush) {
+			flush_write_bio(epd);
+			flush = 1;
+		}
+		while (1) {
+			wait_on_extent_buffer_writeback(eb);
+			btrfs_tree_lock(eb);
+			if (!test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags))
+				break;
+			btrfs_tree_unlock(eb);
+		}
+	}
+
+	if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
+		set_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags);
+		btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
+		spin_lock(&fs_info->delalloc_lock);
+		if (fs_info->dirty_metadata_bytes >= eb->len)
+			fs_info->dirty_metadata_bytes -= eb->len;
+		else
+			WARN_ON(1);
+		spin_unlock(&fs_info->delalloc_lock);
+		ret = 1;
+	}
+
+	btrfs_tree_unlock(eb);
+
+	if (!ret)
+		return ret;
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = 0; i < num_pages; i++) {
+		struct page *p = extent_buffer_page(eb, i);
+
+		if (!trylock_page(p)) {
+			if (!flush) {
+				flush_write_bio(epd);
+				flush = 1;
+			}
+			lock_page(p);
+		}
+	}
+
+	return ret;
+}
+
+static void end_extent_buffer_writeback(struct extent_buffer *eb)
+{
+	clear_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags);
+	smp_mb__after_clear_bit();
+	wake_up_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK);
+}
+
+static void end_bio_extent_buffer_writepage(struct bio *bio, int err)
+{
+	int uptodate = err == 0;
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct extent_buffer *eb;
+	int done;
+
+	do {
+		struct page *page = bvec->bv_page;
+
+		bvec--;
+		eb = (struct extent_buffer *)page->private;
+		BUG_ON(!eb);
+		done = atomic_dec_and_test(&eb->io_pages);
+
+		if (!uptodate || test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) {
+			set_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
+			ClearPageUptodate(page);
+			SetPageError(page);
+		}
+
+		end_page_writeback(page);
+
+		if (!done)
+			continue;
+
+		end_extent_buffer_writeback(eb);
+	} while (bvec >= bio->bi_io_vec);
+
+	bio_put(bio);
+
+}
+
+static int write_one_eb(struct extent_buffer *eb,
+			struct btrfs_fs_info *fs_info,
+			struct writeback_control *wbc,
+			struct extent_page_data *epd)
+{
+	struct block_device *bdev = fs_info->fs_devices->latest_bdev;
+	u64 offset = eb->start;
+	unsigned long i, num_pages;
+	int rw = (epd->sync_io ? WRITE_SYNC : WRITE);
+	int ret;
+
+	clear_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
+	num_pages = num_extent_pages(eb->start, eb->len);
+	atomic_set(&eb->io_pages, num_pages);
+	for (i = 0; i < num_pages; i++) {
+		struct page *p = extent_buffer_page(eb, i);
+
+		clear_page_dirty_for_io(p);
+		set_page_writeback(p);
+		ret = submit_extent_page(rw, eb->tree, p, offset >> 9,
+					 PAGE_CACHE_SIZE, 0, bdev, &epd->bio,
+					 -1, end_bio_extent_buffer_writepage,
+					 0, 0, 0);
+		if (ret) {
+			set_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
+			SetPageError(p);
+			if (atomic_sub_and_test(num_pages - i, &eb->io_pages))
+				end_extent_buffer_writeback(eb);
+			ret = -EIO;
+			break;
+		}
+		offset += PAGE_CACHE_SIZE;
+		update_nr_written(p, wbc, 1);
+		unlock_page(p);
+	}
+
+	if (unlikely(ret)) {
+		for (; i < num_pages; i++) {
+			struct page *p = extent_buffer_page(eb, i);
+			unlock_page(p);
+		}
+	}
+
+	return ret;
+}
+
+int btree_write_cache_pages(struct address_space *mapping,
+				   struct writeback_control *wbc)
+{
+	struct extent_io_tree *tree = &BTRFS_I(mapping->host)->io_tree;
+	struct btrfs_fs_info *fs_info = BTRFS_I(mapping->host)->root->fs_info;
+	struct extent_buffer *eb, *prev_eb = NULL;
+	struct extent_page_data epd = {
+		.bio = NULL,
+		.tree = tree,
+		.extent_locked = 0,
+		.sync_io = wbc->sync_mode == WB_SYNC_ALL,
+	};
+	int ret = 0;
+	int done = 0;
+	int nr_to_write_done = 0;
+	struct pagevec pvec;
+	int nr_pages;
+	pgoff_t index;
+	pgoff_t end;		/* Inclusive */
+	int scanned = 0;
+	int tag;
+
+	pagevec_init(&pvec, 0);
+	if (wbc->range_cyclic) {
+		index = mapping->writeback_index; /* Start from prev offset */
+		end = -1;
+	} else {
+		index = wbc->range_start >> PAGE_CACHE_SHIFT;
+		end = wbc->range_end >> PAGE_CACHE_SHIFT;
+		scanned = 1;
+	}
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		tag = PAGECACHE_TAG_TOWRITE;
+	else
+		tag = PAGECACHE_TAG_DIRTY;
+retry:
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		tag_pages_for_writeback(mapping, index, end);
+	while (!done && !nr_to_write_done && (index <= end) &&
+	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
+			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
+		unsigned i;
+
+		scanned = 1;
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			if (!PagePrivate(page))
+				continue;
+
+			if (!wbc->range_cyclic && page->index > end) {
+				done = 1;
+				break;
+			}
+
+			eb = (struct extent_buffer *)page->private;
+			if (!eb) {
+				WARN_ON(1);
+				continue;
+			}
+
+			if (eb == prev_eb)
+				continue;
+
+			if (!atomic_inc_not_zero(&eb->refs)) {
+				WARN_ON(1);
+				continue;
+			}
+
+			prev_eb = eb;
+			ret = lock_extent_buffer_for_io(eb, fs_info, &epd);
+			if (!ret) {
+				free_extent_buffer(eb);
+				continue;
+			}
+
+			ret = write_one_eb(eb, fs_info, wbc, &epd);
+			if (ret) {
+				done = 1;
+				free_extent_buffer(eb);
+				break;
+			}
+			free_extent_buffer(eb);
+
+			/*
+			 * the filesystem may choose to bump up nr_to_write.
+			 * We have to make sure to honor the new nr_to_write
+			 * at any time
+			 */
+			nr_to_write_done = wbc->nr_to_write <= 0;
+		}
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+	if (!scanned && !done) {
+		/*
+		 * We hit the last page and there is more work to be done: wrap
+		 * back to the start of the file
+		 */
+		scanned = 1;
+		index = 0;
+		goto retry;
+	}
+	flush_write_bio(&epd);
+	return ret;
+}
+
 /**
  * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
  * @mapping: address space structure to write
@@ -3099,10 +3439,14 @@ retry:
 static void flush_epd_write_bio(struct extent_page_data *epd)
 {
 	if (epd->bio) {
+		int rw = WRITE;
+		int ret;
+
 		if (epd->sync_io)
-			submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
-		else
-			submit_one_bio(WRITE, epd->bio, 0, 0);
+			rw = WRITE_SYNC;
+
+		ret = submit_one_bio(rw, epd->bio, 0, 0);
+		BUG_ON(ret < 0); /* -ENOMEM */
 		epd->bio = NULL;
 	}
 }
@@ -3219,7 +3563,7 @@ int extent_readpages(struct extent_io_tree *tree,
 	}
 	BUG_ON(!list_empty(pages));
 	if (bio)
-		submit_one_bio(READ, bio, 0, bio_flags);
+		return submit_one_bio(READ, bio, 0, bio_flags);
 	return 0;
 }
 
@@ -3240,7 +3584,7 @@ int extent_invalidatepage(struct extent_io_tree *tree,
 	if (start > end)
 		return 0;
 
-	lock_extent_bits(tree, start, end, 0, &cached_state, GFP_NOFS);
+	lock_extent_bits(tree, start, end, 0, &cached_state);
 	wait_on_page_writeback(page);
 	clear_extent_bit(tree, start, end,
 			 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
@@ -3454,7 +3798,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	}
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0,
-			 &cached_state, GFP_NOFS);
+			 &cached_state);
 
 	em = get_extent_skip_holes(inode, start, last_for_get_extent,
 				   get_extent);
@@ -3548,26 +3892,7 @@ out:
 inline struct page *extent_buffer_page(struct extent_buffer *eb,
 					      unsigned long i)
 {
-	struct page *p;
-	struct address_space *mapping;
-
-	if (i == 0)
-		return eb->first_page;
-	i += eb->start >> PAGE_CACHE_SHIFT;
-	mapping = eb->first_page->mapping;
-	if (!mapping)
-		return NULL;
-
-	/*
-	 * extent_buffer_page is only called after pinning the page
-	 * by increasing the reference count.  So we know the page must
-	 * be in the radix tree.
-	 */
-	rcu_read_lock();
-	p = radix_tree_lookup(&mapping->page_tree, i);
-	rcu_read_unlock();
-
-	return p;
+	return eb->pages[i];
 }
 
 inline unsigned long num_extent_pages(u64 start, u64 len)
@@ -3576,6 +3901,19 @@ inline unsigned long num_extent_pages(u64 start, u64 len)
 		(start >> PAGE_CACHE_SHIFT);
 }
 
+static void __free_extent_buffer(struct extent_buffer *eb)
+{
+#if LEAK_DEBUG
+	unsigned long flags;
+	spin_lock_irqsave(&leak_lock, flags);
+	list_del(&eb->leak_list);
+	spin_unlock_irqrestore(&leak_lock, flags);
+#endif
+	if (eb->pages && eb->pages != eb->inline_pages)
+		kfree(eb->pages);
+	kmem_cache_free(extent_buffer_cache, eb);
+}
+
 static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 						   u64 start,
 						   unsigned long len,
@@ -3591,6 +3929,7 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 		return NULL;
 	eb->start = start;
 	eb->len = len;
+	eb->tree = tree;
 	rwlock_init(&eb->lock);
 	atomic_set(&eb->write_locks, 0);
 	atomic_set(&eb->read_locks, 0);
@@ -3607,20 +3946,32 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 	list_add(&eb->leak_list, &buffers);
 	spin_unlock_irqrestore(&leak_lock, flags);
 #endif
+	spin_lock_init(&eb->refs_lock);
 	atomic_set(&eb->refs, 1);
+	atomic_set(&eb->io_pages, 0);
+
+	if (len > MAX_INLINE_EXTENT_BUFFER_SIZE) {
+		struct page **pages;
+		int num_pages = (len + PAGE_CACHE_SIZE - 1) >>
+			PAGE_CACHE_SHIFT;
+		pages = kzalloc(num_pages, mask);
+		if (!pages) {
+			__free_extent_buffer(eb);
+			return NULL;
+		}
+		eb->pages = pages;
+	} else {
+		eb->pages = eb->inline_pages;
+	}
 
 	return eb;
 }
 
-static void __free_extent_buffer(struct extent_buffer *eb)
+static int extent_buffer_under_io(struct extent_buffer *eb)
 {
-#if LEAK_DEBUG
-	unsigned long flags;
-	spin_lock_irqsave(&leak_lock, flags);
-	list_del(&eb->leak_list);
-	spin_unlock_irqrestore(&leak_lock, flags);
-#endif
-	kmem_cache_free(extent_buffer_cache, eb);
+	return (atomic_read(&eb->io_pages) ||
+		test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) ||
+		test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
 }
 
 /*
@@ -3632,8 +3983,7 @@ static void btrfs_release_extent_buffer_page(struct extent_buffer *eb,
 	unsigned long index;
 	struct page *page;
 
-	if (!eb->first_page)
-		return;
+	BUG_ON(extent_buffer_under_io(eb));
 
 	index = num_extent_pages(eb->start, eb->len);
 	if (start_idx >= index)
@@ -3642,8 +3992,34 @@ static void btrfs_release_extent_buffer_page(struct extent_buffer *eb,
 	do {
 		index--;
 		page = extent_buffer_page(eb, index);
-		if (page)
+		if (page) {
+			spin_lock(&page->mapping->private_lock);
+			/*
+			 * We do this since we'll remove the pages after we've
+			 * removed the eb from the radix tree, so we could race
+			 * and have this page now attached to the new eb.  So
+			 * only clear page_private if it's still connected to
+			 * this eb.
+			 */
+			if (PagePrivate(page) &&
+			    page->private == (unsigned long)eb) {
+				BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
+				BUG_ON(PageDirty(page));
+				BUG_ON(PageWriteback(page));
+				/*
+				 * We need to make sure we haven't be attached
+				 * to a new eb.
+				 */
+				ClearPagePrivate(page);
+				set_page_private(page, 0);
+				/* One for the page private */
+				page_cache_release(page);
+			}
+			spin_unlock(&page->mapping->private_lock);
+
+			/* One for when we alloced the page */
 			page_cache_release(page);
+		}
 	} while (index != start_idx);
 }
 
@@ -3656,9 +4032,50 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
 	__free_extent_buffer(eb);
 }
 
+static void check_buffer_tree_ref(struct extent_buffer *eb)
+{
+	/* the ref bit is tricky.  We have to make sure it is set
+	 * if we have the buffer dirty.   Otherwise the
+	 * code to free a buffer can end up dropping a dirty
+	 * page
+	 *
+	 * Once the ref bit is set, it won't go away while the
+	 * buffer is dirty or in writeback, and it also won't
+	 * go away while we have the reference count on the
+	 * eb bumped.
+	 *
+	 * We can't just set the ref bit without bumping the
+	 * ref on the eb because free_extent_buffer might
+	 * see the ref bit and try to clear it.  If this happens
+	 * free_extent_buffer might end up dropping our original
+	 * ref by mistake and freeing the page before we are able
+	 * to add one more ref.
+	 *
+	 * So bump the ref count first, then set the bit.  If someone
+	 * beat us to it, drop the ref we added.
+	 */
+	if (!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) {
+		atomic_inc(&eb->refs);
+		if (test_and_set_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags))
+			atomic_dec(&eb->refs);
+	}
+}
+
+static void mark_extent_buffer_accessed(struct extent_buffer *eb)
+{
+	unsigned long num_pages, i;
+
+	check_buffer_tree_ref(eb);
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = 0; i < num_pages; i++) {
+		struct page *p = extent_buffer_page(eb, i);
+		mark_page_accessed(p);
+	}
+}
+
 struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
-					  u64 start, unsigned long len,
-					  struct page *page0)
+					  u64 start, unsigned long len)
 {
 	unsigned long num_pages = num_extent_pages(start, len);
 	unsigned long i;
@@ -3674,7 +4091,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
 	if (eb && atomic_inc_not_zero(&eb->refs)) {
 		rcu_read_unlock();
-		mark_page_accessed(eb->first_page);
+		mark_extent_buffer_accessed(eb);
 		return eb;
 	}
 	rcu_read_unlock();
@@ -3683,32 +4100,44 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 	if (!eb)
 		return NULL;
 
-	if (page0) {
-		eb->first_page = page0;
-		i = 1;
-		index++;
-		page_cache_get(page0);
-		mark_page_accessed(page0);
-		set_page_extent_mapped(page0);
-		set_page_extent_head(page0, len);
-		uptodate = PageUptodate(page0);
-	} else {
-		i = 0;
-	}
-	for (; i < num_pages; i++, index++) {
+	for (i = 0; i < num_pages; i++, index++) {
 		p = find_or_create_page(mapping, index, GFP_NOFS);
 		if (!p) {
 			WARN_ON(1);
 			goto free_eb;
 		}
-		set_page_extent_mapped(p);
-		mark_page_accessed(p);
-		if (i == 0) {
-			eb->first_page = p;
-			set_page_extent_head(p, len);
-		} else {
-			set_page_private(p, EXTENT_PAGE_PRIVATE);
+
+		spin_lock(&mapping->private_lock);
+		if (PagePrivate(p)) {
+			/*
+			 * We could have already allocated an eb for this page
+			 * and attached one so lets see if we can get a ref on
+			 * the existing eb, and if we can we know it's good and
+			 * we can just return that one, else we know we can just
+			 * overwrite page->private.
+			 */
+			exists = (struct extent_buffer *)p->private;
+			if (atomic_inc_not_zero(&exists->refs)) {
+				spin_unlock(&mapping->private_lock);
+				unlock_page(p);
+				page_cache_release(p);
+				mark_extent_buffer_accessed(exists);
+				goto free_eb;
+			}
+
+			/*
+			 * Do this so attach doesn't complain and we need to
+			 * drop the ref the old guy had.
+			 */
+			ClearPagePrivate(p);
+			WARN_ON(PageDirty(p));
+			page_cache_release(p);
 		}
+		attach_extent_buffer_page(eb, p);
+		spin_unlock(&mapping->private_lock);
+		WARN_ON(PageDirty(p));
+		mark_page_accessed(p);
+		eb->pages[i] = p;
 		if (!PageUptodate(p))
 			uptodate = 0;
 
@@ -3716,12 +4145,10 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 		 * see below about how we avoid a nasty race with release page
 		 * and why we unlock later
 		 */
-		if (i != 0)
-			unlock_page(p);
 	}
 	if (uptodate)
 		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
-
+again:
 	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
 	if (ret)
 		goto free_eb;
@@ -3731,14 +4158,21 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 	if (ret == -EEXIST) {
 		exists = radix_tree_lookup(&tree->buffer,
 						start >> PAGE_CACHE_SHIFT);
-		/* add one reference for the caller */
-		atomic_inc(&exists->refs);
+		if (!atomic_inc_not_zero(&exists->refs)) {
+			spin_unlock(&tree->buffer_lock);
+			radix_tree_preload_end();
+			exists = NULL;
+			goto again;
+		}
 		spin_unlock(&tree->buffer_lock);
 		radix_tree_preload_end();
+		mark_extent_buffer_accessed(exists);
 		goto free_eb;
 	}
 	/* add one reference for the tree */
-	atomic_inc(&eb->refs);
+	spin_lock(&eb->refs_lock);
+	check_buffer_tree_ref(eb);
+	spin_unlock(&eb->refs_lock);
 	spin_unlock(&tree->buffer_lock);
 	radix_tree_preload_end();
 
@@ -3751,18 +4185,22 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 	 * after the extent buffer is in the radix tree so
 	 * it doesn't get lost
 	 */
-	set_page_extent_mapped(eb->first_page);
-	set_page_extent_head(eb->first_page, eb->len);
-	if (!page0)
-		unlock_page(eb->first_page);
+	SetPageChecked(eb->pages[0]);
+	for (i = 1; i < num_pages; i++) {
+		p = extent_buffer_page(eb, i);
+		ClearPageChecked(p);
+		unlock_page(p);
+	}
+	unlock_page(eb->pages[0]);
 	return eb;
 
 free_eb:
-	if (eb->first_page && !page0)
-		unlock_page(eb->first_page);
+	for (i = 0; i < num_pages; i++) {
+		if (eb->pages[i])
+			unlock_page(eb->pages[i]);
+	}
 
-	if (!atomic_dec_and_test(&eb->refs))
-		return exists;
+	WARN_ON(!atomic_dec_and_test(&eb->refs));
 	btrfs_release_extent_buffer(eb);
 	return exists;
 }
@@ -3776,7 +4214,7 @@ struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
 	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
 	if (eb && atomic_inc_not_zero(&eb->refs)) {
 		rcu_read_unlock();
-		mark_page_accessed(eb->first_page);
+		mark_extent_buffer_accessed(eb);
 		return eb;
 	}
 	rcu_read_unlock();
@@ -3784,19 +4222,71 @@ struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
 	return NULL;
 }
 
+static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head)
+{
+	struct extent_buffer *eb =
+			container_of(head, struct extent_buffer, rcu_head);
+
+	__free_extent_buffer(eb);
+}
+
+/* Expects to have eb->eb_lock already held */
+static void release_extent_buffer(struct extent_buffer *eb, gfp_t mask)
+{
+	WARN_ON(atomic_read(&eb->refs) == 0);
+	if (atomic_dec_and_test(&eb->refs)) {
+		struct extent_io_tree *tree = eb->tree;
+
+		spin_unlock(&eb->refs_lock);
+
+		spin_lock(&tree->buffer_lock);
+		radix_tree_delete(&tree->buffer,
+				  eb->start >> PAGE_CACHE_SHIFT);
+		spin_unlock(&tree->buffer_lock);
+
+		/* Should be safe to release our pages at this point */
+		btrfs_release_extent_buffer_page(eb, 0);
+
+		call_rcu(&eb->rcu_head, btrfs_release_extent_buffer_rcu);
+		return;
+	}
+	spin_unlock(&eb->refs_lock);
+}
+
 void free_extent_buffer(struct extent_buffer *eb)
 {
 	if (!eb)
 		return;
 
-	if (!atomic_dec_and_test(&eb->refs))
+	spin_lock(&eb->refs_lock);
+	if (atomic_read(&eb->refs) == 2 &&
+	    test_bit(EXTENT_BUFFER_STALE, &eb->bflags) &&
+	    !extent_buffer_under_io(eb) &&
+	    test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags))
+		atomic_dec(&eb->refs);
+
+	/*
+	 * I know this is terrible, but it's temporary until we stop tracking
+	 * the uptodate bits and such for the extent buffers.
+	 */
+	release_extent_buffer(eb, GFP_ATOMIC);
+}
+
+void free_extent_buffer_stale(struct extent_buffer *eb)
+{
+	if (!eb)
 		return;
 
-	WARN_ON(1);
+	spin_lock(&eb->refs_lock);
+	set_bit(EXTENT_BUFFER_STALE, &eb->bflags);
+
+	if (atomic_read(&eb->refs) == 2 && !extent_buffer_under_io(eb) &&
+	    test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags))
+		atomic_dec(&eb->refs);
+	release_extent_buffer(eb, GFP_NOFS);
 }
 
-int clear_extent_buffer_dirty(struct extent_io_tree *tree,
-			      struct extent_buffer *eb)
+void clear_extent_buffer_dirty(struct extent_buffer *eb)
 {
 	unsigned long i;
 	unsigned long num_pages;
@@ -3812,10 +4302,6 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree,
 		lock_page(page);
 		WARN_ON(!PagePrivate(page));
 
-		set_page_extent_mapped(page);
-		if (i == 0)
-			set_page_extent_head(page, eb->len);
-
 		clear_page_dirty_for_io(page);
 		spin_lock_irq(&page->mapping->tree_lock);
 		if (!PageDirty(page)) {
@@ -3827,24 +4313,29 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree,
 		ClearPageError(page);
 		unlock_page(page);
 	}
-	return 0;
+	WARN_ON(atomic_read(&eb->refs) == 0);
 }
 
-int set_extent_buffer_dirty(struct extent_io_tree *tree,
-			     struct extent_buffer *eb)
+int set_extent_buffer_dirty(struct extent_buffer *eb)
 {
 	unsigned long i;
 	unsigned long num_pages;
 	int was_dirty = 0;
 
+	check_buffer_tree_ref(eb);
+
 	was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
+
 	num_pages = num_extent_pages(eb->start, eb->len);
+	WARN_ON(atomic_read(&eb->refs) == 0);
+	WARN_ON(!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags));
+
 	for (i = 0; i < num_pages; i++)
-		__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
+		set_page_dirty(extent_buffer_page(eb, i));
 	return was_dirty;
 }
 
-static int __eb_straddles_pages(u64 start, u64 len)
+static int range_straddles_pages(u64 start, u64 len)
 {
 	if (len < PAGE_CACHE_SIZE)
 		return 1;
@@ -3855,25 +4346,14 @@ static int __eb_straddles_pages(u64 start, u64 len)
 	return 0;
 }
 
-static int eb_straddles_pages(struct extent_buffer *eb)
-{
-	return __eb_straddles_pages(eb->start, eb->len);
-}
-
-int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
-				struct extent_buffer *eb,
-				struct extent_state **cached_state)
+int clear_extent_buffer_uptodate(struct extent_buffer *eb)
 {
 	unsigned long i;
 	struct page *page;
 	unsigned long num_pages;
 
-	num_pages = num_extent_pages(eb->start, eb->len);
 	clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
-
-	clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
-			      cached_state, GFP_NOFS);
-
+	num_pages = num_extent_pages(eb->start, eb->len);
 	for (i = 0; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
 		if (page)
@@ -3882,27 +4362,16 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
 	return 0;
 }
 
-int set_extent_buffer_uptodate(struct extent_io_tree *tree,
-				struct extent_buffer *eb)
+int set_extent_buffer_uptodate(struct extent_buffer *eb)
 {
 	unsigned long i;
 	struct page *page;
 	unsigned long num_pages;
 
+	set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 	num_pages = num_extent_pages(eb->start, eb->len);
-
-	if (eb_straddles_pages(eb)) {
-		set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
-				    NULL, GFP_NOFS);
-	}
 	for (i = 0; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
-		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
-		    ((i == num_pages - 1) &&
-		     ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
-			check_page_uptodate(tree, page);
-			continue;
-		}
 		SetPageUptodate(page);
 	}
 	return 0;
@@ -3917,7 +4386,7 @@ int extent_range_uptodate(struct extent_io_tree *tree,
 	int uptodate;
 	unsigned long index;
 
-	if (__eb_straddles_pages(start, end - start + 1)) {
+	if (range_straddles_pages(start, end - start + 1)) {
 		ret = test_range_bit(tree, start, end,
 				     EXTENT_UPTODATE, 1, NULL);
 		if (ret)
@@ -3939,35 +4408,9 @@ int extent_range_uptodate(struct extent_io_tree *tree,
 	return pg_uptodate;
 }
 
-int extent_buffer_uptodate(struct extent_io_tree *tree,
-			   struct extent_buffer *eb,
-			   struct extent_state *cached_state)
+int extent_buffer_uptodate(struct extent_buffer *eb)
 {
-	int ret = 0;
-	unsigned long num_pages;
-	unsigned long i;
-	struct page *page;
-	int pg_uptodate = 1;
-
-	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
-		return 1;
-
-	if (eb_straddles_pages(eb)) {
-		ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
-				   EXTENT_UPTODATE, 1, cached_state);
-		if (ret)
-			return ret;
-	}
-
-	num_pages = num_extent_pages(eb->start, eb->len);
-	for (i = 0; i < num_pages; i++) {
-		page = extent_buffer_page(eb, i);
-		if (!PageUptodate(page)) {
-			pg_uptodate = 0;
-			break;
-		}
-	}
-	return pg_uptodate;
+	return test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 }
 
 int read_extent_buffer_pages(struct extent_io_tree *tree,
@@ -3981,21 +4424,14 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 	int ret = 0;
 	int locked_pages = 0;
 	int all_uptodate = 1;
-	int inc_all_pages = 0;
 	unsigned long num_pages;
+	unsigned long num_reads = 0;
 	struct bio *bio = NULL;
 	unsigned long bio_flags = 0;
 
 	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 0;
 
-	if (eb_straddles_pages(eb)) {
-		if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
-				   EXTENT_UPTODATE, 1, NULL)) {
-			return 0;
-		}
-	}
-
 	if (start) {
 		WARN_ON(start < eb->start);
 		start_i = (start >> PAGE_CACHE_SHIFT) -
@@ -4014,8 +4450,10 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 			lock_page(page);
 		}
 		locked_pages++;
-		if (!PageUptodate(page))
+		if (!PageUptodate(page)) {
+			num_reads++;
 			all_uptodate = 0;
+		}
 	}
 	if (all_uptodate) {
 		if (start_i == 0)
@@ -4023,20 +4461,12 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 		goto unlock_exit;
 	}
 
+	clear_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
+	eb->read_mirror = 0;
+	atomic_set(&eb->io_pages, num_reads);
 	for (i = start_i; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
-
-		WARN_ON(!PagePrivate(page));
-
-		set_page_extent_mapped(page);
-		if (i == 0)
-			set_page_extent_head(page, eb->len);
-
-		if (inc_all_pages)
-			page_cache_get(page);
 		if (!PageUptodate(page)) {
-			if (start_i == 0)
-				inc_all_pages = 1;
 			ClearPageError(page);
 			err = __extent_read_full_page(tree, page,
 						      get_extent, &bio,
@@ -4048,8 +4478,11 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 		}
 	}
 
-	if (bio)
-		submit_one_bio(READ, bio, mirror_num, bio_flags);
+	if (bio) {
+		err = submit_one_bio(READ, bio, mirror_num, bio_flags);
+		if (err)
+			return err;
+	}
 
 	if (ret || wait != WAIT_COMPLETE)
 		return ret;
@@ -4061,8 +4494,6 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 			ret = -EIO;
 	}
 
-	if (!ret)
-		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 	return ret;
 
 unlock_exit:
@@ -4304,15 +4735,20 @@ static void copy_pages(struct page *dst_page, struct page *src_page,
 {
 	char *dst_kaddr = page_address(dst_page);
 	char *src_kaddr;
+	int must_memmove = 0;
 
 	if (dst_page != src_page) {
 		src_kaddr = page_address(src_page);
 	} else {
 		src_kaddr = dst_kaddr;
-		BUG_ON(areas_overlap(src_off, dst_off, len));
+		if (areas_overlap(src_off, dst_off, len))
+			must_memmove = 1;
 	}
 
-	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
+	if (must_memmove)
+		memmove(dst_kaddr + dst_off, src_kaddr + src_off, len);
+	else
+		memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
 }
 
 void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
@@ -4382,7 +4818,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 		       "len %lu len %lu\n", dst_offset, len, dst->len);
 		BUG_ON(1);
 	}
-	if (!areas_overlap(src_offset, dst_offset, len)) {
+	if (dst_offset < src_offset) {
 		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
 		return;
 	}
@@ -4408,47 +4844,48 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 	}
 }
 
-static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head)
-{
-	struct extent_buffer *eb =
-			container_of(head, struct extent_buffer, rcu_head);
-
-	btrfs_release_extent_buffer(eb);
-}
-
-int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page)
+int try_release_extent_buffer(struct page *page, gfp_t mask)
 {
-	u64 start = page_offset(page);
 	struct extent_buffer *eb;
-	int ret = 1;
 
-	spin_lock(&tree->buffer_lock);
-	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
-	if (!eb) {
-		spin_unlock(&tree->buffer_lock);
-		return ret;
+	/*
+	 * We need to make sure noboody is attaching this page to an eb right
+	 * now.
+	 */
+	spin_lock(&page->mapping->private_lock);
+	if (!PagePrivate(page)) {
+		spin_unlock(&page->mapping->private_lock);
+		return 1;
 	}
 
-	if (test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
-		ret = 0;
-		goto out;
-	}
+	eb = (struct extent_buffer *)page->private;
+	BUG_ON(!eb);
 
 	/*
-	 * set @eb->refs to 0 if it is already 1, and then release the @eb.
-	 * Or go back.
+	 * This is a little awful but should be ok, we need to make sure that
+	 * the eb doesn't disappear out from under us while we're looking at
+	 * this page.
 	 */
-	if (atomic_cmpxchg(&eb->refs, 1, 0) != 1) {
-		ret = 0;
-		goto out;
+	spin_lock(&eb->refs_lock);
+	if (atomic_read(&eb->refs) != 1 || extent_buffer_under_io(eb)) {
+		spin_unlock(&eb->refs_lock);
+		spin_unlock(&page->mapping->private_lock);
+		return 0;
 	}
+	spin_unlock(&page->mapping->private_lock);
 
-	radix_tree_delete(&tree->buffer, start >> PAGE_CACHE_SHIFT);
-out:
-	spin_unlock(&tree->buffer_lock);
+	if ((mask & GFP_NOFS) == GFP_NOFS)
+		mask = GFP_NOFS;
 
-	/* at this point we can safely release the extent buffer */
-	if (atomic_read(&eb->refs) == 0)
-		call_rcu(&eb->rcu_head, btrfs_release_extent_buffer_rcu);
-	return ret;
+	/*
+	 * If tree ref isn't set then we know the ref on this eb is a real ref,
+	 * so just return, this page will likely be freed soon anyway.
+	 */
+	if (!test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) {
+		spin_unlock(&eb->refs_lock);
+		return 0;
+	}
+	release_extent_buffer(eb, mask);
+
+	return 1;
 }
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index cecc3518c12..b516c3b8dec 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -35,6 +35,10 @@
 #define EXTENT_BUFFER_DIRTY 2
 #define EXTENT_BUFFER_CORRUPT 3
 #define EXTENT_BUFFER_READAHEAD 4	/* this got triggered by readahead */
+#define EXTENT_BUFFER_TREE_REF 5
+#define EXTENT_BUFFER_STALE 6
+#define EXTENT_BUFFER_WRITEBACK 7
+#define EXTENT_BUFFER_IOERR 8
 
 /* these are flags for extent_clear_unlock_delalloc */
 #define EXTENT_CLEAR_UNLOCK_PAGE 0x1
@@ -54,6 +58,7 @@
 #define EXTENT_PAGE_PRIVATE_FIRST_PAGE 3
 
 struct extent_state;
+struct btrfs_root;
 
 typedef	int (extent_submit_bio_hook_t)(struct inode *inode, int rw,
 				       struct bio *bio, int mirror_num,
@@ -69,14 +74,12 @@ struct extent_io_ops {
 			      size_t size, struct bio *bio,
 			      unsigned long bio_flags);
 	int (*readpage_io_hook)(struct page *page, u64 start, u64 end);
-	int (*readpage_io_failed_hook)(struct bio *bio, struct page *page,
-				       u64 start, u64 end, int failed_mirror,
-				       struct extent_state *state);
+	int (*readpage_io_failed_hook)(struct page *page, int failed_mirror);
 	int (*writepage_io_failed_hook)(struct bio *bio, struct page *page,
 					u64 start, u64 end,
 				       struct extent_state *state);
 	int (*readpage_end_io_hook)(struct page *page, u64 start, u64 end,
-				    struct extent_state *state);
+				    struct extent_state *state, int mirror);
 	int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end,
 				      struct extent_state *state, int uptodate);
 	void (*set_bit_hook)(struct inode *inode, struct extent_state *state,
@@ -97,6 +100,7 @@ struct extent_io_tree {
 	struct radix_tree_root buffer;
 	struct address_space *mapping;
 	u64 dirty_bytes;
+	int track_uptodate;
 	spinlock_t lock;
 	spinlock_t buffer_lock;
 	struct extent_io_ops *ops;
@@ -119,16 +123,21 @@ struct extent_state {
 	struct list_head leak_list;
 };
 
+#define INLINE_EXTENT_BUFFER_PAGES 16
+#define MAX_INLINE_EXTENT_BUFFER_SIZE (INLINE_EXTENT_BUFFER_PAGES * PAGE_CACHE_SIZE)
 struct extent_buffer {
 	u64 start;
 	unsigned long len;
 	unsigned long map_start;
 	unsigned long map_len;
-	struct page *first_page;
 	unsigned long bflags;
+	struct extent_io_tree *tree;
+	spinlock_t refs_lock;
+	atomic_t refs;
+	atomic_t io_pages;
+	int read_mirror;
 	struct list_head leak_list;
 	struct rcu_head rcu_head;
-	atomic_t refs;
 	pid_t lock_owner;
 
 	/* count of read lock holders on the extent buffer */
@@ -152,6 +161,9 @@ struct extent_buffer {
 	 * to unlock
 	 */
 	wait_queue_head_t read_lock_wq;
+	wait_queue_head_t lock_wq;
+	struct page *inline_pages[INLINE_EXTENT_BUFFER_PAGES];
+	struct page **pages;
 };
 
 static inline void extent_set_compress_type(unsigned long *bio_flags,
@@ -178,18 +190,17 @@ void extent_io_tree_init(struct extent_io_tree *tree,
 int try_release_extent_mapping(struct extent_map_tree *map,
 			       struct extent_io_tree *tree, struct page *page,
 			       gfp_t mask);
-int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page);
+int try_release_extent_buffer(struct page *page, gfp_t mask);
 int try_release_extent_state(struct extent_map_tree *map,
 			     struct extent_io_tree *tree, struct page *page,
 			     gfp_t mask);
-int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask);
+int lock_extent(struct extent_io_tree *tree, u64 start, u64 end);
 int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
-		     int bits, struct extent_state **cached, gfp_t mask);
-int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask);
+		     int bits, struct extent_state **cached);
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end);
 int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
 			 struct extent_state **cached, gfp_t mask);
-int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
-		    gfp_t mask);
+int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end);
 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
 			  get_extent_t *get_extent, int mirror_num);
 int __init extent_io_init(void);
@@ -210,7 +221,7 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		    int bits, gfp_t mask);
 int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
-		   int bits, int exclusive_bits, u64 *failed_start,
+		   int bits, u64 *failed_start,
 		   struct extent_state **cached_state, gfp_t mask);
 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
 			struct extent_state **cached_state, gfp_t mask);
@@ -240,6 +251,8 @@ int extent_writepages(struct extent_io_tree *tree,
 		      struct address_space *mapping,
 		      get_extent_t *get_extent,
 		      struct writeback_control *wbc);
+int btree_write_cache_pages(struct address_space *mapping,
+			    struct writeback_control *wbc);
 int extent_readpages(struct extent_io_tree *tree,
 		     struct address_space *mapping,
 		     struct list_head *pages, unsigned nr_pages,
@@ -251,11 +264,11 @@ int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private);
 void set_page_extent_mapped(struct page *page);
 
 struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
-					  u64 start, unsigned long len,
-					  struct page *page0);
+					  u64 start, unsigned long len);
 struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
 					 u64 start, unsigned long len);
 void free_extent_buffer(struct extent_buffer *eb);
+void free_extent_buffer_stale(struct extent_buffer *eb);
 #define WAIT_NONE	0
 #define WAIT_COMPLETE	1
 #define WAIT_PAGE_LOCK	2
@@ -287,19 +300,12 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 			   unsigned long src_offset, unsigned long len);
 void memset_extent_buffer(struct extent_buffer *eb, char c,
 			  unsigned long start, unsigned long len);
-int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits);
-int clear_extent_buffer_dirty(struct extent_io_tree *tree,
-			      struct extent_buffer *eb);
-int set_extent_buffer_dirty(struct extent_io_tree *tree,
-			     struct extent_buffer *eb);
-int set_extent_buffer_uptodate(struct extent_io_tree *tree,
-			       struct extent_buffer *eb);
-int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
-				struct extent_buffer *eb,
-				struct extent_state **cached_state);
-int extent_buffer_uptodate(struct extent_io_tree *tree,
-			   struct extent_buffer *eb,
-			   struct extent_state *cached_state);
+void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits);
+void clear_extent_buffer_dirty(struct extent_buffer *eb);
+int set_extent_buffer_dirty(struct extent_buffer *eb);
+int set_extent_buffer_uptodate(struct extent_buffer *eb);
+int clear_extent_buffer_uptodate(struct extent_buffer *eb);
+int extent_buffer_uptodate(struct extent_buffer *eb);
 int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset,
 		      unsigned long min_len, char **map,
 		      unsigned long *map_start,
@@ -320,4 +326,6 @@ int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start,
 			u64 length, u64 logical, struct page *page,
 			int mirror_num);
 int end_extent_writepage(struct page *page, int err, u64 start, u64 end);
+int repair_eb_io_failure(struct btrfs_root *root, struct extent_buffer *eb,
+			 int mirror_num);
 #endif
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index c7fb3a4247d..5d158d32023 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -25,10 +25,12 @@
 #include "transaction.h"
 #include "print-tree.h"
 
-#define MAX_CSUM_ITEMS(r, size) ((((BTRFS_LEAF_DATA_SIZE(r) - \
+#define __MAX_CSUM_ITEMS(r, size) ((((BTRFS_LEAF_DATA_SIZE(r) - \
 				   sizeof(struct btrfs_item) * 2) / \
 				  size) - 1))
 
+#define MAX_CSUM_ITEMS(r, size) (min(__MAX_CSUM_ITEMS(r, size), PAGE_CACHE_SIZE))
+
 #define MAX_ORDERED_SUM_BYTES(r) ((PAGE_SIZE - \
 				   sizeof(struct btrfs_ordered_sum)) / \
 				   sizeof(struct btrfs_sector_sum) * \
@@ -59,7 +61,7 @@ int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
 				      sizeof(*item));
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret);
+	BUG_ON(ret); /* Can't happen */
 	leaf = path->nodes[0];
 	item = btrfs_item_ptr(leaf, path->slots[0],
 			      struct btrfs_file_extent_item);
@@ -284,6 +286,7 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 	struct btrfs_ordered_sum *sums;
 	struct btrfs_sector_sum *sector_sum;
 	struct btrfs_csum_item *item;
+	LIST_HEAD(tmplist);
 	unsigned long offset;
 	int ret;
 	size_t size;
@@ -358,7 +361,10 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 					MAX_ORDERED_SUM_BYTES(root));
 			sums = kzalloc(btrfs_ordered_sum_size(root, size),
 					GFP_NOFS);
-			BUG_ON(!sums);
+			if (!sums) {
+				ret = -ENOMEM;
+				goto fail;
+			}
 
 			sector_sum = sums->sums;
 			sums->bytenr = start;
@@ -380,12 +386,19 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 				offset += csum_size;
 				sector_sum++;
 			}
-			list_add_tail(&sums->list, list);
+			list_add_tail(&sums->list, &tmplist);
 		}
 		path->slots[0]++;
 	}
 	ret = 0;
 fail:
+	while (ret < 0 && !list_empty(&tmplist)) {
+		sums = list_entry(&tmplist, struct btrfs_ordered_sum, list);
+		list_del(&sums->list);
+		kfree(sums);
+	}
+	list_splice_tail(&tmplist, list);
+
 	btrfs_free_path(path);
 	return ret;
 }
@@ -420,7 +433,7 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 		offset = page_offset(bvec->bv_page) + bvec->bv_offset;
 
 	ordered = btrfs_lookup_ordered_extent(inode, offset);
-	BUG_ON(!ordered);
+	BUG_ON(!ordered); /* Logic error */
 	sums->bytenr = ordered->start;
 
 	while (bio_index < bio->bi_vcnt) {
@@ -439,21 +452,21 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 
 			sums = kzalloc(btrfs_ordered_sum_size(root, bytes_left),
 				       GFP_NOFS);
-			BUG_ON(!sums);
+			BUG_ON(!sums); /* -ENOMEM */
 			sector_sum = sums->sums;
 			sums->len = bytes_left;
 			ordered = btrfs_lookup_ordered_extent(inode, offset);
-			BUG_ON(!ordered);
+			BUG_ON(!ordered); /* Logic error */
 			sums->bytenr = ordered->start;
 		}
 
-		data = kmap_atomic(bvec->bv_page, KM_USER0);
+		data = kmap_atomic(bvec->bv_page);
 		sector_sum->sum = ~(u32)0;
 		sector_sum->sum = btrfs_csum_data(root,
 						  data + bvec->bv_offset,
 						  sector_sum->sum,
 						  bvec->bv_len);
-		kunmap_atomic(data, KM_USER0);
+		kunmap_atomic(data);
 		btrfs_csum_final(sector_sum->sum,
 				 (char *)&sector_sum->sum);
 		sector_sum->bytenr = disk_bytenr;
@@ -483,18 +496,17 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
  * This calls btrfs_truncate_item with the correct args based on the
  * overlap, and fixes up the key as required.
  */
-static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
-				      struct btrfs_root *root,
-				      struct btrfs_path *path,
-				      struct btrfs_key *key,
-				      u64 bytenr, u64 len)
+static noinline void truncate_one_csum(struct btrfs_trans_handle *trans,
+				       struct btrfs_root *root,
+				       struct btrfs_path *path,
+				       struct btrfs_key *key,
+				       u64 bytenr, u64 len)
 {
 	struct extent_buffer *leaf;
 	u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy);
 	u64 csum_end;
 	u64 end_byte = bytenr + len;
 	u32 blocksize_bits = root->fs_info->sb->s_blocksize_bits;
-	int ret;
 
 	leaf = path->nodes[0];
 	csum_end = btrfs_item_size_nr(leaf, path->slots[0]) / csum_size;
@@ -510,7 +522,7 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
 		 */
 		u32 new_size = (bytenr - key->offset) >> blocksize_bits;
 		new_size *= csum_size;
-		ret = btrfs_truncate_item(trans, root, path, new_size, 1);
+		btrfs_truncate_item(trans, root, path, new_size, 1);
 	} else if (key->offset >= bytenr && csum_end > end_byte &&
 		   end_byte > key->offset) {
 		/*
@@ -522,15 +534,13 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
 		u32 new_size = (csum_end - end_byte) >> blocksize_bits;
 		new_size *= csum_size;
 
-		ret = btrfs_truncate_item(trans, root, path, new_size, 0);
+		btrfs_truncate_item(trans, root, path, new_size, 0);
 
 		key->offset = end_byte;
-		ret = btrfs_set_item_key_safe(trans, root, path, key);
-		BUG_ON(ret);
+		btrfs_set_item_key_safe(trans, root, path, key);
 	} else {
 		BUG();
 	}
-	return 0;
 }
 
 /*
@@ -635,13 +645,14 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 			 * item changed size or key
 			 */
 			ret = btrfs_split_item(trans, root, path, &key, offset);
-			BUG_ON(ret && ret != -EAGAIN);
+			if (ret && ret != -EAGAIN) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto out;
+			}
 
 			key.offset = end_byte - 1;
 		} else {
-			ret = truncate_one_csum(trans, root, path,
-						&key, bytenr, len);
-			BUG_ON(ret);
+			truncate_one_csum(trans, root, path, &key, bytenr, len);
 			if (key.offset < bytenr)
 				break;
 		}
@@ -772,7 +783,7 @@ again:
 		if (diff != csum_size)
 			goto insert;
 
-		ret = btrfs_extend_item(trans, root, path, diff);
+		btrfs_extend_item(trans, root, path, diff);
 		goto csum;
 	}
 
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index e8d06b6b919..53bf2d764bb 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -452,7 +452,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			split = alloc_extent_map();
 		if (!split2)
 			split2 = alloc_extent_map();
-		BUG_ON(!split || !split2);
+		BUG_ON(!split || !split2); /* -ENOMEM */
 
 		write_lock(&em_tree->lock);
 		em = lookup_extent_mapping(em_tree, start, len);
@@ -494,7 +494,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			split->flags = flags;
 			split->compress_type = em->compress_type;
 			ret = add_extent_mapping(em_tree, split);
-			BUG_ON(ret);
+			BUG_ON(ret); /* Logic error */
 			free_extent_map(split);
 			split = split2;
 			split2 = NULL;
@@ -520,7 +520,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			}
 
 			ret = add_extent_mapping(em_tree, split);
-			BUG_ON(ret);
+			BUG_ON(ret); /* Logic error */
 			free_extent_map(split);
 			split = NULL;
 		}
@@ -567,6 +567,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 	int extent_type;
 	int recow;
 	int ret;
+	int modify_tree = -1;
 
 	if (drop_cache)
 		btrfs_drop_extent_cache(inode, start, end - 1, 0);
@@ -575,10 +576,13 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 	if (!path)
 		return -ENOMEM;
 
+	if (start >= BTRFS_I(inode)->disk_i_size)
+		modify_tree = 0;
+
 	while (1) {
 		recow = 0;
 		ret = btrfs_lookup_file_extent(trans, root, path, ino,
-					       search_start, -1);
+					       search_start, modify_tree);
 		if (ret < 0)
 			break;
 		if (ret > 0 && path->slots[0] > 0 && search_start == start) {
@@ -634,7 +638,8 @@ next_slot:
 		}
 
 		search_start = max(key.offset, start);
-		if (recow) {
+		if (recow || !modify_tree) {
+			modify_tree = -1;
 			btrfs_release_path(path);
 			continue;
 		}
@@ -679,7 +684,7 @@ next_slot:
 						root->root_key.objectid,
 						new_key.objectid,
 						start - extent_offset, 0);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 				*hint_byte = disk_bytenr;
 			}
 			key.offset = start;
@@ -754,7 +759,7 @@ next_slot:
 						root->root_key.objectid,
 						key.objectid, key.offset -
 						extent_offset, 0);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 				inode_sub_bytes(inode,
 						extent_end - key.offset);
 				*hint_byte = disk_bytenr;
@@ -770,7 +775,10 @@ next_slot:
 
 			ret = btrfs_del_items(trans, root, path, del_slot,
 					      del_nr);
-			BUG_ON(ret);
+			if (ret) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto out;
+			}
 
 			del_nr = 0;
 			del_slot = 0;
@@ -782,11 +790,13 @@ next_slot:
 		BUG_ON(1);
 	}
 
-	if (del_nr > 0) {
+	if (!ret && del_nr > 0) {
 		ret = btrfs_del_items(trans, root, path, del_slot, del_nr);
-		BUG_ON(ret);
+		if (ret)
+			btrfs_abort_transaction(trans, root, ret);
 	}
 
+out:
 	btrfs_free_path(path);
 	return ret;
 }
@@ -944,7 +954,10 @@ again:
 			btrfs_release_path(path);
 			goto again;
 		}
-		BUG_ON(ret < 0);
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out;
+		}
 
 		leaf = path->nodes[0];
 		fi = btrfs_item_ptr(leaf, path->slots[0] - 1,
@@ -963,7 +976,7 @@ again:
 		ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
 					   root->root_key.objectid,
 					   ino, orig_offset, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 
 		if (split == start) {
 			key.offset = start;
@@ -990,7 +1003,7 @@ again:
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
 					ino, orig_offset, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 	other_start = 0;
 	other_end = start;
@@ -1007,7 +1020,7 @@ again:
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
 					ino, orig_offset, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 	}
 	if (del_nr == 0) {
 		fi = btrfs_item_ptr(leaf, path->slots[0],
@@ -1025,7 +1038,10 @@ again:
 		btrfs_mark_buffer_dirty(leaf);
 
 		ret = btrfs_del_items(trans, root, path, del_slot, del_nr);
-		BUG_ON(ret);
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out;
+		}
 	}
 out:
 	btrfs_free_path(path);
@@ -1105,8 +1121,7 @@ again:
 	if (start_pos < inode->i_size) {
 		struct btrfs_ordered_extent *ordered;
 		lock_extent_bits(&BTRFS_I(inode)->io_tree,
-				 start_pos, last_pos - 1, 0, &cached_state,
-				 GFP_NOFS);
+				 start_pos, last_pos - 1, 0, &cached_state);
 		ordered = btrfs_lookup_first_ordered_extent(inode,
 							    last_pos - 1);
 		if (ordered &&
@@ -1638,7 +1653,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 		 * transaction
 		 */
 		lock_extent_bits(&BTRFS_I(inode)->io_tree, alloc_start,
-				 locked_end, 0, &cached_state, GFP_NOFS);
+				 locked_end, 0, &cached_state);
 		ordered = btrfs_lookup_first_ordered_extent(inode,
 							    alloc_end - 1);
 		if (ordered &&
@@ -1667,7 +1682,13 @@ static long btrfs_fallocate(struct file *file, int mode,
 
 		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
 				      alloc_end - cur_offset, 0);
-		BUG_ON(IS_ERR_OR_NULL(em));
+		if (IS_ERR_OR_NULL(em)) {
+			if (!em)
+				ret = -ENOMEM;
+			else
+				ret = PTR_ERR(em);
+			break;
+		}
 		last_byte = min(extent_map_end(em), alloc_end);
 		actual_end = min_t(u64, extent_map_end(em), offset + len);
 		last_byte = (last_byte + mask) & ~mask;
@@ -1737,7 +1758,7 @@ static int find_desired_extent(struct inode *inode, loff_t *offset, int origin)
 		return -ENXIO;
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, 0,
-			 &cached_state, GFP_NOFS);
+			 &cached_state);
 
 	/*
 	 * Delalloc is such a pain.  If we have a hole and we have pending
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 710ea380c7e..202008ec367 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -230,11 +230,13 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 
 	if (ret) {
 		trans->block_rsv = rsv;
-		WARN_ON(1);
+		btrfs_abort_transaction(trans, root, ret);
 		return ret;
 	}
 
 	ret = btrfs_update_inode(trans, root, inode);
+	if (ret)
+		btrfs_abort_transaction(trans, root, ret);
 	trans->block_rsv = rsv;
 
 	return ret;
@@ -746,13 +748,6 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 	u64 used = btrfs_block_group_used(&block_group->item);
 
 	/*
-	 * If we're unmounting then just return, since this does a search on the
-	 * normal root and not the commit root and we could deadlock.
-	 */
-	if (btrfs_fs_closing(fs_info))
-		return 0;
-
-	/*
 	 * If this block group has been marked to be cleared for one reason or
 	 * another then we can't trust the on disk cache, so just return.
 	 */
@@ -766,6 +761,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 	path = btrfs_alloc_path();
 	if (!path)
 		return 0;
+	path->search_commit_root = 1;
+	path->skip_locking = 1;
 
 	inode = lookup_free_space_inode(root, block_group, path);
 	if (IS_ERR(inode)) {
@@ -869,7 +866,7 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
 	io_ctl_prepare_pages(&io_ctl, inode, 0);
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
-			 0, &cached_state, GFP_NOFS);
+			 0, &cached_state);
 
 	node = rb_first(&ctl->free_space_offset);
 	if (!node && cluster) {
@@ -1068,7 +1065,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		spin_unlock(&block_group->lock);
 		ret = 0;
 #ifdef DEBUG
-		printk(KERN_ERR "btrfs: failed to write free space cace "
+		printk(KERN_ERR "btrfs: failed to write free space cache "
 		       "for block group %llu\n", block_group->key.objectid);
 #endif
 	}
@@ -1948,14 +1945,14 @@ again:
 		 */
 		ret = btrfs_add_free_space(block_group, old_start,
 					   offset - old_start);
-		WARN_ON(ret);
+		WARN_ON(ret); /* -ENOMEM */
 		goto out;
 	}
 
 	ret = remove_from_bitmap(ctl, info, &offset, &bytes);
 	if (ret == -EAGAIN)
 		goto again;
-	BUG_ON(ret);
+	BUG_ON(ret); /* logic error */
 out_lock:
 	spin_unlock(&ctl->tree_lock);
 out:
@@ -2346,7 +2343,7 @@ again:
 	rb_erase(&entry->offset_index, &ctl->free_space_offset);
 	ret = tree_insert_offset(&cluster->root, entry->offset,
 				 &entry->offset_index, 1);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -EEXIST; Logic error */
 
 	trace_btrfs_setup_cluster(block_group, cluster,
 				  total_found * block_group->sectorsize, 1);
@@ -2439,7 +2436,7 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 		ret = tree_insert_offset(&cluster->root, entry->offset,
 					 &entry->offset_index, 0);
 		total_size += entry->bytes;
-		BUG_ON(ret);
+		BUG_ON(ret); /* -EEXIST; Logic error */
 	} while (node && entry != last);
 
 	cluster->max_size = max_extent;
@@ -2830,6 +2827,7 @@ u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
 		int ret;
 
 		ret = search_bitmap(ctl, entry, &offset, &count);
+		/* Logic error; Should be empty if it can't find anything */
 		BUG_ON(ret);
 
 		ino = offset;
diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c
index baa74f3db69..a13cf1a96c7 100644
--- a/fs/btrfs/inode-item.c
+++ b/fs/btrfs/inode-item.c
@@ -19,6 +19,7 @@
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
+#include "print-tree.h"
 
 static int find_name_in_backref(struct btrfs_path *path, const char *name,
 			 int name_len, struct btrfs_inode_ref **ref_ret)
@@ -128,13 +129,14 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 	item_start = btrfs_item_ptr_offset(leaf, path->slots[0]);
 	memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
 			      item_size - (ptr + sub_item_len - item_start));
-	ret = btrfs_truncate_item(trans, root, path,
+	btrfs_truncate_item(trans, root, path,
 				  item_size - sub_item_len, 1);
 out:
 	btrfs_free_path(path);
 	return ret;
 }
 
+/* Will return 0, -ENOMEM, -EMLINK, or -EEXIST or anything from the CoW path */
 int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   const char *name, int name_len,
@@ -165,7 +167,7 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
 			goto out;
 
 		old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
-		ret = btrfs_extend_item(trans, root, path, ins_len);
+		btrfs_extend_item(trans, root, path, ins_len);
 		ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
 				     struct btrfs_inode_ref);
 		ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size);
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index ee15d88b33d..b1a1c929ba8 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -178,7 +178,7 @@ static void start_caching(struct btrfs_root *root)
 
 	tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n",
 			  root->root_key.objectid);
-	BUG_ON(IS_ERR(tsk));
+	BUG_ON(IS_ERR(tsk)); /* -ENOMEM */
 }
 
 int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid)
@@ -271,7 +271,7 @@ void btrfs_unpin_free_ino(struct btrfs_root *root)
 			break;
 
 		info = rb_entry(n, struct btrfs_free_space, offset_index);
-		BUG_ON(info->bitmap);
+		BUG_ON(info->bitmap); /* Logic error */
 
 		if (info->offset > root->cache_progress)
 			goto free;
@@ -439,17 +439,16 @@ int btrfs_save_ino_cache(struct btrfs_root *root,
 	if (ret)
 		goto out;
 	trace_btrfs_space_reservation(root->fs_info, "ino_cache",
-				      (u64)(unsigned long)trans,
-				      trans->bytes_reserved, 1);
+				      trans->transid, trans->bytes_reserved, 1);
 again:
 	inode = lookup_free_ino_inode(root, path);
-	if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) {
+	if (IS_ERR(inode) && (PTR_ERR(inode) != -ENOENT || retry)) {
 		ret = PTR_ERR(inode);
 		goto out_release;
 	}
 
 	if (IS_ERR(inode)) {
-		BUG_ON(retry);
+		BUG_ON(retry); /* Logic error */
 		retry = true;
 
 		ret = create_free_ino_inode(root, trans, path);
@@ -460,12 +459,17 @@ again:
 
 	BTRFS_I(inode)->generation = 0;
 	ret = btrfs_update_inode(trans, root, inode);
-	WARN_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out_put;
+	}
 
 	if (i_size_read(inode) > 0) {
 		ret = btrfs_truncate_free_space_cache(root, trans, path, inode);
-		if (ret)
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
 			goto out_put;
+		}
 	}
 
 	spin_lock(&root->cache_lock);
@@ -502,8 +506,7 @@ out_put:
 	iput(inode);
 out_release:
 	trace_btrfs_space_reservation(root->fs_info, "ino_cache",
-				      (u64)(unsigned long)trans,
-				      trans->bytes_reserved, 0);
+				      trans->transid, trans->bytes_reserved, 0);
 	btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved);
 out:
 	trans->block_rsv = rsv;
@@ -532,7 +535,7 @@ static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid)
 	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
 	if (ret < 0)
 		goto error;
-	BUG_ON(ret == 0);
+	BUG_ON(ret == 0); /* Corruption */
 	if (path->slots[0] > 0) {
 		slot = path->slots[0] - 1;
 		l = path->nodes[0];
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 892b34785cc..61b16c641ce 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -150,7 +150,6 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
 	inode_add_bytes(inode, size);
 	ret = btrfs_insert_empty_item(trans, root, path, &key,
 				      datasize);
-	BUG_ON(ret);
 	if (ret) {
 		err = ret;
 		goto fail;
@@ -173,9 +172,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
 			cur_size = min_t(unsigned long, compressed_size,
 				       PAGE_CACHE_SIZE);
 
-			kaddr = kmap_atomic(cpage, KM_USER0);
+			kaddr = kmap_atomic(cpage);
 			write_extent_buffer(leaf, kaddr, ptr, cur_size);
-			kunmap_atomic(kaddr, KM_USER0);
+			kunmap_atomic(kaddr);
 
 			i++;
 			ptr += cur_size;
@@ -187,10 +186,10 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
 		page = find_get_page(inode->i_mapping,
 				     start >> PAGE_CACHE_SHIFT);
 		btrfs_set_file_extent_compression(leaf, ei, 0);
-		kaddr = kmap_atomic(page, KM_USER0);
+		kaddr = kmap_atomic(page);
 		offset = start & (PAGE_CACHE_SIZE - 1);
 		write_extent_buffer(leaf, kaddr + offset, ptr, size);
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 		page_cache_release(page);
 	}
 	btrfs_mark_buffer_dirty(leaf);
@@ -206,9 +205,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
 	 * could end up racing with unlink.
 	 */
 	BTRFS_I(inode)->disk_i_size = inode->i_size;
-	btrfs_update_inode(trans, root, inode);
+	ret = btrfs_update_inode(trans, root, inode);
 
-	return 0;
+	return ret;
 fail:
 	btrfs_free_path(path);
 	return err;
@@ -250,14 +249,18 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_drop_extents(trans, inode, start, aligned_end,
 				 &hint_byte, 1);
-	BUG_ON(ret);
+	if (ret)
+		return ret;
 
 	if (isize > actual_end)
 		inline_len = min_t(u64, isize, actual_end);
 	ret = insert_inline_extent(trans, root, inode, start,
 				   inline_len, compressed_size,
 				   compress_type, compressed_pages);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		return ret;
+	}
 	btrfs_delalloc_release_metadata(inode, end + 1 - start);
 	btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
 	return 0;
@@ -293,7 +296,7 @@ static noinline int add_async_extent(struct async_cow *cow,
 	struct async_extent *async_extent;
 
 	async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS);
-	BUG_ON(!async_extent);
+	BUG_ON(!async_extent); /* -ENOMEM */
 	async_extent->start = start;
 	async_extent->ram_size = ram_size;
 	async_extent->compressed_size = compressed_size;
@@ -344,8 +347,9 @@ static noinline int compress_file_range(struct inode *inode,
 	int will_compress;
 	int compress_type = root->fs_info->compress_type;
 
-	/* if this is a small write inside eof, kick off a defragbot */
-	if (end <= BTRFS_I(inode)->disk_i_size && (end - start + 1) < 16 * 1024)
+	/* if this is a small write inside eof, kick off a defrag */
+	if ((end - start + 1) < 16 * 1024 &&
+	    (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size))
 		btrfs_add_inode_defrag(NULL, inode);
 
 	actual_end = min_t(u64, isize, end + 1);
@@ -422,10 +426,10 @@ again:
 			 * sending it down to disk
 			 */
 			if (offset) {
-				kaddr = kmap_atomic(page, KM_USER0);
+				kaddr = kmap_atomic(page);
 				memset(kaddr + offset, 0,
 				       PAGE_CACHE_SIZE - offset);
-				kunmap_atomic(kaddr, KM_USER0);
+				kunmap_atomic(kaddr);
 			}
 			will_compress = 1;
 		}
@@ -433,7 +437,11 @@ again:
 cont:
 	if (start == 0) {
 		trans = btrfs_join_transaction(root);
-		BUG_ON(IS_ERR(trans));
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+			trans = NULL;
+			goto cleanup_and_out;
+		}
 		trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 		/* lets try to make an inline extent */
@@ -450,11 +458,11 @@ cont:
 						    total_compressed,
 						    compress_type, pages);
 		}
-		if (ret == 0) {
+		if (ret <= 0) {
 			/*
-			 * inline extent creation worked, we don't need
-			 * to create any more async work items.  Unlock
-			 * and free up our temp pages.
+			 * inline extent creation worked or returned error,
+			 * we don't need to create any more async work items.
+			 * Unlock and free up our temp pages.
 			 */
 			extent_clear_unlock_delalloc(inode,
 			     &BTRFS_I(inode)->io_tree,
@@ -547,7 +555,7 @@ cleanup_and_bail_uncompressed:
 	}
 
 out:
-	return 0;
+	return ret;
 
 free_pages_out:
 	for (i = 0; i < nr_pages_ret; i++) {
@@ -557,6 +565,20 @@ free_pages_out:
 	kfree(pages);
 
 	goto out;
+
+cleanup_and_out:
+	extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
+				     start, end, NULL,
+				     EXTENT_CLEAR_UNLOCK_PAGE |
+				     EXTENT_CLEAR_DIRTY |
+				     EXTENT_CLEAR_DELALLOC |
+				     EXTENT_SET_WRITEBACK |
+				     EXTENT_END_WRITEBACK);
+	if (!trans || IS_ERR(trans))
+		btrfs_error(root->fs_info, ret, "Failed to join transaction");
+	else
+		btrfs_abort_transaction(trans, root, ret);
+	goto free_pages_out;
 }
 
 /*
@@ -597,7 +619,7 @@ retry:
 
 			lock_extent(io_tree, async_extent->start,
 					 async_extent->start +
-					 async_extent->ram_size - 1, GFP_NOFS);
+					 async_extent->ram_size - 1);
 
 			/* allocate blocks */
 			ret = cow_file_range(inode, async_cow->locked_page,
@@ -606,6 +628,8 @@ retry:
 					     async_extent->ram_size - 1,
 					     &page_started, &nr_written, 0);
 
+			/* JDM XXX */
+
 			/*
 			 * if page_started, cow_file_range inserted an
 			 * inline extent and took care of all the unlocking
@@ -625,18 +649,21 @@ retry:
 		}
 
 		lock_extent(io_tree, async_extent->start,
-			    async_extent->start + async_extent->ram_size - 1,
-			    GFP_NOFS);
+			    async_extent->start + async_extent->ram_size - 1);
 
 		trans = btrfs_join_transaction(root);
-		BUG_ON(IS_ERR(trans));
-		trans->block_rsv = &root->fs_info->delalloc_block_rsv;
-		ret = btrfs_reserve_extent(trans, root,
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+		} else {
+			trans->block_rsv = &root->fs_info->delalloc_block_rsv;
+			ret = btrfs_reserve_extent(trans, root,
 					   async_extent->compressed_size,
 					   async_extent->compressed_size,
-					   0, alloc_hint,
-					   (u64)-1, &ins, 1);
-		btrfs_end_transaction(trans, root);
+					   0, alloc_hint, &ins, 1);
+			if (ret)
+				btrfs_abort_transaction(trans, root, ret);
+			btrfs_end_transaction(trans, root);
+		}
 
 		if (ret) {
 			int i;
@@ -649,8 +676,10 @@ retry:
 			async_extent->pages = NULL;
 			unlock_extent(io_tree, async_extent->start,
 				      async_extent->start +
-				      async_extent->ram_size - 1, GFP_NOFS);
-			goto retry;
+				      async_extent->ram_size - 1);
+			if (ret == -ENOSPC)
+				goto retry;
+			goto out_free; /* JDM: Requeue? */
 		}
 
 		/*
@@ -662,7 +691,7 @@ retry:
 					async_extent->ram_size - 1, 0);
 
 		em = alloc_extent_map();
-		BUG_ON(!em);
+		BUG_ON(!em); /* -ENOMEM */
 		em->start = async_extent->start;
 		em->len = async_extent->ram_size;
 		em->orig_start = em->start;
@@ -694,7 +723,7 @@ retry:
 						ins.offset,
 						BTRFS_ORDERED_COMPRESSED,
 						async_extent->compress_type);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 
 		/*
 		 * clear dirty, set writeback and unlock the pages.
@@ -716,13 +745,17 @@ retry:
 				    ins.offset, async_extent->pages,
 				    async_extent->nr_pages);
 
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 		alloc_hint = ins.objectid + ins.offset;
 		kfree(async_extent);
 		cond_resched();
 	}
-
-	return 0;
+	ret = 0;
+out:
+	return ret;
+out_free:
+	kfree(async_extent);
+	goto out;
 }
 
 static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
@@ -791,7 +824,18 @@ static noinline int cow_file_range(struct inode *inode,
 
 	BUG_ON(btrfs_is_free_space_inode(root, inode));
 	trans = btrfs_join_transaction(root);
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans)) {
+		extent_clear_unlock_delalloc(inode,
+			     &BTRFS_I(inode)->io_tree,
+			     start, end, NULL,
+			     EXTENT_CLEAR_UNLOCK_PAGE |
+			     EXTENT_CLEAR_UNLOCK |
+			     EXTENT_CLEAR_DELALLOC |
+			     EXTENT_CLEAR_DIRTY |
+			     EXTENT_SET_WRITEBACK |
+			     EXTENT_END_WRITEBACK);
+		return PTR_ERR(trans);
+	}
 	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 	num_bytes = (end - start + blocksize) & ~(blocksize - 1);
@@ -800,7 +844,8 @@ static noinline int cow_file_range(struct inode *inode,
 	ret = 0;
 
 	/* if this is a small write inside eof, kick off defrag */
-	if (end <= BTRFS_I(inode)->disk_i_size && num_bytes < 64 * 1024)
+	if (num_bytes < 64 * 1024 &&
+	    (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size))
 		btrfs_add_inode_defrag(trans, inode);
 
 	if (start == 0) {
@@ -821,8 +866,10 @@ static noinline int cow_file_range(struct inode *inode,
 			*nr_written = *nr_written +
 			     (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
 			*page_started = 1;
-			ret = 0;
 			goto out;
+		} else if (ret < 0) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out_unlock;
 		}
 	}
 
@@ -838,11 +885,14 @@ static noinline int cow_file_range(struct inode *inode,
 		cur_alloc_size = disk_num_bytes;
 		ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
 					   root->sectorsize, 0, alloc_hint,
-					   (u64)-1, &ins, 1);
-		BUG_ON(ret);
+					   &ins, 1);
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out_unlock;
+		}
 
 		em = alloc_extent_map();
-		BUG_ON(!em);
+		BUG_ON(!em); /* -ENOMEM */
 		em->start = start;
 		em->orig_start = em->start;
 		ram_size = ins.offset;
@@ -868,13 +918,16 @@ static noinline int cow_file_range(struct inode *inode,
 		cur_alloc_size = ins.offset;
 		ret = btrfs_add_ordered_extent(inode, start, ins.objectid,
 					       ram_size, cur_alloc_size, 0);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 
 		if (root->root_key.objectid ==
 		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
 			ret = btrfs_reloc_clone_csums(inode, start,
 						      cur_alloc_size);
-			BUG_ON(ret);
+			if (ret) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto out_unlock;
+			}
 		}
 
 		if (disk_num_bytes < cur_alloc_size)
@@ -899,11 +952,23 @@ static noinline int cow_file_range(struct inode *inode,
 		alloc_hint = ins.objectid + ins.offset;
 		start += cur_alloc_size;
 	}
-out:
 	ret = 0;
+out:
 	btrfs_end_transaction(trans, root);
 
 	return ret;
+out_unlock:
+	extent_clear_unlock_delalloc(inode,
+		     &BTRFS_I(inode)->io_tree,
+		     start, end, NULL,
+		     EXTENT_CLEAR_UNLOCK_PAGE |
+		     EXTENT_CLEAR_UNLOCK |
+		     EXTENT_CLEAR_DELALLOC |
+		     EXTENT_CLEAR_DIRTY |
+		     EXTENT_SET_WRITEBACK |
+		     EXTENT_END_WRITEBACK);
+
+	goto out;
 }
 
 /*
@@ -969,7 +1034,7 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page,
 			 1, 0, NULL, GFP_NOFS);
 	while (start < end) {
 		async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS);
-		BUG_ON(!async_cow);
+		BUG_ON(!async_cow); /* -ENOMEM */
 		async_cow->inode = inode;
 		async_cow->root = root;
 		async_cow->locked_page = locked_page;
@@ -1060,7 +1125,7 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 	u64 disk_bytenr;
 	u64 num_bytes;
 	int extent_type;
-	int ret;
+	int ret, err;
 	int type;
 	int nocow;
 	int check_prev = 1;
@@ -1078,7 +1143,11 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 	else
 		trans = btrfs_join_transaction(root);
 
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans)) {
+		btrfs_free_path(path);
+		return PTR_ERR(trans);
+	}
+
 	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 	cow_start = (u64)-1;
@@ -1086,7 +1155,10 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 	while (1) {
 		ret = btrfs_lookup_file_extent(trans, root, path, ino,
 					       cur_offset, 0);
-		BUG_ON(ret < 0);
+		if (ret < 0) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto error;
+		}
 		if (ret > 0 && path->slots[0] > 0 && check_prev) {
 			leaf = path->nodes[0];
 			btrfs_item_key_to_cpu(leaf, &found_key,
@@ -1100,8 +1172,10 @@ next_slot:
 		leaf = path->nodes[0];
 		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
 			ret = btrfs_next_leaf(root, path);
-			if (ret < 0)
-				BUG_ON(1);
+			if (ret < 0) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto error;
+			}
 			if (ret > 0)
 				break;
 			leaf = path->nodes[0];
@@ -1189,7 +1263,10 @@ out_check:
 			ret = cow_file_range(inode, locked_page, cow_start,
 					found_key.offset - 1, page_started,
 					nr_written, 1);
-			BUG_ON(ret);
+			if (ret) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto error;
+			}
 			cow_start = (u64)-1;
 		}
 
@@ -1198,7 +1275,7 @@ out_check:
 			struct extent_map_tree *em_tree;
 			em_tree = &BTRFS_I(inode)->extent_tree;
 			em = alloc_extent_map();
-			BUG_ON(!em);
+			BUG_ON(!em); /* -ENOMEM */
 			em->start = cur_offset;
 			em->orig_start = em->start;
 			em->len = num_bytes;
@@ -1224,13 +1301,16 @@ out_check:
 
 		ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr,
 					       num_bytes, num_bytes, type);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM */
 
 		if (root->root_key.objectid ==
 		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
 			ret = btrfs_reloc_clone_csums(inode, cur_offset,
 						      num_bytes);
-			BUG_ON(ret);
+			if (ret) {
+				btrfs_abort_transaction(trans, root, ret);
+				goto error;
+			}
 		}
 
 		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
@@ -1249,18 +1329,23 @@ out_check:
 	if (cow_start != (u64)-1) {
 		ret = cow_file_range(inode, locked_page, cow_start, end,
 				     page_started, nr_written, 1);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto error;
+		}
 	}
 
+error:
 	if (nolock) {
-		ret = btrfs_end_transaction_nolock(trans, root);
-		BUG_ON(ret);
+		err = btrfs_end_transaction_nolock(trans, root);
 	} else {
-		ret = btrfs_end_transaction(trans, root);
-		BUG_ON(ret);
+		err = btrfs_end_transaction(trans, root);
 	}
+	if (!ret)
+		ret = err;
+
 	btrfs_free_path(path);
-	return 0;
+	return ret;
 }
 
 /*
@@ -1425,10 +1510,11 @@ int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
 	map_length = length;
 	ret = btrfs_map_block(map_tree, READ, logical,
 			      &map_length, NULL, 0);
-
+	/* Will always return 0 or 1 with map_multi == NULL */
+	BUG_ON(ret < 0);
 	if (map_length < length + size)
 		return 1;
-	return ret;
+	return 0;
 }
 
 /*
@@ -1448,7 +1534,7 @@ static int __btrfs_submit_bio_start(struct inode *inode, int rw,
 	int ret = 0;
 
 	ret = btrfs_csum_one_bio(root, inode, bio, 0, 0);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 	return 0;
 }
 
@@ -1479,14 +1565,16 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret = 0;
 	int skip_sum;
+	int metadata = 0;
 
 	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
 	if (btrfs_is_free_space_inode(root, inode))
-		ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2);
-	else
-		ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
-	BUG_ON(ret);
+		metadata = 2;
+
+	ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata);
+	if (ret)
+		return ret;
 
 	if (!(rw & REQ_WRITE)) {
 		if (bio_flags & EXTENT_BIO_COMPRESSED) {
@@ -1571,7 +1659,7 @@ again:
 	page_end = page_offset(page) + PAGE_CACHE_SIZE - 1;
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0,
-			 &cached_state, GFP_NOFS);
+			 &cached_state);
 
 	/* already ordered? We're done */
 	if (PagePrivate2(page))
@@ -1675,13 +1763,15 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 	 */
 	ret = btrfs_drop_extents(trans, inode, file_pos, file_pos + num_bytes,
 				 &hint, 0);
-	BUG_ON(ret);
+	if (ret)
+		goto out;
 
 	ins.objectid = btrfs_ino(inode);
 	ins.offset = file_pos;
 	ins.type = BTRFS_EXTENT_DATA_KEY;
 	ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi));
-	BUG_ON(ret);
+	if (ret)
+		goto out;
 	leaf = path->nodes[0];
 	fi = btrfs_item_ptr(leaf, path->slots[0],
 			    struct btrfs_file_extent_item);
@@ -1709,10 +1799,10 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 	ret = btrfs_alloc_reserved_file_extent(trans, root,
 					root->root_key.objectid,
 					btrfs_ino(inode), file_pos, &ins);
-	BUG_ON(ret);
+out:
 	btrfs_free_path(path);
 
-	return 0;
+	return ret;
 }
 
 /*
@@ -1740,35 +1830,41 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 					     end - start + 1);
 	if (!ret)
 		return 0;
-	BUG_ON(!ordered_extent);
+	BUG_ON(!ordered_extent); /* Logic error */
 
 	nolock = btrfs_is_free_space_inode(root, inode);
 
 	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
-		BUG_ON(!list_empty(&ordered_extent->list));
+		BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */
 		ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
 		if (!ret) {
 			if (nolock)
 				trans = btrfs_join_transaction_nolock(root);
 			else
 				trans = btrfs_join_transaction(root);
-			BUG_ON(IS_ERR(trans));
+			if (IS_ERR(trans))
+				return PTR_ERR(trans);
 			trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 			ret = btrfs_update_inode_fallback(trans, root, inode);
-			BUG_ON(ret);
+			if (ret) /* -ENOMEM or corruption */
+				btrfs_abort_transaction(trans, root, ret);
 		}
 		goto out;
 	}
 
 	lock_extent_bits(io_tree, ordered_extent->file_offset,
 			 ordered_extent->file_offset + ordered_extent->len - 1,
-			 0, &cached_state, GFP_NOFS);
+			 0, &cached_state);
 
 	if (nolock)
 		trans = btrfs_join_transaction_nolock(root);
 	else
 		trans = btrfs_join_transaction(root);
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		trans = NULL;
+		goto out_unlock;
+	}
 	trans->block_rsv = &root->fs_info->delalloc_block_rsv;
 
 	if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
@@ -1779,7 +1875,6 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 						ordered_extent->file_offset,
 						ordered_extent->file_offset +
 						ordered_extent->len);
-		BUG_ON(ret);
 	} else {
 		BUG_ON(root == root->fs_info->tree_root);
 		ret = insert_reserved_file_extent(trans, inode,
@@ -1793,11 +1888,14 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 		unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
 				   ordered_extent->file_offset,
 				   ordered_extent->len);
-		BUG_ON(ret);
 	}
 	unlock_extent_cached(io_tree, ordered_extent->file_offset,
 			     ordered_extent->file_offset +
 			     ordered_extent->len - 1, &cached_state, GFP_NOFS);
+	if (ret < 0) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out;
+	}
 
 	add_pending_csums(trans, inode, ordered_extent->file_offset,
 			  &ordered_extent->list);
@@ -1805,7 +1903,10 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 	ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
 	if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) {
 		ret = btrfs_update_inode_fallback(trans, root, inode);
-		BUG_ON(ret);
+		if (ret) { /* -ENOMEM or corruption */
+			btrfs_abort_transaction(trans, root, ret);
+			goto out;
+		}
 	}
 	ret = 0;
 out:
@@ -1824,6 +1925,11 @@ out:
 	btrfs_put_ordered_extent(ordered_extent);
 
 	return 0;
+out_unlock:
+	unlock_extent_cached(io_tree, ordered_extent->file_offset,
+			     ordered_extent->file_offset +
+			     ordered_extent->len - 1, &cached_state, GFP_NOFS);
+	goto out;
 }
 
 static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
@@ -1841,7 +1947,7 @@ static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
  * extent_io.c will try to find good copies for us.
  */
 static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
-			       struct extent_state *state)
+			       struct extent_state *state, int mirror)
 {
 	size_t offset = start - ((u64)page->index << PAGE_CACHE_SHIFT);
 	struct inode *inode = page->mapping->host;
@@ -1873,7 +1979,7 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 	} else {
 		ret = get_state_private(io_tree, start, &private);
 	}
-	kaddr = kmap_atomic(page, KM_USER0);
+	kaddr = kmap_atomic(page);
 	if (ret)
 		goto zeroit;
 
@@ -1882,7 +1988,7 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 	if (csum != private)
 		goto zeroit;
 
-	kunmap_atomic(kaddr, KM_USER0);
+	kunmap_atomic(kaddr);
 good:
 	return 0;
 
@@ -1894,7 +2000,7 @@ zeroit:
 		       (unsigned long long)private);
 	memset(kaddr + offset, 1, end - start + 1);
 	flush_dcache_page(page);
-	kunmap_atomic(kaddr, KM_USER0);
+	kunmap_atomic(kaddr);
 	if (private == 0)
 		return 0;
 	return -EIO;
@@ -1905,6 +2011,8 @@ struct delayed_iput {
 	struct inode *inode;
 };
 
+/* JDM: If this is fs-wide, why can't we add a pointer to
+ * btrfs_inode instead and avoid the allocation? */
 void btrfs_add_delayed_iput(struct inode *inode)
 {
 	struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
@@ -2051,20 +2159,27 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
 	/* grab metadata reservation from transaction handle */
 	if (reserve) {
 		ret = btrfs_orphan_reserve_metadata(trans, inode);
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */
 	}
 
 	/* insert an orphan item to track this unlinked/truncated file */
 	if (insert >= 1) {
 		ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode));
-		BUG_ON(ret && ret != -EEXIST);
+		if (ret && ret != -EEXIST) {
+			btrfs_abort_transaction(trans, root, ret);
+			return ret;
+		}
+		ret = 0;
 	}
 
 	/* insert an orphan item to track subvolume contains orphan files */
 	if (insert >= 2) {
 		ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
 					       root->root_key.objectid);
-		BUG_ON(ret);
+		if (ret && ret != -EEXIST) {
+			btrfs_abort_transaction(trans, root, ret);
+			return ret;
+		}
 	}
 	return 0;
 }
@@ -2094,7 +2209,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
 
 	if (trans && delete_item) {
 		ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode));
-		BUG_ON(ret);
+		BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */
 	}
 
 	if (release_rsv)
@@ -2228,7 +2343,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
 			}
 			ret = btrfs_del_orphan_item(trans, root,
 						    found_key.objectid);
-			BUG_ON(ret);
+			BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */
 			btrfs_end_transaction(trans, root);
 			continue;
 		}
@@ -2610,16 +2725,22 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 		printk(KERN_INFO "btrfs failed to delete reference to %.*s, "
 		       "inode %llu parent %llu\n", name_len, name,
 		       (unsigned long long)ino, (unsigned long long)dir_ino);
+		btrfs_abort_transaction(trans, root, ret);
 		goto err;
 	}
 
 	ret = btrfs_delete_delayed_dir_index(trans, root, dir, index);
-	if (ret)
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
 		goto err;
+	}
 
 	ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
 					 inode, dir_ino);
-	BUG_ON(ret != 0 && ret != -ENOENT);
+	if (ret != 0 && ret != -ENOENT) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto err;
+	}
 
 	ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
 					   dir, index);
@@ -2777,7 +2898,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 			err = ret;
 			goto out;
 		}
-		BUG_ON(ret == 0);
+		BUG_ON(ret == 0); /* Corruption */
 		if (check_path_shared(root, path))
 			goto out;
 		btrfs_release_path(path);
@@ -2810,7 +2931,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 		err = PTR_ERR(ref);
 		goto out;
 	}
-	BUG_ON(!ref);
+	BUG_ON(!ref); /* Logic error */
 	if (check_path_shared(root, path))
 		goto out;
 	index = btrfs_inode_ref_index(path->nodes[0], ref);
@@ -2917,23 +3038,42 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 
 	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
 				   name, name_len, -1);
-	BUG_ON(IS_ERR_OR_NULL(di));
+	if (IS_ERR_OR_NULL(di)) {
+		if (!di)
+			ret = -ENOENT;
+		else
+			ret = PTR_ERR(di);
+		goto out;
+	}
 
 	leaf = path->nodes[0];
 	btrfs_dir_item_key_to_cpu(leaf, di, &key);
 	WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid);
 	ret = btrfs_delete_one_dir_name(trans, root, path, di);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out;
+	}
 	btrfs_release_path(path);
 
 	ret = btrfs_del_root_ref(trans, root->fs_info->tree_root,
 				 objectid, root->root_key.objectid,
 				 dir_ino, &index, name, name_len);
 	if (ret < 0) {
-		BUG_ON(ret != -ENOENT);
+		if (ret != -ENOENT) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out;
+		}
 		di = btrfs_search_dir_index_item(root, path, dir_ino,
 						 name, name_len);
-		BUG_ON(IS_ERR_OR_NULL(di));
+		if (IS_ERR_OR_NULL(di)) {
+			if (!di)
+				ret = -ENOENT;
+			else
+				ret = PTR_ERR(di);
+			btrfs_abort_transaction(trans, root, ret);
+			goto out;
+		}
 
 		leaf = path->nodes[0];
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
@@ -2943,15 +3083,19 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 	btrfs_release_path(path);
 
 	ret = btrfs_delete_delayed_dir_index(trans, root, dir, index);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out;
+	}
 
 	btrfs_i_size_write(dir, dir->i_size - name_len * 2);
 	dir->i_mtime = dir->i_ctime = CURRENT_TIME;
 	ret = btrfs_update_inode(trans, root, dir);
-	BUG_ON(ret);
-
+	if (ret)
+		btrfs_abort_transaction(trans, root, ret);
+out:
 	btrfs_free_path(path);
-	return 0;
+	return ret;
 }
 
 static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
@@ -3161,8 +3305,8 @@ search_again:
 				}
 				size =
 				    btrfs_file_extent_calc_inline_size(size);
-				ret = btrfs_truncate_item(trans, root, path,
-							  size, 1);
+				btrfs_truncate_item(trans, root, path,
+						    size, 1);
 			} else if (root->ref_cows) {
 				inode_sub_bytes(inode, item_end + 1 -
 						found_key.offset);
@@ -3210,7 +3354,11 @@ delete:
 				ret = btrfs_del_items(trans, root, path,
 						pending_del_slot,
 						pending_del_nr);
-				BUG_ON(ret);
+				if (ret) {
+					btrfs_abort_transaction(trans,
+								root, ret);
+					goto error;
+				}
 				pending_del_nr = 0;
 			}
 			btrfs_release_path(path);
@@ -3223,8 +3371,10 @@ out:
 	if (pending_del_nr) {
 		ret = btrfs_del_items(trans, root, path, pending_del_slot,
 				      pending_del_nr);
-		BUG_ON(ret);
+		if (ret)
+			btrfs_abort_transaction(trans, root, ret);
 	}
+error:
 	btrfs_free_path(path);
 	return err;
 }
@@ -3282,8 +3432,7 @@ again:
 	}
 	wait_on_page_writeback(page);
 
-	lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
-			 GFP_NOFS);
+	lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state);
 	set_page_extent_mapped(page);
 
 	ordered = btrfs_lookup_ordered_extent(inode, page_start);
@@ -3359,7 +3508,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 		btrfs_wait_ordered_range(inode, hole_start,
 					 block_end - hole_start);
 		lock_extent_bits(io_tree, hole_start, block_end - 1, 0,
-				 &cached_state, GFP_NOFS);
+				 &cached_state);
 		ordered = btrfs_lookup_ordered_extent(inode, hole_start);
 		if (!ordered)
 			break;
@@ -3372,7 +3521,10 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 	while (1) {
 		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
 				block_end - cur_offset, 0);
-		BUG_ON(IS_ERR_OR_NULL(em));
+		if (IS_ERR(em)) {
+			err = PTR_ERR(em);
+			break;
+		}
 		last_byte = min(extent_map_end(em), block_end);
 		last_byte = (last_byte + mask) & ~mask;
 		if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {
@@ -3389,7 +3541,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 						 cur_offset + hole_size,
 						 &hint_byte, 1);
 			if (err) {
-				btrfs_update_inode(trans, root, inode);
+				btrfs_abort_transaction(trans, root, err);
 				btrfs_end_transaction(trans, root);
 				break;
 			}
@@ -3399,7 +3551,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 					0, hole_size, 0, hole_size,
 					0, 0, 0);
 			if (err) {
-				btrfs_update_inode(trans, root, inode);
+				btrfs_abort_transaction(trans, root, err);
 				btrfs_end_transaction(trans, root);
 				break;
 			}
@@ -3779,7 +3931,7 @@ static void inode_tree_del(struct inode *inode)
 	}
 }
 
-int btrfs_invalidate_inodes(struct btrfs_root *root)
+void btrfs_invalidate_inodes(struct btrfs_root *root)
 {
 	struct rb_node *node;
 	struct rb_node *prev;
@@ -3839,7 +3991,6 @@ again:
 		node = rb_next(node);
 	}
 	spin_unlock(&root->inode_lock);
-	return 0;
 }
 
 static int btrfs_init_locked_inode(struct inode *inode, void *p)
@@ -3918,7 +4069,7 @@ static struct inode *new_simple_dir(struct super_block *s,
 	BTRFS_I(inode)->dummy_inode = 1;
 
 	inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID;
-	inode->i_op = &simple_dir_inode_operations;
+	inode->i_op = &btrfs_dir_ro_inode_operations;
 	inode->i_fop = &simple_dir_operations;
 	inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
@@ -3989,14 +4140,18 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
 static int btrfs_dentry_delete(const struct dentry *dentry)
 {
 	struct btrfs_root *root;
+	struct inode *inode = dentry->d_inode;
 
-	if (!dentry->d_inode && !IS_ROOT(dentry))
-		dentry = dentry->d_parent;
+	if (!inode && !IS_ROOT(dentry))
+		inode = dentry->d_parent->d_inode;
 
-	if (dentry->d_inode) {
-		root = BTRFS_I(dentry->d_inode)->root;
+	if (inode) {
+		root = BTRFS_I(inode)->root;
 		if (btrfs_root_refs(&root->root_item) == 0)
 			return 1;
+
+		if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+			return 1;
 	}
 	return 0;
 }
@@ -4037,7 +4192,6 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 	struct btrfs_path *path;
 	struct list_head ins_list;
 	struct list_head del_list;
-	struct qstr q;
 	int ret;
 	struct extent_buffer *leaf;
 	int slot;
@@ -4128,7 +4282,6 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 
 		while (di_cur < di_total) {
 			struct btrfs_key location;
-			struct dentry *tmp;
 
 			if (verify_dir_item(root, leaf, di))
 				break;
@@ -4149,35 +4302,15 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 			d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)];
 			btrfs_dir_item_key_to_cpu(leaf, di, &location);
 
-			q.name = name_ptr;
-			q.len = name_len;
-			q.hash = full_name_hash(q.name, q.len);
-			tmp = d_lookup(filp->f_dentry, &q);
-			if (!tmp) {
-				struct btrfs_key *newkey;
-
-				newkey = kzalloc(sizeof(struct btrfs_key),
-						 GFP_NOFS);
-				if (!newkey)
-					goto no_dentry;
-				tmp = d_alloc(filp->f_dentry, &q);
-				if (!tmp) {
-					kfree(newkey);
-					dput(tmp);
-					goto no_dentry;
-				}
-				memcpy(newkey, &location,
-				       sizeof(struct btrfs_key));
-				tmp->d_fsdata = newkey;
-				tmp->d_flags |= DCACHE_NEED_LOOKUP;
-				d_rehash(tmp);
-				dput(tmp);
-			} else {
-				dput(tmp);
-			}
-no_dentry:
+
 			/* is this a reference to our own snapshot? If so
-			 * skip it
+			 * skip it.
+			 *
+			 * In contrast to old kernels, we insert the snapshot's
+			 * dir item and dir index after it has been created, so
+			 * we won't find a reference to our own snapshot. We
+			 * still keep the following code for backward
+			 * compatibility.
 			 */
 			if (location.type == BTRFS_ROOT_ITEM_KEY &&
 			    location.objectid == root->root_key.objectid) {
@@ -4581,18 +4714,26 @@ int btrfs_add_link(struct btrfs_trans_handle *trans,
 					     parent_ino, index);
 	}
 
-	if (ret == 0) {
-		ret = btrfs_insert_dir_item(trans, root, name, name_len,
-					    parent_inode, &key,
-					    btrfs_inode_type(inode), index);
-		if (ret)
-			goto fail_dir_item;
+	/* Nothing to clean up yet */
+	if (ret)
+		return ret;
 
-		btrfs_i_size_write(parent_inode, parent_inode->i_size +
-				   name_len * 2);
-		parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
-		ret = btrfs_update_inode(trans, root, parent_inode);
+	ret = btrfs_insert_dir_item(trans, root, name, name_len,
+				    parent_inode, &key,
+				    btrfs_inode_type(inode), index);
+	if (ret == -EEXIST)
+		goto fail_dir_item;
+	else if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		return ret;
 	}
+
+	btrfs_i_size_write(parent_inode, parent_inode->i_size +
+			   name_len * 2);
+	parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
+	ret = btrfs_update_inode(trans, root, parent_inode);
+	if (ret)
+		btrfs_abort_transaction(trans, root, ret);
 	return ret;
 
 fail_dir_item:
@@ -4806,7 +4947,8 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	} else {
 		struct dentry *parent = dentry->d_parent;
 		err = btrfs_update_inode(trans, root, inode);
-		BUG_ON(err);
+		if (err)
+			goto fail;
 		d_instantiate(dentry, inode);
 		btrfs_log_new_name(trans, inode, NULL, parent);
 	}
@@ -4937,12 +5079,12 @@ static noinline int uncompress_inline(struct btrfs_path *path,
 	ret = btrfs_decompress(compress_type, tmp, page,
 			       extent_offset, inline_size, max_size);
 	if (ret) {
-		char *kaddr = kmap_atomic(page, KM_USER0);
+		char *kaddr = kmap_atomic(page);
 		unsigned long copy_size = min_t(u64,
 				  PAGE_CACHE_SIZE - pg_offset,
 				  max_size - extent_offset);
 		memset(kaddr + pg_offset, 0, copy_size);
-		kunmap_atomic(kaddr, KM_USER0);
+		kunmap_atomic(kaddr);
 	}
 	kfree(tmp);
 	return 0;
@@ -5137,7 +5279,7 @@ again:
 				ret = uncompress_inline(path, inode, page,
 							pg_offset,
 							extent_offset, item);
-				BUG_ON(ret);
+				BUG_ON(ret); /* -ENOMEM */
 			} else {
 				map = kmap(page);
 				read_extent_buffer(leaf, map + pg_offset, ptr,
@@ -5252,6 +5394,7 @@ out:
 		free_extent_map(em);
 		return ERR_PTR(err);
 	}
+	BUG_ON(!em); /* Error is always set */
 	return em;
 }
 
@@ -5414,7 +5557,7 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
 
 	alloc_hint = get_extent_allocation_hint(inode, start, len);
 	ret = btrfs_reserve_extent(trans, root, len, root->sectorsize, 0,
-				   alloc_hint, (u64)-1, &ins, 1);
+				   alloc_hint, &ins, 1);
 	if (ret) {
 		em = ERR_PTR(ret);
 		goto out;
@@ -5602,7 +5745,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 		free_extent_map(em);
 		/* DIO will do one hole at a time, so just unlock a sector */
 		unlock_extent(&BTRFS_I(inode)->io_tree, start,
-			      start + root->sectorsize - 1, GFP_NOFS);
+			      start + root->sectorsize - 1);
 		return 0;
 	}
 
@@ -5719,11 +5862,11 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 			unsigned long flags;
 
 			local_irq_save(flags);
-			kaddr = kmap_atomic(page, KM_IRQ0);
+			kaddr = kmap_atomic(page);
 			csum = btrfs_csum_data(root, kaddr + bvec->bv_offset,
 					       csum, bvec->bv_len);
 			btrfs_csum_final(csum, (char *)&csum);
-			kunmap_atomic(kaddr, KM_IRQ0);
+			kunmap_atomic(kaddr);
 			local_irq_restore(flags);
 
 			flush_dcache_page(bvec->bv_page);
@@ -5743,7 +5886,7 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 	} while (bvec <= bvec_end);
 
 	unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset,
-		      dip->logical_offset + dip->bytes - 1, GFP_NOFS);
+		      dip->logical_offset + dip->bytes - 1);
 	bio->bi_private = dip->private;
 
 	kfree(dip->csums);
@@ -5794,7 +5937,7 @@ again:
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree, ordered->file_offset,
 			 ordered->file_offset + ordered->len - 1, 0,
-			 &cached_state, GFP_NOFS);
+			 &cached_state);
 
 	if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) {
 		ret = btrfs_mark_extent_written(trans, inode,
@@ -5868,7 +6011,7 @@ static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw,
 	int ret;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	ret = btrfs_csum_one_bio(root, inode, bio, offset, 1);
-	BUG_ON(ret);
+	BUG_ON(ret); /* -ENOMEM */
 	return 0;
 }
 
@@ -6209,7 +6352,7 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
 
 	while (1) {
 		lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend,
-				 0, &cached_state, GFP_NOFS);
+				 0, &cached_state);
 		/*
 		 * We're concerned with the entire range that we're going to be
 		 * doing DIO to, so we need to make sure theres no ordered
@@ -6233,7 +6376,7 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
 	if (writing) {
 		write_bits = EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING;
 		ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
-				     EXTENT_DELALLOC, 0, NULL, &cached_state,
+				     EXTENT_DELALLOC, NULL, &cached_state,
 				     GFP_NOFS);
 		if (ret) {
 			clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart,
@@ -6363,8 +6506,7 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
 		btrfs_releasepage(page, GFP_NOFS);
 		return;
 	}
-	lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
-			 GFP_NOFS);
+	lock_extent_bits(tree, page_start, page_end, 0, &cached_state);
 	ordered = btrfs_lookup_ordered_extent(page->mapping->host,
 					   page_offset(page));
 	if (ordered) {
@@ -6386,8 +6528,7 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
 		}
 		btrfs_put_ordered_extent(ordered);
 		cached_state = NULL;
-		lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
-				 GFP_NOFS);
+		lock_extent_bits(tree, page_start, page_end, 0, &cached_state);
 	}
 	clear_extent_bit(tree, page_start, page_end,
 		 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
@@ -6462,8 +6603,7 @@ again:
 	}
 	wait_on_page_writeback(page);
 
-	lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
-			 GFP_NOFS);
+	lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state);
 	set_page_extent_mapped(page);
 
 	/*
@@ -6737,10 +6877,9 @@ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
 	btrfs_i_size_write(inode, 0);
 
 	err = btrfs_update_inode(trans, new_root, inode);
-	BUG_ON(err);
 
 	iput(inode);
-	return 0;
+	return err;
 }
 
 struct inode *btrfs_alloc_inode(struct super_block *sb)
@@ -6783,6 +6922,8 @@ struct inode *btrfs_alloc_inode(struct super_block *sb)
 	extent_map_tree_init(&ei->extent_tree);
 	extent_io_tree_init(&ei->io_tree, &inode->i_data);
 	extent_io_tree_init(&ei->io_failure_tree, &inode->i_data);
+	ei->io_tree.track_uptodate = 1;
+	ei->io_failure_tree.track_uptodate = 1;
 	mutex_init(&ei->log_mutex);
 	mutex_init(&ei->delalloc_mutex);
 	btrfs_ordered_inode_tree_init(&ei->ordered_tree);
@@ -7072,7 +7213,10 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		if (!ret)
 			ret = btrfs_update_inode(trans, root, old_inode);
 	}
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out_fail;
+	}
 
 	if (new_inode) {
 		new_inode->i_ctime = CURRENT_TIME;
@@ -7090,11 +7234,14 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 						 new_dentry->d_name.name,
 						 new_dentry->d_name.len);
 		}
-		BUG_ON(ret);
-		if (new_inode->i_nlink == 0) {
+		if (!ret && new_inode->i_nlink == 0) {
 			ret = btrfs_orphan_add(trans, new_dentry->d_inode);
 			BUG_ON(ret);
 		}
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			goto out_fail;
+		}
 	}
 
 	fixup_inode_flags(new_dir, old_inode);
@@ -7102,7 +7249,10 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	ret = btrfs_add_link(trans, new_dir, old_inode,
 			     new_dentry->d_name.name,
 			     new_dentry->d_name.len, 0, index);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto out_fail;
+	}
 
 	if (old_ino != BTRFS_FIRST_FREE_OBJECTID) {
 		struct dentry *parent = new_dentry->d_parent;
@@ -7315,7 +7465,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 		}
 
 		ret = btrfs_reserve_extent(trans, root, num_bytes, min_size,
-					   0, *alloc_hint, (u64)-1, &ins, 1);
+					   0, *alloc_hint, &ins, 1);
 		if (ret) {
 			if (own_trans)
 				btrfs_end_transaction(trans, root);
@@ -7327,7 +7477,12 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 						  ins.offset, ins.offset,
 						  ins.offset, 0, 0, 0,
 						  BTRFS_FILE_EXTENT_PREALLOC);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			if (own_trans)
+				btrfs_end_transaction(trans, root);
+			break;
+		}
 		btrfs_drop_extent_cache(inode, cur_offset,
 					cur_offset + ins.offset -1, 0);
 
@@ -7349,7 +7504,13 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 		}
 
 		ret = btrfs_update_inode(trans, root, inode);
-		BUG_ON(ret);
+
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			if (own_trans)
+				btrfs_end_transaction(trans, root);
+			break;
+		}
 
 		if (own_trans)
 			btrfs_end_transaction(trans, root);
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 1b36f1932ef..14f8e1faa46 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -425,22 +425,37 @@ static noinline int create_subvol(struct btrfs_root *root,
 
 	key.offset = (u64)-1;
 	new_root = btrfs_read_fs_root_no_name(root->fs_info, &key);
-	BUG_ON(IS_ERR(new_root));
+	if (IS_ERR(new_root)) {
+		btrfs_abort_transaction(trans, root, PTR_ERR(new_root));
+		ret = PTR_ERR(new_root);
+		goto fail;
+	}
 
 	btrfs_record_root_in_trans(trans, new_root);
 
 	ret = btrfs_create_subvol_root(trans, new_root, new_dirid);
+	if (ret) {
+		/* We potentially lose an unused inode item here */
+		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
+
 	/*
 	 * insert the directory item
 	 */
 	ret = btrfs_set_inode_index(dir, &index);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
 
 	ret = btrfs_insert_dir_item(trans, root,
 				    name, namelen, dir, &key,
 				    BTRFS_FT_DIR, index);
-	if (ret)
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
 		goto fail;
+	}
 
 	btrfs_i_size_write(dir, dir->i_size + namelen * 2);
 	ret = btrfs_update_inode(trans, root, dir);
@@ -769,6 +784,31 @@ none:
 	return -ENOENT;
 }
 
+/*
+ * Validaty check of prev em and next em:
+ * 1) no prev/next em
+ * 2) prev/next em is an hole/inline extent
+ */
+static int check_adjacent_extents(struct inode *inode, struct extent_map *em)
+{
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_map *prev = NULL, *next = NULL;
+	int ret = 0;
+
+	read_lock(&em_tree->lock);
+	prev = lookup_extent_mapping(em_tree, em->start - 1, (u64)-1);
+	next = lookup_extent_mapping(em_tree, em->start + em->len, (u64)-1);
+	read_unlock(&em_tree->lock);
+
+	if ((!prev || prev->block_start >= EXTENT_MAP_LAST_BYTE) &&
+	    (!next || next->block_start >= EXTENT_MAP_LAST_BYTE))
+		ret = 1;
+	free_extent_map(prev);
+	free_extent_map(next);
+
+	return ret;
+}
+
 static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 			       int thresh, u64 *last_len, u64 *skip,
 			       u64 *defrag_end)
@@ -797,17 +837,25 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 
 	if (!em) {
 		/* get the big lock and read metadata off disk */
-		lock_extent(io_tree, start, start + len - 1, GFP_NOFS);
+		lock_extent(io_tree, start, start + len - 1);
 		em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
-		unlock_extent(io_tree, start, start + len - 1, GFP_NOFS);
+		unlock_extent(io_tree, start, start + len - 1);
 
 		if (IS_ERR(em))
 			return 0;
 	}
 
 	/* this will cover holes, and inline extents */
-	if (em->block_start >= EXTENT_MAP_LAST_BYTE)
+	if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
 		ret = 0;
+		goto out;
+	}
+
+	/* If we have nothing to merge with us, just skip. */
+	if (check_adjacent_extents(inode, em)) {
+		ret = 0;
+		goto out;
+	}
 
 	/*
 	 * we hit a real extent, if it is big don't bother defragging it again
@@ -815,6 +863,7 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 	if ((*last_len == 0 || *last_len >= thresh) && em->len >= thresh)
 		ret = 0;
 
+out:
 	/*
 	 * last_len ends up being a counter of how many bytes we've defragged.
 	 * every time we choose not to defrag an extent, we reset *last_len
@@ -856,6 +905,7 @@ static int cluster_pages_for_defrag(struct inode *inode,
 	u64 isize = i_size_read(inode);
 	u64 page_start;
 	u64 page_end;
+	u64 page_cnt;
 	int ret;
 	int i;
 	int i_done;
@@ -864,19 +914,21 @@ static int cluster_pages_for_defrag(struct inode *inode,
 	struct extent_io_tree *tree;
 	gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
 
-	if (isize == 0)
-		return 0;
 	file_end = (isize - 1) >> PAGE_CACHE_SHIFT;
+	if (!isize || start_index > file_end)
+		return 0;
+
+	page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1);
 
 	ret = btrfs_delalloc_reserve_space(inode,
-					   num_pages << PAGE_CACHE_SHIFT);
+					   page_cnt << PAGE_CACHE_SHIFT);
 	if (ret)
 		return ret;
 	i_done = 0;
 	tree = &BTRFS_I(inode)->io_tree;
 
 	/* step one, lock all the pages */
-	for (i = 0; i < num_pages; i++) {
+	for (i = 0; i < page_cnt; i++) {
 		struct page *page;
 again:
 		page = find_or_create_page(inode->i_mapping,
@@ -887,10 +939,10 @@ again:
 		page_start = page_offset(page);
 		page_end = page_start + PAGE_CACHE_SIZE - 1;
 		while (1) {
-			lock_extent(tree, page_start, page_end, GFP_NOFS);
+			lock_extent(tree, page_start, page_end);
 			ordered = btrfs_lookup_ordered_extent(inode,
 							      page_start);
-			unlock_extent(tree, page_start, page_end, GFP_NOFS);
+			unlock_extent(tree, page_start, page_end);
 			if (!ordered)
 				break;
 
@@ -898,6 +950,15 @@ again:
 			btrfs_start_ordered_extent(inode, ordered, 1);
 			btrfs_put_ordered_extent(ordered);
 			lock_page(page);
+			/*
+			 * we unlocked the page above, so we need check if
+			 * it was released or not.
+			 */
+			if (page->mapping != inode->i_mapping) {
+				unlock_page(page);
+				page_cache_release(page);
+				goto again;
+			}
 		}
 
 		if (!PageUptodate(page)) {
@@ -911,15 +972,6 @@ again:
 			}
 		}
 
-		isize = i_size_read(inode);
-		file_end = (isize - 1) >> PAGE_CACHE_SHIFT;
-		if (!isize || page->index > file_end) {
-			/* whoops, we blew past eof, skip this page */
-			unlock_page(page);
-			page_cache_release(page);
-			break;
-		}
-
 		if (page->mapping != inode->i_mapping) {
 			unlock_page(page);
 			page_cache_release(page);
@@ -946,19 +998,18 @@ again:
 	page_end = page_offset(pages[i_done - 1]) + PAGE_CACHE_SIZE;
 
 	lock_extent_bits(&BTRFS_I(inode)->io_tree,
-			 page_start, page_end - 1, 0, &cached_state,
-			 GFP_NOFS);
+			 page_start, page_end - 1, 0, &cached_state);
 	clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start,
 			  page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC |
 			  EXTENT_DO_ACCOUNTING, 0, 0, &cached_state,
 			  GFP_NOFS);
 
-	if (i_done != num_pages) {
+	if (i_done != page_cnt) {
 		spin_lock(&BTRFS_I(inode)->lock);
 		BTRFS_I(inode)->outstanding_extents++;
 		spin_unlock(&BTRFS_I(inode)->lock);
 		btrfs_delalloc_release_space(inode,
-				     (num_pages - i_done) << PAGE_CACHE_SHIFT);
+				     (page_cnt - i_done) << PAGE_CACHE_SHIFT);
 	}
 
 
@@ -983,7 +1034,7 @@ out:
 		unlock_page(pages[i]);
 		page_cache_release(pages[i]);
 	}
-	btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT);
+	btrfs_delalloc_release_space(inode, page_cnt << PAGE_CACHE_SHIFT);
 	return ret;
 
 }
@@ -1089,12 +1140,9 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 		if (!(inode->i_sb->s_flags & MS_ACTIVE))
 			break;
 
-		if (!newer_than &&
-		    !should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT,
-					PAGE_CACHE_SIZE,
-					extent_thresh,
-					&last_len, &skip,
-					&defrag_end)) {
+		if (!should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT,
+					 PAGE_CACHE_SIZE, extent_thresh,
+					 &last_len, &skip, &defrag_end)) {
 			unsigned long next;
 			/*
 			 * the should_defrag function tells us how much to skip
@@ -1123,17 +1171,24 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			ra_index += max_cluster;
 		}
 
+		mutex_lock(&inode->i_mutex);
 		ret = cluster_pages_for_defrag(inode, pages, i, cluster);
-		if (ret < 0)
+		if (ret < 0) {
+			mutex_unlock(&inode->i_mutex);
 			goto out_ra;
+		}
 
 		defrag_count += ret;
 		balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret);
+		mutex_unlock(&inode->i_mutex);
 
 		if (newer_than) {
 			if (newer_off == (u64)-1)
 				break;
 
+			if (ret > 0)
+				i += ret;
+
 			newer_off = max(newer_off + 1,
 					(u64)i << PAGE_CACHE_SHIFT);
 
@@ -1966,7 +2021,11 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 				dest->root_key.objectid,
 				dentry->d_name.name,
 				dentry->d_name.len);
-	BUG_ON(ret);
+	if (ret) {
+		err = ret;
+		btrfs_abort_transaction(trans, root, ret);
+		goto out_end_trans;
+	}
 
 	btrfs_record_root_in_trans(trans, dest);
 
@@ -1979,11 +2038,16 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 		ret = btrfs_insert_orphan_item(trans,
 					root->fs_info->tree_root,
 					dest->root_key.objectid);
-		BUG_ON(ret);
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
+			err = ret;
+			goto out_end_trans;
+		}
 	}
-
+out_end_trans:
 	ret = btrfs_end_transaction(trans, root);
-	BUG_ON(ret);
+	if (ret && !err)
+		err = ret;
 	inode->i_flags |= S_DEAD;
 out_up_write:
 	up_write(&root->fs_info->subvol_sem);
@@ -2198,7 +2262,10 @@ static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
 	di_args->bytes_used = dev->bytes_used;
 	di_args->total_bytes = dev->total_bytes;
 	memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid));
-	strncpy(di_args->path, dev->name, sizeof(di_args->path));
+	if (dev->name)
+		strncpy(di_args->path, dev->name, sizeof(di_args->path));
+	else
+		di_args->path[0] = '\0';
 
 out:
 	if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args)))
@@ -2326,13 +2393,13 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	   another, and lock file content */
 	while (1) {
 		struct btrfs_ordered_extent *ordered;
-		lock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+		lock_extent(&BTRFS_I(src)->io_tree, off, off+len);
 		ordered = btrfs_lookup_first_ordered_extent(src, off+len);
 		if (!ordered &&
 		    !test_range_bit(&BTRFS_I(src)->io_tree, off, off+len,
 				   EXTENT_DELALLOC, 0, NULL))
 			break;
-		unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+		unlock_extent(&BTRFS_I(src)->io_tree, off, off+len);
 		if (ordered)
 			btrfs_put_ordered_extent(ordered);
 		btrfs_wait_ordered_range(src, off, len);
@@ -2447,11 +2514,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 							 new_key.offset,
 							 new_key.offset + datal,
 							 &hint_byte, 1);
-				BUG_ON(ret);
+				if (ret) {
+					btrfs_abort_transaction(trans, root,
+								ret);
+					btrfs_end_transaction(trans, root);
+					goto out;
+				}
 
 				ret = btrfs_insert_empty_item(trans, root, path,
 							      &new_key, size);
-				BUG_ON(ret);
+				if (ret) {
+					btrfs_abort_transaction(trans, root,
+								ret);
+					btrfs_end_transaction(trans, root);
+					goto out;
+				}
 
 				leaf = path->nodes[0];
 				slot = path->slots[0];
@@ -2478,7 +2555,15 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 							btrfs_ino(inode),
 							new_key.offset - datao,
 							0);
-					BUG_ON(ret);
+					if (ret) {
+						btrfs_abort_transaction(trans,
+									root,
+									ret);
+						btrfs_end_transaction(trans,
+								      root);
+						goto out;
+
+					}
 				}
 			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
 				u64 skip = 0;
@@ -2503,11 +2588,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 							 new_key.offset,
 							 new_key.offset + datal,
 							 &hint_byte, 1);
-				BUG_ON(ret);
+				if (ret) {
+					btrfs_abort_transaction(trans, root,
+								ret);
+					btrfs_end_transaction(trans, root);
+					goto out;
+				}
 
 				ret = btrfs_insert_empty_item(trans, root, path,
 							      &new_key, size);
-				BUG_ON(ret);
+				if (ret) {
+					btrfs_abort_transaction(trans, root,
+								ret);
+					btrfs_end_transaction(trans, root);
+					goto out;
+				}
 
 				if (skip) {
 					u32 start =
@@ -2541,8 +2636,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 				btrfs_i_size_write(inode, endoff);
 
 			ret = btrfs_update_inode(trans, root, inode);
-			BUG_ON(ret);
-			btrfs_end_transaction(trans, root);
+			if (ret) {
+				btrfs_abort_transaction(trans, root, ret);
+				btrfs_end_transaction(trans, root);
+				goto out;
+			}
+			ret = btrfs_end_transaction(trans, root);
 		}
 next:
 		btrfs_release_path(path);
@@ -2551,7 +2650,7 @@ next:
 	ret = 0;
 out:
 	btrfs_release_path(path);
-	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len);
 out_unlock:
 	mutex_unlock(&src->i_mutex);
 	mutex_unlock(&inode->i_mutex);
diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h
index 4f69028a68c..086e6bdae1c 100644
--- a/fs/btrfs/ioctl.h
+++ b/fs/btrfs/ioctl.h
@@ -252,7 +252,7 @@ struct btrfs_data_container {
 
 struct btrfs_ioctl_ino_path_args {
 	__u64				inum;		/* in */
-	__u32				size;		/* in */
+	__u64				size;		/* in */
 	__u64				reserved[4];
 	/* struct btrfs_data_container	*fspath;	   out */
 	__u64				fspath;		/* out */
@@ -260,7 +260,7 @@ struct btrfs_ioctl_ino_path_args {
 
 struct btrfs_ioctl_logical_ino_args {
 	__u64				logical;	/* in */
-	__u32				size;		/* in */
+	__u64				size;		/* in */
 	__u64				reserved[4];
 	/* struct btrfs_data_container	*inodes;	out   */
 	__u64				inodes;
diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c
index 5e178d8f716..272f911203f 100644
--- a/fs/btrfs/locking.c
+++ b/fs/btrfs/locking.c
@@ -208,7 +208,7 @@ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb)
  * take a spinning write lock.  This will wait for both
  * blocking readers or writers
  */
-int btrfs_tree_lock(struct extent_buffer *eb)
+void btrfs_tree_lock(struct extent_buffer *eb)
 {
 again:
 	wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0);
@@ -230,13 +230,12 @@ again:
 	atomic_inc(&eb->spinning_writers);
 	atomic_inc(&eb->write_locks);
 	eb->lock_owner = current->pid;
-	return 0;
 }
 
 /*
  * drop a spinning or a blocking write lock.
  */
-int btrfs_tree_unlock(struct extent_buffer *eb)
+void btrfs_tree_unlock(struct extent_buffer *eb)
 {
 	int blockers = atomic_read(&eb->blocking_writers);
 
@@ -255,7 +254,6 @@ int btrfs_tree_unlock(struct extent_buffer *eb)
 		atomic_dec(&eb->spinning_writers);
 		write_unlock(&eb->lock);
 	}
-	return 0;
 }
 
 void btrfs_assert_tree_locked(struct extent_buffer *eb)
diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h
index 17247ddb81a..ca52681e5f4 100644
--- a/fs/btrfs/locking.h
+++ b/fs/btrfs/locking.h
@@ -24,8 +24,8 @@
 #define BTRFS_WRITE_LOCK_BLOCKING 3
 #define BTRFS_READ_LOCK_BLOCKING 4
 
-int btrfs_tree_lock(struct extent_buffer *eb);
-int btrfs_tree_unlock(struct extent_buffer *eb);
+void btrfs_tree_lock(struct extent_buffer *eb);
+void btrfs_tree_unlock(struct extent_buffer *eb);
 int btrfs_try_spin_lock(struct extent_buffer *eb);
 
 void btrfs_tree_read_lock(struct extent_buffer *eb);
diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c
index a178f5ebea7..743b86fa4fc 100644
--- a/fs/btrfs/lzo.c
+++ b/fs/btrfs/lzo.c
@@ -411,9 +411,9 @@ static int lzo_decompress(struct list_head *ws, unsigned char *data_in,
 
 	bytes = min_t(unsigned long, destlen, out_len - start_byte);
 
-	kaddr = kmap_atomic(dest_page, KM_USER0);
+	kaddr = kmap_atomic(dest_page);
 	memcpy(kaddr, workspace->buf + start_byte, bytes);
-	kunmap_atomic(kaddr, KM_USER0);
+	kunmap_atomic(kaddr);
 out:
 	return ret;
 }
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index a1c94042530..bbf6d0d9aeb 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -59,6 +59,14 @@ static struct rb_node *tree_insert(struct rb_root *root, u64 file_offset,
 	return NULL;
 }
 
+static void ordered_data_tree_panic(struct inode *inode, int errno,
+					       u64 offset)
+{
+	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
+	btrfs_panic(fs_info, errno, "Inconsistency in ordered tree at offset "
+		    "%llu\n", (unsigned long long)offset);
+}
+
 /*
  * look for a given offset in the tree, and if it can't be found return the
  * first lesser offset
@@ -207,7 +215,8 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 	spin_lock(&tree->lock);
 	node = tree_insert(&tree->tree, file_offset,
 			   &entry->rb_node);
-	BUG_ON(node);
+	if (node)
+		ordered_data_tree_panic(inode, -EEXIST, file_offset);
 	spin_unlock(&tree->lock);
 
 	spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
@@ -215,7 +224,6 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 		      &BTRFS_I(inode)->root->fs_info->ordered_extents);
 	spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
 
-	BUG_ON(node);
 	return 0;
 }
 
@@ -249,9 +257,9 @@ int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset,
  * when an ordered extent is finished.  If the list covers more than one
  * ordered extent, it is split across multiples.
  */
-int btrfs_add_ordered_sum(struct inode *inode,
-			  struct btrfs_ordered_extent *entry,
-			  struct btrfs_ordered_sum *sum)
+void btrfs_add_ordered_sum(struct inode *inode,
+			   struct btrfs_ordered_extent *entry,
+			   struct btrfs_ordered_sum *sum)
 {
 	struct btrfs_ordered_inode_tree *tree;
 
@@ -259,7 +267,6 @@ int btrfs_add_ordered_sum(struct inode *inode,
 	spin_lock(&tree->lock);
 	list_add_tail(&sum->list, &entry->list);
 	spin_unlock(&tree->lock);
-	return 0;
 }
 
 /*
@@ -384,7 +391,7 @@ out:
  * used to drop a reference on an ordered extent.  This will free
  * the extent if the last reference is dropped
  */
-int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
+void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
 {
 	struct list_head *cur;
 	struct btrfs_ordered_sum *sum;
@@ -400,7 +407,6 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
 		}
 		kfree(entry);
 	}
-	return 0;
 }
 
 /*
@@ -408,8 +414,8 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
  * and you must wake_up entry->wait.  You must hold the tree lock
  * while you call this function.
  */
-static int __btrfs_remove_ordered_extent(struct inode *inode,
-				struct btrfs_ordered_extent *entry)
+static void __btrfs_remove_ordered_extent(struct inode *inode,
+					  struct btrfs_ordered_extent *entry)
 {
 	struct btrfs_ordered_inode_tree *tree;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -436,35 +442,30 @@ static int __btrfs_remove_ordered_extent(struct inode *inode,
 		list_del_init(&BTRFS_I(inode)->ordered_operations);
 	}
 	spin_unlock(&root->fs_info->ordered_extent_lock);
-
-	return 0;
 }
 
 /*
  * remove an ordered extent from the tree.  No references are dropped
  * but any waiters are woken.
  */
-int btrfs_remove_ordered_extent(struct inode *inode,
-				struct btrfs_ordered_extent *entry)
+void btrfs_remove_ordered_extent(struct inode *inode,
+				 struct btrfs_ordered_extent *entry)
 {
 	struct btrfs_ordered_inode_tree *tree;
-	int ret;
 
 	tree = &BTRFS_I(inode)->ordered_tree;
 	spin_lock(&tree->lock);
-	ret = __btrfs_remove_ordered_extent(inode, entry);
+	__btrfs_remove_ordered_extent(inode, entry);
 	spin_unlock(&tree->lock);
 	wake_up(&entry->wait);
-
-	return ret;
 }
 
 /*
  * wait for all the ordered extents in a root.  This is done when balancing
  * space between drives.
  */
-int btrfs_wait_ordered_extents(struct btrfs_root *root,
-			       int nocow_only, int delay_iput)
+void btrfs_wait_ordered_extents(struct btrfs_root *root,
+				int nocow_only, int delay_iput)
 {
 	struct list_head splice;
 	struct list_head *cur;
@@ -512,7 +513,6 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root,
 		spin_lock(&root->fs_info->ordered_extent_lock);
 	}
 	spin_unlock(&root->fs_info->ordered_extent_lock);
-	return 0;
 }
 
 /*
@@ -525,7 +525,7 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root,
  * extra check to make sure the ordered operation list really is empty
  * before we return
  */
-int btrfs_run_ordered_operations(struct btrfs_root *root, int wait)
+void btrfs_run_ordered_operations(struct btrfs_root *root, int wait)
 {
 	struct btrfs_inode *btrfs_inode;
 	struct inode *inode;
@@ -573,8 +573,6 @@ again:
 
 	spin_unlock(&root->fs_info->ordered_extent_lock);
 	mutex_unlock(&root->fs_info->ordered_operations_mutex);
-
-	return 0;
 }
 
 /*
@@ -609,7 +607,7 @@ void btrfs_start_ordered_extent(struct inode *inode,
 /*
  * Used to wait on ordered extents across a large range of bytes.
  */
-int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len)
+void btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len)
 {
 	u64 end;
 	u64 orig_end;
@@ -664,7 +662,6 @@ again:
 		schedule_timeout(1);
 		goto again;
 	}
-	return 0;
 }
 
 /*
@@ -948,9 +945,8 @@ out:
  * If trans is not null, we'll do a friendly check for a transaction that
  * is already flushing things and force the IO down ourselves.
  */
-int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
-				struct btrfs_root *root,
-				struct inode *inode)
+void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root, struct inode *inode)
 {
 	u64 last_mod;
 
@@ -961,7 +957,7 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 	 * commit, we can safely return without doing anything
 	 */
 	if (last_mod < root->fs_info->last_trans_committed)
-		return 0;
+		return;
 
 	/*
 	 * the transaction is already committing.  Just start the IO and
@@ -969,7 +965,7 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 	 */
 	if (trans && root->fs_info->running_transaction->blocked) {
 		btrfs_wait_ordered_range(inode, 0, (u64)-1);
-		return 0;
+		return;
 	}
 
 	spin_lock(&root->fs_info->ordered_extent_lock);
@@ -978,6 +974,4 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 			      &root->fs_info->ordered_operations);
 	}
 	spin_unlock(&root->fs_info->ordered_extent_lock);
-
-	return 0;
 }
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
index ff1f69aa188..c355ad4dc1a 100644
--- a/fs/btrfs/ordered-data.h
+++ b/fs/btrfs/ordered-data.h
@@ -138,8 +138,8 @@ btrfs_ordered_inode_tree_init(struct btrfs_ordered_inode_tree *t)
 	t->last = NULL;
 }
 
-int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry);
-int btrfs_remove_ordered_extent(struct inode *inode,
+void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry);
+void btrfs_remove_ordered_extent(struct inode *inode,
 				struct btrfs_ordered_extent *entry);
 int btrfs_dec_test_ordered_pending(struct inode *inode,
 				   struct btrfs_ordered_extent **cached,
@@ -154,14 +154,14 @@ int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset,
 int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset,
 				      u64 start, u64 len, u64 disk_len,
 				      int type, int compress_type);
-int btrfs_add_ordered_sum(struct inode *inode,
-			  struct btrfs_ordered_extent *entry,
-			  struct btrfs_ordered_sum *sum);
+void btrfs_add_ordered_sum(struct inode *inode,
+			   struct btrfs_ordered_extent *entry,
+			   struct btrfs_ordered_sum *sum);
 struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode,
 							 u64 file_offset);
 void btrfs_start_ordered_extent(struct inode *inode,
 				struct btrfs_ordered_extent *entry, int wait);
-int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len);
+void btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len);
 struct btrfs_ordered_extent *
 btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset);
 struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
@@ -170,10 +170,10 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
 int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
 				struct btrfs_ordered_extent *ordered);
 int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum);
-int btrfs_run_ordered_operations(struct btrfs_root *root, int wait);
-int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
-				struct btrfs_root *root,
-				struct inode *inode);
-int btrfs_wait_ordered_extents(struct btrfs_root *root,
-			       int nocow_only, int delay_iput);
+void btrfs_run_ordered_operations(struct btrfs_root *root, int wait);
+void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct inode *inode);
+void btrfs_wait_ordered_extents(struct btrfs_root *root,
+				int nocow_only, int delay_iput);
 #endif
diff --git a/fs/btrfs/orphan.c b/fs/btrfs/orphan.c
index f8be250963a..24cad1695af 100644
--- a/fs/btrfs/orphan.c
+++ b/fs/btrfs/orphan.c
@@ -58,7 +58,7 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 	if (ret < 0)
 		goto out;
-	if (ret) {
+	if (ret) { /* JDM: Really? */
 		ret = -ENOENT;
 		goto out;
 	}
diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c
index 22db04550f6..ac5d0108588 100644
--- a/fs/btrfs/reada.c
+++ b/fs/btrfs/reada.c
@@ -54,7 +54,6 @@
  * than the 2 started one after another.
  */
 
-#define MAX_MIRRORS 2
 #define MAX_IN_FLIGHT 6
 
 struct reada_extctl {
@@ -71,7 +70,7 @@ struct reada_extent {
 	struct list_head	extctl;
 	struct kref		refcnt;
 	spinlock_t		lock;
-	struct reada_zone	*zones[MAX_MIRRORS];
+	struct reada_zone	*zones[BTRFS_MAX_MIRRORS];
 	int			nzones;
 	struct btrfs_device	*scheduled_for;
 };
@@ -84,7 +83,8 @@ struct reada_zone {
 	spinlock_t		lock;
 	int			locked;
 	struct btrfs_device	*device;
-	struct btrfs_device	*devs[MAX_MIRRORS]; /* full list, incl self */
+	struct btrfs_device	*devs[BTRFS_MAX_MIRRORS]; /* full list, incl
+							   * self */
 	int			ndevs;
 	struct kref		refcnt;
 };
@@ -250,14 +250,12 @@ static struct reada_zone *reada_find_zone(struct btrfs_fs_info *fs_info,
 					  struct btrfs_bio *bbio)
 {
 	int ret;
-	int looped = 0;
 	struct reada_zone *zone;
 	struct btrfs_block_group_cache *cache = NULL;
 	u64 start;
 	u64 end;
 	int i;
 
-again:
 	zone = NULL;
 	spin_lock(&fs_info->reada_lock);
 	ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone,
@@ -274,9 +272,6 @@ again:
 		spin_unlock(&fs_info->reada_lock);
 	}
 
-	if (looped)
-		return NULL;
-
 	cache = btrfs_lookup_block_group(fs_info, logical);
 	if (!cache)
 		return NULL;
@@ -307,13 +302,15 @@ again:
 	ret = radix_tree_insert(&dev->reada_zones,
 				(unsigned long)(zone->end >> PAGE_CACHE_SHIFT),
 				zone);
-	spin_unlock(&fs_info->reada_lock);
 
-	if (ret) {
+	if (ret == -EEXIST) {
 		kfree(zone);
-		looped = 1;
-		goto again;
+		ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone,
+					     logical >> PAGE_CACHE_SHIFT, 1);
+		if (ret == 1)
+			kref_get(&zone->refcnt);
 	}
+	spin_unlock(&fs_info->reada_lock);
 
 	return zone;
 }
@@ -323,26 +320,26 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root,
 					      struct btrfs_key *top, int level)
 {
 	int ret;
-	int looped = 0;
 	struct reada_extent *re = NULL;
+	struct reada_extent *re_exist = NULL;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
 	struct btrfs_bio *bbio = NULL;
 	struct btrfs_device *dev;
+	struct btrfs_device *prev_dev;
 	u32 blocksize;
 	u64 length;
 	int nzones = 0;
 	int i;
 	unsigned long index = logical >> PAGE_CACHE_SHIFT;
 
-again:
 	spin_lock(&fs_info->reada_lock);
 	re = radix_tree_lookup(&fs_info->reada_tree, index);
 	if (re)
 		kref_get(&re->refcnt);
 	spin_unlock(&fs_info->reada_lock);
 
-	if (re || looped)
+	if (re)
 		return re;
 
 	re = kzalloc(sizeof(*re), GFP_NOFS);
@@ -365,9 +362,9 @@ again:
 	if (ret || !bbio || length < blocksize)
 		goto error;
 
-	if (bbio->num_stripes > MAX_MIRRORS) {
+	if (bbio->num_stripes > BTRFS_MAX_MIRRORS) {
 		printk(KERN_ERR "btrfs readahead: more than %d copies not "
-				"supported", MAX_MIRRORS);
+				"supported", BTRFS_MAX_MIRRORS);
 		goto error;
 	}
 
@@ -398,16 +395,31 @@ again:
 	/* insert extent in reada_tree + all per-device trees, all or nothing */
 	spin_lock(&fs_info->reada_lock);
 	ret = radix_tree_insert(&fs_info->reada_tree, index, re);
+	if (ret == -EEXIST) {
+		re_exist = radix_tree_lookup(&fs_info->reada_tree, index);
+		BUG_ON(!re_exist);
+		kref_get(&re_exist->refcnt);
+		spin_unlock(&fs_info->reada_lock);
+		goto error;
+	}
 	if (ret) {
 		spin_unlock(&fs_info->reada_lock);
-		if (ret != -ENOMEM) {
-			/* someone inserted the extent in the meantime */
-			looped = 1;
-		}
 		goto error;
 	}
+	prev_dev = NULL;
 	for (i = 0; i < nzones; ++i) {
 		dev = bbio->stripes[i].dev;
+		if (dev == prev_dev) {
+			/*
+			 * in case of DUP, just add the first zone. As both
+			 * are on the same device, there's nothing to gain
+			 * from adding both.
+			 * Also, it wouldn't work, as the tree is per device
+			 * and adding would fail with EEXIST
+			 */
+			continue;
+		}
+		prev_dev = dev;
 		ret = radix_tree_insert(&dev->reada_extents, index, re);
 		if (ret) {
 			while (--i >= 0) {
@@ -450,9 +462,7 @@ error:
 	}
 	kfree(bbio);
 	kfree(re);
-	if (looped)
-		goto again;
-	return NULL;
+	return re_exist;
 }
 
 static void reada_kref_dummy(struct kref *kr)
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 8c1aae2c845..646ee21bb03 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -326,6 +326,19 @@ static struct rb_node *tree_search(struct rb_root *root, u64 bytenr)
 	return NULL;
 }
 
+void backref_tree_panic(struct rb_node *rb_node, int errno,
+					  u64 bytenr)
+{
+
+	struct btrfs_fs_info *fs_info = NULL;
+	struct backref_node *bnode = rb_entry(rb_node, struct backref_node,
+					      rb_node);
+	if (bnode->root)
+		fs_info = bnode->root->fs_info;
+	btrfs_panic(fs_info, errno, "Inconsistency in backref cache "
+		    "found at offset %llu\n", (unsigned long long)bytenr);
+}
+
 /*
  * walk up backref nodes until reach node presents tree root
  */
@@ -452,7 +465,8 @@ static void update_backref_node(struct backref_cache *cache,
 	rb_erase(&node->rb_node, &cache->rb_root);
 	node->bytenr = bytenr;
 	rb_node = tree_insert(&cache->rb_root, node->bytenr, &node->rb_node);
-	BUG_ON(rb_node);
+	if (rb_node)
+		backref_tree_panic(rb_node, -EEXIST, bytenr);
 }
 
 /*
@@ -999,7 +1013,8 @@ next:
 	if (!cowonly) {
 		rb_node = tree_insert(&cache->rb_root, node->bytenr,
 				      &node->rb_node);
-		BUG_ON(rb_node);
+		if (rb_node)
+			backref_tree_panic(rb_node, -EEXIST, node->bytenr);
 		list_add_tail(&node->lower, &cache->leaves);
 	}
 
@@ -1034,7 +1049,9 @@ next:
 		if (!cowonly) {
 			rb_node = tree_insert(&cache->rb_root, upper->bytenr,
 					      &upper->rb_node);
-			BUG_ON(rb_node);
+			if (rb_node)
+				backref_tree_panic(rb_node, -EEXIST,
+						   upper->bytenr);
 		}
 
 		list_add_tail(&edge->list[UPPER], &upper->lower);
@@ -1180,7 +1197,8 @@ static int clone_backref_node(struct btrfs_trans_handle *trans,
 
 	rb_node = tree_insert(&cache->rb_root, new_node->bytenr,
 			      &new_node->rb_node);
-	BUG_ON(rb_node);
+	if (rb_node)
+		backref_tree_panic(rb_node, -EEXIST, new_node->bytenr);
 
 	if (!new_node->lowest) {
 		list_for_each_entry(new_edge, &new_node->lower, list[UPPER]) {
@@ -1203,14 +1221,15 @@ fail:
 /*
  * helper to add 'address of tree root -> reloc tree' mapping
  */
-static int __add_reloc_root(struct btrfs_root *root)
+static int __must_check __add_reloc_root(struct btrfs_root *root)
 {
 	struct rb_node *rb_node;
 	struct mapping_node *node;
 	struct reloc_control *rc = root->fs_info->reloc_ctl;
 
 	node = kmalloc(sizeof(*node), GFP_NOFS);
-	BUG_ON(!node);
+	if (!node)
+		return -ENOMEM;
 
 	node->bytenr = root->node->start;
 	node->data = root;
@@ -1219,7 +1238,12 @@ static int __add_reloc_root(struct btrfs_root *root)
 	rb_node = tree_insert(&rc->reloc_root_tree.rb_root,
 			      node->bytenr, &node->rb_node);
 	spin_unlock(&rc->reloc_root_tree.lock);
-	BUG_ON(rb_node);
+	if (rb_node) {
+		kfree(node);
+		btrfs_panic(root->fs_info, -EEXIST, "Duplicate root found "
+			    "for start=%llu while inserting into relocation "
+			    "tree\n");
+	}
 
 	list_add_tail(&root->root_list, &rc->reloc_roots);
 	return 0;
@@ -1252,9 +1276,12 @@ static int __update_reloc_root(struct btrfs_root *root, int del)
 		rb_node = tree_insert(&rc->reloc_root_tree.rb_root,
 				      node->bytenr, &node->rb_node);
 		spin_unlock(&rc->reloc_root_tree.lock);
-		BUG_ON(rb_node);
+		if (rb_node)
+			backref_tree_panic(rb_node, -EEXIST, node->bytenr);
 	} else {
+		spin_lock(&root->fs_info->trans_lock);
 		list_del_init(&root->root_list);
+		spin_unlock(&root->fs_info->trans_lock);
 		kfree(node);
 	}
 	return 0;
@@ -1334,6 +1361,7 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 	struct btrfs_root *reloc_root;
 	struct reloc_control *rc = root->fs_info->reloc_ctl;
 	int clear_rsv = 0;
+	int ret;
 
 	if (root->reloc_root) {
 		reloc_root = root->reloc_root;
@@ -1353,7 +1381,8 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
 	if (clear_rsv)
 		trans->block_rsv = NULL;
 
-	__add_reloc_root(reloc_root);
+	ret = __add_reloc_root(reloc_root);
+	BUG_ON(ret < 0);
 	root->reloc_root = reloc_root;
 	return 0;
 }
@@ -1577,15 +1606,14 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 				WARN_ON(!IS_ALIGNED(end, root->sectorsize));
 				end--;
 				ret = try_lock_extent(&BTRFS_I(inode)->io_tree,
-						      key.offset, end,
-						      GFP_NOFS);
+						      key.offset, end);
 				if (!ret)
 					continue;
 
 				btrfs_drop_extent_cache(inode, key.offset, end,
 							1);
 				unlock_extent(&BTRFS_I(inode)->io_tree,
-					      key.offset, end, GFP_NOFS);
+					      key.offset, end);
 			}
 		}
 
@@ -1956,9 +1984,9 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		}
 
 		/* the lock_extent waits for readpage to complete */
-		lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		lock_extent(&BTRFS_I(inode)->io_tree, start, end);
 		btrfs_drop_extent_cache(inode, start, end, 1);
-		unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		unlock_extent(&BTRFS_I(inode)->io_tree, start, end);
 	}
 	return 0;
 }
@@ -2246,7 +2274,8 @@ again:
 		} else {
 			list_del_init(&reloc_root->root_list);
 		}
-		btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0, 1);
+		ret = btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0, 1);
+		BUG_ON(ret < 0);
 	}
 
 	if (found) {
@@ -2862,12 +2891,12 @@ int prealloc_file_extent_cluster(struct inode *inode,
 		else
 			end = cluster->end - offset;
 
-		lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		lock_extent(&BTRFS_I(inode)->io_tree, start, end);
 		num_bytes = end + 1 - start;
 		ret = btrfs_prealloc_file_range(inode, 0, start,
 						num_bytes, num_bytes,
 						end + 1, &alloc_hint);
-		unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+		unlock_extent(&BTRFS_I(inode)->io_tree, start, end);
 		if (ret)
 			break;
 		nr++;
@@ -2899,7 +2928,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end,
 	em->bdev = root->fs_info->fs_devices->latest_bdev;
 	set_bit(EXTENT_FLAG_PINNED, &em->flags);
 
-	lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+	lock_extent(&BTRFS_I(inode)->io_tree, start, end);
 	while (1) {
 		write_lock(&em_tree->lock);
 		ret = add_extent_mapping(em_tree, em);
@@ -2910,7 +2939,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end,
 		}
 		btrfs_drop_extent_cache(inode, start, end, 0);
 	}
-	unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS);
+	unlock_extent(&BTRFS_I(inode)->io_tree, start, end);
 	return ret;
 }
 
@@ -2990,8 +3019,7 @@ static int relocate_file_extent_cluster(struct inode *inode,
 		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
 		page_end = page_start + PAGE_CACHE_SIZE - 1;
 
-		lock_extent(&BTRFS_I(inode)->io_tree,
-			    page_start, page_end, GFP_NOFS);
+		lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end);
 
 		set_page_extent_mapped(page);
 
@@ -3007,7 +3035,7 @@ static int relocate_file_extent_cluster(struct inode *inode,
 		set_page_dirty(page);
 
 		unlock_extent(&BTRFS_I(inode)->io_tree,
-			      page_start, page_end, GFP_NOFS);
+			      page_start, page_end);
 		unlock_page(page);
 		page_cache_release(page);
 
@@ -3154,7 +3182,8 @@ static int add_tree_block(struct reloc_control *rc,
 	block->key_ready = 0;
 
 	rb_node = tree_insert(blocks, block->bytenr, &block->rb_node);
-	BUG_ON(rb_node);
+	if (rb_node)
+		backref_tree_panic(rb_node, -EEXIST, block->bytenr);
 
 	return 0;
 }
@@ -3426,7 +3455,9 @@ static int find_data_references(struct reloc_control *rc,
 			block->key_ready = 1;
 			rb_node = tree_insert(blocks, block->bytenr,
 					      &block->rb_node);
-			BUG_ON(rb_node);
+			if (rb_node)
+				backref_tree_panic(rb_node, -EEXIST,
+						   block->bytenr);
 		}
 		if (counted)
 			added = 1;
@@ -3782,7 +3813,7 @@ restart:
 
 		ret = btrfs_block_rsv_check(rc->extent_root, rc->block_rsv, 5);
 		if (ret < 0) {
-			if (ret != -EAGAIN) {
+			if (ret != -ENOSPC) {
 				err = ret;
 				WARN_ON(1);
 				break;
@@ -4073,10 +4104,11 @@ out:
 static noinline_for_stack int mark_garbage_root(struct btrfs_root *root)
 {
 	struct btrfs_trans_handle *trans;
-	int ret;
+	int ret, err;
 
 	trans = btrfs_start_transaction(root->fs_info->tree_root, 0);
-	BUG_ON(IS_ERR(trans));
+	if (IS_ERR(trans))
+		return PTR_ERR(trans);
 
 	memset(&root->root_item.drop_progress, 0,
 		sizeof(root->root_item.drop_progress));
@@ -4084,11 +4116,11 @@ static noinline_for_stack int mark_garbage_root(struct btrfs_root *root)
 	btrfs_set_root_refs(&root->root_item, 0);
 	ret = btrfs_update_root(trans, root->fs_info->tree_root,
 				&root->root_key, &root->root_item);
-	BUG_ON(ret);
 
-	ret = btrfs_end_transaction(trans, root->fs_info->tree_root);
-	BUG_ON(ret);
-	return 0;
+	err = btrfs_end_transaction(trans, root->fs_info->tree_root);
+	if (err)
+		return err;
+	return ret;
 }
 
 /*
@@ -4156,7 +4188,11 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 					err = ret;
 					goto out;
 				}
-				mark_garbage_root(reloc_root);
+				ret = mark_garbage_root(reloc_root);
+				if (ret < 0) {
+					err = ret;
+					goto out;
+				}
 			}
 		}
 
@@ -4202,13 +4238,19 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 		fs_root = read_fs_root(root->fs_info,
 				       reloc_root->root_key.offset);
-		BUG_ON(IS_ERR(fs_root));
+		if (IS_ERR(fs_root)) {
+			err = PTR_ERR(fs_root);
+			goto out_free;
+		}
 
-		__add_reloc_root(reloc_root);
+		err = __add_reloc_root(reloc_root);
+		BUG_ON(err < 0); /* -ENOMEM or logic error */
 		fs_root->reloc_root = reloc_root;
 	}
 
-	btrfs_commit_transaction(trans, rc->extent_root);
+	err = btrfs_commit_transaction(trans, rc->extent_root);
+	if (err)
+		goto out_free;
 
 	merge_reloc_roots(rc);
 
@@ -4218,7 +4260,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	if (IS_ERR(trans))
 		err = PTR_ERR(trans);
 	else
-		btrfs_commit_transaction(trans, rc->extent_root);
+		err = btrfs_commit_transaction(trans, rc->extent_root);
 out_free:
 	kfree(rc);
 out:
@@ -4267,6 +4309,8 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 	disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
 	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
 				       disk_bytenr + len - 1, &list, 0);
+	if (ret)
+		goto out;
 
 	while (!list_empty(&list)) {
 		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
@@ -4284,6 +4328,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 
 		btrfs_add_ordered_sum(inode, ordered, sums);
 	}
+out:
 	btrfs_put_ordered_extent(ordered);
 	return ret;
 }
@@ -4380,7 +4425,7 @@ void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
  * called after snapshot is created. migrate block reservation
  * and create reloc root for the newly created snapshot
  */
-void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
+int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 			       struct btrfs_pending_snapshot *pending)
 {
 	struct btrfs_root *root = pending->root;
@@ -4390,7 +4435,7 @@ void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 	int ret;
 
 	if (!root->reloc_root)
-		return;
+		return 0;
 
 	rc = root->fs_info->reloc_ctl;
 	rc->merging_rsv_size += rc->nodes_relocated;
@@ -4399,18 +4444,21 @@ void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 		ret = btrfs_block_rsv_migrate(&pending->block_rsv,
 					      rc->block_rsv,
 					      rc->nodes_relocated);
-		BUG_ON(ret);
+		if (ret)
+			return ret;
 	}
 
 	new_root = pending->snap;
 	reloc_root = create_reloc_root(trans, root->reloc_root,
 				       new_root->root_key.objectid);
+	if (IS_ERR(reloc_root))
+		return PTR_ERR(reloc_root);
 
-	__add_reloc_root(reloc_root);
+	ret = __add_reloc_root(reloc_root);
+	BUG_ON(ret < 0);
 	new_root->reloc_root = reloc_root;
 
-	if (rc->create_reloc_tree) {
+	if (rc->create_reloc_tree)
 		ret = clone_backref_node(trans, rc, root, reloc_root);
-		BUG_ON(ret);
-	}
+	return ret;
 }
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index f4099904565..24fb8ce4e07 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -93,10 +93,14 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 	unsigned long ptr;
 
 	path = btrfs_alloc_path();
-	BUG_ON(!path);
+	if (!path)
+		return -ENOMEM;
+
 	ret = btrfs_search_slot(trans, root, key, path, 0, 1);
-	if (ret < 0)
+	if (ret < 0) {
+		btrfs_abort_transaction(trans, root, ret);
 		goto out;
+	}
 
 	if (ret != 0) {
 		btrfs_print_leaf(root, path->nodes[0]);
@@ -116,13 +120,10 @@ out:
 	return ret;
 }
 
-int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
-		      *root, struct btrfs_key *key, struct btrfs_root_item
-		      *item)
+int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		      struct btrfs_key *key, struct btrfs_root_item *item)
 {
-	int ret;
-	ret = btrfs_insert_item(trans, root, key, item, sizeof(*item));
-	return ret;
+	return btrfs_insert_item(trans, root, key, item, sizeof(*item));
 }
 
 /*
@@ -384,6 +385,8 @@ int btrfs_find_root_ref(struct btrfs_root *tree_root,
  *
  * For a back ref the root_id is the id of the subvol or snapshot and
  * ref_id is the id of the tree referencing it.
+ *
+ * Will return 0, -ENOMEM, or anything from the CoW path
  */
 int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *tree_root,
@@ -407,7 +410,11 @@ int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
 again:
 	ret = btrfs_insert_empty_item(trans, tree_root, path, &key,
 				      sizeof(*ref) + name_len);
-	BUG_ON(ret);
+	if (ret) {
+		btrfs_abort_transaction(trans, tree_root, ret);
+		btrfs_free_path(path);
+		return ret;
+	}
 
 	leaf = path->nodes[0];
 	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index b9b84cdfc35..2f3d6f917fb 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -36,37 +36,30 @@
  * Future enhancements:
  *  - In case an unrepairable extent is encountered, track which files are
  *    affected and report them
- *  - In case of a read error on files with nodatasum, map the file and read
- *    the extent to trigger a writeback of the good copy
  *  - track and record media errors, throw out bad devices
  *  - add a mode to also read unallocated space
  */
 
-struct scrub_bio;
-struct scrub_page;
+struct scrub_block;
 struct scrub_dev;
-static void scrub_bio_end_io(struct bio *bio, int err);
-static void scrub_checksum(struct btrfs_work *work);
-static int scrub_checksum_data(struct scrub_dev *sdev,
-			       struct scrub_page *spag, void *buffer);
-static int scrub_checksum_tree_block(struct scrub_dev *sdev,
-				     struct scrub_page *spag, u64 logical,
-				     void *buffer);
-static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer);
-static int scrub_fixup_check(struct scrub_bio *sbio, int ix);
-static void scrub_fixup_end_io(struct bio *bio, int err);
-static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
-			  struct page *page);
-static void scrub_fixup(struct scrub_bio *sbio, int ix);
 
 #define SCRUB_PAGES_PER_BIO	16	/* 64k per bio */
 #define SCRUB_BIOS_PER_DEV	16	/* 1 MB per device in flight */
+#define SCRUB_MAX_PAGES_PER_BLOCK	16	/* 64k per node/leaf/sector */
 
 struct scrub_page {
+	struct scrub_block	*sblock;
+	struct page		*page;
+	struct block_device	*bdev;
 	u64			flags;  /* extent flags */
 	u64			generation;
-	int			mirror_num;
-	int			have_csum;
+	u64			logical;
+	u64			physical;
+	struct {
+		unsigned int	mirror_num:8;
+		unsigned int	have_csum:1;
+		unsigned int	io_error:1;
+	};
 	u8			csum[BTRFS_CSUM_SIZE];
 };
 
@@ -77,12 +70,25 @@ struct scrub_bio {
 	int			err;
 	u64			logical;
 	u64			physical;
-	struct scrub_page	spag[SCRUB_PAGES_PER_BIO];
-	u64			count;
+	struct scrub_page	*pagev[SCRUB_PAGES_PER_BIO];
+	int			page_count;
 	int			next_free;
 	struct btrfs_work	work;
 };
 
+struct scrub_block {
+	struct scrub_page	pagev[SCRUB_MAX_PAGES_PER_BLOCK];
+	int			page_count;
+	atomic_t		outstanding_pages;
+	atomic_t		ref_count; /* free mem on transition to zero */
+	struct scrub_dev	*sdev;
+	struct {
+		unsigned int	header_error:1;
+		unsigned int	checksum_error:1;
+		unsigned int	no_io_error_seen:1;
+	};
+};
+
 struct scrub_dev {
 	struct scrub_bio	*bios[SCRUB_BIOS_PER_DEV];
 	struct btrfs_device	*dev;
@@ -96,6 +102,10 @@ struct scrub_dev {
 	struct list_head	csum_list;
 	atomic_t		cancel_req;
 	int			readonly;
+	int			pages_per_bio; /* <= SCRUB_PAGES_PER_BIO */
+	u32			sectorsize;
+	u32			nodesize;
+	u32			leafsize;
 	/*
 	 * statistics
 	 */
@@ -124,6 +134,43 @@ struct scrub_warning {
 	int			scratch_bufsize;
 };
 
+
+static int scrub_handle_errored_block(struct scrub_block *sblock_to_check);
+static int scrub_setup_recheck_block(struct scrub_dev *sdev,
+				     struct btrfs_mapping_tree *map_tree,
+				     u64 length, u64 logical,
+				     struct scrub_block *sblock);
+static int scrub_recheck_block(struct btrfs_fs_info *fs_info,
+			       struct scrub_block *sblock, int is_metadata,
+			       int have_csum, u8 *csum, u64 generation,
+			       u16 csum_size);
+static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info,
+					 struct scrub_block *sblock,
+					 int is_metadata, int have_csum,
+					 const u8 *csum, u64 generation,
+					 u16 csum_size);
+static void scrub_complete_bio_end_io(struct bio *bio, int err);
+static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad,
+					     struct scrub_block *sblock_good,
+					     int force_write);
+static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
+					    struct scrub_block *sblock_good,
+					    int page_num, int force_write);
+static int scrub_checksum_data(struct scrub_block *sblock);
+static int scrub_checksum_tree_block(struct scrub_block *sblock);
+static int scrub_checksum_super(struct scrub_block *sblock);
+static void scrub_block_get(struct scrub_block *sblock);
+static void scrub_block_put(struct scrub_block *sblock);
+static int scrub_add_page_to_bio(struct scrub_dev *sdev,
+				 struct scrub_page *spage);
+static int scrub_pages(struct scrub_dev *sdev, u64 logical, u64 len,
+		       u64 physical, u64 flags, u64 gen, int mirror_num,
+		       u8 *csum, int force);
+static void scrub_bio_end_io(struct bio *bio, int err);
+static void scrub_bio_end_io_worker(struct btrfs_work *work);
+static void scrub_block_complete(struct scrub_block *sblock);
+
+
 static void scrub_free_csums(struct scrub_dev *sdev)
 {
 	while (!list_empty(&sdev->csum_list)) {
@@ -135,23 +182,6 @@ static void scrub_free_csums(struct scrub_dev *sdev)
 	}
 }
 
-static void scrub_free_bio(struct bio *bio)
-{
-	int i;
-	struct page *last_page = NULL;
-
-	if (!bio)
-		return;
-
-	for (i = 0; i < bio->bi_vcnt; ++i) {
-		if (bio->bi_io_vec[i].bv_page == last_page)
-			continue;
-		last_page = bio->bi_io_vec[i].bv_page;
-		__free_page(last_page);
-	}
-	bio_put(bio);
-}
-
 static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev)
 {
 	int i;
@@ -159,13 +189,23 @@ static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev)
 	if (!sdev)
 		return;
 
+	/* this can happen when scrub is cancelled */
+	if (sdev->curr != -1) {
+		struct scrub_bio *sbio = sdev->bios[sdev->curr];
+
+		for (i = 0; i < sbio->page_count; i++) {
+			BUG_ON(!sbio->pagev[i]);
+			BUG_ON(!sbio->pagev[i]->page);
+			scrub_block_put(sbio->pagev[i]->sblock);
+		}
+		bio_put(sbio->bio);
+	}
+
 	for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
 		struct scrub_bio *sbio = sdev->bios[i];
 
 		if (!sbio)
 			break;
-
-		scrub_free_bio(sbio->bio);
 		kfree(sbio);
 	}
 
@@ -179,11 +219,16 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
 	struct scrub_dev *sdev;
 	int		i;
 	struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
+	int pages_per_bio;
 
+	pages_per_bio = min_t(int, SCRUB_PAGES_PER_BIO,
+			      bio_get_nr_vecs(dev->bdev));
 	sdev = kzalloc(sizeof(*sdev), GFP_NOFS);
 	if (!sdev)
 		goto nomem;
 	sdev->dev = dev;
+	sdev->pages_per_bio = pages_per_bio;
+	sdev->curr = -1;
 	for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
 		struct scrub_bio *sbio;
 
@@ -194,8 +239,8 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
 
 		sbio->index = i;
 		sbio->sdev = sdev;
-		sbio->count = 0;
-		sbio->work.func = scrub_checksum;
+		sbio->page_count = 0;
+		sbio->work.func = scrub_bio_end_io_worker;
 
 		if (i != SCRUB_BIOS_PER_DEV-1)
 			sdev->bios[i]->next_free = i + 1;
@@ -203,7 +248,9 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
 			sdev->bios[i]->next_free = -1;
 	}
 	sdev->first_free = 0;
-	sdev->curr = -1;
+	sdev->nodesize = dev->dev_root->nodesize;
+	sdev->leafsize = dev->dev_root->leafsize;
+	sdev->sectorsize = dev->dev_root->sectorsize;
 	atomic_set(&sdev->in_flight, 0);
 	atomic_set(&sdev->fixup_cnt, 0);
 	atomic_set(&sdev->cancel_req, 0);
@@ -294,10 +341,9 @@ err:
 	return 0;
 }
 
-static void scrub_print_warning(const char *errstr, struct scrub_bio *sbio,
-				int ix)
+static void scrub_print_warning(const char *errstr, struct scrub_block *sblock)
 {
-	struct btrfs_device *dev = sbio->sdev->dev;
+	struct btrfs_device *dev = sblock->sdev->dev;
 	struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
 	struct btrfs_path *path;
 	struct btrfs_key found_key;
@@ -316,8 +362,9 @@ static void scrub_print_warning(const char *errstr, struct scrub_bio *sbio,
 
 	swarn.scratch_buf = kmalloc(bufsize, GFP_NOFS);
 	swarn.msg_buf = kmalloc(bufsize, GFP_NOFS);
-	swarn.sector = (sbio->physical + ix * PAGE_SIZE) >> 9;
-	swarn.logical = sbio->logical + ix * PAGE_SIZE;
+	BUG_ON(sblock->page_count < 1);
+	swarn.sector = (sblock->pagev[0].physical) >> 9;
+	swarn.logical = sblock->pagev[0].logical;
 	swarn.errstr = errstr;
 	swarn.dev = dev;
 	swarn.msg_bufsize = bufsize;
@@ -342,7 +389,8 @@ static void scrub_print_warning(const char *errstr, struct scrub_bio *sbio,
 		do {
 			ret = tree_backref_for_extent(&ptr, eb, ei, item_size,
 							&ref_root, &ref_level);
-			printk(KERN_WARNING "%s at logical %llu on dev %s, "
+			printk(KERN_WARNING
+				"btrfs: %s at logical %llu on dev %s, "
 				"sector %llu: metadata %s (level %d) in tree "
 				"%llu\n", errstr, swarn.logical, dev->name,
 				(unsigned long long)swarn.sector,
@@ -531,9 +579,9 @@ out:
 		spin_lock(&sdev->stat_lock);
 		++sdev->stat.uncorrectable_errors;
 		spin_unlock(&sdev->stat_lock);
-		printk_ratelimited(KERN_ERR "btrfs: unable to fixup "
-					"(nodatasum) error at logical %llu\n",
-					fixup->logical);
+		printk_ratelimited(KERN_ERR
+			"btrfs: unable to fixup (nodatasum) error at logical %llu on dev %s\n",
+			(unsigned long long)fixup->logical, sdev->dev->name);
 	}
 
 	btrfs_free_path(path);
@@ -550,91 +598,168 @@ out:
 }
 
 /*
- * scrub_recheck_error gets called when either verification of the page
- * failed or the bio failed to read, e.g. with EIO. In the latter case,
- * recheck_error gets called for every page in the bio, even though only
- * one may be bad
+ * scrub_handle_errored_block gets called when either verification of the
+ * pages failed or the bio failed to read, e.g. with EIO. In the latter
+ * case, this function handles all pages in the bio, even though only one
+ * may be bad.
+ * The goal of this function is to repair the errored block by using the
+ * contents of one of the mirrors.
  */
-static int scrub_recheck_error(struct scrub_bio *sbio, int ix)
+static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 {
-	struct scrub_dev *sdev = sbio->sdev;
-	u64 sector = (sbio->physical + ix * PAGE_SIZE) >> 9;
+	struct scrub_dev *sdev = sblock_to_check->sdev;
+	struct btrfs_fs_info *fs_info;
+	u64 length;
+	u64 logical;
+	u64 generation;
+	unsigned int failed_mirror_index;
+	unsigned int is_metadata;
+	unsigned int have_csum;
+	u8 *csum;
+	struct scrub_block *sblocks_for_recheck; /* holds one for each mirror */
+	struct scrub_block *sblock_bad;
+	int ret;
+	int mirror_index;
+	int page_num;
+	int success;
 	static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL,
-					DEFAULT_RATELIMIT_BURST);
+				      DEFAULT_RATELIMIT_BURST);
+
+	BUG_ON(sblock_to_check->page_count < 1);
+	fs_info = sdev->dev->dev_root->fs_info;
+	length = sblock_to_check->page_count * PAGE_SIZE;
+	logical = sblock_to_check->pagev[0].logical;
+	generation = sblock_to_check->pagev[0].generation;
+	BUG_ON(sblock_to_check->pagev[0].mirror_num < 1);
+	failed_mirror_index = sblock_to_check->pagev[0].mirror_num - 1;
+	is_metadata = !(sblock_to_check->pagev[0].flags &
+			BTRFS_EXTENT_FLAG_DATA);
+	have_csum = sblock_to_check->pagev[0].have_csum;
+	csum = sblock_to_check->pagev[0].csum;
 
-	if (sbio->err) {
-		if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, sector,
-				   sbio->bio->bi_io_vec[ix].bv_page) == 0) {
-			if (scrub_fixup_check(sbio, ix) == 0)
-				return 0;
-		}
-		if (__ratelimit(&_rs))
-			scrub_print_warning("i/o error", sbio, ix);
-	} else {
-		if (__ratelimit(&_rs))
-			scrub_print_warning("checksum error", sbio, ix);
+	/*
+	 * read all mirrors one after the other. This includes to
+	 * re-read the extent or metadata block that failed (that was
+	 * the cause that this fixup code is called) another time,
+	 * page by page this time in order to know which pages
+	 * caused I/O errors and which ones are good (for all mirrors).
+	 * It is the goal to handle the situation when more than one
+	 * mirror contains I/O errors, but the errors do not
+	 * overlap, i.e. the data can be repaired by selecting the
+	 * pages from those mirrors without I/O error on the
+	 * particular pages. One example (with blocks >= 2 * PAGE_SIZE)
+	 * would be that mirror #1 has an I/O error on the first page,
+	 * the second page is good, and mirror #2 has an I/O error on
+	 * the second page, but the first page is good.
+	 * Then the first page of the first mirror can be repaired by
+	 * taking the first page of the second mirror, and the
+	 * second page of the second mirror can be repaired by
+	 * copying the contents of the 2nd page of the 1st mirror.
+	 * One more note: if the pages of one mirror contain I/O
+	 * errors, the checksum cannot be verified. In order to get
+	 * the best data for repairing, the first attempt is to find
+	 * a mirror without I/O errors and with a validated checksum.
+	 * Only if this is not possible, the pages are picked from
+	 * mirrors with I/O errors without considering the checksum.
+	 * If the latter is the case, at the end, the checksum of the
+	 * repaired area is verified in order to correctly maintain
+	 * the statistics.
+	 */
+
+	sblocks_for_recheck = kzalloc(BTRFS_MAX_MIRRORS *
+				     sizeof(*sblocks_for_recheck),
+				     GFP_NOFS);
+	if (!sblocks_for_recheck) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.malloc_errors++;
+		sdev->stat.read_errors++;
+		sdev->stat.uncorrectable_errors++;
+		spin_unlock(&sdev->stat_lock);
+		goto out;
 	}
 
-	spin_lock(&sdev->stat_lock);
-	++sdev->stat.read_errors;
-	spin_unlock(&sdev->stat_lock);
+	/* setup the context, map the logical blocks and alloc the pages */
+	ret = scrub_setup_recheck_block(sdev, &fs_info->mapping_tree, length,
+					logical, sblocks_for_recheck);
+	if (ret) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.read_errors++;
+		sdev->stat.uncorrectable_errors++;
+		spin_unlock(&sdev->stat_lock);
+		goto out;
+	}
+	BUG_ON(failed_mirror_index >= BTRFS_MAX_MIRRORS);
+	sblock_bad = sblocks_for_recheck + failed_mirror_index;
 
-	scrub_fixup(sbio, ix);
-	return 1;
-}
+	/* build and submit the bios for the failed mirror, check checksums */
+	ret = scrub_recheck_block(fs_info, sblock_bad, is_metadata, have_csum,
+				  csum, generation, sdev->csum_size);
+	if (ret) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.read_errors++;
+		sdev->stat.uncorrectable_errors++;
+		spin_unlock(&sdev->stat_lock);
+		goto out;
+	}
 
-static int scrub_fixup_check(struct scrub_bio *sbio, int ix)
-{
-	int ret = 1;
-	struct page *page;
-	void *buffer;
-	u64 flags = sbio->spag[ix].flags;
+	if (!sblock_bad->header_error && !sblock_bad->checksum_error &&
+	    sblock_bad->no_io_error_seen) {
+		/*
+		 * the error disappeared after reading page by page, or
+		 * the area was part of a huge bio and other parts of the
+		 * bio caused I/O errors, or the block layer merged several
+		 * read requests into one and the error is caused by a
+		 * different bio (usually one of the two latter cases is
+		 * the cause)
+		 */
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.unverified_errors++;
+		spin_unlock(&sdev->stat_lock);
 
-	page = sbio->bio->bi_io_vec[ix].bv_page;
-	buffer = kmap_atomic(page, KM_USER0);
-	if (flags & BTRFS_EXTENT_FLAG_DATA) {
-		ret = scrub_checksum_data(sbio->sdev,
-					  sbio->spag + ix, buffer);
-	} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
-		ret = scrub_checksum_tree_block(sbio->sdev,
-						sbio->spag + ix,
-						sbio->logical + ix * PAGE_SIZE,
-						buffer);
-	} else {
-		WARN_ON(1);
+		goto out;
 	}
-	kunmap_atomic(buffer, KM_USER0);
 
-	return ret;
-}
+	if (!sblock_bad->no_io_error_seen) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.read_errors++;
+		spin_unlock(&sdev->stat_lock);
+		if (__ratelimit(&_rs))
+			scrub_print_warning("i/o error", sblock_to_check);
+	} else if (sblock_bad->checksum_error) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.csum_errors++;
+		spin_unlock(&sdev->stat_lock);
+		if (__ratelimit(&_rs))
+			scrub_print_warning("checksum error", sblock_to_check);
+	} else if (sblock_bad->header_error) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.verify_errors++;
+		spin_unlock(&sdev->stat_lock);
+		if (__ratelimit(&_rs))
+			scrub_print_warning("checksum/header error",
+					    sblock_to_check);
+	}
 
-static void scrub_fixup_end_io(struct bio *bio, int err)
-{
-	complete((struct completion *)bio->bi_private);
-}
+	if (sdev->readonly)
+		goto did_not_correct_error;
 
-static void scrub_fixup(struct scrub_bio *sbio, int ix)
-{
-	struct scrub_dev *sdev = sbio->sdev;
-	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
-	struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
-	struct btrfs_bio *bbio = NULL;
-	struct scrub_fixup_nodatasum *fixup;
-	u64 logical = sbio->logical + ix * PAGE_SIZE;
-	u64 length;
-	int i;
-	int ret;
-	DECLARE_COMPLETION_ONSTACK(complete);
-
-	if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) &&
-	    (sbio->spag[ix].have_csum == 0)) {
-		fixup = kzalloc(sizeof(*fixup), GFP_NOFS);
-		if (!fixup)
-			goto uncorrectable;
-		fixup->sdev = sdev;
-		fixup->logical = logical;
-		fixup->root = fs_info->extent_root;
-		fixup->mirror_num = sbio->spag[ix].mirror_num;
+	if (!is_metadata && !have_csum) {
+		struct scrub_fixup_nodatasum *fixup_nodatasum;
+
+		/*
+		 * !is_metadata and !have_csum, this means that the data
+		 * might not be COW'ed, that it might be modified
+		 * concurrently. The general strategy to work on the
+		 * commit root does not help in the case when COW is not
+		 * used.
+		 */
+		fixup_nodatasum = kzalloc(sizeof(*fixup_nodatasum), GFP_NOFS);
+		if (!fixup_nodatasum)
+			goto did_not_correct_error;
+		fixup_nodatasum->sdev = sdev;
+		fixup_nodatasum->logical = logical;
+		fixup_nodatasum->root = fs_info->extent_root;
+		fixup_nodatasum->mirror_num = failed_mirror_index + 1;
 		/*
 		 * increment scrubs_running to prevent cancel requests from
 		 * completing as long as a fixup worker is running. we must also
@@ -649,235 +774,533 @@ static void scrub_fixup(struct scrub_bio *sbio, int ix)
 		atomic_inc(&fs_info->scrubs_paused);
 		mutex_unlock(&fs_info->scrub_lock);
 		atomic_inc(&sdev->fixup_cnt);
-		fixup->work.func = scrub_fixup_nodatasum;
-		btrfs_queue_worker(&fs_info->scrub_workers, &fixup->work);
-		return;
+		fixup_nodatasum->work.func = scrub_fixup_nodatasum;
+		btrfs_queue_worker(&fs_info->scrub_workers,
+				   &fixup_nodatasum->work);
+		goto out;
 	}
 
-	length = PAGE_SIZE;
-	ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length,
-			      &bbio, 0);
-	if (ret || !bbio || length < PAGE_SIZE) {
-		printk(KERN_ERR
-		       "scrub_fixup: btrfs_map_block failed us for %llu\n",
-		       (unsigned long long)logical);
-		WARN_ON(1);
-		kfree(bbio);
-		return;
+	/*
+	 * now build and submit the bios for the other mirrors, check
+	 * checksums
+	 */
+	for (mirror_index = 0;
+	     mirror_index < BTRFS_MAX_MIRRORS &&
+	     sblocks_for_recheck[mirror_index].page_count > 0;
+	     mirror_index++) {
+		if (mirror_index == failed_mirror_index)
+			continue;
+
+		/* build and submit the bios, check checksums */
+		ret = scrub_recheck_block(fs_info,
+					  sblocks_for_recheck + mirror_index,
+					  is_metadata, have_csum, csum,
+					  generation, sdev->csum_size);
+		if (ret)
+			goto did_not_correct_error;
 	}
 
-	if (bbio->num_stripes == 1)
-		/* there aren't any replicas */
-		goto uncorrectable;
+	/*
+	 * first try to pick the mirror which is completely without I/O
+	 * errors and also does not have a checksum error.
+	 * If one is found, and if a checksum is present, the full block
+	 * that is known to contain an error is rewritten. Afterwards
+	 * the block is known to be corrected.
+	 * If a mirror is found which is completely correct, and no
+	 * checksum is present, only those pages are rewritten that had
+	 * an I/O error in the block to be repaired, since it cannot be
+	 * determined, which copy of the other pages is better (and it
+	 * could happen otherwise that a correct page would be
+	 * overwritten by a bad one).
+	 */
+	for (mirror_index = 0;
+	     mirror_index < BTRFS_MAX_MIRRORS &&
+	     sblocks_for_recheck[mirror_index].page_count > 0;
+	     mirror_index++) {
+		struct scrub_block *sblock_other = sblocks_for_recheck +
+						   mirror_index;
+
+		if (!sblock_other->header_error &&
+		    !sblock_other->checksum_error &&
+		    sblock_other->no_io_error_seen) {
+			int force_write = is_metadata || have_csum;
+
+			ret = scrub_repair_block_from_good_copy(sblock_bad,
+								sblock_other,
+								force_write);
+			if (0 == ret)
+				goto corrected_error;
+		}
+	}
 
 	/*
-	 * first find a good copy
+	 * in case of I/O errors in the area that is supposed to be
+	 * repaired, continue by picking good copies of those pages.
+	 * Select the good pages from mirrors to rewrite bad pages from
+	 * the area to fix. Afterwards verify the checksum of the block
+	 * that is supposed to be repaired. This verification step is
+	 * only done for the purpose of statistic counting and for the
+	 * final scrub report, whether errors remain.
+	 * A perfect algorithm could make use of the checksum and try
+	 * all possible combinations of pages from the different mirrors
+	 * until the checksum verification succeeds. For example, when
+	 * the 2nd page of mirror #1 faces I/O errors, and the 2nd page
+	 * of mirror #2 is readable but the final checksum test fails,
+	 * then the 2nd page of mirror #3 could be tried, whether now
+	 * the final checksum succeedes. But this would be a rare
+	 * exception and is therefore not implemented. At least it is
+	 * avoided that the good copy is overwritten.
+	 * A more useful improvement would be to pick the sectors
+	 * without I/O error based on sector sizes (512 bytes on legacy
+	 * disks) instead of on PAGE_SIZE. Then maybe 512 byte of one
+	 * mirror could be repaired by taking 512 byte of a different
+	 * mirror, even if other 512 byte sectors in the same PAGE_SIZE
+	 * area are unreadable.
 	 */
-	for (i = 0; i < bbio->num_stripes; ++i) {
-		if (i + 1 == sbio->spag[ix].mirror_num)
-			continue;
 
-		if (scrub_fixup_io(READ, bbio->stripes[i].dev->bdev,
-				   bbio->stripes[i].physical >> 9,
-				   sbio->bio->bi_io_vec[ix].bv_page)) {
-			/* I/O-error, this is not a good copy */
+	/* can only fix I/O errors from here on */
+	if (sblock_bad->no_io_error_seen)
+		goto did_not_correct_error;
+
+	success = 1;
+	for (page_num = 0; page_num < sblock_bad->page_count; page_num++) {
+		struct scrub_page *page_bad = sblock_bad->pagev + page_num;
+
+		if (!page_bad->io_error)
 			continue;
+
+		for (mirror_index = 0;
+		     mirror_index < BTRFS_MAX_MIRRORS &&
+		     sblocks_for_recheck[mirror_index].page_count > 0;
+		     mirror_index++) {
+			struct scrub_block *sblock_other = sblocks_for_recheck +
+							   mirror_index;
+			struct scrub_page *page_other = sblock_other->pagev +
+							page_num;
+
+			if (!page_other->io_error) {
+				ret = scrub_repair_page_from_good_copy(
+					sblock_bad, sblock_other, page_num, 0);
+				if (0 == ret) {
+					page_bad->io_error = 0;
+					break; /* succeeded for this page */
+				}
+			}
 		}
 
-		if (scrub_fixup_check(sbio, ix) == 0)
-			break;
+		if (page_bad->io_error) {
+			/* did not find a mirror to copy the page from */
+			success = 0;
+		}
 	}
-	if (i == bbio->num_stripes)
-		goto uncorrectable;
 
-	if (!sdev->readonly) {
-		/*
-		 * bi_io_vec[ix].bv_page now contains good data, write it back
-		 */
-		if (scrub_fixup_io(WRITE, sdev->dev->bdev,
-				   (sbio->physical + ix * PAGE_SIZE) >> 9,
-				   sbio->bio->bi_io_vec[ix].bv_page)) {
-			/* I/O-error, writeback failed, give up */
-			goto uncorrectable;
+	if (success) {
+		if (is_metadata || have_csum) {
+			/*
+			 * need to verify the checksum now that all
+			 * sectors on disk are repaired (the write
+			 * request for data to be repaired is on its way).
+			 * Just be lazy and use scrub_recheck_block()
+			 * which re-reads the data before the checksum
+			 * is verified, but most likely the data comes out
+			 * of the page cache.
+			 */
+			ret = scrub_recheck_block(fs_info, sblock_bad,
+						  is_metadata, have_csum, csum,
+						  generation, sdev->csum_size);
+			if (!ret && !sblock_bad->header_error &&
+			    !sblock_bad->checksum_error &&
+			    sblock_bad->no_io_error_seen)
+				goto corrected_error;
+			else
+				goto did_not_correct_error;
+		} else {
+corrected_error:
+			spin_lock(&sdev->stat_lock);
+			sdev->stat.corrected_errors++;
+			spin_unlock(&sdev->stat_lock);
+			printk_ratelimited(KERN_ERR
+				"btrfs: fixed up error at logical %llu on dev %s\n",
+				(unsigned long long)logical, sdev->dev->name);
 		}
+	} else {
+did_not_correct_error:
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.uncorrectable_errors++;
+		spin_unlock(&sdev->stat_lock);
+		printk_ratelimited(KERN_ERR
+			"btrfs: unable to fixup (regular) error at logical %llu on dev %s\n",
+			(unsigned long long)logical, sdev->dev->name);
 	}
 
-	kfree(bbio);
-	spin_lock(&sdev->stat_lock);
-	++sdev->stat.corrected_errors;
-	spin_unlock(&sdev->stat_lock);
+out:
+	if (sblocks_for_recheck) {
+		for (mirror_index = 0; mirror_index < BTRFS_MAX_MIRRORS;
+		     mirror_index++) {
+			struct scrub_block *sblock = sblocks_for_recheck +
+						     mirror_index;
+			int page_index;
+
+			for (page_index = 0; page_index < SCRUB_PAGES_PER_BIO;
+			     page_index++)
+				if (sblock->pagev[page_index].page)
+					__free_page(
+						sblock->pagev[page_index].page);
+		}
+		kfree(sblocks_for_recheck);
+	}
 
-	printk_ratelimited(KERN_ERR "btrfs: fixed up error at logical %llu\n",
-			       (unsigned long long)logical);
-	return;
+	return 0;
+}
 
-uncorrectable:
-	kfree(bbio);
-	spin_lock(&sdev->stat_lock);
-	++sdev->stat.uncorrectable_errors;
-	spin_unlock(&sdev->stat_lock);
+static int scrub_setup_recheck_block(struct scrub_dev *sdev,
+				     struct btrfs_mapping_tree *map_tree,
+				     u64 length, u64 logical,
+				     struct scrub_block *sblocks_for_recheck)
+{
+	int page_index;
+	int mirror_index;
+	int ret;
+
+	/*
+	 * note: the three members sdev, ref_count and outstanding_pages
+	 * are not used (and not set) in the blocks that are used for
+	 * the recheck procedure
+	 */
+
+	page_index = 0;
+	while (length > 0) {
+		u64 sublen = min_t(u64, length, PAGE_SIZE);
+		u64 mapped_length = sublen;
+		struct btrfs_bio *bbio = NULL;
+
+		/*
+		 * with a length of PAGE_SIZE, each returned stripe
+		 * represents one mirror
+		 */
+		ret = btrfs_map_block(map_tree, WRITE, logical, &mapped_length,
+				      &bbio, 0);
+		if (ret || !bbio || mapped_length < sublen) {
+			kfree(bbio);
+			return -EIO;
+		}
 
-	printk_ratelimited(KERN_ERR "btrfs: unable to fixup (regular) error at "
-				"logical %llu\n", (unsigned long long)logical);
+		BUG_ON(page_index >= SCRUB_PAGES_PER_BIO);
+		for (mirror_index = 0; mirror_index < (int)bbio->num_stripes;
+		     mirror_index++) {
+			struct scrub_block *sblock;
+			struct scrub_page *page;
+
+			if (mirror_index >= BTRFS_MAX_MIRRORS)
+				continue;
+
+			sblock = sblocks_for_recheck + mirror_index;
+			page = sblock->pagev + page_index;
+			page->logical = logical;
+			page->physical = bbio->stripes[mirror_index].physical;
+			/* for missing devices, bdev is NULL */
+			page->bdev = bbio->stripes[mirror_index].dev->bdev;
+			page->mirror_num = mirror_index + 1;
+			page->page = alloc_page(GFP_NOFS);
+			if (!page->page) {
+				spin_lock(&sdev->stat_lock);
+				sdev->stat.malloc_errors++;
+				spin_unlock(&sdev->stat_lock);
+				return -ENOMEM;
+			}
+			sblock->page_count++;
+		}
+		kfree(bbio);
+		length -= sublen;
+		logical += sublen;
+		page_index++;
+	}
+
+	return 0;
 }
 
-static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
-			 struct page *page)
+/*
+ * this function will check the on disk data for checksum errors, header
+ * errors and read I/O errors. If any I/O errors happen, the exact pages
+ * which are errored are marked as being bad. The goal is to enable scrub
+ * to take those pages that are not errored from all the mirrors so that
+ * the pages that are errored in the just handled mirror can be repaired.
+ */
+static int scrub_recheck_block(struct btrfs_fs_info *fs_info,
+			       struct scrub_block *sblock, int is_metadata,
+			       int have_csum, u8 *csum, u64 generation,
+			       u16 csum_size)
 {
-	struct bio *bio = NULL;
-	int ret;
-	DECLARE_COMPLETION_ONSTACK(complete);
+	int page_num;
 
-	bio = bio_alloc(GFP_NOFS, 1);
-	bio->bi_bdev = bdev;
-	bio->bi_sector = sector;
-	bio_add_page(bio, page, PAGE_SIZE, 0);
-	bio->bi_end_io = scrub_fixup_end_io;
-	bio->bi_private = &complete;
-	btrfsic_submit_bio(rw, bio);
+	sblock->no_io_error_seen = 1;
+	sblock->header_error = 0;
+	sblock->checksum_error = 0;
 
-	/* this will also unplug the queue */
-	wait_for_completion(&complete);
+	for (page_num = 0; page_num < sblock->page_count; page_num++) {
+		struct bio *bio;
+		int ret;
+		struct scrub_page *page = sblock->pagev + page_num;
+		DECLARE_COMPLETION_ONSTACK(complete);
 
-	ret = !test_bit(BIO_UPTODATE, &bio->bi_flags);
-	bio_put(bio);
-	return ret;
+		if (page->bdev == NULL) {
+			page->io_error = 1;
+			sblock->no_io_error_seen = 0;
+			continue;
+		}
+
+		BUG_ON(!page->page);
+		bio = bio_alloc(GFP_NOFS, 1);
+		if (!bio)
+			return -EIO;
+		bio->bi_bdev = page->bdev;
+		bio->bi_sector = page->physical >> 9;
+		bio->bi_end_io = scrub_complete_bio_end_io;
+		bio->bi_private = &complete;
+
+		ret = bio_add_page(bio, page->page, PAGE_SIZE, 0);
+		if (PAGE_SIZE != ret) {
+			bio_put(bio);
+			return -EIO;
+		}
+		btrfsic_submit_bio(READ, bio);
+
+		/* this will also unplug the queue */
+		wait_for_completion(&complete);
+
+		page->io_error = !test_bit(BIO_UPTODATE, &bio->bi_flags);
+		if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
+			sblock->no_io_error_seen = 0;
+		bio_put(bio);
+	}
+
+	if (sblock->no_io_error_seen)
+		scrub_recheck_block_checksum(fs_info, sblock, is_metadata,
+					     have_csum, csum, generation,
+					     csum_size);
+
+	return 0;
 }
 
-static void scrub_bio_end_io(struct bio *bio, int err)
+static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info,
+					 struct scrub_block *sblock,
+					 int is_metadata, int have_csum,
+					 const u8 *csum, u64 generation,
+					 u16 csum_size)
 {
-	struct scrub_bio *sbio = bio->bi_private;
-	struct scrub_dev *sdev = sbio->sdev;
-	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+	int page_num;
+	u8 calculated_csum[BTRFS_CSUM_SIZE];
+	u32 crc = ~(u32)0;
+	struct btrfs_root *root = fs_info->extent_root;
+	void *mapped_buffer;
+
+	BUG_ON(!sblock->pagev[0].page);
+	if (is_metadata) {
+		struct btrfs_header *h;
+
+		mapped_buffer = kmap_atomic(sblock->pagev[0].page);
+		h = (struct btrfs_header *)mapped_buffer;
+
+		if (sblock->pagev[0].logical != le64_to_cpu(h->bytenr) ||
+		    generation != le64_to_cpu(h->generation) ||
+		    memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE) ||
+		    memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
+			   BTRFS_UUID_SIZE))
+			sblock->header_error = 1;
+		csum = h->csum;
+	} else {
+		if (!have_csum)
+			return;
 
-	sbio->err = err;
-	sbio->bio = bio;
+		mapped_buffer = kmap_atomic(sblock->pagev[0].page);
+	}
 
-	btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work);
+	for (page_num = 0;;) {
+		if (page_num == 0 && is_metadata)
+			crc = btrfs_csum_data(root,
+				((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE,
+				crc, PAGE_SIZE - BTRFS_CSUM_SIZE);
+		else
+			crc = btrfs_csum_data(root, mapped_buffer, crc,
+					      PAGE_SIZE);
+
+		kunmap_atomic(mapped_buffer);
+		page_num++;
+		if (page_num >= sblock->page_count)
+			break;
+		BUG_ON(!sblock->pagev[page_num].page);
+
+		mapped_buffer = kmap_atomic(sblock->pagev[page_num].page);
+	}
+
+	btrfs_csum_final(crc, calculated_csum);
+	if (memcmp(calculated_csum, csum, csum_size))
+		sblock->checksum_error = 1;
 }
 
-static void scrub_checksum(struct btrfs_work *work)
+static void scrub_complete_bio_end_io(struct bio *bio, int err)
 {
-	struct scrub_bio *sbio = container_of(work, struct scrub_bio, work);
-	struct scrub_dev *sdev = sbio->sdev;
-	struct page *page;
-	void *buffer;
-	int i;
-	u64 flags;
-	u64 logical;
-	int ret;
+	complete((struct completion *)bio->bi_private);
+}
 
-	if (sbio->err) {
-		ret = 0;
-		for (i = 0; i < sbio->count; ++i)
-			ret |= scrub_recheck_error(sbio, i);
-		if (!ret) {
-			spin_lock(&sdev->stat_lock);
-			++sdev->stat.unverified_errors;
-			spin_unlock(&sdev->stat_lock);
-		}
+static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad,
+					     struct scrub_block *sblock_good,
+					     int force_write)
+{
+	int page_num;
+	int ret = 0;
 
-		sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1);
-		sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
-		sbio->bio->bi_phys_segments = 0;
-		sbio->bio->bi_idx = 0;
+	for (page_num = 0; page_num < sblock_bad->page_count; page_num++) {
+		int ret_sub;
 
-		for (i = 0; i < sbio->count; i++) {
-			struct bio_vec *bi;
-			bi = &sbio->bio->bi_io_vec[i];
-			bi->bv_offset = 0;
-			bi->bv_len = PAGE_SIZE;
-		}
-		goto out;
+		ret_sub = scrub_repair_page_from_good_copy(sblock_bad,
+							   sblock_good,
+							   page_num,
+							   force_write);
+		if (ret_sub)
+			ret = ret_sub;
 	}
-	for (i = 0; i < sbio->count; ++i) {
-		page = sbio->bio->bi_io_vec[i].bv_page;
-		buffer = kmap_atomic(page, KM_USER0);
-		flags = sbio->spag[i].flags;
-		logical = sbio->logical + i * PAGE_SIZE;
-		ret = 0;
-		if (flags & BTRFS_EXTENT_FLAG_DATA) {
-			ret = scrub_checksum_data(sdev, sbio->spag + i, buffer);
-		} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
-			ret = scrub_checksum_tree_block(sdev, sbio->spag + i,
-							logical, buffer);
-		} else if (flags & BTRFS_EXTENT_FLAG_SUPER) {
-			BUG_ON(i);
-			(void)scrub_checksum_super(sbio, buffer);
-		} else {
-			WARN_ON(1);
-		}
-		kunmap_atomic(buffer, KM_USER0);
-		if (ret) {
-			ret = scrub_recheck_error(sbio, i);
-			if (!ret) {
-				spin_lock(&sdev->stat_lock);
-				++sdev->stat.unverified_errors;
-				spin_unlock(&sdev->stat_lock);
-			}
+
+	return ret;
+}
+
+static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
+					    struct scrub_block *sblock_good,
+					    int page_num, int force_write)
+{
+	struct scrub_page *page_bad = sblock_bad->pagev + page_num;
+	struct scrub_page *page_good = sblock_good->pagev + page_num;
+
+	BUG_ON(sblock_bad->pagev[page_num].page == NULL);
+	BUG_ON(sblock_good->pagev[page_num].page == NULL);
+	if (force_write || sblock_bad->header_error ||
+	    sblock_bad->checksum_error || page_bad->io_error) {
+		struct bio *bio;
+		int ret;
+		DECLARE_COMPLETION_ONSTACK(complete);
+
+		bio = bio_alloc(GFP_NOFS, 1);
+		if (!bio)
+			return -EIO;
+		bio->bi_bdev = page_bad->bdev;
+		bio->bi_sector = page_bad->physical >> 9;
+		bio->bi_end_io = scrub_complete_bio_end_io;
+		bio->bi_private = &complete;
+
+		ret = bio_add_page(bio, page_good->page, PAGE_SIZE, 0);
+		if (PAGE_SIZE != ret) {
+			bio_put(bio);
+			return -EIO;
 		}
+		btrfsic_submit_bio(WRITE, bio);
+
+		/* this will also unplug the queue */
+		wait_for_completion(&complete);
+		bio_put(bio);
 	}
 
-out:
-	scrub_free_bio(sbio->bio);
-	sbio->bio = NULL;
-	spin_lock(&sdev->list_lock);
-	sbio->next_free = sdev->first_free;
-	sdev->first_free = sbio->index;
-	spin_unlock(&sdev->list_lock);
-	atomic_dec(&sdev->in_flight);
-	wake_up(&sdev->list_wait);
+	return 0;
 }
 
-static int scrub_checksum_data(struct scrub_dev *sdev,
-			       struct scrub_page *spag, void *buffer)
+static void scrub_checksum(struct scrub_block *sblock)
 {
+	u64 flags;
+	int ret;
+
+	BUG_ON(sblock->page_count < 1);
+	flags = sblock->pagev[0].flags;
+	ret = 0;
+	if (flags & BTRFS_EXTENT_FLAG_DATA)
+		ret = scrub_checksum_data(sblock);
+	else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
+		ret = scrub_checksum_tree_block(sblock);
+	else if (flags & BTRFS_EXTENT_FLAG_SUPER)
+		(void)scrub_checksum_super(sblock);
+	else
+		WARN_ON(1);
+	if (ret)
+		scrub_handle_errored_block(sblock);
+}
+
+static int scrub_checksum_data(struct scrub_block *sblock)
+{
+	struct scrub_dev *sdev = sblock->sdev;
 	u8 csum[BTRFS_CSUM_SIZE];
+	u8 *on_disk_csum;
+	struct page *page;
+	void *buffer;
 	u32 crc = ~(u32)0;
 	int fail = 0;
 	struct btrfs_root *root = sdev->dev->dev_root;
+	u64 len;
+	int index;
 
-	if (!spag->have_csum)
+	BUG_ON(sblock->page_count < 1);
+	if (!sblock->pagev[0].have_csum)
 		return 0;
 
-	crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE);
+	on_disk_csum = sblock->pagev[0].csum;
+	page = sblock->pagev[0].page;
+	buffer = kmap_atomic(page);
+
+	len = sdev->sectorsize;
+	index = 0;
+	for (;;) {
+		u64 l = min_t(u64, len, PAGE_SIZE);
+
+		crc = btrfs_csum_data(root, buffer, crc, l);
+		kunmap_atomic(buffer);
+		len -= l;
+		if (len == 0)
+			break;
+		index++;
+		BUG_ON(index >= sblock->page_count);
+		BUG_ON(!sblock->pagev[index].page);
+		page = sblock->pagev[index].page;
+		buffer = kmap_atomic(page);
+	}
+
 	btrfs_csum_final(crc, csum);
-	if (memcmp(csum, spag->csum, sdev->csum_size))
+	if (memcmp(csum, on_disk_csum, sdev->csum_size))
 		fail = 1;
 
-	spin_lock(&sdev->stat_lock);
-	++sdev->stat.data_extents_scrubbed;
-	sdev->stat.data_bytes_scrubbed += PAGE_SIZE;
-	if (fail)
-		++sdev->stat.csum_errors;
-	spin_unlock(&sdev->stat_lock);
-
 	return fail;
 }
 
-static int scrub_checksum_tree_block(struct scrub_dev *sdev,
-				     struct scrub_page *spag, u64 logical,
-				     void *buffer)
+static int scrub_checksum_tree_block(struct scrub_block *sblock)
 {
+	struct scrub_dev *sdev = sblock->sdev;
 	struct btrfs_header *h;
 	struct btrfs_root *root = sdev->dev->dev_root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	u8 csum[BTRFS_CSUM_SIZE];
+	u8 calculated_csum[BTRFS_CSUM_SIZE];
+	u8 on_disk_csum[BTRFS_CSUM_SIZE];
+	struct page *page;
+	void *mapped_buffer;
+	u64 mapped_size;
+	void *p;
 	u32 crc = ~(u32)0;
 	int fail = 0;
 	int crc_fail = 0;
+	u64 len;
+	int index;
+
+	BUG_ON(sblock->page_count < 1);
+	page = sblock->pagev[0].page;
+	mapped_buffer = kmap_atomic(page);
+	h = (struct btrfs_header *)mapped_buffer;
+	memcpy(on_disk_csum, h->csum, sdev->csum_size);
 
 	/*
 	 * we don't use the getter functions here, as we
 	 * a) don't have an extent buffer and
 	 * b) the page is already kmapped
 	 */
-	h = (struct btrfs_header *)buffer;
 
-	if (logical != le64_to_cpu(h->bytenr))
+	if (sblock->pagev[0].logical != le64_to_cpu(h->bytenr))
 		++fail;
 
-	if (spag->generation != le64_to_cpu(h->generation))
+	if (sblock->pagev[0].generation != le64_to_cpu(h->generation))
 		++fail;
 
 	if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
@@ -887,51 +1310,90 @@ static int scrub_checksum_tree_block(struct scrub_dev *sdev,
 		   BTRFS_UUID_SIZE))
 		++fail;
 
-	crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
-			      PAGE_SIZE - BTRFS_CSUM_SIZE);
-	btrfs_csum_final(crc, csum);
-	if (memcmp(csum, h->csum, sdev->csum_size))
-		++crc_fail;
+	BUG_ON(sdev->nodesize != sdev->leafsize);
+	len = sdev->nodesize - BTRFS_CSUM_SIZE;
+	mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE;
+	p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE;
+	index = 0;
+	for (;;) {
+		u64 l = min_t(u64, len, mapped_size);
 
-	spin_lock(&sdev->stat_lock);
-	++sdev->stat.tree_extents_scrubbed;
-	sdev->stat.tree_bytes_scrubbed += PAGE_SIZE;
-	if (crc_fail)
-		++sdev->stat.csum_errors;
-	if (fail)
-		++sdev->stat.verify_errors;
-	spin_unlock(&sdev->stat_lock);
+		crc = btrfs_csum_data(root, p, crc, l);
+		kunmap_atomic(mapped_buffer);
+		len -= l;
+		if (len == 0)
+			break;
+		index++;
+		BUG_ON(index >= sblock->page_count);
+		BUG_ON(!sblock->pagev[index].page);
+		page = sblock->pagev[index].page;
+		mapped_buffer = kmap_atomic(page);
+		mapped_size = PAGE_SIZE;
+		p = mapped_buffer;
+	}
+
+	btrfs_csum_final(crc, calculated_csum);
+	if (memcmp(calculated_csum, on_disk_csum, sdev->csum_size))
+		++crc_fail;
 
 	return fail || crc_fail;
 }
 
-static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer)
+static int scrub_checksum_super(struct scrub_block *sblock)
 {
 	struct btrfs_super_block *s;
-	u64 logical;
-	struct scrub_dev *sdev = sbio->sdev;
+	struct scrub_dev *sdev = sblock->sdev;
 	struct btrfs_root *root = sdev->dev->dev_root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	u8 csum[BTRFS_CSUM_SIZE];
+	u8 calculated_csum[BTRFS_CSUM_SIZE];
+	u8 on_disk_csum[BTRFS_CSUM_SIZE];
+	struct page *page;
+	void *mapped_buffer;
+	u64 mapped_size;
+	void *p;
 	u32 crc = ~(u32)0;
 	int fail = 0;
+	u64 len;
+	int index;
 
-	s = (struct btrfs_super_block *)buffer;
-	logical = sbio->logical;
+	BUG_ON(sblock->page_count < 1);
+	page = sblock->pagev[0].page;
+	mapped_buffer = kmap_atomic(page);
+	s = (struct btrfs_super_block *)mapped_buffer;
+	memcpy(on_disk_csum, s->csum, sdev->csum_size);
 
-	if (logical != le64_to_cpu(s->bytenr))
+	if (sblock->pagev[0].logical != le64_to_cpu(s->bytenr))
 		++fail;
 
-	if (sbio->spag[0].generation != le64_to_cpu(s->generation))
+	if (sblock->pagev[0].generation != le64_to_cpu(s->generation))
 		++fail;
 
 	if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
 		++fail;
 
-	crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
-			      PAGE_SIZE - BTRFS_CSUM_SIZE);
-	btrfs_csum_final(crc, csum);
-	if (memcmp(csum, s->csum, sbio->sdev->csum_size))
+	len = BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE;
+	mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE;
+	p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE;
+	index = 0;
+	for (;;) {
+		u64 l = min_t(u64, len, mapped_size);
+
+		crc = btrfs_csum_data(root, p, crc, l);
+		kunmap_atomic(mapped_buffer);
+		len -= l;
+		if (len == 0)
+			break;
+		index++;
+		BUG_ON(index >= sblock->page_count);
+		BUG_ON(!sblock->pagev[index].page);
+		page = sblock->pagev[index].page;
+		mapped_buffer = kmap_atomic(page);
+		mapped_size = PAGE_SIZE;
+		p = mapped_buffer;
+	}
+
+	btrfs_csum_final(crc, calculated_csum);
+	if (memcmp(calculated_csum, on_disk_csum, sdev->csum_size))
 		++fail;
 
 	if (fail) {
@@ -948,29 +1410,42 @@ static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer)
 	return fail;
 }
 
-static int scrub_submit(struct scrub_dev *sdev)
+static void scrub_block_get(struct scrub_block *sblock)
+{
+	atomic_inc(&sblock->ref_count);
+}
+
+static void scrub_block_put(struct scrub_block *sblock)
+{
+	if (atomic_dec_and_test(&sblock->ref_count)) {
+		int i;
+
+		for (i = 0; i < sblock->page_count; i++)
+			if (sblock->pagev[i].page)
+				__free_page(sblock->pagev[i].page);
+		kfree(sblock);
+	}
+}
+
+static void scrub_submit(struct scrub_dev *sdev)
 {
 	struct scrub_bio *sbio;
 
 	if (sdev->curr == -1)
-		return 0;
+		return;
 
 	sbio = sdev->bios[sdev->curr];
-	sbio->err = 0;
 	sdev->curr = -1;
 	atomic_inc(&sdev->in_flight);
 
 	btrfsic_submit_bio(READ, sbio->bio);
-
-	return 0;
 }
 
-static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len,
-		      u64 physical, u64 flags, u64 gen, int mirror_num,
-		      u8 *csum, int force)
+static int scrub_add_page_to_bio(struct scrub_dev *sdev,
+				 struct scrub_page *spage)
 {
+	struct scrub_block *sblock = spage->sblock;
 	struct scrub_bio *sbio;
-	struct page *page;
 	int ret;
 
 again:
@@ -983,7 +1458,7 @@ again:
 		if (sdev->curr != -1) {
 			sdev->first_free = sdev->bios[sdev->curr]->next_free;
 			sdev->bios[sdev->curr]->next_free = -1;
-			sdev->bios[sdev->curr]->count = 0;
+			sdev->bios[sdev->curr]->page_count = 0;
 			spin_unlock(&sdev->list_lock);
 		} else {
 			spin_unlock(&sdev->list_lock);
@@ -991,62 +1466,200 @@ again:
 		}
 	}
 	sbio = sdev->bios[sdev->curr];
-	if (sbio->count == 0) {
+	if (sbio->page_count == 0) {
 		struct bio *bio;
 
-		sbio->physical = physical;
-		sbio->logical = logical;
-		bio = bio_alloc(GFP_NOFS, SCRUB_PAGES_PER_BIO);
-		if (!bio)
-			return -ENOMEM;
+		sbio->physical = spage->physical;
+		sbio->logical = spage->logical;
+		bio = sbio->bio;
+		if (!bio) {
+			bio = bio_alloc(GFP_NOFS, sdev->pages_per_bio);
+			if (!bio)
+				return -ENOMEM;
+			sbio->bio = bio;
+		}
 
 		bio->bi_private = sbio;
 		bio->bi_end_io = scrub_bio_end_io;
 		bio->bi_bdev = sdev->dev->bdev;
-		bio->bi_sector = sbio->physical >> 9;
+		bio->bi_sector = spage->physical >> 9;
 		sbio->err = 0;
-		sbio->bio = bio;
-	} else if (sbio->physical + sbio->count * PAGE_SIZE != physical ||
-		   sbio->logical + sbio->count * PAGE_SIZE != logical) {
-		ret = scrub_submit(sdev);
-		if (ret)
-			return ret;
+	} else if (sbio->physical + sbio->page_count * PAGE_SIZE !=
+		   spage->physical ||
+		   sbio->logical + sbio->page_count * PAGE_SIZE !=
+		   spage->logical) {
+		scrub_submit(sdev);
 		goto again;
 	}
-	sbio->spag[sbio->count].flags = flags;
-	sbio->spag[sbio->count].generation = gen;
-	sbio->spag[sbio->count].have_csum = 0;
-	sbio->spag[sbio->count].mirror_num = mirror_num;
-
-	page = alloc_page(GFP_NOFS);
-	if (!page)
-		return -ENOMEM;
 
-	ret = bio_add_page(sbio->bio, page, PAGE_SIZE, 0);
-	if (!ret) {
-		__free_page(page);
-		ret = scrub_submit(sdev);
-		if (ret)
-			return ret;
+	sbio->pagev[sbio->page_count] = spage;
+	ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0);
+	if (ret != PAGE_SIZE) {
+		if (sbio->page_count < 1) {
+			bio_put(sbio->bio);
+			sbio->bio = NULL;
+			return -EIO;
+		}
+		scrub_submit(sdev);
 		goto again;
 	}
 
-	if (csum) {
-		sbio->spag[sbio->count].have_csum = 1;
-		memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size);
+	scrub_block_get(sblock); /* one for the added page */
+	atomic_inc(&sblock->outstanding_pages);
+	sbio->page_count++;
+	if (sbio->page_count == sdev->pages_per_bio)
+		scrub_submit(sdev);
+
+	return 0;
+}
+
+static int scrub_pages(struct scrub_dev *sdev, u64 logical, u64 len,
+		       u64 physical, u64 flags, u64 gen, int mirror_num,
+		       u8 *csum, int force)
+{
+	struct scrub_block *sblock;
+	int index;
+
+	sblock = kzalloc(sizeof(*sblock), GFP_NOFS);
+	if (!sblock) {
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.malloc_errors++;
+		spin_unlock(&sdev->stat_lock);
+		return -ENOMEM;
 	}
-	++sbio->count;
-	if (sbio->count == SCRUB_PAGES_PER_BIO || force) {
+
+	/* one ref inside this function, plus one for each page later on */
+	atomic_set(&sblock->ref_count, 1);
+	sblock->sdev = sdev;
+	sblock->no_io_error_seen = 1;
+
+	for (index = 0; len > 0; index++) {
+		struct scrub_page *spage = sblock->pagev + index;
+		u64 l = min_t(u64, len, PAGE_SIZE);
+
+		BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK);
+		spage->page = alloc_page(GFP_NOFS);
+		if (!spage->page) {
+			spin_lock(&sdev->stat_lock);
+			sdev->stat.malloc_errors++;
+			spin_unlock(&sdev->stat_lock);
+			while (index > 0) {
+				index--;
+				__free_page(sblock->pagev[index].page);
+			}
+			kfree(sblock);
+			return -ENOMEM;
+		}
+		spage->sblock = sblock;
+		spage->bdev = sdev->dev->bdev;
+		spage->flags = flags;
+		spage->generation = gen;
+		spage->logical = logical;
+		spage->physical = physical;
+		spage->mirror_num = mirror_num;
+		if (csum) {
+			spage->have_csum = 1;
+			memcpy(spage->csum, csum, sdev->csum_size);
+		} else {
+			spage->have_csum = 0;
+		}
+		sblock->page_count++;
+		len -= l;
+		logical += l;
+		physical += l;
+	}
+
+	BUG_ON(sblock->page_count == 0);
+	for (index = 0; index < sblock->page_count; index++) {
+		struct scrub_page *spage = sblock->pagev + index;
 		int ret;
 
-		ret = scrub_submit(sdev);
-		if (ret)
+		ret = scrub_add_page_to_bio(sdev, spage);
+		if (ret) {
+			scrub_block_put(sblock);
 			return ret;
+		}
 	}
 
+	if (force)
+		scrub_submit(sdev);
+
+	/* last one frees, either here or in bio completion for last page */
+	scrub_block_put(sblock);
 	return 0;
 }
 
+static void scrub_bio_end_io(struct bio *bio, int err)
+{
+	struct scrub_bio *sbio = bio->bi_private;
+	struct scrub_dev *sdev = sbio->sdev;
+	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+
+	sbio->err = err;
+	sbio->bio = bio;
+
+	btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work);
+}
+
+static void scrub_bio_end_io_worker(struct btrfs_work *work)
+{
+	struct scrub_bio *sbio = container_of(work, struct scrub_bio, work);
+	struct scrub_dev *sdev = sbio->sdev;
+	int i;
+
+	BUG_ON(sbio->page_count > SCRUB_PAGES_PER_BIO);
+	if (sbio->err) {
+		for (i = 0; i < sbio->page_count; i++) {
+			struct scrub_page *spage = sbio->pagev[i];
+
+			spage->io_error = 1;
+			spage->sblock->no_io_error_seen = 0;
+		}
+	}
+
+	/* now complete the scrub_block items that have all pages completed */
+	for (i = 0; i < sbio->page_count; i++) {
+		struct scrub_page *spage = sbio->pagev[i];
+		struct scrub_block *sblock = spage->sblock;
+
+		if (atomic_dec_and_test(&sblock->outstanding_pages))
+			scrub_block_complete(sblock);
+		scrub_block_put(sblock);
+	}
+
+	if (sbio->err) {
+		/* what is this good for??? */
+		sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1);
+		sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
+		sbio->bio->bi_phys_segments = 0;
+		sbio->bio->bi_idx = 0;
+
+		for (i = 0; i < sbio->page_count; i++) {
+			struct bio_vec *bi;
+			bi = &sbio->bio->bi_io_vec[i];
+			bi->bv_offset = 0;
+			bi->bv_len = PAGE_SIZE;
+		}
+	}
+
+	bio_put(sbio->bio);
+	sbio->bio = NULL;
+	spin_lock(&sdev->list_lock);
+	sbio->next_free = sdev->first_free;
+	sdev->first_free = sbio->index;
+	spin_unlock(&sdev->list_lock);
+	atomic_dec(&sdev->in_flight);
+	wake_up(&sdev->list_wait);
+}
+
+static void scrub_block_complete(struct scrub_block *sblock)
+{
+	if (!sblock->no_io_error_seen)
+		scrub_handle_errored_block(sblock);
+	else
+		scrub_checksum(sblock);
+}
+
 static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
 			   u8 *csum)
 {
@@ -1054,7 +1667,6 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
 	int ret = 0;
 	unsigned long i;
 	unsigned long num_sectors;
-	u32 sectorsize = sdev->dev->dev_root->sectorsize;
 
 	while (!list_empty(&sdev->csum_list)) {
 		sum = list_first_entry(&sdev->csum_list,
@@ -1072,7 +1684,7 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
 	if (!sum)
 		return 0;
 
-	num_sectors = sum->len / sectorsize;
+	num_sectors = sum->len / sdev->sectorsize;
 	for (i = 0; i < num_sectors; ++i) {
 		if (sum->sums[i].bytenr == logical) {
 			memcpy(csum, &sum->sums[i].sum, sdev->csum_size);
@@ -1093,9 +1705,28 @@ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len,
 {
 	int ret;
 	u8 csum[BTRFS_CSUM_SIZE];
+	u32 blocksize;
+
+	if (flags & BTRFS_EXTENT_FLAG_DATA) {
+		blocksize = sdev->sectorsize;
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.data_extents_scrubbed++;
+		sdev->stat.data_bytes_scrubbed += len;
+		spin_unlock(&sdev->stat_lock);
+	} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+		BUG_ON(sdev->nodesize != sdev->leafsize);
+		blocksize = sdev->nodesize;
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.tree_extents_scrubbed++;
+		sdev->stat.tree_bytes_scrubbed += len;
+		spin_unlock(&sdev->stat_lock);
+	} else {
+		blocksize = sdev->sectorsize;
+		BUG_ON(1);
+	}
 
 	while (len) {
-		u64 l = min_t(u64, len, PAGE_SIZE);
+		u64 l = min_t(u64, len, blocksize);
 		int have_csum = 0;
 
 		if (flags & BTRFS_EXTENT_FLAG_DATA) {
@@ -1104,8 +1735,8 @@ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len,
 			if (have_csum == 0)
 				++sdev->stat.no_csum;
 		}
-		ret = scrub_page(sdev, logical, l, physical, flags, gen,
-				 mirror_num, have_csum ? csum : NULL, 0);
+		ret = scrub_pages(sdev, logical, l, physical, flags, gen,
+				  mirror_num, have_csum ? csum : NULL, 0);
 		if (ret)
 			return ret;
 		len -= l;
@@ -1170,6 +1801,11 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
 	if (!path)
 		return -ENOMEM;
 
+	/*
+	 * work on commit root. The related disk blocks are static as
+	 * long as COW is applied. This means, it is save to rewrite
+	 * them to repair disk errors without any race conditions
+	 */
 	path->search_commit_root = 1;
 	path->skip_locking = 1;
 
@@ -1516,15 +2152,18 @@ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev)
 	struct btrfs_device *device = sdev->dev;
 	struct btrfs_root *root = device->dev_root;
 
+	if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
+		return -EIO;
+
 	gen = root->fs_info->last_trans_committed;
 
 	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
 		bytenr = btrfs_sb_offset(i);
-		if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes)
+		if (bytenr + BTRFS_SUPER_INFO_SIZE > device->total_bytes)
 			break;
 
-		ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr,
-				 BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1);
+		ret = scrub_pages(sdev, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr,
+				     BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1);
 		if (ret)
 			return ret;
 	}
@@ -1583,10 +2222,30 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
 	/*
 	 * check some assumptions
 	 */
-	if (root->sectorsize != PAGE_SIZE ||
-	    root->sectorsize != root->leafsize ||
-	    root->sectorsize != root->nodesize) {
-		printk(KERN_ERR "btrfs_scrub: size assumptions fail\n");
+	if (root->nodesize != root->leafsize) {
+		printk(KERN_ERR
+		       "btrfs_scrub: size assumption nodesize == leafsize (%d == %d) fails\n",
+		       root->nodesize, root->leafsize);
+		return -EINVAL;
+	}
+
+	if (root->nodesize > BTRFS_STRIPE_LEN) {
+		/*
+		 * in this case scrub is unable to calculate the checksum
+		 * the way scrub is implemented. Do not handle this
+		 * situation at all because it won't ever happen.
+		 */
+		printk(KERN_ERR
+		       "btrfs_scrub: size assumption nodesize <= BTRFS_STRIPE_LEN (%d <= %d) fails\n",
+		       root->nodesize, BTRFS_STRIPE_LEN);
+		return -EINVAL;
+	}
+
+	if (root->sectorsize != PAGE_SIZE) {
+		/* not supported for data w/o checksums */
+		printk(KERN_ERR
+		       "btrfs_scrub: size assumption sectorsize != PAGE_SIZE (%d != %lld) fails\n",
+		       root->sectorsize, (unsigned long long)PAGE_SIZE);
 		return -EINVAL;
 	}
 
@@ -1656,7 +2315,7 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
 	return ret;
 }
 
-int btrfs_scrub_pause(struct btrfs_root *root)
+void btrfs_scrub_pause(struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
@@ -1671,34 +2330,28 @@ int btrfs_scrub_pause(struct btrfs_root *root)
 		mutex_lock(&fs_info->scrub_lock);
 	}
 	mutex_unlock(&fs_info->scrub_lock);
-
-	return 0;
 }
 
-int btrfs_scrub_continue(struct btrfs_root *root)
+void btrfs_scrub_continue(struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	atomic_dec(&fs_info->scrub_pause_req);
 	wake_up(&fs_info->scrub_pause_wait);
-	return 0;
 }
 
-int btrfs_scrub_pause_super(struct btrfs_root *root)
+void btrfs_scrub_pause_super(struct btrfs_root *root)
 {
 	down_write(&root->fs_info->scrub_super_lock);
-	return 0;
 }
 
-int btrfs_scrub_continue_super(struct btrfs_root *root)
+void btrfs_scrub_continue_super(struct btrfs_root *root)
 {
 	up_write(&root->fs_info->scrub_super_lock);
-	return 0;
 }
 
-int btrfs_scrub_cancel(struct btrfs_root *root)
+int __btrfs_scrub_cancel(struct btrfs_fs_info *fs_info)
 {
-	struct btrfs_fs_info *fs_info = root->fs_info;
 
 	mutex_lock(&fs_info->scrub_lock);
 	if (!atomic_read(&fs_info->scrubs_running)) {
@@ -1719,6 +2372,11 @@ int btrfs_scrub_cancel(struct btrfs_root *root)
 	return 0;
 }
 
+int btrfs_scrub_cancel(struct btrfs_root *root)
+{
+	return __btrfs_scrub_cancel(root->fs_info);
+}
+