Merge branch 'master' into next

117 files changed, 3255 insertions, 1204 deletions

diff --git a/fs/Makefile b/fs/Makefile
index 38bc735c67a..dc20db34867 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -69,10 +69,12 @@ obj-$(CONFIG_DLM)		+= dlm/
 # Do not add any filesystems before this line
 obj-$(CONFIG_REISERFS_FS)	+= reiserfs/
 obj-$(CONFIG_EXT3_FS)		+= ext3/ # Before ext2 so root fs can be ext3
-obj-$(CONFIG_EXT4_FS)		+= ext4/ # Before ext2 so root fs can be ext4
+obj-$(CONFIG_EXT2_FS)		+= ext2/
+# We place ext4 after ext2 so plain ext2 root fs's are mounted using ext2
+# unless explicitly requested by rootfstype
+obj-$(CONFIG_EXT4_FS)		+= ext4/
 obj-$(CONFIG_JBD)		+= jbd/
 obj-$(CONFIG_JBD2)		+= jbd2/
-obj-$(CONFIG_EXT2_FS)		+= ext2/
 obj-$(CONFIG_CRAMFS)		+= cramfs/
 obj-$(CONFIG_SQUASHFS)		+= squashfs/
 obj-y				+= ramfs/
diff --git a/fs/aio.c b/fs/aio.c
index 8fa77e23394..76da1253795 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -443,7 +443,7 @@ static struct kiocb *__aio_get_req(struct kioctx *ctx)
 	req->private = NULL;
 	req->ki_iovec = NULL;
 	INIT_LIST_HEAD(&req->ki_run_list);
-	req->ki_eventfd = ERR_PTR(-EINVAL);
+	req->ki_eventfd = NULL;
 
 	/* Check if the completion queue has enough free space to
 	 * accept an event from this io.
@@ -485,8 +485,6 @@ static inline void really_put_req(struct kioctx *ctx, struct kiocb *req)
 {
 	assert_spin_locked(&ctx->ctx_lock);
 
-	if (!IS_ERR(req->ki_eventfd))
-		fput(req->ki_eventfd);
 	if (req->ki_dtor)
 		req->ki_dtor(req);
 	if (req->ki_iovec != &req->ki_inline_vec)
@@ -508,8 +506,11 @@ static void aio_fput_routine(struct work_struct *data)
 		list_del(&req->ki_list);
 		spin_unlock_irq(&fput_lock);
 
-		/* Complete the fput */
-		__fput(req->ki_filp);
+		/* Complete the fput(s) */
+		if (req->ki_filp != NULL)
+			__fput(req->ki_filp);
+		if (req->ki_eventfd != NULL)
+			__fput(req->ki_eventfd);
 
 		/* Link the iocb into the context's free list */
 		spin_lock_irq(&ctx->ctx_lock);
@@ -527,12 +528,14 @@ static void aio_fput_routine(struct work_struct *data)
  */
 static int __aio_put_req(struct kioctx *ctx, struct kiocb *req)
 {
+	int schedule_putreq = 0;
+
 	dprintk(KERN_DEBUG "aio_put(%p): f_count=%ld\n",
 		req, atomic_long_read(&req->ki_filp->f_count));
 
 	assert_spin_locked(&ctx->ctx_lock);
 
-	req->ki_users --;
+	req->ki_users--;
 	BUG_ON(req->ki_users < 0);
 	if (likely(req->ki_users))
 		return 0;
@@ -540,10 +543,23 @@ static int __aio_put_req(struct kioctx *ctx, struct kiocb *req)
 	req->ki_cancel = NULL;
 	req->ki_retry = NULL;
 
-	/* Must be done under the lock to serialise against cancellation.
-	 * Call this aio_fput as it duplicates fput via the fput_work.
+	/*
+	 * Try to optimize the aio and eventfd file* puts, by avoiding to
+	 * schedule work in case it is not __fput() time. In normal cases,
+	 * we would not be holding the last reference to the file*, so
+	 * this function will be executed w/out any aio kthread wakeup.
 	 */
-	if (unlikely(atomic_long_dec_and_test(&req->ki_filp->f_count))) {
+	if (unlikely(atomic_long_dec_and_test(&req->ki_filp->f_count)))
+		schedule_putreq++;
+	else
+		req->ki_filp = NULL;
+	if (req->ki_eventfd != NULL) {
+		if (unlikely(atomic_long_dec_and_test(&req->ki_eventfd->f_count)))
+			schedule_putreq++;
+		else
+			req->ki_eventfd = NULL;
+	}
+	if (unlikely(schedule_putreq)) {
 		get_ioctx(ctx);
 		spin_lock(&fput_lock);
 		list_add(&req->ki_list, &fput_head);
@@ -571,7 +587,7 @@ int aio_put_req(struct kiocb *req)
 static struct kioctx *lookup_ioctx(unsigned long ctx_id)
 {
 	struct mm_struct *mm = current->mm;
-	struct kioctx *ctx = NULL;
+	struct kioctx *ctx, *ret = NULL;
 	struct hlist_node *n;
 
 	rcu_read_lock();
@@ -579,12 +595,13 @@ static struct kioctx *lookup_ioctx(unsigned long ctx_id)
 	hlist_for_each_entry_rcu(ctx, n, &mm->ioctx_list, list) {
 		if (ctx->user_id == ctx_id && !ctx->dead) {
 			get_ioctx(ctx);
+			ret = ctx;
 			break;
 		}
 	}
 
 	rcu_read_unlock();
-	return ctx;
+	return ret;
 }
 
 /*
@@ -1009,7 +1026,7 @@ int aio_complete(struct kiocb *iocb, long res, long res2)
 	 * eventfd. The eventfd_signal() function is safe to be called
 	 * from IRQ context.
 	 */
-	if (!IS_ERR(iocb->ki_eventfd))
+	if (iocb->ki_eventfd != NULL)
 		eventfd_signal(iocb->ki_eventfd, 1);
 
 put_rq:
@@ -1608,6 +1625,7 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 		req->ki_eventfd = eventfd_fget((int) iocb->aio_resfd);
 		if (IS_ERR(req->ki_eventfd)) {
 			ret = PTR_ERR(req->ki_eventfd);
+			req->ki_eventfd = NULL;
 			goto out_put_req;
 		}
 	}
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index e3ff2b9e602..33b7235f853 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1208,9 +1208,11 @@ static unsigned long vma_dump_size(struct vm_area_struct *vma,
 	 * check for an ELF header.  If we find one, dump the first page to
 	 * aid in determining what was mapped here.
 	 */
-	if (FILTER(ELF_HEADERS) && vma->vm_file != NULL && vma->vm_pgoff == 0) {
+	if (FILTER(ELF_HEADERS) &&
+	    vma->vm_pgoff == 0 && (vma->vm_flags & VM_READ)) {
 		u32 __user *header = (u32 __user *) vma->vm_start;
 		u32 word;
+		mm_segment_t fs = get_fs();
 		/*
 		 * Doing it this way gets the constant folded by GCC.
 		 */
@@ -1223,7 +1225,15 @@ static unsigned long vma_dump_size(struct vm_area_struct *vma,
 		magic.elfmag[EI_MAG1] = ELFMAG1;
 		magic.elfmag[EI_MAG2] = ELFMAG2;
 		magic.elfmag[EI_MAG3] = ELFMAG3;
-		if (get_user(word, header) == 0 && word == magic.cmp)
+		/*
+		 * Switch to the user "segment" for get_user(),
+		 * then put back what elf_core_dump() had in place.
+		 */
+		set_fs(USER_DS);
+		if (unlikely(get_user(word, header)))
+			word = 0;
+		set_fs(fs);
+		if (word == magic.cmp)
 			return PAGE_SIZE;
 	}
 
diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c
index 549b0144da1..fe2b1aa2464 100644
--- a/fs/bio-integrity.c
+++ b/fs/bio-integrity.c
@@ -685,19 +685,20 @@ EXPORT_SYMBOL(bio_integrity_split);
  * bio_integrity_clone - Callback for cloning bios with integrity metadata
  * @bio:	New bio
  * @bio_src:	Original bio
+ * @gfp_mask:	Memory allocation mask
  * @bs:		bio_set to allocate bip from
  *
  * Description:	Called to allocate a bip when cloning a bio
  */
 int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
-			struct bio_set *bs)
+			gfp_t gfp_mask, struct bio_set *bs)
 {
 	struct bio_integrity_payload *bip_src = bio_src->bi_integrity;
 	struct bio_integrity_payload *bip;
 
 	BUG_ON(bip_src == NULL);
 
-	bip = bio_integrity_alloc_bioset(bio, GFP_NOIO, bip_src->bip_vcnt, bs);
+	bip = bio_integrity_alloc_bioset(bio, gfp_mask, bip_src->bip_vcnt, bs);
 
 	if (bip == NULL)
 		return -EIO;
diff --git a/fs/bio.c b/fs/bio.c
index 062299acbcc..d4f06327c81 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -302,9 +302,10 @@ void bio_init(struct bio *bio)
 struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 {
 	struct bio *bio = NULL;
+	void *uninitialized_var(p);
 
 	if (bs) {
-		void *p = mempool_alloc(bs->bio_pool, gfp_mask);
+		p = mempool_alloc(bs->bio_pool, gfp_mask);
 
 		if (p)
 			bio = p + bs->front_pad;
@@ -329,7 +330,7 @@ struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 			}
 			if (unlikely(!bvl)) {
 				if (bs)
-					mempool_free(bio, bs->bio_pool);
+					mempool_free(p, bs->bio_pool);
 				else
 					kfree(bio);
 				bio = NULL;
@@ -462,10 +463,12 @@ struct bio *bio_clone(struct bio *bio, gfp_t gfp_mask)
 	if (bio_integrity(bio)) {
 		int ret;
 
-		ret = bio_integrity_clone(b, bio, fs_bio_set);
+		ret = bio_integrity_clone(b, bio, gfp_mask, fs_bio_set);
 
-		if (ret < 0)
+		if (ret < 0) {
+			bio_put(b);
 			return NULL;
+		}
 	}
 
 	return b;
diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig
index f8fcf999ea1..7bb3c020e57 100644
--- a/fs/btrfs/Kconfig
+++ b/fs/btrfs/Kconfig
@@ -16,3 +16,16 @@ config BTRFS_FS
 	  module will be called btrfs.
 
 	  If unsure, say N.
+
+config BTRFS_FS_POSIX_ACL
+	bool "Btrfs POSIX Access Control Lists"
+	depends on BTRFS_FS
+	select FS_POSIX_ACL
+	help
+	  POSIX Access Control Lists (ACLs) support permissions for users and
+	  groups beyond the owner/group/world scheme.
+
+	  To learn more about Access Control Lists, visit the POSIX ACLs for
+	  Linux website <http://acl.bestbits.at/>.
+
+	  If you don't know what Access Control Lists are, say N
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
index 8e2fec05dbe..c84ca1f5259 100644
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -16,11 +16,11 @@
  * Boston, MA 021110-1307, USA.
  */
 
-#include <linux/version.h>
 #include <linux/kthread.h>
 #include <linux/list.h>
 #include <linux/spinlock.h>
-# include <linux/freezer.h>
+#include <linux/freezer.h>
+#include <linux/ftrace.h>
 #include "async-thread.h"
 
 #define WORK_QUEUED_BIT 0
@@ -143,6 +143,7 @@ static int worker_loop(void *arg)
 	struct btrfs_work *work;
 	do {
 		spin_lock_irq(&worker->lock);
+again_locked:
 		while (!list_empty(&worker->pending)) {
 			cur = worker->pending.next;
 			work = list_entry(cur, struct btrfs_work, list);
@@ -165,14 +166,50 @@ static int worker_loop(void *arg)
 			check_idle_worker(worker);
 
 		}
-		worker->working = 0;
 		if (freezing(current)) {
+			worker->working = 0;
+			spin_unlock_irq(&worker->lock);
 			refrigerator();
 		} else {
-			set_current_state(TASK_INTERRUPTIBLE);
 			spin_unlock_irq(&worker->lock);
-			if (!kthread_should_stop())
+			if (!kthread_should_stop()) {
+				cpu_relax();
+				/*
+				 * we've dropped the lock, did someone else
+				 * jump_in?
+				 */
+				smp_mb();
+				if (!list_empty(&worker->pending))
+					continue;
+
+				/*
+				 * this short schedule allows more work to
+				 * come in without the queue functions
+				 * needing to go through wake_up_process()
+				 *
+				 * worker->working is still 1, so nobody
+				 * is going to try and wake us up
+				 */
+				schedule_timeout(1);
+				smp_mb();
+				if (!list_empty(&worker->pending))
+					continue;
+
+				/* still no more work?, sleep for real */
+				spin_lock_irq(&worker->lock);
+				set_current_state(TASK_INTERRUPTIBLE);
+				if (!list_empty(&worker->pending))
+					goto again_locked;
+
+				/*
+				 * this makes sure we get a wakeup when someone
+				 * adds something new to the queue
+				 */
+				worker->working = 0;
+				spin_unlock_irq(&worker->lock);
+
 				schedule();
+			}
 			__set_current_state(TASK_RUNNING);
 		}
 	} while (!kthread_should_stop());
@@ -350,13 +387,14 @@ int 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;
 
 	spin_lock_irqsave(&worker->lock, flags);
-	atomic_inc(&worker->num_pending);
 	list_add_tail(&work->list, &worker->pending);
+	atomic_inc(&worker->num_pending);
 
 	/* by definition we're busy, take ourselves off the idle
 	 * list
@@ -368,10 +406,16 @@ int btrfs_requeue_work(struct btrfs_work *work)
 			       &worker->workers->worker_list);
 		spin_unlock_irqrestore(&worker->workers->lock, flags);
 	}
+	if (!worker->working) {
+		wake = 1;
+		worker->working = 1;
+	}
 
 	spin_unlock_irqrestore(&worker->lock, flags);
-
+	if (wake)
+		wake_up_process(worker->task);
 out:
+
 	return 0;
 }
 
@@ -398,9 +442,10 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work)
 	}
 
 	spin_lock_irqsave(&worker->lock, flags);
+
+	list_add_tail(&work->list, &worker->pending);
 	atomic_inc(&worker->num_pending);
 	check_busy_worker(worker);
-	list_add_tail(&work->list, &worker->pending);
 
 	/*
 	 * avoid calling into wake_up_process if this thread has already
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index a8c9693b75a..72677ce2b74 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -66,6 +66,9 @@ struct btrfs_inode {
 	 */
 	struct list_head delalloc_inodes;
 
+	/* the space_info for where this inode's data allocations are done */
+	struct btrfs_space_info *space_info;
+
 	/* full 64 bit generation number, struct vfs_inode doesn't have a big
 	 * enough field for this.
 	 */
@@ -94,6 +97,11 @@ struct btrfs_inode {
 	 */
 	u64 delalloc_bytes;
 
+	/* total number of bytes that may be used for this inode for
+	 * delalloc
+	 */
+	u64 reserved_bytes;
+
 	/*
 	 * the size of the file stored in the metadata on disk.  data=ordered
 	 * means the in-memory i_size might be larger than the size on disk
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index ee848d8585d..ab07627084f 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -32,7 +32,6 @@
 #include <linux/swap.h>
 #include <linux/writeback.h>
 #include <linux/bit_spinlock.h>
-#include <linux/version.h>
 #include <linux/pagevec.h>
 #include "compat.h"
 #include "ctree.h"
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 9e46c077681..37f31b5529a 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -38,22 +38,64 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		   struct btrfs_path *path, int level, int slot);
 
-inline void btrfs_init_path(struct btrfs_path *p)
-{
-	memset(p, 0, sizeof(*p));
-}
-
 struct btrfs_path *btrfs_alloc_path(void)
 {
 	struct btrfs_path *path;
-	path = kmem_cache_alloc(btrfs_path_cachep, GFP_NOFS);
-	if (path) {
-		btrfs_init_path(path);
+	path = kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS);
+	if (path)
 		path->reada = 1;
-	}
 	return path;
 }
 
+/*
+ * set all locked nodes in the path to blocking locks.  This should
+ * be done before scheduling
+ */
+noinline void btrfs_set_path_blocking(struct btrfs_path *p)
+{
+	int i;
+	for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+		if (p->nodes[i] && p->locks[i])
+			btrfs_set_lock_blocking(p->nodes[i]);
+	}
+}
+
+/*
+ * reset all the locked nodes in the patch to spinning locks.
+ *
+ * held is used to keep lockdep happy, when lockdep is enabled
+ * we set held to a blocking lock before we go around and
+ * retake all the spinlocks in the path.  You can safely use NULL
+ * for held
+ */
+noinline void btrfs_clear_path_blocking(struct btrfs_path *p,
+					struct extent_buffer *held)
+{
+	int i;
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	/* lockdep really cares that we take all of these spinlocks
+	 * in the right order.  If any of the locks in the path are not
+	 * currently blocking, it is going to complain.  So, make really
+	 * really sure by forcing the path to blocking before we clear
+	 * the path blocking.
+	 */
+	if (held)
+		btrfs_set_lock_blocking(held);
+	btrfs_set_path_blocking(p);
+#endif
+
+	for (i = BTRFS_MAX_LEVEL - 1; i >= 0; i--) {
+		if (p->nodes[i] && p->locks[i])
+			btrfs_clear_lock_blocking(p->nodes[i]);
+	}
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	if (held)
+		btrfs_clear_lock_blocking(held);
+#endif
+}
+
 /* this also releases the path */
 void btrfs_free_path(struct btrfs_path *p)
 {
@@ -235,7 +277,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 	if (*cow_ret == buf)
 		unlock_orig = 1;
 
-	WARN_ON(!btrfs_tree_locked(buf));
+	btrfs_assert_tree_locked(buf);
 
 	if (parent)
 		parent_start = parent->start;
@@ -261,7 +303,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 						  trans->transid, level, &ins);
 		BUG_ON(ret);
 		cow = btrfs_init_new_buffer(trans, root, prealloc_dest,
-					    buf->len);
+					    buf->len, level);
 	} else {
 		cow = btrfs_alloc_free_block(trans, root, buf->len,
 					     parent_start,
@@ -272,6 +314,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 	if (IS_ERR(cow))
 		return PTR_ERR(cow);
 
+	/* cow is set to blocking by btrfs_init_new_buffer */
+
 	copy_extent_buffer(cow, buf, 0, 0, cow->len);
 	btrfs_set_header_bytenr(cow, cow->start);
 	btrfs_set_header_generation(cow, trans->transid);
@@ -388,17 +432,20 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
 		WARN_ON(1);
 	}
 
-	spin_lock(&root->fs_info->hash_lock);
 	if (btrfs_header_generation(buf) == trans->transid &&
 	    btrfs_header_owner(buf) == root->root_key.objectid &&
 	    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
 		*cow_ret = buf;
-		spin_unlock(&root->fs_info->hash_lock);
 		WARN_ON(prealloc_dest);
 		return 0;
 	}
-	spin_unlock(&root->fs_info->hash_lock);
+
 	search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1);
+
+	if (parent)
+		btrfs_set_lock_blocking(parent);
+	btrfs_set_lock_blocking(buf);
+
 	ret = __btrfs_cow_block(trans, root, buf, parent,
 				 parent_slot, cow_ret, search_start, 0,
 				 prealloc_dest);
@@ -504,6 +551,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	if (parent_nritems == 1)
 		return 0;
 
+	btrfs_set_lock_blocking(parent);
+
 	for (i = start_slot; i < end_slot; i++) {
 		int close = 1;
 
@@ -564,6 +613,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 			search_start = last_block;
 
 		btrfs_tree_lock(cur);
+		btrfs_set_lock_blocking(cur);
 		err = __btrfs_cow_block(trans, root, cur, parent, i,
 					&cur, search_start,
 					min(16 * blocksize,
@@ -862,6 +912,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		return 0;
 
 	mid = path->nodes[level];
+
 	WARN_ON(!path->locks[level]);
 	WARN_ON(btrfs_header_generation(mid) != trans->transid);
 
@@ -883,8 +934,9 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 
 		/* promote the child to a root */
 		child = read_node_slot(root, mid, 0);
-		btrfs_tree_lock(child);
 		BUG_ON(!child);
+		btrfs_tree_lock(child);
+		btrfs_set_lock_blocking(child);
 		ret = btrfs_cow_block(trans, root, child, mid, 0, &child, 0);
 		BUG_ON(ret);
 
@@ -900,6 +952,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 
 		add_root_to_dirty_list(root);
 		btrfs_tree_unlock(child);
+
 		path->locks[level] = 0;
 		path->nodes[level] = NULL;
 		clean_tree_block(trans, root, mid);
@@ -924,6 +977,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	left = read_node_slot(root, parent, pslot - 1);
 	if (left) {
 		btrfs_tree_lock(left);
+		btrfs_set_lock_blocking(left);
 		wret = btrfs_cow_block(trans, root, left,
 				       parent, pslot - 1, &left, 0);
 		if (wret) {
@@ -934,6 +988,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	right = read_node_slot(root, parent, pslot + 1);
 	if (right) {
 		btrfs_tree_lock(right);
+		btrfs_set_lock_blocking(right);
 		wret = btrfs_cow_block(trans, root, right,
 				       parent, pslot + 1, &right, 0);
 		if (wret) {
@@ -1109,6 +1164,8 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 		u32 left_nr;
 
 		btrfs_tree_lock(left);
+		btrfs_set_lock_blocking(left);
+
 		left_nr = btrfs_header_nritems(left);
 		if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
 			wret = 1;
@@ -1155,7 +1212,10 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 	 */
 	if (right) {
 		u32 right_nr;
+
 		btrfs_tree_lock(right);
+		btrfs_set_lock_blocking(right);
+
 		right_nr = btrfs_header_nritems(right);
 		if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
 			wret = 1;
@@ -1210,8 +1270,7 @@ static noinline void reada_for_search(struct btrfs_root *root,
 	struct btrfs_disk_key disk_key;
 	u32 nritems;
 	u64 search;
-	u64 lowest_read;
-	u64 highest_read;
+	u64 target;
 	u64 nread = 0;
 	int direction = path->reada;
 	struct extent_buffer *eb;
@@ -1235,8 +1294,7 @@ static noinline void reada_for_search(struct btrfs_root *root,
 		return;
 	}
 
-	highest_read = search;
-	lowest_read = search;
+	target = search;
 
 	nritems = btrfs_header_nritems(node);
 	nr = slot;
@@ -1256,27 +1314,80 @@ static noinline void reada_for_search(struct btrfs_root *root,
 				break;
 		}
 		search = btrfs_node_blockptr(node, nr);
-		if ((search >= lowest_read && search <= highest_read) ||
-		    (search < lowest_read && lowest_read - search <= 16384) ||
-		    (search > highest_read && search - highest_read <= 16384)) {
+		if ((search <= target && target - search <= 65536) ||
+		    (search > target && search - target <= 65536)) {
 			readahead_tree_block(root, search, blocksize,
 				     btrfs_node_ptr_generation(node, nr));
 			nread += blocksize;
 		}
 		nscan++;
-		if (path->reada < 2 && (nread > (64 * 1024) || nscan > 32))
+		if ((nread > 65536 || nscan > 32))
 			break;
+	}
+}
 
-		if (nread > (256 * 1024) || nscan > 128)
-			break;
+/*
+ * returns -EAGAIN if it had to drop the path, or zero if everything was in
+ * cache
+ */
+static noinline int reada_for_balance(struct btrfs_root *root,
+				      struct btrfs_path *path, int level)
+{
+	int slot;
+	int nritems;
+	struct extent_buffer *parent;
+	struct extent_buffer *eb;
+	u64 gen;
+	u64 block1 = 0;
+	u64 block2 = 0;
+	int ret = 0;
+	int blocksize;
+
+	parent = path->nodes[level - 1];
+	if (!parent)
+		return 0;
 
-		if (search < lowest_read)
-			lowest_read = search;
-		if (search > highest_read)
-			highest_read = search;
+	nritems = btrfs_header_nritems(parent);
+	slot = path->slots[level];
+	blocksize = btrfs_level_size(root, level);
+
+	if (slot > 0) {
+		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))
+			block1 = 0;
+		free_extent_buffer(eb);
+	}
+	if (slot < nritems) {
+		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))
+			block2 = 0;
+		free_extent_buffer(eb);
 	}
+	if (block1 || block2) {
+		ret = -EAGAIN;
+		btrfs_release_path(root, path);
+		if (block1)
+			readahead_tree_block(root, block1, blocksize, 0);
+		if (block2)
+			readahead_tree_block(root, block2, blocksize, 0);
+
+		if (block1) {
+			eb = read_tree_block(root, block1, blocksize, 0);
+			free_extent_buffer(eb);
+		}
+		if (block1) {
+			eb = read_tree_block(root, block2, blocksize, 0);
+			free_extent_buffer(eb);
+		}
+	}
+	return ret;
 }
 
+
 /*
  * when we walk down the tree, it is usually safe to unlock the higher layers
  * in the tree.  The exceptions are when our path goes through slot 0, because
@@ -1328,6 +1439,32 @@ static noinline void unlock_up(struct btrfs_path *path, int level,
 }
 
 /*
+ * This releases any locks held in the path starting at level and
+ * going all the way up to the root.
+ *
+ * btrfs_search_slot will keep the lock held on higher nodes in a few
+ * corner cases, such as COW of the block at slot zero in the node.  This
+ * ignores those rules, and it should only be called when there are no
+ * more updates to be done higher up in the tree.
+ */
+noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
+{
+	int i;
+
+	if (path->keep_locks || path->lowest_level)
+		return;
+
+	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+		if (!path->nodes[i])
+			continue;
+		if (!path->locks[i])
+			continue;
+		btrfs_tree_unlock(path->nodes[i]);
+		path->locks[i] = 0;
+	}
+}
+
+/*
  * look for key in the tree.  path is filled in with nodes along the way
  * if key is found, we return zero and you can find the item in the leaf
  * level of the path (level 0)
@@ -1387,32 +1524,30 @@ again:
 			int wret;
 
 			/* is a cow on this block not required */
-			spin_lock(&root->fs_info->hash_lock);
 			if (btrfs_header_generation(b) == trans->transid &&
 			    btrfs_header_owner(b) == root->root_key.objectid &&
 			    !btrfs_header_flag(b, BTRFS_HEADER_FLAG_WRITTEN)) {
-				spin_unlock(&root->fs_info->hash_lock);
 				goto cow_done;
 			}
-			spin_unlock(&root->fs_info->hash_lock);
 
 			/* ok, we have to cow, is our old prealloc the right
 			 * size?
 			 */
 			if (prealloc_block.objectid &&
 			    prealloc_block.offset != b->len) {
+				btrfs_release_path(root, p);
 				btrfs_free_reserved_extent(root,
 					   prealloc_block.objectid,
 					   prealloc_block.offset);
 				prealloc_block.objectid = 0;
+				goto again;
 			}
 
 			/*
 			 * for higher level blocks, try not to allocate blocks
 			 * with the block and the parent locks held.
 			 */
-			if (level > 1 && !prealloc_block.objectid &&
-			    btrfs_path_lock_waiting(p, level)) {
+			if (level > 0 && !prealloc_block.objectid) {
 				u32 size = b->len;
 				u64 hint = b->start;
 
@@ -1425,6 +1560,8 @@ again:
 				goto again;
 			}
 
+			btrfs_set_path_blocking(p);
+
 			wret = btrfs_cow_block(trans, root, b,
 					       p->nodes[level + 1],
 					       p->slots[level + 1],
@@ -1446,6 +1583,22 @@ cow_done:
 		if (!p->skip_locking)
 			p->locks[level] = 1;
 
+		btrfs_clear_path_blocking(p, NULL);
+
+		/*
+		 * we have a lock on b and as long as we aren't changing
+		 * the tree, there is no way to for the items in b to change.
+		 * It is safe to drop the lock on our parent before we
+		 * go through the expensive btree search on b.
+		 *
+		 * If cow is true, then we might be changing slot zero,
+		 * which may require changing the parent.  So, we can't
+		 * drop the lock until after we know which slot we're
+		 * operating on.
+		 */
+		if (!cow)
+			btrfs_unlock_up_safe(p, level + 1);
+
 		ret = check_block(root, p, level);
 		if (ret) {
 			ret = -1;
@@ -1453,6 +1606,7 @@ cow_done:
 		}
 
 		ret = bin_search(b, key, level, &slot);
+
 		if (level != 0) {
 			if (ret && slot > 0)
 				slot -= 1;
@@ -1460,7 +1614,16 @@ cow_done:
 			if ((p->search_for_split || ins_len > 0) &&
 			    btrfs_header_nritems(b) >=
 			    BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
-				int sret = split_node(trans, root, p, level);
+				int sret;
+
+				sret = reada_for_balance(root, p, level);
+				if (sret)
+					goto again;
+
+				btrfs_set_path_blocking(p);
+				sret = split_node(trans, root, p, level);
+				btrfs_clear_path_blocking(p, NULL);
+
 				BUG_ON(sret > 0);
 				if (sret) {
 					ret = sret;
@@ -1468,9 +1631,19 @@ cow_done:
 				}
 				b = p->nodes[level];
 				slot = p->slots[level];
-			} else if (ins_len < 0) {
-				int sret = balance_level(trans, root, p,
-							 level);
+			} else if (ins_len < 0 &&
+				   btrfs_header_nritems(b) <
+				   BTRFS_NODEPTRS_PER_BLOCK(root) / 4) {
+				int sret;
+
+				sret = reada_for_balance(root, p, level);
+				if (sret)
+					goto again;
+
+				btrfs_set_path_blocking(p);
+				sret = balance_level(trans, root, p, level);
+				btrfs_clear_path_blocking(p, NULL);
+
 				if (sret) {
 					ret = sret;
 					goto done;
@@ -1504,7 +1677,7 @@ cow_done:
 				 * of the btree by dropping locks before
 				 * we read.
 				 */
-				if (level > 1) {
+				if (level > 0) {
 					btrfs_release_path(NULL, p);
 					if (tmp)
 						free_extent_buffer(tmp);
@@ -1519,6 +1692,7 @@ cow_done:
 						free_extent_buffer(tmp);
 					goto again;
 				} else {
+					btrfs_set_path_blocking(p);
 					if (tmp)
 						free_extent_buffer(tmp);
 					if (should_reada)
@@ -1528,14 +1702,29 @@ cow_done:
 					b = read_node_slot(root, b, slot);
 				}
 			}
-			if (!p->skip_locking)
-				btrfs_tree_lock(b);
+			if (!p->skip_locking) {
+				int lret;
+
+				btrfs_clear_path_blocking(p, NULL);
+				lret = btrfs_try_spin_lock(b);
+
+				if (!lret) {
+					btrfs_set_path_blocking(p);
+					btrfs_tree_lock(b);
+					btrfs_clear_path_blocking(p, b);
+				}
+			}
 		} else {
 			p->slots[level] = slot;
 			if (ins_len > 0 &&
 			    btrfs_leaf_free_space(root, b) < ins_len) {
-				int sret = split_leaf(trans, root, key,
+				int sret;
+
+				btrfs_set_path_blocking(p);
+				sret = split_leaf(trans, root, key,
 						      p, ins_len, ret == 0);
+				btrfs_clear_path_blocking(p, NULL);
+
 				BUG_ON(sret > 0);
 				if (sret) {
 					ret = sret;
@@ -1549,12 +1738,16 @@ cow_done:
 	}
 	ret = 1;
 done:
+	/*
+	 * we don't really know what they plan on doing with the path
+	 * from here on, so for now just mark it as blocking
+	 */
+	btrfs_set_path_blocking(p);
 	if (prealloc_block.objectid) {
 		btrfs_free_reserved_extent(root,
 			   prealloc_block.objectid,
 			   prealloc_block.offset);
 	}
-
 	return ret;
 }
 
@@ -1578,6 +1771,8 @@ int btrfs_merge_path(struct btrfs_trans_handle *trans,
 	ret = btrfs_cow_block(trans, root, eb, NULL, 0, &eb, 0);
 	BUG_ON(ret);
 
+	btrfs_set_lock_blocking(eb);
+
 	parent = eb;
 	while (1) {
 		level = btrfs_header_level(parent);
@@ -1602,6 +1797,7 @@ int btrfs_merge_path(struct btrfs_trans_handle *trans,
 			eb = read_tree_block(root, bytenr, blocksize,
 					     generation);
 			btrfs_tree_lock(eb);
+			btrfs_set_lock_blocking(eb);
 		}
 
 		/*
@@ -1626,6 +1822,7 @@ int btrfs_merge_path(struct btrfs_trans_handle *trans,
 				eb = read_tree_block(root, bytenr, blocksize,
 						generation);
 				btrfs_tree_lock(eb);
+				btrfs_set_lock_blocking(eb);
 			}
 
 			ret = btrfs_cow_block(trans, root, eb, parent, slot,
@@ -2168,10 +2365,12 @@ static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (slot >= btrfs_header_nritems(upper) - 1)
 		return 1;
 
-	WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+	btrfs_assert_tree_locked(path->nodes[1]);
 
 	right = read_node_slot(root, upper, slot + 1);
 	btrfs_tree_lock(right);
+	btrfs_set_lock_blocking(right);
+
 	free_space = btrfs_leaf_free_space(root, right);
 	if (free_space < data_size)
 		goto out_unlock;
@@ -2363,10 +2562,12 @@ static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (right_nritems == 0)
 		return 1;
 
-	WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+	btrfs_assert_tree_locked(path->nodes[1]);
 
 	left = read_node_slot(root, path->nodes[1], slot - 1);
 	btrfs_tree_lock(left);
+	btrfs_set_lock_blocking(left);
+
 	free_space = btrfs_leaf_free_space(root, left);
 	if (free_space < data_size) {
 		ret = 1;
@@ -2825,6 +3026,12 @@ int btrfs_split_item(struct btrfs_trans_handle *trans,
 	path->keep_locks = 0;
 	BUG_ON(ret);
 
+	/*
+	 * make sure any changes to the path from split_leaf leave it
+	 * in a blocking state
+	 */
+	btrfs_set_path_blocking(path);
+
 	leaf = path->nodes[0];
 	BUG_ON(btrfs_leaf_free_space(root, leaf) < sizeof(struct btrfs_item));
 
@@ -3354,6 +3561,7 @@ int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
 		BUG();
 	}
 out:
+	btrfs_unlock_up_safe(path, 1);
 	return ret;
 }
 
@@ -3441,15 +3649,22 @@ noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
 {
 	int ret;
 	u64 root_gen = btrfs_header_generation(path->nodes[1]);
+	u64 parent_start = path->nodes[1]->start;
+	u64 parent_owner = btrfs_header_owner(path->nodes[1]);
 
 	ret = del_ptr(trans, root, path, 1, path->slots[1]);
 	if (ret)
 		return ret;
 
+	/*
+	 * btrfs_free_extent is expensive, we want to make sure we
+	 * aren't holding any locks when we call it
+	 */
+	btrfs_unlock_up_safe(path, 0);
+
 	ret = btrfs_free_extent(trans, root, bytenr,
 				btrfs_level_size(root, 0),
-				path->nodes[1]->start,
-				btrfs_header_owner(path->nodes[1]),
+				parent_start, parent_owner,
 				root_gen, 0, 1);
 	return ret;
 }
@@ -3721,6 +3936,7 @@ find_next_key:
 		 */
 		if (slot >= nritems) {
 			path->slots[level] = slot;
+			btrfs_set_path_blocking(path);
 			sret = btrfs_find_next_key(root, path, min_key, level,
 						  cache_only, min_trans);
 			if (sret == 0) {
@@ -3738,16 +3954,20 @@ find_next_key:
 			unlock_up(path, level, 1);
 			goto out;
 		}
+		btrfs_set_path_blocking(path);
 		cur = read_node_slot(root, cur, slot);
 
 		btrfs_tree_lock(cur);
+
 		path->locks[level - 1] = 1;
 		path->nodes[level - 1] = cur;
 		unlock_up(path, level, 1);
+		btrfs_clear_path_blocking(path, NULL);
 	}
 out:
 	if (ret == 0)
 		memcpy(min_key, &found_key, sizeof(found_key));
+	btrfs_set_path_blocking(path);
 	return ret;
 }
 
@@ -3843,6 +4063,7 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 	if (ret < 0)
 		return ret;
 
+	btrfs_set_path_blocking(path);
 	nritems = btrfs_header_nritems(path->nodes[0]);
 	/*
 	 * by releasing the path above we dropped all our locks.  A balance
@@ -3873,14 +4094,16 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 			free_extent_buffer(next);
 		}
 
+		/* the path was set to blocking above */
 		if (level == 1 && (path->locks[1] || path->skip_locking) &&
 		    path->reada)
 			reada_for_search(root, path, level, slot, 0);
 
 		next = read_node_slot(root, c, slot);
 		if (!path->skip_locking) {
-			WARN_ON(!btrfs_tree_locked(c));
+			btrfs_assert_tree_locked(c);
 			btrfs_tree_lock(next);
+			btrfs_set_lock_blocking(next);
 		}
 		break;
 	}
@@ -3897,12 +4120,15 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 			path->locks[level] = 1;
 		if (!level)
 			break;
+
+		btrfs_set_path_blocking(path);
 		if (level == 1 && path->locks[1] && path->reada)
 			reada_for_search(root, path, level, slot, 0);
 		next = read_node_slot(root, next, 0);
 		if (!path->skip_locking) {
-			WARN_ON(!btrfs_tree_locked(path->nodes[level]));
+			btrfs_assert_tree_locked(path->nodes[level]);
 			btrfs_tree_lock(next);
+			btrfs_set_lock_blocking(next);
 		}
 	}
 done:
@@ -3927,6 +4153,7 @@ int btrfs_previous_item(struct btrfs_root *root,
 
 	while (1) {
 		if (path->slots[0] == 0) {
+			btrfs_set_path_blocking(path);
 			ret = btrfs_prev_leaf(root, path);
 			if (ret != 0)
 				return ret;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index eee060f8811..5e1d4e30e9d 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -43,11 +43,7 @@ struct btrfs_ordered_sum;
 
 #define BTRFS_ACL_NOT_CACHED    ((void *)-1)
 
-#ifdef CONFIG_LOCKDEP
-# define BTRFS_MAX_LEVEL 7
-#else
-# define BTRFS_MAX_LEVEL 8
-#endif
+#define BTRFS_MAX_LEVEL 8
 
 /* holds pointers to all of the tree roots */
 #define BTRFS_ROOT_TREE_OBJECTID 1ULL
@@ -454,17 +450,11 @@ struct btrfs_timespec {
 	__le32 nsec;
 } __attribute__ ((__packed__));
 
-typedef enum {
+enum btrfs_compression_type {
 	BTRFS_COMPRESS_NONE = 0,
 	BTRFS_COMPRESS_ZLIB = 1,
 	BTRFS_COMPRESS_LAST = 2,
-} btrfs_compression_type;
-
-/* we don't understand any encryption methods right now */
-typedef enum {
-	BTRFS_ENCRYPTION_NONE = 0,
-	BTRFS_ENCRYPTION_LAST = 1,
-} btrfs_encryption_type;
+};
 
 struct btrfs_inode_item {
 	/* nfs style generation number */
@@ -606,13 +596,27 @@ struct btrfs_block_group_item {
 
 struct btrfs_space_info {
 	u64 flags;
-	u64 total_bytes;
-	u64 bytes_used;
-	u64 bytes_pinned;
-	u64 bytes_reserved;
-	u64 bytes_readonly;
-	int full;
-	int force_alloc;
+
+	u64 total_bytes;	/* total bytes in the space */
+	u64 bytes_used;		/* total bytes used on disk */
+	u64 bytes_pinned;	/* total bytes pinned, will be freed when the
+				   transaction finishes */
+	u64 bytes_reserved;	/* total bytes the allocator has reserved for
+				   current allocations */
+	u64 bytes_readonly;	/* total bytes that are read only */
+
+	/* delalloc accounting */
+	u64 bytes_delalloc;	/* number of bytes reserved for allocation,
+				   this space is not necessarily reserved yet
+				   by the allocator */
+	u64 bytes_may_use;	/* number of bytes that may be used for
+				   delalloc */
+
+	int full;		/* indicates that we cannot allocate any more
+				   chunks for this space */
+	int force_alloc;	/* set if we need to force a chunk alloc for
+				   this space */
+
 	struct list_head list;
 
 	/* for block groups in our same type */
@@ -701,9 +705,7 @@ struct btrfs_fs_info {
 	struct btrfs_transaction *running_transaction;
 	wait_queue_head_t transaction_throttle;
 	wait_queue_head_t transaction_wait;
-
 	wait_queue_head_t async_submit_wait;
-	wait_queue_head_t tree_log_wait;
 
 	struct btrfs_super_block super_copy;
 	struct btrfs_super_block super_for_commit;
@@ -711,7 +713,6 @@ struct btrfs_fs_info {
 	struct super_block *sb;
 	struct inode *btree_inode;
 	struct backing_dev_info bdi;
-	spinlock_t hash_lock;
 	struct mutex trans_mutex;
 	struct mutex tree_log_mutex;
 	struct mutex transaction_kthread_mutex;
@@ -730,10 +731,6 @@ struct btrfs_fs_info {
 	atomic_t async_submit_draining;
 	atomic_t nr_async_bios;
 	atomic_t async_delalloc_pages;
-	atomic_t tree_log_writers;
-	atomic_t tree_log_commit;
-	unsigned long tree_log_batch;
-	u64 tree_log_transid;
 
 	/*
 	 * this is used by the balancing code to wait for all the pending
@@ -787,7 +784,14 @@ struct btrfs_fs_info {
 	struct list_head dirty_cowonly_roots;
 
 	struct btrfs_fs_devices *fs_devices;
+
+	/*
+	 * the space_info list is almost entirely read only.  It only changes
+	 * when we add a new raid type to the FS, and that happens
+	 * very rarely.  RCU is used to protect it.
+	 */
 	struct list_head space_info;
+
 	spinlock_t delalloc_lock;
 	spinlock_t new_trans_lock;
 	u64 delalloc_bytes;
@@ -833,7 +837,14 @@ struct btrfs_root {
 	struct kobject root_kobj;
 	struct completion kobj_unregister;
 	struct mutex objectid_mutex;
+
 	struct mutex log_mutex;
+	wait_queue_head_t log_writer_wait;
+	wait_queue_head_t log_commit_wait[2];
+	atomic_t log_writers;
+	atomic_t log_commit[2];
+	unsigned long log_transid;
+	unsigned long log_batch;
 
 	u64 objectid;
 	u64 last_trans;
@@ -1721,7 +1732,8 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 					     u64 empty_size);
 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 					    struct btrfs_root *root,
-					    u64 bytenr, u32 blocksize);
+					    u64 bytenr, u32 blocksize,
+					    int level);
 int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root,
 		       u64 num_bytes, u64 parent, u64 min_bytes,
@@ -1791,6 +1803,18 @@ int btrfs_add_dead_reloc_root(struct btrfs_root *root);
 int btrfs_cleanup_reloc_trees(struct btrfs_root *root);
 int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags);
+void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde);
+void btrfs_clear_space_info_full(struct btrfs_fs_info *info);
+
+int btrfs_check_metadata_free_space(struct btrfs_root *root);
+int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
+				u64 bytes);
+void btrfs_free_reserved_data_space(struct btrfs_root *root,
+				    struct inode *inode, u64 bytes);
+void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
+				 u64 bytes);
+void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
+			      u64 bytes);
 /* ctree.c */
 int btrfs_previous_item(struct btrfs_root *root,
 			struct btrfs_path *path, u64 min_objectid,
@@ -1840,7 +1864,9 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p);
 struct btrfs_path *btrfs_alloc_path(void);
 void btrfs_free_path(struct btrfs_path *p);
-void btrfs_init_path(struct btrfs_path *p);
+void btrfs_set_path_blocking(struct btrfs_path *p);
+void btrfs_unlock_up_safe(struct btrfs_path *p, int level);
+
 int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		   struct btrfs_path *path, int slot, int nr);
 int btrfs_del_leaf(struct btrfs_trans_handle *trans,
@@ -2034,8 +2060,6 @@ int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
 unsigned long btrfs_force_ra(struct address_space *mapping,
 			      struct file_ra_state *ra, struct file *file,
 			      pgoff_t offset, pgoff_t last_index);
-int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
-			   int for_del);
 int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page);
 int btrfs_readpage(struct file *file, struct page *page);
 void btrfs_delete_inode(struct inode *inode);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 81a313874ae..3e18175248e 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -16,7 +16,6 @@
  * Boston, MA 021110-1307, USA.
  */
 
-#include <linux/version.h>
 #include <linux/fs.h>
 #include <linux/blkdev.h>
 #include <linux/scatterlist.h>
@@ -76,6 +75,40 @@ struct async_submit_bio {
 	struct btrfs_work work;
 };
 
+/* These are used to set the lockdep class on the extent buffer locks.
+ * The class is set by the readpage_end_io_hook after the buffer has
+ * passed csum validation but before the pages are unlocked.
+ *
+ * The lockdep class is also set by btrfs_init_new_buffer on freshly
+ * allocated blocks.
+ *
+ * The class is based on the level in the tree block, which allows lockdep
+ * to know that lower nodes nest inside the locks of higher nodes.
+ *
+ * We also add a check to make sure the highest level of the tree is
+ * the same as our lockdep setup here.  If BTRFS_MAX_LEVEL changes, this
+ * code needs update as well.
+ */
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+# if BTRFS_MAX_LEVEL != 8
+#  error
+# endif
+static struct lock_class_key btrfs_eb_class[BTRFS_MAX_LEVEL + 1];
+static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = {
+	/* leaf */
+	"btrfs-extent-00",
+	"btrfs-extent-01",
+	"btrfs-extent-02",
+	"btrfs-extent-03",
+	"btrfs-extent-04",
+	"btrfs-extent-05",
+	"btrfs-extent-06",
+	"btrfs-extent-07",
+	/* highest possible level */
+	"btrfs-extent-08",
+};
+#endif
+
 /*
  * extents on the btree inode are pretty simple, there's one extent
  * that covers the entire device
@@ -348,6 +381,15 @@ static int check_tree_block_fsid(struct btrfs_root *root,
 	return ret;
 }
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level)
+{
+	lockdep_set_class_and_name(&eb->lock,
+			   &btrfs_eb_class[level],
+			   btrfs_eb_name[level]);
+}
+#endif
+
 static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 			       struct extent_state *state)
 {
@@ -393,6 +435,8 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 	}
 	found_level = btrfs_header_level(eb);
 
+	btrfs_set_buffer_lockdep_class(eb, found_level);
+
 	ret = csum_tree_block(root, eb, 1);
 	if (ret)
 		ret = -EIO;
@@ -800,7 +844,7 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
 
 	if (ret == 0)
-		buf->flags |= EXTENT_UPTODATE;
+		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
 	else
 		WARN_ON(1);
 	return buf;
@@ -813,7 +857,11 @@ int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	if (btrfs_header_generation(buf) ==
 	    root->fs_info->running_transaction->transid) {
-		WARN_ON(!btrfs_tree_locked(buf));
+		btrfs_assert_tree_locked(buf);
+
+		/* ugh, clear_extent_buffer_dirty can be expensive */
+		btrfs_set_lock_blocking(buf);
+
 		clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
 					  buf);
 	}
@@ -850,6 +898,14 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	spin_lock_init(&root->list_lock);
 	mutex_init(&root->objectid_mutex);
 	mutex_init(&root->log_mutex);
+	init_waitqueue_head(&root->log_writer_wait);
+	init_waitqueue_head(&root->log_commit_wait[0]);
+	init_waitqueue_head(&root->log_commit_wait[1]);
+	atomic_set(&root->log_commit[0], 0);
+	atomic_set(&root->log_commit[1], 0);
+	atomic_set(&root->log_writers, 0);
+	root->log_batch = 0;
+	root->log_transid = 0;
 	extent_io_tree_init(&root->dirty_log_pages,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
 
@@ -934,15 +990,16 @@ int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
 	return 0;
 }
 
-int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
-			     struct btrfs_fs_info *fs_info)
+static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
+					 struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root;
 	struct btrfs_root *tree_root = fs_info->tree_root;
+	struct extent_buffer *leaf;
 
 	root = kzalloc(sizeof(*root), GFP_NOFS);
 	if (!root)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	__setup_root(tree_root->nodesize, tree_root->leafsize,
 		     tree_root->sectorsize, tree_root->stripesize,
@@ -951,12 +1008,23 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
 	root->root_key.objectid = BTRFS_TREE_LOG_OBJECTID;
 	root->root_key.type = BTRFS_ROOT_ITEM_KEY;
 	root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
+	/*
+	 * log trees do not get reference counted because they go away
+	 * before a real commit is actually done.  They do store pointers
+	 * to file data extents, and those reference counts still get
+	 * updated (along with back refs to the log tree).
+	 */
 	root->ref_cows = 0;
 
-	root->node = btrfs_alloc_free_block(trans, root, root->leafsize,
-					    0, BTRFS_TREE_LOG_OBJECTID,
-					    trans->transid, 0, 0, 0);
+	leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
+				      0, BTRFS_TREE_LOG_OBJECTID,
+				      trans->transid, 0, 0, 0);
+	if (IS_ERR(leaf)) {
+		kfree(root);
+		return ERR_CAST(leaf);
+	}
 
+	root->node = leaf;
 	btrfs_set_header_nritems(root->node, 0);
 	btrfs_set_header_level(root->node, 0);
 	btrfs_set_header_bytenr(root->node, root->node->start);
@@ -968,7 +1036,48 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
 			    BTRFS_FSID_SIZE);
 	btrfs_mark_buffer_dirty(root->node);
 	btrfs_tree_unlock(root->node);
-	fs_info->log_root_tree = root;
+	return root;
+}
+
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *log_root;
+
+	log_root = alloc_log_tree(trans, fs_info);
+	if (IS_ERR(log_root))
+		return PTR_ERR(log_root);
+	WARN_ON(fs_info->log_root_tree);
+	fs_info->log_root_tree = log_root;
+	return 0;
+}
+
+int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root)
+{
+	struct btrfs_root *log_root;
+	struct btrfs_inode_item *inode_item;
+
+	log_root = alloc_log_tree(trans, root->fs_info);
+	if (IS_ERR(log_root))
+		return PTR_ERR(log_root);
+
+	log_root->last_trans = trans->transid;
+	log_root->root_key.offset = root->root_key.objectid;
+
+	inode_item = &log_root->root_item.inode;
+	inode_item->generation = cpu_to_le64(1);
+	inode_item->size = cpu_to_le64(3);
+	inode_item->nlink = cpu_to_le32(1);
+	inode_item->nbytes = cpu_to_le64(root->leafsize);
+	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+
+	btrfs_set_root_bytenr(&log_root->root_item, log_root->node->start);
+	btrfs_set_root_generation(&log_root->root_item, trans->transid);
+
+	WARN_ON(root->log_root);
+	root->log_root = log_root;
+	root->log_transid = 0;
 	return 0;
 }
 
@@ -1136,7 +1245,6 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 {
 	struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
 	int ret = 0;
-	struct list_head *cur;
 	struct btrfs_device *device;
 	struct backing_dev_info *bdi;
 #if 0
@@ -1144,8 +1252,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 	    btrfs_congested_async(info, 0))
 		return 1;
 #endif
-	list_for_each(cur, &info->fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
 		bdi = blk_get_backing_dev_info(device->bdev);
@@ -1163,13 +1270,11 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
  */
 static void __unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
 {
-	struct list_head *cur;
 	struct btrfs_device *device;
 	struct btrfs_fs_info *info;
 
 	info = (struct btrfs_fs_info *)bdi->unplug_io_data;
-	list_for_each(cur, &info->fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
 
@@ -1447,7 +1552,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->dead_roots);
 	INIT_LIST_HEAD(&fs_info->hashers);
 	INIT_LIST_HEAD(&fs_info->delalloc_inodes);
-	spin_lock_init(&fs_info->hash_lock);
 	spin_lock_init(&fs_info->delalloc_lock);
 	spin_lock_init(&fs_info->new_trans_lock);
 	spin_lock_init(&fs_info->ref_cache_lock);
@@ -1535,10 +1639,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	init_waitqueue_head(&fs_info->transaction_throttle);
 	init_waitqueue_head(&fs_info->transaction_wait);
 	init_waitqueue_head(&fs_info->async_submit_wait);
-	init_waitqueue_head(&fs_info->tree_log_wait);
-	atomic_set(&fs_info->tree_log_commit, 0);
-	atomic_set(&fs_info->tree_log_writers, 0);
-	fs_info->tree_log_transid = 0;
 
 	__setup_root(4096, 4096, 4096, 4096, tree_root,
 		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
@@ -1627,6 +1727,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	 * low idle thresh
 	 */
 	fs_info->endio_workers.idle_thresh = 4;
+	fs_info->endio_meta_workers.idle_thresh = 4;
+
 	fs_info->endio_write_workers.idle_thresh = 64;
 	fs_info->endio_meta_write_workers.idle_thresh = 64;
 
@@ -1720,7 +1822,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	ret = find_and_setup_root(tree_root, fs_info,
 				  BTRFS_DEV_TREE_OBJECTID, dev_root);
 	dev_root->track_dirty = 1;
-
 	if (ret)
 		goto fail_extent_root;
 
@@ -1740,13 +1841,13 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->system_alloc_profile = fs_info->metadata_alloc_profile;
 	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
 					       "btrfs-cleaner");
-	if (!fs_info->cleaner_kthread)
+	if (IS_ERR(fs_info->cleaner_kthread))
 		goto fail_csum_root;
 
 	fs_info->transaction_kthread = kthread_run(transaction_kthread,
 						   tree_root,
 						   "btrfs-transaction");
-	if (!fs_info->transaction_kthread)
+	if (IS_ERR(fs_info->transaction_kthread))
 		goto fail_cleaner;
 
 	if (btrfs_super_log_root(disk_super) != 0) {
@@ -1828,13 +1929,14 @@ fail_sb_buffer:
 fail_iput:
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	iput(fs_info->btree_inode);
-fail:
+
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
+	bdi_destroy(&fs_info->bdi);
 
+fail:
 	kfree(extent_root);
 	kfree(tree_root);
-	bdi_destroy(&fs_info->bdi);
 	kfree(fs_info);
 	kfree(chunk_root);
 	kfree(dev_root);
@@ -1995,7 +2097,6 @@ static int write_dev_supers(struct btrfs_device *device,
 
 int write_all_supers(struct btrfs_root *root, int max_mirrors)
 {
-	struct list_head *cur;
 	struct list_head *head = &root->fs_info->fs_devices->devices;
 	struct btrfs_device *dev;
 	struct btrfs_super_block *sb;
@@ -2011,8 +2112,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 
 	sb = &root->fs_info->super_for_commit;
 	dev_item = &sb->dev_item;
-	list_for_each(cur, head) {
-		dev = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(dev, head, dev_list) {
 		if (!dev->bdev) {
 			total_errors++;
 			continue;
@@ -2045,8 +2145,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	}
 
 	total_errors = 0;
-	list_for_each(cur, head) {
-		dev = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(dev, head, dev_list) {
 		if (!dev->bdev)
 			continue;
 		if (!dev->in_fs_metadata || !dev->writeable)
@@ -2260,7 +2359,9 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 	u64 transid = btrfs_header_generation(buf);
 	struct inode *btree_inode = root->fs_info->btree_inode;
 
-	WARN_ON(!btrfs_tree_locked(buf));
+	btrfs_set_lock_blocking(buf);
+
+	btrfs_assert_tree_locked(buf);
 	if (transid != root->fs_info->generation) {
 		printk(KERN_CRIT "btrfs transid mismatch buffer %llu, "
 		       "found %llu running %llu\n",
@@ -2302,14 +2403,13 @@ int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
 	int ret;
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
 	if (ret == 0)
-		buf->flags |= EXTENT_UPTODATE;
+		set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags);
 	return ret;
 }
 
 int btree_lock_page_hook(struct page *page)
 {
 	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;
@@ -2324,9 +2424,7 @@ int btree_lock_page_hook(struct page *page)
 		goto out;
 
 	btrfs_tree_lock(eb);
-	spin_lock(&root->fs_info->hash_lock);
 	btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
-	spin_unlock(&root->fs_info->hash_lock);
 	btrfs_tree_unlock(eb);
 	free_extent_buffer(eb);
 out:
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index c0ff404c31b..95029db227b 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -98,5 +98,17 @@ int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
 			     struct btrfs_fs_info *fs_info);
 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 btree_lock_page_hook(struct page *page);
+
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level);
+#else
+static inline void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb,
+						 int level)
+{
+}
+#endif
 #endif
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 293da650873..fefe83ad205 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -19,7 +19,8 @@
 #include <linux/pagemap.h>
 #include <linux/writeback.h>
 #include <linux/blkdev.h>
-#include <linux/version.h>
+#include <linux/sort.h>
+#include <linux/rcupdate.h>
 #include "compat.h"
 #include "hash.h"
 #include "crc32c.h"
@@ -30,7 +31,6 @@
 #include "volumes.h"
 #include "locking.h"
 #include "ref-cache.h"
-#include "compat.h"
 
 #define PENDING_EXTENT_INSERT 0
 #define PENDING_EXTENT_DELETE 1
@@ -61,6 +61,10 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 			      u64 bytenr, u64 num_bytes, int alloc,
 			      int mark_free);
 
+static int do_chunk_alloc(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *extent_root, u64 alloc_bytes,
+			  u64 flags, int force);
+
 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
 {
 	return (cache->flags & bits) == bits;
@@ -326,16 +330,34 @@ static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
 						  u64 flags)
 {
 	struct list_head *head = &info->space_info;
-	struct list_head *cur;
 	struct btrfs_space_info *found;
-	list_for_each(cur, head) {
-		found = list_entry(cur, struct btrfs_space_info, list);
-		if (found->flags == flags)
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(found, head, list) {
+		if (found->flags == flags) {
+			rcu_read_unlock();
 			return found;
+		}
 	}
+	rcu_read_unlock();
 	return NULL;
 }
 
+/*
+ * after adding space to the filesystem, we need to clear the full flags
+ * on all the space infos.
+ */
+void btrfs_clear_space_info_full(struct btrfs_fs_info *info)
+{
+	struct list_head *head = &info->space_info;
+	struct btrfs_space_info *found;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(found, head, list)
+		found->full = 0;
+	rcu_read_unlock();
+}
+
 static u64 div_factor(u64 num, int factor)
 {
 	if (factor == 10)
@@ -1326,8 +1348,25 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root)
 {
-	finish_current_insert(trans, root->fs_info->extent_root, 1);
-	del_pending_extents(trans, root->fs_info->extent_root, 1);
+	u64 start;
+	u64 end;
+	int ret;
+
+	while(1) {
+		finish_current_insert(trans, root->fs_info->extent_root, 1);
+		del_pending_extents(trans, root->fs_info->extent_root, 1);
+
+		/* is there more work to do? */
+		ret = find_first_extent_bit(&root->fs_info->pending_del,
+					    0, &start, &end, EXTENT_WRITEBACK);
+		if (!ret)
+			continue;
+		ret = find_first_extent_bit(&root->fs_info->extent_ins,
+					    0, &start, &end, EXTENT_WRITEBACK);
+		if (!ret)
+			continue;
+		break;
+	}
 	return 0;
 }
 
@@ -1525,15 +1564,55 @@ out:
 	return ret;
 }
 
-int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		  struct extent_buffer *orig_buf, struct extent_buffer *buf,
-		  u32 *nr_extents)
+/* when a block goes through cow, we update the reference counts of
+ * everything that block points to.  The internal pointers of the block
+ * can be in just about any order, and it is likely to have clusters of
+ * things that are close together and clusters of things that are not.
+ *
+ * To help reduce the seeks that come with updating all of these reference
+ * counts, sort them by byte number before actual updates are done.
+ *
+ * struct refsort is used to match byte number to slot in the btree block.
+ * we sort based on the byte number and then use the slot to actually
+ * find the item.
+ *
+ * struct refsort is smaller than strcut btrfs_item and smaller than
+ * struct btrfs_key_ptr.  Since we're currently limited to the page size
+ * for a btree block, there's no way for a kmalloc of refsorts for a
+ * single node to be bigger than a page.
+ */
+struct refsort {
+	u64 bytenr;
+	u32 slot;
+};
+
+/*
+ * for passing into sort()
+ */
+static int refsort_cmp(const void *a_void, const void *b_void)
+{
+	const struct refsort *a = a_void;
+	const struct refsort *b = b_void;
+
+	if (a->bytenr < b->bytenr)
+		return -1;
+	if (a->bytenr > b->bytenr)
+		return 1;
+	return 0;
+}
+
+
+noinline int btrfs_inc_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct extent_buffer *orig_buf,
+			   struct extent_buffer *buf, u32 *nr_extents)
 {
 	u64 bytenr;
 	u64 ref_root;
 	u64 orig_root;
 	u64 ref_generation;
 	u64 orig_generation;
+	struct refsort *sorted;
 	u32 nritems;
 	u32 nr_file_extents = 0;
 	struct btrfs_key key;
@@ -1542,6 +1621,8 @@ int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	int level;
 	int ret = 0;
 	int faili = 0;
+	int refi = 0;
+	int slot;
 	int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
 			    u64, u64, u64, u64, u64, u64, u64, u64);
 
@@ -1553,6 +1634,9 @@ int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	nritems = btrfs_header_nritems(buf);
 	level = btrfs_header_level(buf);
 
+	sorted = kmalloc(sizeof(struct refsort) * nritems, GFP_NOFS);
+	BUG_ON(!sorted);
+
 	if (root->ref_cows) {
 		process_func = __btrfs_inc_extent_ref;
 	} else {
@@ -1565,6 +1649,11 @@ int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		process_func = __btrfs_update_extent_ref;
 	}
 
+	/*
+	 * we make two passes through the items.  In the first pass we
+	 * only record the byte number and slot.  Then we sort based on
+	 * byte number and do the actual work based on the sorted results
+	 */
 	for (i = 0; i < nritems; i++) {
 		cond_resched();
 		if (level == 0) {
@@ -1581,6 +1670,32 @@ int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 				continue;
 
 			nr_file_extents++;
+			sorted[refi].bytenr = bytenr;
+			sorted[refi].slot = i;
+			refi++;
+		} else {
+			bytenr = btrfs_node_blockptr(buf, i);
+			sorted[refi].bytenr = bytenr;
+			sorted[refi].slot = i;
+			refi++;
+		}
+	}
+	/*
+	 * if refi == 0, we didn't actually put anything into the sorted
+	 * array and we're done
+	 */
+	if (refi == 0)
+		goto out;
+
+	sort(sorted, refi, sizeof(struct refsort), refsort_cmp, NULL);
+
+	for (i = 0; i < refi; i++) {
+		cond_resched();
+		slot = sorted[i].slot;
+		bytenr = sorted[i].bytenr;
+
+		if (level == 0) {
+			btrfs_item_key_to_cpu(buf, &key, slot);
 
 			ret = process_func(trans, root, bytenr,
 					   orig_buf->start, buf->start,
@@ -1589,25 +1704,25 @@ int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 					   key.objectid);
 
 			if (ret) {
-				faili = i;
+				faili = slot;
 				WARN_ON(1);
 				goto fail;
 			}
 		} else {
-			bytenr = btrfs_node_blockptr(buf, i);
 			ret = process_func(trans, root, bytenr,
 					   orig_buf->start, buf->start,
 					   orig_root, ref_root,
 					   orig_generation, ref_generation,
 					   level - 1);
 			if (ret) {
-				faili = i;
+				faili = slot;
 				WARN_ON(1);
 				goto fail;
 			}
 		}
 	}
 out:
+	kfree(sorted);
 	if (nr_extents) {
 		if (level == 0)
 			*nr_extents = nr_file_extents;
@@ -1616,6 +1731,7 @@ out:
 	}
 	return 0;
 fail:
+	kfree(sorted);
 	WARN_ON(1);
 	return ret;
 }
@@ -1808,7 +1924,6 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	if (!found)
 		return -ENOMEM;
 
-	list_add(&found->list, &info->space_info);
 	INIT_LIST_HEAD(&found->block_groups);
 	init_rwsem(&found->groups_sem);
 	spin_lock_init(&found->lock);
@@ -1818,9 +1933,11 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->bytes_pinned = 0;
 	found->bytes_reserved = 0;
 	found->bytes_readonly = 0;
+	found->bytes_delalloc = 0;
 	found->full = 0;
 	found->force_alloc = 0;
 	*space_info = found;
+	list_add_rcu(&found->list, &info->space_info);
 	return 0;
 }
 
@@ -1881,6 +1998,233 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 	return flags;
 }
 
+static u64 btrfs_get_alloc_profile(struct btrfs_root *root, u64 data)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	u64 alloc_profile;
+
+	if (data) {
+		alloc_profile = info->avail_data_alloc_bits &
+			info->data_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
+	} else if (root == root->fs_info->chunk_root) {
+		alloc_profile = info->avail_system_alloc_bits &
+			info->system_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
+	} else {
+		alloc_profile = info->avail_metadata_alloc_bits &
+			info->metadata_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
+	}
+
+	return btrfs_reduce_alloc_profile(root, data);
+}
+
+void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
+{
+	u64 alloc_target;
+
+	alloc_target = btrfs_get_alloc_profile(root, 1);
+	BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
+						       alloc_target);
+}
+
+/*
+ * for now this just makes sure we have at least 5% of our metadata space free
+ * for use.
+ */
+int btrfs_check_metadata_free_space(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	struct btrfs_space_info *meta_sinfo;
+	u64 alloc_target, thresh;
+	int committed = 0, ret;
+
+	/* get the space info for where the metadata will live */
+	alloc_target = btrfs_get_alloc_profile(root, 0);
+	meta_sinfo = __find_space_info(info, alloc_target);
+
+again:
+	spin_lock(&meta_sinfo->lock);
+	if (!meta_sinfo->full)
+		thresh = meta_sinfo->total_bytes * 80;
+	else
+		thresh = meta_sinfo->total_bytes * 95;
+
+	do_div(thresh, 100);
+
+	if (meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
+	    meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly > thresh) {
+		struct btrfs_trans_handle *trans;
+		if (!meta_sinfo->full) {
+			meta_sinfo->force_alloc = 1;
+			spin_unlock(&meta_sinfo->lock);
+
+			trans = btrfs_start_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+					     2 * 1024 * 1024, alloc_target, 0);
+			btrfs_end_transaction(trans, root);
+			goto again;
+		}
+		spin_unlock(&meta_sinfo->lock);
+
+		if (!committed) {
+			committed = 1;
+			trans = btrfs_join_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+			ret = btrfs_commit_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+		return -ENOSPC;
+	}
+	spin_unlock(&meta_sinfo->lock);
+
+	return 0;
+}
+
+/*
+ * This will check the space that the inode allocates from to make sure we have
+ * enough space for bytes.
+ */
+int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
+				u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+	int ret = 0, committed = 0;
+
+	/* make sure bytes are sectorsize aligned */
+	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+
+	data_sinfo = BTRFS_I(inode)->space_info;
+again:
+	/* make sure we have enough space to handle the data first */
+	spin_lock(&data_sinfo->lock);
+	if (data_sinfo->total_bytes - data_sinfo->bytes_used -
+	    data_sinfo->bytes_delalloc - data_sinfo->bytes_reserved -
+	    data_sinfo->bytes_pinned - data_sinfo->bytes_readonly -
+	    data_sinfo->bytes_may_use < bytes) {
+		struct btrfs_trans_handle *trans;
+
+		/*
+		 * if we don't have enough free bytes in this space then we need
+		 * to alloc a new chunk.
+		 */
+		if (!data_sinfo->full) {
+			u64 alloc_target;
+
+			data_sinfo->force_alloc = 1;
+			spin_unlock(&data_sinfo->lock);
+
+			alloc_target = btrfs_get_alloc_profile(root, 1);
+			trans = btrfs_start_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+					     bytes + 2 * 1024 * 1024,
+					     alloc_target, 0);
+			btrfs_end_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+		spin_unlock(&data_sinfo->lock);
+
+		/* commit the current transaction and try again */
+		if (!committed) {
+			committed = 1;
+			trans = btrfs_join_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+			ret = btrfs_commit_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+
+		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
+		       ", %llu bytes_used, %llu bytes_reserved, "
+		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use"
+		       "%llu total\n", bytes, data_sinfo->bytes_delalloc,
+		       data_sinfo->bytes_used, data_sinfo->bytes_reserved,
+		       data_sinfo->bytes_pinned, data_sinfo->bytes_readonly,
+		       data_sinfo->bytes_may_use, data_sinfo->total_bytes);
+		return -ENOSPC;
+	}
+	data_sinfo->bytes_may_use += bytes;
+	BTRFS_I(inode)->reserved_bytes += bytes;
+	spin_unlock(&data_sinfo->lock);
+
+	return btrfs_check_metadata_free_space(root);
+}
+
+/*
+ * if there was an error for whatever reason after calling
+ * btrfs_check_data_free_space, call this so we can cleanup the counters.
+ */
+void btrfs_free_reserved_data_space(struct btrfs_root *root,
+				    struct inode *inode, u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+
+	/* make sure bytes are sectorsize aligned */
+	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+
+	data_sinfo = BTRFS_I(inode)->space_info;
+	spin_lock(&data_sinfo->lock);
+	data_sinfo->bytes_may_use -= bytes;
+	BTRFS_I(inode)->reserved_bytes -= bytes;
+	spin_unlock(&data_sinfo->lock);
+}
+
+/* called when we are adding a delalloc extent to the inode's io_tree */
+void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
+				  u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+
+	/* get the space info for where this inode will be storing its data */
+	data_sinfo = BTRFS_I(inode)->space_info;
+
+	/* make sure we have enough space to handle the data first */
+	spin_lock(&data_sinfo->lock);
+	data_sinfo->bytes_delalloc += bytes;
+
+	/*
+	 * we are adding a delalloc extent without calling
+	 * btrfs_check_data_free_space first.  This happens on a weird
+	 * writepage condition, but shouldn't hurt our accounting
+	 */
+	if (unlikely(bytes > BTRFS_I(inode)->reserved_bytes)) {
+		data_sinfo->bytes_may_use -= BTRFS_I(inode)->reserved_bytes;
+		BTRFS_I(inode)->reserved_bytes = 0;
+	} else {
+		data_sinfo->bytes_may_use -= bytes;
+		BTRFS_I(inode)->reserved_bytes -= bytes;
+	}
+
+	spin_unlock(&data_sinfo->lock);
+}
+
+/* called when we are clearing an delalloc extent from the inode's io_tree */
+void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
+			      u64 bytes)
+{
+	struct btrfs_space_info *info;
+
+	info = BTRFS_I(inode)->space_info;
+
+	spin_lock(&info->lock);
+	info->bytes_delalloc -= bytes;
+	spin_unlock(&info->lock);
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
@@ -2137,13 +2481,12 @@ static int finish_current_insert(struct btrfs_trans_handle *trans,
 	u64 end;
 	u64 priv;
 	u64 search = 0;
-	u64 skipped = 0;
 	struct btrfs_fs_info *info = extent_root->fs_info;
 	struct btrfs_path *path;
 	struct pending_extent_op *extent_op, *tmp;
 	struct list_head insert_list, update_list;
 	int ret;
-	int num_inserts = 0, max_inserts;
+	int num_inserts = 0, max_inserts, restart = 0;
 
 	path = btrfs_alloc_path();
 	INIT_LIST_HEAD(&insert_list);
@@ -2159,18 +2502,19 @@ again:
 		ret = find_first_extent_bit(&info->extent_ins, search, &start,
 					    &end, EXTENT_WRITEBACK);
 		if (ret) {
-			if (skipped && all && !num_inserts) {
-				skipped = 0;
+			if (restart && !num_inserts &&
+			    list_empty(&update_list)) {
+				restart = 0;
 				search = 0;
 				continue;
 			}
-			mutex_unlock(&info->extent_ins_mutex);
 			break;
 		}
 
 		ret = try_lock_extent(&info->extent_ins, start, end, GFP_NOFS);
 		if (!ret) {
-			skipped = 1;
+			if (all)
+				restart = 1;
 			search = end + 1;
 			if (need_resched()) {
 				mutex_unlock(&info->extent_ins_mutex);
@@ -2189,7 +2533,7 @@ again:
 			list_add_tail(&extent_op->list, &insert_list);
 			search = end + 1;
 			if (num_inserts == max_inserts) {
-				mutex_unlock(&info->extent_ins_mutex);
+				restart = 1;
 				break;
 			}
 		} else if (extent_op->type == PENDING_BACKREF_UPDATE) {
@@ -2205,7 +2549,6 @@ again:
 	 * somebody marked this thing for deletion then just unlock it and be
 	 * done, the free_extents will handle it
 	 */
-	mutex_lock(&info->extent_ins_mutex);
 	list_for_each_entry_safe(extent_op, tmp, &update_list, list) {
 		clear_extent_bits(&info->extent_ins, extent_op->bytenr,
 				  extent_op->bytenr + extent_op->num_bytes - 1,
@@ -2227,6 +2570,10 @@ again:
 	if (!list_empty(&update_list)) {
 		ret = update_backrefs(trans, extent_root, path, &update_list);
 		BUG_ON(ret);
+
+		/* we may have COW'ed new blocks, so lets start over */
+		if (all)
+			restart = 1;
 	}
 
 	/*
@@ -2234,9 +2581,9 @@ again:
 	 * need to make sure everything is cleaned then reset everything and
 	 * go back to the beginning
 	 */
-	if (!num_inserts && all && skipped) {
+	if (!num_inserts && restart) {
 		search = 0;
-		skipped = 0;
+		restart = 0;
 		INIT_LIST_HEAD(&update_list);
 		INIT_LIST_HEAD(&insert_list);
 		goto again;
@@ -2293,27 +2640,19 @@ again:
 	BUG_ON(ret);
 
 	/*
-	 * if we broke out of the loop in order to insert stuff because we hit
-	 * the maximum number of inserts at a time we can handle, then loop
-	 * back and pick up where we left off
-	 */
-	if (num_inserts == max_inserts) {
-		INIT_LIST_HEAD(&insert_list);
-		INIT_LIST_HEAD(&update_list);
-		num_inserts = 0;
-		goto again;
-	}
-
-	/*
-	 * again, if we need to make absolutely sure there are no more pending
-	 * extent operations left and we know that we skipped some, go back to
-	 * the beginning and do it all again
+	 * if restart is set for whatever reason we need to go back and start
+	 * searching through the pending list again.
+	 *
+	 * We just inserted some extents, which could have resulted in new
+	 * blocks being allocated, which would result in new blocks needing
+	 * updates, so if all is set we _must_ restart to get the updated
+	 * blocks.
 	 */
-	if (all && skipped) {
+	if (restart || all) {
 		INIT_LIST_HEAD(&insert_list);
 		INIT_LIST_HEAD(&update_list);
 		search = 0;
-		skipped = 0;
+		restart = 0;
 		num_inserts = 0;
 		goto again;
 	}
@@ -2547,6 +2886,7 @@ again:
 		if (ret) {
 			if (all && skipped && !nr) {
 				search = 0;
+				skipped = 0;
 				continue;
 			}
 			mutex_unlock(&info->extent_ins_mutex);
@@ -2633,6 +2973,8 @@ again:
 		goto again;
 	}
 
+	if (!err)
+		finish_current_insert(trans, extent_root, 0);
 	return err;
 }
 
@@ -2700,13 +3042,9 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	/* if metadata always pin */
 	if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID) {
 		if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
-			struct btrfs_block_group_cache *cache;
-
-			/* btrfs_free_reserved_extent */
-			cache = btrfs_lookup_block_group(root->fs_info, bytenr);
-			BUG_ON(!cache);
-			btrfs_add_free_space(cache, bytenr, num_bytes);
-			put_block_group(cache);
+			mutex_lock(&root->fs_info->pinned_mutex);
+			btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
+			mutex_unlock(&root->fs_info->pinned_mutex);
 			update_reserved_extents(root, bytenr, num_bytes, 0);
 			return 0;
 		}
@@ -2787,7 +3125,8 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 
 	if (data & BTRFS_BLOCK_GROUP_METADATA) {
 		last_ptr = &root->fs_info->last_alloc;
-		empty_cluster = 64 * 1024;
+		if (!btrfs_test_opt(root, SSD))
+			empty_cluster = 64 * 1024;
 	}
 
 	if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD))
@@ -3014,16 +3353,18 @@ loop_check:
 static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
 {
 	struct btrfs_block_group_cache *cache;
-	struct list_head *l;
 
 	printk(KERN_INFO "space_info has %llu free, is %sfull\n",
 	       (unsigned long long)(info->total_bytes - info->bytes_used -
 				    info->bytes_pinned - info->bytes_reserved),
 	       (info->full) ? "" : "not ");
+	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
+	       " may_use=%llu, used=%llu\n", info->total_bytes,
+	       info->bytes_pinned, info->bytes_delalloc, info->bytes_may_use,
+	       info->bytes_used);
 
 	down_read(&info->groups_sem);
-	list_for_each(l, &info->block_groups) {
-		cache = list_entry(l, struct btrfs_block_group_cache, list);
+	list_for_each_entry(cache, &info->block_groups, list) {
 		spin_lock(&cache->lock);
 		printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
 		       "%llu pinned %llu reserved\n",
@@ -3047,24 +3388,10 @@ static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 {
 	int ret;
 	u64 search_start = 0;
-	u64 alloc_profile;
 	struct btrfs_fs_info *info = root->fs_info;
 
-	if (data) {
-		alloc_profile = info->avail_data_alloc_bits &
-			info->data_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
-	} else if (root == root->fs_info->chunk_root) {
-		alloc_profile = info->avail_system_alloc_bits &
-			info->system_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
-	} else {
-		alloc_profile = info->avail_metadata_alloc_bits &
-			info->metadata_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
-	}
+	data = btrfs_get_alloc_profile(root, data);
 again:
-	data = btrfs_reduce_alloc_profile(root, data);
 	/*
 	 * the only place that sets empty_size is btrfs_realloc_node, which
 	 * is not called recursively on allocations
@@ -3332,7 +3659,8 @@ int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
 
 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 					    struct btrfs_root *root,
-					    u64 bytenr, u32 blocksize)
+					    u64 bytenr, u32 blocksize,
+					    int level)
 {
 	struct extent_buffer *buf;
 
@@ -3340,9 +3668,13 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 	if (!buf)
 		return ERR_PTR(-ENOMEM);
 	btrfs_set_header_generation(buf, trans->transid);
+	btrfs_set_buffer_lockdep_class(buf, level);
 	btrfs_tree_lock(buf);
 	clean_tree_block(trans, root, buf);
+
+	btrfs_set_lock_blocking(buf);
 	btrfs_set_buffer_uptodate(buf);
+
 	if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
 		set_extent_dirty(&root->dirty_log_pages, buf->start,
 			 buf->start + buf->len - 1, GFP_NOFS);
@@ -3351,6 +3683,7 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
 			 buf->start + buf->len - 1, GFP_NOFS);
 	}
 	trans->blocks_used++;
+	/* this returns a buffer locked for blocking */
 	return buf;
 }
 
@@ -3379,7 +3712,8 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 		return ERR_PTR(ret);
 	}
 
-	buf = btrfs_init_new_buffer(trans, root, ins.objectid, blocksize);
+	buf = btrfs_init_new_buffer(trans, root, ins.objectid,
+				    blocksize, level);
 	return buf;
 }
 
@@ -3388,36 +3722,73 @@ int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
 {
 	u64 leaf_owner;
 	u64 leaf_generation;
+	struct refsort *sorted;
 	struct btrfs_key key;
 	struct btrfs_file_extent_item *fi;
 	int i;
 	int nritems;
 	int ret;
+	int refi = 0;
+	int slot;
 
 	BUG_ON(!btrfs_is_leaf(leaf));
 	nritems = btrfs_header_nritems(leaf);
 	leaf_owner = btrfs_header_owner(leaf);
 	leaf_generation = btrfs_header_generation(leaf);
 
+	sorted = kmalloc(sizeof(*sorted) * nritems, GFP_NOFS);
+	/* we do this loop twice.  The first time we build a list
+	 * of the extents we have a reference on, then we sort the list
+	 * by bytenr.  The second time around we actually do the
+	 * extent freeing.
+	 */
 	for (i = 0; i < nritems; i++) {
 		u64 disk_bytenr;
 		cond_resched();
 
 		btrfs_item_key_to_cpu(leaf, &key, i);
+
+		/* only extents have references, skip everything else */
 		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
 			continue;
+
 		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
+
+		/* inline extents live in the btree, they don't have refs */
 		if (btrfs_file_extent_type(leaf, fi) ==
 		    BTRFS_FILE_EXTENT_INLINE)
 			continue;
-		/*
-		 * FIXME make sure to insert a trans record that
-		 * repeats the snapshot del on crash
-		 */
+
 		disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+
+		/* holes don't have refs */
 		if (disk_bytenr == 0)
 			continue;
 
+		sorted[refi].bytenr = disk_bytenr;
+		sorted[refi].slot = i;
+		refi++;
+	}
+
+	if (refi == 0)
+		goto out;
+
+	sort(sorted, refi, sizeof(struct refsort), refsort_cmp, NULL);
+
+	for (i = 0; i < refi; i++) {
+		u64 disk_bytenr;
+
+		disk_bytenr = sorted[i].bytenr;
+		slot = sorted[i].slot;
+
+		cond_resched();
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+			continue;
+
+		fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+
 		ret = __btrfs_free_extent(trans, root, disk_bytenr,
 				btrfs_file_extent_disk_num_bytes(leaf, fi),
 				leaf->start, leaf_owner, leaf_generation,
@@ -3428,6 +3799,8 @@ int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
 		wake_up(&root->fs_info->transaction_throttle);
 		cond_resched();
 	}
+out:
+	kfree(sorted);
 	return 0;
 }
 
@@ -3437,9 +3810,25 @@ static noinline int cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
 {
 	int i;
 	int ret;
-	struct btrfs_extent_info *info = ref->extents;
+	struct btrfs_extent_info *info;
+	struct refsort *sorted;
+
+	if (ref->nritems == 0)
+		return 0;
 
+	sorted = kmalloc(sizeof(*sorted) * ref->nritems, GFP_NOFS);
 	for (i = 0; i < ref->nritems; i++) {
+		sorted[i].bytenr = ref->extents[i].bytenr;
+		sorted[i].slot = i;
+	}
+	sort(sorted, ref->nritems, sizeof(struct refsort), refsort_cmp, NULL);
+
+	/*
+	 * the items in the ref were sorted when the ref was inserted
+	 * into the ref cache, so this is already in order
+	 */
+	for (i = 0; i < ref->nritems; i++) {
+		info = ref->extents + sorted[i].slot;
 		ret = __btrfs_free_extent(trans, root, info->bytenr,
 					  info->num_bytes, ref->bytenr,
 					  ref->owner, ref->generation,
@@ -3453,6 +3842,7 @@ static noinline int cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
 		info++;
 	}
 
+	kfree(sorted);
 	return 0;
 }
 
@@ -3497,6 +3887,152 @@ static int drop_snap_lookup_refcount(struct btrfs_root *root, u64 start,
 }
 
 /*
+ * this is used while deleting old snapshots, and it drops the refs
+ * on a whole subtree starting from a level 1 node.
+ *
+ * The idea is to sort all the leaf pointers, and then drop the
+ * ref on all the leaves in order.  Most of the time the leaves
+ * will have ref cache entries, so no leaf IOs will be required to
+ * find the extents they have references on.
+ *
+ * For each leaf, any references it has are also dropped in order
+ *
+ * This ends up dropping the references in something close to optimal
+ * order for reading and modifying the extent allocation tree.
+ */
+static noinline int drop_level_one_refs(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_path *path)
+{
+	u64 bytenr;
+	u64 root_owner;
+	u64 root_gen;
+	struct extent_buffer *eb = path->nodes[1];
+	struct extent_buffer *leaf;
+	struct btrfs_leaf_ref *ref;
+	struct refsort *sorted = NULL;
+	int nritems = btrfs_header_nritems(eb);
+	int ret;
+	int i;
+	int refi = 0;
+	int slot = path->slots[1];
+	u32 blocksize = btrfs_level_size(root, 0);
+	u32 refs;
+
+	if (nritems == 0)
+		goto out;
+
+	root_owner = btrfs_header_owner(eb);
+	root_gen = btrfs_header_generation(eb);
+	sorted = kmalloc(sizeof(*sorted) * nritems, GFP_NOFS);
+
+	/*
+	 * step one, sort all the leaf pointers so we don't scribble
+	 * randomly into the extent allocation tree
+	 */
+	for (i = slot; i < nritems; i++) {
+		sorted[refi].bytenr = btrfs_node_blockptr(eb, i);
+		sorted[refi].slot = i;
+		refi++;
+	}
+
+	/*
+	 * nritems won't be zero, but if we're picking up drop_snapshot
+	 * after a crash, slot might be > 0, so double check things
+	 * just in case.
+	 */
+	if (refi == 0)
+		goto out;
+
+	sort(sorted, refi, sizeof(struct refsort), refsort_cmp, NULL);
+
+	/*
+	 * the first loop frees everything the leaves point to
+	 */
+	for (i = 0; i < refi; i++) {
+		u64 ptr_gen;
+
+		bytenr = sorted[i].bytenr;
+
+		/*
+		 * check the reference count on this leaf.  If it is > 1
+		 * we just decrement it below and don't update any
+		 * of the refs the leaf points to.
+		 */
+		ret = drop_snap_lookup_refcount(root, bytenr, blocksize, &refs);
+		BUG_ON(ret);
+		if (refs != 1)
+			continue;
+
+		ptr_gen = btrfs_node_ptr_generation(eb, sorted[i].slot);
+
+		/*
+		 * the leaf only had one reference, which means the
+		 * only thing pointing to this leaf is the snapshot
+		 * we're deleting.  It isn't possible for the reference
+		 * count to increase again later
+		 *
+		 * The reference cache is checked for the leaf,
+		 * and if found we'll be able to drop any refs held by
+		 * the leaf without needing to read it in.
+		 */
+		ref = btrfs_lookup_leaf_ref(root, bytenr);
+		if (ref && ref->generation != ptr_gen) {
+			btrfs_free_leaf_ref(root, ref);
+			ref = NULL;
+		}
+		if (ref) {
+			ret = cache_drop_leaf_ref(trans, root, ref);
+			BUG_ON(ret);
+			btrfs_remove_leaf_ref(root, ref);
+			btrfs_free_leaf_ref(root, ref);
+		} else {
+			/*
+			 * the leaf wasn't in the reference cache, so
+			 * we have to read it.
+			 */
+			leaf = read_tree_block(root, bytenr, blocksize,
+					       ptr_gen);
+			ret = btrfs_drop_leaf_ref(trans, root, leaf);
+			BUG_ON(ret);
+			free_extent_buffer(leaf);
+		}
+		atomic_inc(&root->fs_info->throttle_gen);
+		wake_up(&root->fs_info->transaction_throttle);
+		cond_resched();
+	}
+
+	/*
+	 * run through the loop again to free the refs on the leaves.
+	 * This is faster than doing it in the loop above because
+	 * the leaves are likely to be clustered together.  We end up
+	 * working in nice chunks on the extent allocation tree.
+	 */
+	for (i = 0; i < refi; i++) {
+		bytenr = sorted[i].bytenr;
+		ret = __btrfs_free_extent(trans, root, bytenr,
+					blocksize, eb->start,
+					root_owner, root_gen, 0, 1);
+		BUG_ON(ret);
+
+		atomic_inc(&root->fs_info->throttle_gen);
+		wake_up(&root->fs_info->transaction_throttle);
+		cond_resched();
+	}
+out:
+	kfree(sorted);
+
+	/*
+	 * update the path to show we've processed the entire level 1
+	 * node.  This will get saved into the root's drop_snapshot_progress
+	 * field so these drops are not repeated again if this transaction
+	 * commits.
+	 */
+	path->slots[1] = nritems;
+	return 0;
+}
+
+/*
  * helper function for drop_snapshot, this walks down the tree dropping ref
  * counts as it goes.
  */
@@ -3511,7 +4047,6 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
 	struct extent_buffer *next;
 	struct extent_buffer *cur;
 	struct extent_buffer *parent;
-	struct btrfs_leaf_ref *ref;
 	u32 blocksize;
 	int ret;
 	u32 refs;
@@ -3538,17 +4073,46 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
 		if (path->slots[*level] >=
 		    btrfs_header_nritems(cur))
 			break;
+
+		/* the new code goes down to level 1 and does all the
+		 * leaves pointed to that node in bulk.  So, this check
+		 * for level 0 will always be false.
+		 *
+		 * But, the disk format allows the drop_snapshot_progress
+		 * field in the root to leave things in a state where
+		 * a leaf will need cleaning up here.  If someone crashes
+		 * with the old code and then boots with the new code,
+		 * we might find a leaf here.
+		 */
 		if (*level == 0) {
 			ret = btrfs_drop_leaf_ref(trans, root, cur);
 			BUG_ON(ret);
 			break;
 		}
+
+		/*
+		 * once we get to level one, process the whole node
+		 * at once, including everything below it.
+		 */
+		if (*level == 1) {
+			ret = drop_level_one_refs(trans, root, path);
+			BUG_ON(ret);
+			break;
+		}
+
 		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
 		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
 		blocksize = btrfs_level_size(root, *level - 1);
 
 		ret = drop_snap_lookup_refcount(root, bytenr, blocksize, &refs);
 		BUG_ON(ret);
+
+		/*
+		 * if there is more than one reference, we don't need
+		 * to read that node to drop any references it has.  We
+		 * just drop the ref we hold on that node and move on to the
+		 * next slot in this level.
+		 */
 		if (refs != 1) {
 			parent = path->nodes[*level];
 			root_owner = btrfs_header_owner(parent);
@@ -3567,46 +4131,12 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
 
 			continue;
 		}
+
 		/*
-		 * at this point, we have a single ref, and since the
-		 * only place referencing this extent is a dead root
-		 * the reference count should never go higher.
-		 * So, we don't need to check it again
+		 * we need to keep freeing things in the next level down.
+		 * read the block and loop around to process it
 		 */
-		if (*level == 1) {
-			ref = btrfs_lookup_leaf_ref(root, bytenr);
-			if (ref && ref->generation != ptr_gen) {
-				btrfs_free_leaf_ref(root, ref);
-				ref = NULL;
-			}
-			if (ref) {
-				ret = cache_drop_leaf_ref(trans, root, ref);
-				BUG_ON(ret);
-				btrfs_remove_leaf_ref(root, ref);
-				btrfs_free_leaf_ref(root, ref);
-				*level = 0;
-				break;
-			}
-		}
-		next = btrfs_find_tree_block(root, bytenr, blocksize);
-		if (!next || !btrfs_buffer_uptodate(next, ptr_gen)) {
-			free_extent_buffer(next);
-
-			next = read_tree_block(root, bytenr, blocksize,
-					       ptr_gen);
-			cond_resched();
-#if 0
-			/*
-			 * this is a debugging check and can go away
-			 * the ref should never go all the way down to 1
-			 * at this point
-			 */
-			ret = lookup_extent_ref(NULL, root, bytenr, blocksize,
-						&refs);
-			BUG_ON(ret);
-			WARN_ON(refs != 1);
-#endif
-		}
+		next = read_tree_block(root, bytenr, blocksize, ptr_gen);
 		WARN_ON(*level <= 0);
 		if (path->nodes[*level-1])
 			free_extent_buffer(path->nodes[*level-1]);
@@ -3631,11 +4161,16 @@ out:
 	root_owner = btrfs_header_owner(parent);
 	root_gen = btrfs_header_generation(parent);
 
+	/*
+	 * cleanup and free the reference on the last node
+	 * we processed
+	 */
 	ret = __btrfs_free_extent(trans, root, bytenr, blocksize,
 				  parent->start, root_owner, root_gen,
 				  *level, 1);
 	free_extent_buffer(path->nodes[*level]);
 	path->nodes[*level] = NULL;
+
 	*level += 1;
 	BUG_ON(ret);
 
@@ -3687,6 +4222,7 @@ static noinline int walk_down_subtree(struct btrfs_trans_handle *trans,
 
 		next = read_tree_block(root, bytenr, blocksize, ptr_gen);
 		btrfs_tree_lock(next);
+		btrfs_set_lock_blocking(next);
 
 		ret = btrfs_lookup_extent_ref(trans, root, bytenr, blocksize,
 					      &refs);
@@ -3754,6 +4290,13 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
 		if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
 			struct extent_buffer *node;
 			struct btrfs_disk_key disk_key;
+
+			/*
+			 * there is more work to do in this level.
+			 * Update the drop_progress marker to reflect
+			 * the work we've done so far, and then bump
+			 * the slot number
+			 */
 			node = path->nodes[i];
 			path->slots[i]++;
 			*level = i;
@@ -3765,6 +4308,11 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
 			return 0;
 		} else {
 			struct extent_buffer *parent;
+
+			/*
+			 * this whole node is done, free our reference
+			 * on it and go up one level
+			 */
 			if (path->nodes[*level] == root->node)
 				parent = path->nodes[*level];
 			else
@@ -3891,13 +4439,13 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
 
-	BUG_ON(!btrfs_tree_locked(parent));
+	btrfs_assert_tree_locked(parent);
 	parent_level = btrfs_header_level(parent);
 	extent_buffer_get(parent);
 	path->nodes[parent_level] = parent;
 	path->slots[parent_level] = btrfs_header_nritems(parent);
 
-	BUG_ON(!btrfs_tree_locked(node));
+	btrfs_assert_tree_locked(node);
 	level = btrfs_header_level(node);
 	extent_buffer_get(node);
 	path->nodes[level] = node;
@@ -4444,7 +4992,7 @@ static noinline int replace_one_extent(struct btrfs_trans_handle *trans,
 	u64 lock_end = 0;
 	u64 num_bytes;
 	u64 ext_offset;
-	u64 first_pos;
+	u64 search_end = (u64)-1;
 	u32 nritems;
 	int nr_scaned = 0;
 	int extent_locked = 0;
@@ -4452,7 +5000,6 @@ static noinline int replace_one_extent(struct btrfs_trans_handle *trans,
 	int ret;
 
 	memcpy(&key, leaf_key, sizeof(key));
-	first_pos = INT_LIMIT(loff_t) - extent_key->offset;
 	if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
 		if (key.objectid < ref_path->owner_objectid ||
 		    (key.objectid == ref_path->owner_objectid &&
@@ -4501,7 +5048,7 @@ next:
 			if ((key.objectid > ref_path->owner_objectid) ||
 			    (key.objectid == ref_path->owner_objectid &&
 			     key.type > BTRFS_EXTENT_DATA_KEY) ||
-			    (key.offset >= first_pos + extent_key->offset))
+			    key.offset >= search_end)
 				break;
 		}
 
@@ -4534,8 +5081,10 @@ next:
 		num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
 		ext_offset = btrfs_file_extent_offset(leaf, fi);
 
-		if (first_pos > key.offset - ext_offset)
-			first_pos = key.offset - ext_offset;
+		if (search_end == (u64)-1) {
+			search_end = key.offset - ext_offset +
+				btrfs_file_extent_ram_bytes(leaf, fi);
+		}
 
 		if (!extent_locked) {
 			lock_start = key.offset;
@@ -4724,7 +5273,7 @@ next:
 		}
 skip:
 		if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS &&
-		    key.offset >= first_pos + extent_key->offset)
+		    key.offset >= search_end)
 			break;
 
 		cond_resched();
@@ -4778,6 +5327,7 @@ int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
 		ref->bytenr = buf->start;
 		ref->owner = btrfs_header_owner(buf);
 		ref->generation = btrfs_header_generation(buf);
+
 		ret = btrfs_add_leaf_ref(root, ref, 0);
 		WARN_ON(ret);
 		btrfs_free_leaf_ref(root, ref);
@@ -5351,7 +5901,9 @@ static noinline int relocate_one_extent(struct btrfs_root *extent_root,
 			prev_block = block_start;
 		}
 
+		mutex_lock(&extent_root->fs_info->trans_mutex);
 		btrfs_record_root_in_trans(found_root);
+		mutex_unlock(&extent_root->fs_info->trans_mutex);
 		if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
 			/*
 			 * try to update data extent references while
@@ -5789,6 +6341,7 @@ out:
 int btrfs_free_block_groups(struct btrfs_fs_info *info)
 {
 	struct btrfs_block_group_cache *block_group;
+	struct btrfs_space_info *space_info;
 	struct rb_node *n;
 
 	spin_lock(&info->block_group_cache_lock);
@@ -5810,6 +6363,23 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 		spin_lock(&info->block_group_cache_lock);
 	}
 	spin_unlock(&info->block_group_cache_lock);
+
+	/* now that all the block groups are freed, go through and
+	 * free all the space_info structs.  This is only called during
+	 * the final stages of unmount, and so we know nobody is
+	 * using them.  We call synchronize_rcu() once before we start,
+	 * just to be on the safe side.
+	 */
+	synchronize_rcu();
+
+	while(!list_empty(&info->space_info)) {
+		space_info = list_entry(info->space_info.next,
+					struct btrfs_space_info,
+					list);
+
+		list_del(&space_info->list);
+		kfree(space_info);
+	}
 	return 0;
 }
 
@@ -5957,9 +6527,11 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
 
-	btrfs_remove_free_space_cache(block_group);
+	spin_lock(&root->fs_info->block_group_cache_lock);
 	rb_erase(&block_group->cache_node,
 		 &root->fs_info->block_group_cache_tree);
+	spin_unlock(&root->fs_info->block_group_cache_lock);
+	btrfs_remove_free_space_cache(block_group);
 	down_write(&block_group->space_info->groups_sem);
 	list_del(&block_group->list);
 	up_write(&block_group->space_info->groups_sem);
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index e086d407f1f..ebe6b29e606 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -9,7 +9,6 @@
 #include <linux/spinlock.h>
 #include <linux/blkdev.h>
 #include <linux/swap.h>
-#include <linux/version.h>
 #include <linux/writeback.h>
 #include <linux/pagevec.h>
 #include "extent_io.h"
@@ -31,7 +30,7 @@ static LIST_HEAD(buffers);
 static LIST_HEAD(states);
 
 #define LEAK_DEBUG 0
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 static DEFINE_SPINLOCK(leak_lock);
 #endif
 
@@ -120,7 +119,7 @@ void extent_io_tree_init(struct extent_io_tree *tree,
 static struct extent_state *alloc_extent_state(gfp_t mask)
 {
 	struct extent_state *state;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 	unsigned long flags;
 #endif
 
@@ -130,7 +129,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask)
 	state->state = 0;
 	state->private = 0;
 	state->tree = NULL;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 	spin_lock_irqsave(&leak_lock, flags);
 	list_add(&state->leak_list, &states);
 	spin_unlock_irqrestore(&leak_lock, flags);
@@ -145,11 +144,11 @@ static void free_extent_state(struct extent_state *state)
 	if (!state)
 		return;
 	if (atomic_dec_and_test(&state->refs)) {
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 		unsigned long flags;
 #endif
 		WARN_ON(state->tree);
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 		spin_lock_irqsave(&leak_lock, flags);
 		list_del(&state->leak_list);
 		spin_unlock_irqrestore(&leak_lock, flags);
@@ -416,8 +415,6 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
 
 	node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
 	if (node) {
-		struct extent_state *found;
-		found = rb_entry(node, struct extent_state, rb_node);
 		free_extent_state(prealloc);
 		return -EEXIST;
 	}
@@ -2378,11 +2375,6 @@ static int extent_write_cache_pages(struct extent_io_tree *tree,
 	int scanned = 0;
 	int range_whole = 0;
 
-	if (wbc->nonblocking && bdi_write_congested(bdi)) {
-		wbc->encountered_congestion = 1;
-		return 0;
-	}
-
 	pagevec_init(&pvec, 0);
 	if (wbc->range_cyclic) {
 		index = mapping->writeback_index; /* Start from prev offset */
@@ -2855,6 +2847,98 @@ out:
 	return sector;
 }
 
+int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+		__u64 start, __u64 len, get_extent_t *get_extent)
+{
+	int ret;
+	u64 off = start;
+	u64 max = start + len;
+	u32 flags = 0;
+	u64 disko = 0;
+	struct extent_map *em = NULL;
+	int end = 0;
+	u64 em_start = 0, em_len = 0;
+	unsigned long emflags;
+	ret = 0;
+
+	if (len == 0)
+		return -EINVAL;
+
+	lock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
+		GFP_NOFS);
+	em = get_extent(inode, NULL, 0, off, max - off, 0);
+	if (!em)
+		goto out;
+	if (IS_ERR(em)) {
+		ret = PTR_ERR(em);
+		goto out;
+	}
+	while (!end) {
+		off = em->start + em->len;
+		if (off >= max)
+			end = 1;
+
+		em_start = em->start;
+		em_len = em->len;
+
+		disko = 0;
+		flags = 0;
+
+		switch (em->block_start) {
+		case EXTENT_MAP_LAST_BYTE:
+			end = 1;
+			flags |= FIEMAP_EXTENT_LAST;
+			break;
+		case EXTENT_MAP_HOLE:
+			flags |= FIEMAP_EXTENT_UNWRITTEN;
+			break;
+		case EXTENT_MAP_INLINE:
+			flags |= (FIEMAP_EXTENT_DATA_INLINE |
+				  FIEMAP_EXTENT_NOT_ALIGNED);
+			break;
+		case EXTENT_MAP_DELALLOC:
+			flags |= (FIEMAP_EXTENT_DELALLOC |
+				  FIEMAP_EXTENT_UNKNOWN);
+			break;
+		default:
+			disko = em->block_start;
+			break;
+		}
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
+			flags |= FIEMAP_EXTENT_ENCODED;
+
+		emflags = em->flags;
+		free_extent_map(em);
+		em = NULL;
+
+		if (!end) {
+			em = get_extent(inode, NULL, 0, off, max - off, 0);
+			if (!em)
+				goto out;
+			if (IS_ERR(em)) {
+				ret = PTR_ERR(em);
+				goto out;
+			}
+			emflags = em->flags;
+		}
+		if (test_bit(EXTENT_FLAG_VACANCY, &emflags)) {
+			flags |= FIEMAP_EXTENT_LAST;
+			end = 1;
+		}
+
+		ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
+					em_len, flags);
+		if (ret)
+			goto out_free;
+	}
+out_free:
+	free_extent_map(em);
+out:
+	unlock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
+			GFP_NOFS);
+	return ret;
+}
+
 static inline struct page *extent_buffer_page(struct extent_buffer *eb,
 					      unsigned long i)
 {
@@ -2892,15 +2976,17 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 						   gfp_t mask)
 {
 	struct extent_buffer *eb = NULL;
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 	unsigned long flags;
 #endif
 
 	eb = kmem_cache_zalloc(extent_buffer_cache, mask);
 	eb->start = start;
 	eb->len = len;
-	mutex_init(&eb->mutex);
-#ifdef LEAK_DEBUG
+	spin_lock_init(&eb->lock);
+	init_waitqueue_head(&eb->lock_wq);
+
+#if LEAK_DEBUG
 	spin_lock_irqsave(&leak_lock, flags);
 	list_add(&eb->leak_list, &buffers);
 	spin_unlock_irqrestore(&leak_lock, flags);
@@ -2912,7 +2998,7 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 
 static void __free_extent_buffer(struct extent_buffer *eb)
 {
-#ifdef LEAK_DEBUG
+#if LEAK_DEBUG
 	unsigned long flags;
 	spin_lock_irqsave(&leak_lock, flags);
 	list_del(&eb->leak_list);
@@ -2980,8 +3066,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 		unlock_page(p);
 	}
 	if (uptodate)
-		eb->flags |= EXTENT_UPTODATE;
-	eb->flags |= EXTENT_BUFFER_FILLED;
+		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 
 	spin_lock(&tree->buffer_lock);
 	exists = buffer_tree_insert(tree, start, &eb->rb_node);
@@ -3135,7 +3220,7 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
 	unsigned long num_pages;
 
 	num_pages = num_extent_pages(eb->start, eb->len);
-	eb->flags &= ~EXTENT_UPTODATE;
+	clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 
 	clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
 			      GFP_NOFS);
@@ -3206,7 +3291,7 @@ int extent_buffer_uptodate(struct extent_io_tree *tree,
 	struct page *page;
 	int pg_uptodate = 1;
 
-	if (eb->flags & EXTENT_UPTODATE)
+	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 1;
 
 	ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
@@ -3242,7 +3327,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 	struct bio *bio = NULL;
 	unsigned long bio_flags = 0;
 
-	if (eb->flags & EXTENT_UPTODATE)
+	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 0;
 
 	if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
@@ -3273,7 +3358,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 	}
 	if (all_uptodate) {
 		if (start_i == 0)
-			eb->flags |= EXTENT_UPTODATE;
+			set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 		goto unlock_exit;
 	}
 
@@ -3309,7 +3394,7 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 	}
 
 	if (!ret)
-		eb->flags |= EXTENT_UPTODATE;
+		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 	return ret;
 
 unlock_exit:
@@ -3406,7 +3491,6 @@ int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
 		unmap_extent_buffer(eb, eb->map_token, km);
 		eb->map_token = NULL;
 		save = 1;
-		WARN_ON(!mutex_is_locked(&eb->mutex));
 	}
 	err = map_private_extent_buffer(eb, start, min_len, token, map,
 				       map_start, map_len, km);
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index c5b483a7913..1f9df88afbf 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -22,6 +22,10 @@
 /* flags for bio submission */
 #define EXTENT_BIO_COMPRESSED 1
 
+/* these are bit numbers for test/set bit */
+#define EXTENT_BUFFER_UPTODATE 0
+#define EXTENT_BUFFER_BLOCKING 1
+
 /*
  * page->private values.  Every page that is controlled by the extent
  * map has page->private set to one.
@@ -95,11 +99,19 @@ struct extent_buffer {
 	unsigned long map_start;
 	unsigned long map_len;
 	struct page *first_page;
+	unsigned long bflags;
 	atomic_t refs;
-	int flags;
 	struct list_head leak_list;
 	struct rb_node rb_node;
-	struct mutex mutex;
+
+	/* the spinlock is used to protect most operations */
+	spinlock_t lock;
+
+	/*
+	 * when we keep the lock held while blocking, waiters go onto
+	 * the wq
+	 */
+	wait_queue_head_t lock_wq;
 };
 
 struct extent_map_tree;
@@ -193,6 +205,8 @@ int extent_commit_write(struct extent_io_tree *tree,
 			unsigned from, unsigned to);
 sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
 		get_extent_t *get_extent);
+int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+		__u64 start, __u64 len, get_extent_t *get_extent);
 int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end);
 int set_state_private(struct extent_io_tree *tree, u64 start, u64 private);
 int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private);
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index 4a83e33ada3..50da69da20c 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -3,7 +3,6 @@
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <linux/spinlock.h>
-#include <linux/version.h>
 #include <linux/hardirq.h>
 #include "extent_map.h"
 
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 90268334145..dc78954861b 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -29,7 +29,6 @@
 #include <linux/writeback.h>
 #include <linux/statfs.h>
 #include <linux/compat.h>
-#include <linux/version.h>
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
@@ -1092,19 +1091,24 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		WARN_ON(num_pages > nrptrs);
 		memset(pages, 0, sizeof(struct page *) * nrptrs);
 
-		ret = btrfs_check_free_space(root, write_bytes, 0);
+		ret = btrfs_check_data_free_space(root, inode, write_bytes);
 		if (ret)
 			goto out;
 
 		ret = prepare_pages(root, file, pages, num_pages,
 				    pos, first_index, last_index,
 				    write_bytes);
-		if (ret)
+		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			goto out;
+		}
 
 		ret = btrfs_copy_from_user(pos, num_pages,
 					   write_bytes, pages, buf);
 		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			btrfs_drop_pages(pages, num_pages);
 			goto out;
 		}
@@ -1112,8 +1116,11 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		ret = dirty_and_release_pages(NULL, root, file, pages,
 					      num_pages, pos, write_bytes);
 		btrfs_drop_pages(pages, num_pages);
-		if (ret)
+		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			goto out;
+		}
 
 		if (will_write) {
 			btrfs_fdatawrite_range(inode->i_mapping, pos,
@@ -1137,6 +1144,8 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 	}
 out:
 	mutex_unlock(&inode->i_mutex);
+	if (ret)
+		err = ret;
 
 out_nolock:
 	kfree(pages);
@@ -1215,15 +1224,15 @@ int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 	}
 	mutex_unlock(&root->fs_info->trans_mutex);
 
-	root->fs_info->tree_log_batch++;
+	root->log_batch++;
 	filemap_fdatawrite(inode->i_mapping);
 	btrfs_wait_ordered_range(inode, 0, (u64)-1);
-	root->fs_info->tree_log_batch++;
+	root->log_batch++;
 
 	/*
 	 * ok we haven't committed the transaction yet, lets do a commit
 	 */
-	if (file->private_data)
+	if (file && file->private_data)
 		btrfs_ioctl_trans_end(file);
 
 	trans = btrfs_start_transaction(root, 1);
@@ -1232,7 +1241,7 @@ int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 		goto out;
 	}
 
-	ret = btrfs_log_dentry_safe(trans, root, file->f_dentry);
+	ret = btrfs_log_dentry_safe(trans, root, dentry);
 	if (ret < 0)
 		goto out;
 
@@ -1246,7 +1255,7 @@ int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 	 * file again, but that will end up using the synchronization
 	 * inside btrfs_sync_log to keep things safe.
 	 */
-	mutex_unlock(&file->f_dentry->d_inode->i_mutex);
+	mutex_unlock(&dentry->d_inode->i_mutex);
 
 	if (ret > 0) {
 		ret = btrfs_commit_transaction(trans, root);
@@ -1254,7 +1263,7 @@ int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
 		btrfs_sync_log(trans, root);
 		ret = btrfs_end_transaction(trans, root);
 	}
-	mutex_lock(&file->f_dentry->d_inode->i_mutex);
+	mutex_lock(&dentry->d_inode->i_mutex);
 out:
 	return ret > 0 ? EIO : ret;
 }
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index 2aa79873eb4..cc7334d833c 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -84,7 +84,6 @@ int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
 	search_key.type = 0;
 	search_key.offset = 0;
 
-	btrfs_init_path(path);
 	start_found = 0;
 	ret = btrfs_search_slot(trans, root, &search_key, path, 0, 0);
 	if (ret < 0)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 8adfe059ab4..7d4f948bc22 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -34,7 +34,6 @@
 #include <linux/statfs.h>
 #include <linux/compat.h>
 #include <linux/bit_spinlock.h>
-#include <linux/version.h>
 #include <linux/xattr.h>
 #include <linux/posix_acl.h>
 #include <linux/falloc.h>
@@ -51,6 +50,7 @@
 #include "tree-log.h"
 #include "ref-cache.h"
 #include "compression.h"
+#include "locking.h"
 
 struct btrfs_iget_args {
 	u64 ino;
@@ -91,32 +91,14 @@ static noinline int cow_file_range(struct inode *inode,
 				   u64 start, u64 end, int *page_started,
 				   unsigned long *nr_written, int unlock);
 
-/*
- * a very lame attempt at stopping writes when the FS is 85% full.  There
- * are countless ways this is incorrect, but it is better than nothing.
- */
-int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
-			   int for_del)
+static int btrfs_init_inode_security(struct inode *inode,  struct inode *dir)
 {
-	u64 total;
-	u64 used;
-	u64 thresh;
-	int ret = 0;
-
-	spin_lock(&root->fs_info->delalloc_lock);
-	total = btrfs_super_total_bytes(&root->fs_info->super_copy);
-	used = btrfs_super_bytes_used(&root->fs_info->super_copy);
-	if (for_del)
-		thresh = total * 90;
-	else
-		thresh = total * 85;
-
-	do_div(thresh, 100);
+	int err;
 
-	if (used + root->fs_info->delalloc_bytes + num_required > thresh)
-		ret = -ENOSPC;
-	spin_unlock(&root->fs_info->delalloc_lock);
-	return ret;
+	err = btrfs_init_acl(inode, dir);
+	if (!err)
+		err = btrfs_xattr_security_init(inode, dir);
+	return err;
 }
 
 /*
@@ -350,6 +332,19 @@ again:
 	nr_pages = (end >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT) + 1;
 	nr_pages = min(nr_pages, (128 * 1024UL) / PAGE_CACHE_SIZE);
 
+	/*
+	 * we don't want to send crud past the end of i_size through
+	 * compression, that's just a waste of CPU time.  So, if the
+	 * end of the file is before the start of our current
+	 * requested range of bytes, we bail out to the uncompressed
+	 * cleanup code that can deal with all of this.
+	 *
+	 * It isn't really the fastest way to fix things, but this is a
+	 * very uncommon corner.
+	 */
+	if (actual_end <= start)
+		goto cleanup_and_bail_uncompressed;
+
 	total_compressed = actual_end - start;
 
 	/* we want to make sure that amount of ram required to uncompress
@@ -494,6 +489,7 @@ again:
 			goto again;
 		}
 	} else {
+cleanup_and_bail_uncompressed:
 		/*
 		 * No compression, but we still need to write the pages in
 		 * the file we've been given so far.  redirty the locked
@@ -1166,6 +1162,7 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
 	 */
 	if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
+		btrfs_delalloc_reserve_space(root, inode, end - start + 1);
 		spin_lock(&root->fs_info->delalloc_lock);
 		BTRFS_I(inode)->delalloc_bytes += end - start + 1;
 		root->fs_info->delalloc_bytes += end - start + 1;
@@ -1199,9 +1196,12 @@ static int btrfs_clear_bit_hook(struct inode *inode, u64 start, u64 end,
 			       (unsigned long long)end - start + 1,
 			       (unsigned long long)
 			       root->fs_info->delalloc_bytes);
+			btrfs_delalloc_free_space(root, inode, (u64)-1);
 			root->fs_info->delalloc_bytes = 0;
 			BTRFS_I(inode)->delalloc_bytes = 0;
 		} else {
+			btrfs_delalloc_free_space(root, inode,
+						  end - start + 1);
 			root->fs_info->delalloc_bytes -= end - start + 1;
 			BTRFS_I(inode)->delalloc_bytes -= end - start + 1;
 		}
@@ -1324,12 +1324,11 @@ static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
 			     struct inode *inode, u64 file_offset,
 			     struct list_head *list)
 {
-	struct list_head *cur;
 	struct btrfs_ordered_sum *sum;
 
 	btrfs_set_trans_block_group(trans, inode);
-	list_for_each(cur, list) {
-		sum = list_entry(cur, struct btrfs_ordered_sum, list);
+
+	list_for_each_entry(sum, list, list) {
 		btrfs_csum_file_blocks(trans,
 		       BTRFS_I(inode)->root->fs_info->csum_root, sum);
 	}
@@ -2013,6 +2012,7 @@ void btrfs_read_locked_inode(struct inode *inode)
 	BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);
 
 	alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
+
 	BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
 						alloc_group_block, 0);
 	btrfs_free_path(path);
@@ -2039,6 +2039,7 @@ void btrfs_read_locked_inode(struct inode *inode)
 		inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
 		break;
 	default:
+		inode->i_op = &btrfs_special_inode_operations;
 		init_special_inode(inode, inode->i_mode, rdev);
 		break;
 	}
@@ -2108,6 +2109,7 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans,
 		goto failed;
 	}
 
+	btrfs_unlock_up_safe(path, 1);
 	leaf = path->nodes[0];
 	inode_item = btrfs_item_ptr(leaf, path->slots[0],
 				  struct btrfs_inode_item);
@@ -2219,10 +2221,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 
 	root = BTRFS_I(dir)->root;
 
-	ret = btrfs_check_free_space(root, 1, 1);
-	if (ret)
-		goto fail;
-
 	trans = btrfs_start_transaction(root, 1);
 
 	btrfs_set_trans_block_group(trans, dir);
@@ -2235,7 +2233,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 	nr = trans->blocks_used;
 
 	btrfs_end_transaction_throttle(trans, root);
-fail:
 	btrfs_btree_balance_dirty(root, nr);
 	return ret;
 }
@@ -2258,10 +2255,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 		return -ENOTEMPTY;
 	}
 
-	ret = btrfs_check_free_space(root, 1, 1);
-	if (ret)
-		goto fail;
-
 	trans = btrfs_start_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, dir);
 
@@ -2278,7 +2271,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 fail_trans:
 	nr = trans->blocks_used;
 	ret = btrfs_end_transaction_throttle(trans, root);
-fail:
 	btrfs_btree_balance_dirty(root, nr);
 
 	if (ret && !err)
@@ -2429,6 +2421,8 @@ next_node:
 			ref->generation = leaf_gen;
 			ref->nritems = 0;
 
+			btrfs_sort_leaf_ref(ref);
+
 			ret = btrfs_add_leaf_ref(root, ref, 0);
 			WARN_ON(ret);
 			btrfs_free_leaf_ref(root, ref);
@@ -2476,7 +2470,7 @@ noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	struct btrfs_path *path;
 	struct btrfs_key key;
 	struct btrfs_key found_key;
-	u32 found_type;
+	u32 found_type = (u8)-1;
 	struct extent_buffer *leaf;
 	struct btrfs_file_extent_item *fi;
 	u64 extent_start = 0;
@@ -2503,8 +2497,6 @@ noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	key.offset = (u64)-1;
 	key.type = (u8)-1;
 
-	btrfs_init_path(path);
-
 search_again:
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 	if (ret < 0)
@@ -2663,6 +2655,8 @@ next:
 			if (pending_del_nr)
 				goto del_pending;
 			btrfs_release_path(root, path);
+			if (found_type == BTRFS_INODE_ITEM_KEY)
+				break;
 			goto search_again;
 		}
 
@@ -2679,6 +2673,8 @@ del_pending:
 			BUG_ON(ret);
 			pending_del_nr = 0;
 			btrfs_release_path(root, path);
+			if (found_type == BTRFS_INODE_ITEM_KEY)
+				break;
 			goto search_again;
 		}
 	}
@@ -2788,7 +2784,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 	if (size <= hole_start)
 		return 0;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		return err;
 
@@ -2984,6 +2980,7 @@ static noinline void init_btrfs_i(struct inode *inode)
 	bi->last_trans = 0;
 	bi->logged_trans = 0;
 	bi->delalloc_bytes = 0;
+	bi->reserved_bytes = 0;
 	bi->disk_i_size = 0;
 	bi->flags = 0;
 	bi->index_cnt = (u64)-1;
@@ -3005,6 +3002,7 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p)
 	inode->i_ino = args->ino;
 	init_btrfs_i(inode);
 	BTRFS_I(inode)->root = args->root;
+	btrfs_set_inode_space_info(args->root, inode);
 	return 0;
 }
 
@@ -3265,7 +3263,7 @@ skip:
 
 	/* Reached end of directory/root. Bump pos past the last item. */
 	if (key_type == BTRFS_DIR_INDEX_KEY)
-		filp->f_pos = INT_LIMIT(typeof(filp->f_pos));
+		filp->f_pos = INT_LIMIT(off_t);
 	else
 		filp->f_pos++;
 nopos:
@@ -3425,6 +3423,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	BTRFS_I(inode)->index_cnt = 2;
 	BTRFS_I(inode)->root = root;
 	BTRFS_I(inode)->generation = trans->transid;
+	btrfs_set_inode_space_info(root, inode);
 
 	if (mode & S_IFDIR)
 		owner = 0;
@@ -3458,7 +3457,14 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 		root->highest_inode = objectid;
 
 	inode->i_uid = current_fsuid();
-	inode->i_gid = current_fsgid();
+
+	if (dir && (dir->i_mode & S_ISGID)) {
+		inode->i_gid = dir->i_gid;
+		if (S_ISDIR(mode))
+			mode |= S_ISGID;
+	} else
+		inode->i_gid = current_fsgid();
+
 	inode->i_mode = mode;
 	inode->i_ino = objectid;
 	inode_set_bytes(inode, 0);
@@ -3565,7 +3571,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (!new_valid_dev(rdev))
 		return -EINVAL;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 
@@ -3586,7 +3592,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (IS_ERR(inode))
 		goto out_unlock;
 
-	err = btrfs_init_acl(inode, dir);
+	err = btrfs_init_inode_security(inode, dir);
 	if (err) {
 		drop_inode = 1;
 		goto out_unlock;
@@ -3628,7 +3634,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 	u64 objectid;
 	u64 index = 0;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 	trans = btrfs_start_transaction(root, 1);
@@ -3649,7 +3655,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 	if (IS_ERR(inode))
 		goto out_unlock;
 
-	err = btrfs_init_acl(inode, dir);
+	err = btrfs_init_inode_security(inode, dir);
 	if (err) {
 		drop_inode = 1;
 		goto out_unlock;
@@ -3696,7 +3702,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 		return -ENOENT;
 
 	btrfs_inc_nlink(inode);
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 	err = btrfs_set_inode_index(dir, &index);
@@ -3742,7 +3748,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	u64 index = 0;
 	unsigned long nr = 1;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto out_unlock;
 
@@ -3772,7 +3778,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
 	drop_on_err = 1;
 
-	err = btrfs_init_acl(inode, dir);
+	err = btrfs_init_inode_security(inode, dir);
 	if (err)
 		goto out_fail;
 
@@ -4158,9 +4164,10 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
 	return -EINVAL;
 }
 
-static sector_t btrfs_bmap(struct address_space *mapping, sector_t iblock)
+static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+		__u64 start, __u64 len)
 {
-	return extent_bmap(mapping, iblock, btrfs_get_extent);
+	return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent);
 }
 
 int btrfs_readpage(struct file *file, struct page *page)
@@ -4223,7 +4230,7 @@ static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
 {
 	if (PageWriteback(page) || PageDirty(page))
 		return 0;
-	return __btrfs_releasepage(page, gfp_flags);
+	return __btrfs_releasepage(page, gfp_flags & GFP_NOFS);
 }
 
 static void btrfs_invalidatepage(struct page *page, unsigned long offset)
@@ -4298,7 +4305,7 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
 	u64 page_start;
 	u64 page_end;
 
-	ret = btrfs_check_free_space(root, PAGE_CACHE_SIZE, 0);
+	ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
 	if (ret)
 		goto out;
 
@@ -4311,6 +4318,7 @@ again:
 
 	if ((page->mapping != inode->i_mapping) ||
 	    (page_start >= size)) {
+		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
 		/* page got truncated out from underneath us */
 		goto out_unlock;
 	}
@@ -4593,7 +4601,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		return -EXDEV;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto out_unlock;
 
@@ -4711,7 +4719,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
 		return -ENAMETOOLONG;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto out_fail;
 
@@ -4733,7 +4741,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	if (IS_ERR(inode))
 		goto out_unlock;
 
-	err = btrfs_init_acl(inode, dir);
+	err = btrfs_init_inode_security(inode, dir);
 	if (err) {
 		drop_inode = 1;
 		goto out_unlock;
@@ -4987,13 +4995,24 @@ static struct extent_io_ops btrfs_extent_io_ops = {
 	.clear_bit_hook = btrfs_clear_bit_hook,
 };
 
+/*
+ * btrfs doesn't support the bmap operation because swapfiles
+ * use bmap to make a mapping of extents in the file.  They assume
+ * these extents won't change over the life of the file and they
+ * use the bmap result to do IO directly to the drive.
+ *
+ * the btrfs bmap call would return logical addresses that aren't
+ * suitable for IO and they also will change frequently as COW
+ * operations happen.  So, swapfile + btrfs == corruption.
+ *
+ * For now we're avoiding this by dropping bmap.
+ */
 static struct address_space_operations btrfs_aops = {
 	.readpage	= btrfs_readpage,
 	.writepage	= btrfs_writepage,
 	.writepages	= btrfs_writepages,
 	.readpages	= btrfs_readpages,
 	.sync_page	= block_sync_page,
-	.bmap		= btrfs_bmap,
 	.direct_IO	= btrfs_direct_IO,
 	.invalidatepage = btrfs_invalidatepage,
 	.releasepage	= btrfs_releasepage,
@@ -5017,6 +5036,7 @@ static struct inode_operations btrfs_file_inode_operations = {
 	.removexattr	= btrfs_removexattr,
 	.permission	= btrfs_permission,
 	.fallocate	= btrfs_fallocate,
+	.fiemap		= btrfs_fiemap,
 };
 static struct inode_operations btrfs_special_inode_operations = {
 	.getattr	= btrfs_getattr,
@@ -5032,4 +5052,8 @@ static struct inode_operations btrfs_symlink_inode_operations = {
 	.follow_link	= page_follow_link_light,
 	.put_link	= page_put_link,
 	.permission	= btrfs_permission,
+	.setxattr	= btrfs_setxattr,
+	.getxattr	= btrfs_getxattr,
+	.listxattr	= btrfs_listxattr,
+	.removexattr	= btrfs_removexattr,
 };
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index c2aa33e3feb..bca729fc80c 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -38,7 +38,6 @@
 #include <linux/compat.h>
 #include <linux/bit_spinlock.h>
 #include <linux/security.h>
-#include <linux/version.h>
 #include <linux/xattr.h>
 #include <linux/vmalloc.h>
 #include "compat.h"
@@ -71,7 +70,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	u64 index = 0;
 	unsigned long nr = 1;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto fail_commit;
 
@@ -204,7 +203,7 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
 	if (!root->ref_cows)
 		return -EINVAL;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto fail_unlock;
 
@@ -375,7 +374,7 @@ static int btrfs_defrag_file(struct file *file)
 	unsigned long i;
 	int ret;
 
-	ret = btrfs_check_free_space(root, inode->i_size, 0);
+	ret = btrfs_check_data_free_space(root, inode, inode->i_size);
 	if (ret)
 		return -ENOSPC;
 
diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c
index 39bae7761db..47b0a88c12a 100644
--- a/fs/btrfs/locking.c
+++ b/fs/btrfs/locking.c
@@ -25,64 +25,203 @@
 #include "extent_io.h"
 #include "locking.h"
 
+static inline void spin_nested(struct extent_buffer *eb)
+{
+	spin_lock(&eb->lock);
+}
+
 /*
- * locks the per buffer mutex in an extent buffer.  This uses adaptive locks
- * and the spin is not tuned very extensively.  The spinning does make a big
- * difference in almost every workload, but spinning for the right amount of
- * time needs some help.
- *
- * In general, we want to spin as long as the lock holder is doing btree
- * searches, and we should give up if they are in more expensive code.
+ * Setting a lock to blocking will drop the spinlock and set the
+ * flag that forces other procs who want the lock to wait.  After
+ * this you can safely schedule with the lock held.
  */
+void btrfs_set_lock_blocking(struct extent_buffer *eb)
+{
+	if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) {
+		set_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags);
+		spin_unlock(&eb->lock);
+	}
+	/* exit with the spin lock released and the bit set */
+}
 
-int btrfs_tree_lock(struct extent_buffer *eb)
+/*
+ * clearing the blocking flag will take the spinlock again.
+ * After this you can't safely schedule
+ */
+void btrfs_clear_lock_blocking(struct extent_buffer *eb)
 {
-	int i;
+	if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) {
+		spin_nested(eb);
+		clear_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags);
+		smp_mb__after_clear_bit();
+	}
+	/* exit with the spin lock held */
+}
 
-	if (mutex_trylock(&eb->mutex))
-		return 0;
+/*
+ * unfortunately, many of the places that currently set a lock to blocking
+ * don't end up blocking for every long, and often they don't block
+ * at all.  For a dbench 50 run, if we don't spin one the blocking bit
+ * at all, the context switch rate can jump up to 400,000/sec or more.
+ *
+ * So, we're still stuck with this crummy spin on the blocking bit,
+ * at least until the most common causes of the short blocks
+ * can be dealt with.
+ */
+static int btrfs_spin_on_block(struct extent_buffer *eb)
+{
+	int i;
 	for (i = 0; i < 512; i++) {
 		cpu_relax();
-		if (mutex_trylock(&eb->mutex))
-			return 0;
+		if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+			return 1;
+		if (need_resched())
+			break;
 	}
-	cpu_relax();
-	mutex_lock_nested(&eb->mutex, BTRFS_MAX_LEVEL - btrfs_header_level(eb));
 	return 0;
 }
 
-int btrfs_try_tree_lock(struct extent_buffer *eb)
+/*
+ * This is somewhat different from trylock.  It will take the
+ * spinlock but if it finds the lock is set to blocking, it will
+ * return without the lock held.
+ *
+ * returns 1 if it was able to take the lock and zero otherwise
+ *
+ * After this call, scheduling is not safe without first calling
+ * btrfs_set_lock_blocking()
+ */
+int btrfs_try_spin_lock(struct extent_buffer *eb)
 {
-	return mutex_trylock(&eb->mutex);
+	int i;
+
+	spin_nested(eb);
+	if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+		return 1;
+	spin_unlock(&eb->lock);
+
+	/* spin for a bit on the BLOCKING flag */
+	for (i = 0; i < 2; i++) {
+		if (!btrfs_spin_on_block(eb))
+			break;
+
+		spin_nested(eb);
+		if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+			return 1;
+		spin_unlock(&eb->lock);
+	}
+	return 0;
 }
 
-int btrfs_tree_unlock(struct extent_buffer *eb)
+/*
+ * the autoremove wake function will return 0 if it tried to wake up
+ * a process that was already awake, which means that process won't
+ * count as an exclusive wakeup.  The waitq code will continue waking
+ * procs until it finds one that was actually sleeping.
+ *
+ * For btrfs, this isn't quite what we want.  We want a single proc
+ * to be notified that the lock is ready for taking.  If that proc
+ * already happen to be awake, great, it will loop around and try for
+ * the lock.
+ *
+ * So, btrfs_wake_function always returns 1, even when the proc that we
+ * tried to wake up was already awake.
+ */
+static int btrfs_wake_function(wait_queue_t *wait, unsigned mode,
+			       int sync, void *key)
 {
-	mutex_unlock(&eb->mutex);
-	return 0;
+	autoremove_wake_function(wait, mode, sync, key);
+	return 1;
 }
 
-int btrfs_tree_locked(struct extent_buffer *eb)
+/*
+ * returns with the extent buffer spinlocked.
+ *
+ * This will spin and/or wait as required to take the lock, and then
+ * return with the spinlock held.
+ *
+ * After this call, scheduling is not safe without first calling
+ * btrfs_set_lock_blocking()
+ */
+int btrfs_tree_lock(struct extent_buffer *eb)
 {
-	return mutex_is_locked(&eb->mutex);
+	DEFINE_WAIT(wait);
+	wait.func = btrfs_wake_function;
+
+	while(1) {
+		spin_nested(eb);
+
+		/* nobody is blocking, exit with the spinlock held */
+		if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+			return 0;
+
+		/*
+		 * we have the spinlock, but the real owner is blocking.
+		 * wait for them
+		 */
+		spin_unlock(&eb->lock);
+
+		/*
+		 * spin for a bit, and if the blocking flag goes away,
+		 * loop around
+		 */
+		if (btrfs_spin_on_block(eb))
+			continue;
+
+		prepare_to_wait_exclusive(&eb->lock_wq, &wait,
+					  TASK_UNINTERRUPTIBLE);
+
+		if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+			schedule();
+
+		finish_wait(&eb->lock_wq, &wait);
+	}
+	return 0;
 }
 
 /*
- * btrfs_search_slot uses this to decide if it should drop its locks
- * before doing something expensive like allocating free blocks for cow.
+ * Very quick trylock, this does not spin or schedule.  It returns
+ * 1 with the spinlock held if it was able to take the lock, or it
+ * returns zero if it was unable to take the lock.
+ *
+ * After this call, scheduling is not safe without first calling
+ * btrfs_set_lock_blocking()
  */
-int btrfs_path_lock_waiting(struct btrfs_path *path, int level)
+int btrfs_try_tree_lock(struct extent_buffer *eb)
 {
-	int i;
-	struct extent_buffer *eb;
-	for (i = level; i <= level + 1 && i < BTRFS_MAX_LEVEL; i++) {
-		eb = path->nodes[i];
-		if (!eb)
-			break;
-		smp_mb();
-		if (!list_empty(&eb->mutex.wait_list))
-			return 1;
+	if (spin_trylock(&eb->lock)) {
+		if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) {
+			/*
+			 * we've got the spinlock, but the real owner is
+			 * blocking.  Drop the spinlock and return failure
+			 */
+			spin_unlock(&eb->lock);
+			return 0;
+		}
+		return 1;
 	}
+	/* someone else has the spinlock giveup */
 	return 0;
 }
 
+int btrfs_tree_unlock(struct extent_buffer *eb)
+{
+	/*
+	 * if we were a blocking owner, we don't have the spinlock held
+	 * just clear the bit and look for waiters
+	 */
+	if (test_and_clear_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+		smp_mb__after_clear_bit();
+	else
+		spin_unlock(&eb->lock);
+
+	if (waitqueue_active(&eb->lock_wq))
+		wake_up(&eb->lock_wq);
+	return 0;
+}
+
+void btrfs_assert_tree_locked(struct extent_buffer *eb)
+{
+	if (!test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags))
+		assert_spin_locked(&eb->lock);
+}
diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h
index bc1faef1251..6c4ce457168 100644
--- a/fs/btrfs/locking.h
+++ b/fs/btrfs/locking.h
@@ -21,7 +21,11 @@
 
 int btrfs_tree_lock(struct extent_buffer *eb);
 int btrfs_tree_unlock(struct extent_buffer *eb);
-int btrfs_tree_locked(struct extent_buffer *eb);
+
 int btrfs_try_tree_lock(struct extent_buffer *eb);
-int btrfs_path_lock_waiting(struct btrfs_path *path, int level);
+int btrfs_try_spin_lock(struct extent_buffer *eb);
+
+void btrfs_set_lock_blocking(struct extent_buffer *eb);
+void btrfs_clear_lock_blocking(struct extent_buffer *eb);
+void btrfs_assert_tree_locked(struct extent_buffer *eb);
 #endif
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index a2094017027..77c2411a5f0 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -613,7 +613,6 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
 	struct btrfs_sector_sum *sector_sums;
 	struct btrfs_ordered_extent *ordered;
 	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
-	struct list_head *cur;
 	unsigned long num_sectors;
 	unsigned long i;
 	u32 sectorsize = BTRFS_I(inode)->root->sectorsize;
@@ -624,8 +623,7 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
 		return 1;
 
 	mutex_lock(&tree->mutex);
-	list_for_each_prev(cur, &ordered->list) {
-		ordered_sum = list_entry(cur, struct btrfs_ordered_sum, list);
+	list_for_each_entry_reverse(ordered_sum, &ordered->list, list) {
 		if (disk_bytenr >= ordered_sum->bytenr) {
 			num_sectors = ordered_sum->len / sectorsize;
 			sector_sums = ordered_sum->sums;
diff --git a/fs/btrfs/ref-cache.c b/fs/btrfs/ref-cache.c
index 6f0acc4c9ea..d0cc62bccb9 100644
--- a/fs/btrfs/ref-cache.c
+++ b/fs/btrfs/ref-cache.c
@@ -17,6 +17,7 @@
  */
 
 #include <linux/sched.h>
+#include <linux/sort.h>
 #include "ctree.h"
 #include "ref-cache.h"
 #include "transaction.h"
diff --git a/fs/btrfs/ref-cache.h b/fs/btrfs/ref-cache.h
index 16f3183d7c5..bc283ad2db7 100644
--- a/fs/btrfs/ref-cache.h
+++ b/fs/btrfs/ref-cache.h
@@ -73,5 +73,4 @@ int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref,
 int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen,
 			   int shared);
 int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref);
-
 #endif
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index db9fb3bc1e3..19a4daf03cc 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -37,7 +37,6 @@
 #include <linux/ctype.h>
 #include <linux/namei.h>
 #include <linux/miscdevice.h>
-#include <linux/version.h>
 #include <linux/magic.h>
 #include "compat.h"
 #include "ctree.h"
@@ -380,7 +379,6 @@ int btrfs_sync_fs(struct super_block *sb, int wait)
 	btrfs_start_delalloc_inodes(root);
 	btrfs_wait_ordered_extents(root, 0);
 
-	btrfs_clean_old_snapshots(root);
 	trans = btrfs_start_transaction(root, 1);
 	ret = btrfs_commit_transaction(trans, root);
 	sb->s_dirt = 0;
@@ -512,6 +510,10 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 	struct btrfs_root *root = btrfs_sb(sb);
 	int ret;
 
+	ret = btrfs_parse_options(root, data);
+	if (ret)
+		return -EINVAL;
+
 	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
 		return 0;
 
@@ -583,17 +585,18 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
 	struct btrfs_ioctl_vol_args *vol;
 	struct btrfs_fs_devices *fs_devices;
 	int ret = -ENOTTY;
-	int len;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
 	vol = kmalloc(sizeof(*vol), GFP_KERNEL);
+	if (!vol)
+		return -ENOMEM;
+
 	if (copy_from_user(vol, (void __user *)arg, sizeof(*vol))) {
 		ret = -EFAULT;
 		goto out;
 	}
-	len = strnlen(vol->name, BTRFS_PATH_NAME_MAX);
 
 	switch (cmd) {
 	case BTRFS_IOC_SCAN_DEV:
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 8a08f944334..4112d53d4f4 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -688,7 +688,9 @@ static noinline int drop_dirty_roots(struct btrfs_root *tree_root,
 		num_bytes -= btrfs_root_used(&dirty->root->root_item);
 		bytes_used = btrfs_root_used(&root->root_item);
 		if (num_bytes) {
+			mutex_lock(&root->fs_info->trans_mutex);
 			btrfs_record_root_in_trans(root);
+			mutex_unlock(&root->fs_info->trans_mutex);
 			btrfs_set_root_used(&root->root_item,
 					    bytes_used - num_bytes);
 		}
@@ -852,11 +854,9 @@ static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_pending_snapshot *pending;
 	struct list_head *head = &trans->transaction->pending_snapshots;
-	struct list_head *cur;
 	int ret;
 
-	list_for_each(cur, head) {
-		pending = list_entry(cur, struct btrfs_pending_snapshot, list);
+	list_for_each_entry(pending, head, list) {
 		ret = create_pending_snapshot(trans, fs_info, pending);
 		BUG_ON(ret);
 	}
diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c
index 3e8358c3616..98d25fa4570 100644
--- a/fs/btrfs/tree-defrag.c
+++ b/fs/btrfs/tree-defrag.c
@@ -74,6 +74,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 		u32 nritems;
 
 		root_node = btrfs_lock_root_node(root);
+		btrfs_set_lock_blocking(root_node);
 		nritems = btrfs_header_nritems(root_node);
 		root->defrag_max.objectid = 0;
 		/* from above we know this is not a leaf */
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index d81cda2e077..9c462fbd60f 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -78,104 +78,6 @@ static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
  */
 
 /*
- * btrfs_add_log_tree adds a new per-subvolume log tree into the
- * tree of log tree roots.  This must be called with a tree log transaction
- * running (see start_log_trans).
- */
-static int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
-		      struct btrfs_root *root)
-{
-	struct btrfs_key key;
-	struct btrfs_root_item root_item;
-	struct btrfs_inode_item *inode_item;
-	struct extent_buffer *leaf;
-	struct btrfs_root *new_root = root;
-	int ret;
-	u64 objectid = root->root_key.objectid;
-
-	leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
-				      BTRFS_TREE_LOG_OBJECTID,
-				      trans->transid, 0, 0, 0);
-	if (IS_ERR(leaf)) {
-		ret = PTR_ERR(leaf);
-		return ret;
-	}
-
-	btrfs_set_header_nritems(leaf, 0);
-	btrfs_set_header_level(leaf, 0);
-	btrfs_set_header_bytenr(leaf, leaf->start);
-	btrfs_set_header_generation(leaf, trans->transid);
-	btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
-
-	write_extent_buffer(leaf, root->fs_info->fsid,
-			    (unsigned long)btrfs_header_fsid(leaf),
-			    BTRFS_FSID_SIZE);
-	btrfs_mark_buffer_dirty(leaf);
-
-	inode_item = &root_item.inode;
-	memset(inode_item, 0, sizeof(*inode_item));
-	inode_item->generation = cpu_to_le64(1);
-	inode_item->size = cpu_to_le64(3);
-	inode_item->nlink = cpu_to_le32(1);
-	inode_item->nbytes = cpu_to_le64(root->leafsize);
-	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
-
-	btrfs_set_root_bytenr(&root_item, leaf->start);
-	btrfs_set_root_generation(&root_item, trans->transid);
-	btrfs_set_root_level(&root_item, 0);
-	btrfs_set_root_refs(&root_item, 0);
-	btrfs_set_root_used(&root_item, 0);
-
-	memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
-	root_item.drop_level = 0;
-
-	btrfs_tree_unlock(leaf);
-	free_extent_buffer(leaf);
-	leaf = NULL;
-
-	btrfs_set_root_dirid(&root_item, 0);
-
-	key.objectid = BTRFS_TREE_LOG_OBJECTID;
-	key.offset = objectid;
-	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
-	ret = btrfs_insert_root(trans, root->fs_info->log_root_tree, &key,
-				&root_item);
-	if (ret)
-		goto fail;
-
-	new_root = btrfs_read_fs_root_no_radix(root->fs_info->log_root_tree,
-					       &key);
-	BUG_ON(!new_root);
-
-	WARN_ON(root->log_root);
-	root->log_root = new_root;
-
-	/*
-	 * log trees do not get reference counted because they go away
-	 * before a real commit is actually done.  They do store pointers
-	 * to file data extents, and those reference counts still get
-	 * updated (along with back refs to the log tree).
-	 */
-	new_root->ref_cows = 0;
-	new_root->last_trans = trans->transid;
-
-	/*
-	 * we need to make sure the root block for this new tree
-	 * is marked as dirty in the dirty_log_pages tree.  This
-	 * is how it gets flushed down to disk at tree log commit time.
-	 *
-	 * the tree logging mutex keeps others from coming in and changing
-	 * the new_root->node, so we can safely access it here
-	 */
-	set_extent_dirty(&new_root->dirty_log_pages, new_root->node->start,
-			 new_root->node->start + new_root->node->len - 1,
-			 GFP_NOFS);
-
-fail:
-	return ret;
-}
-
-/*
  * start a sub transaction and setup the log tree
  * this increments the log tree writer count to make the people
  * syncing the tree wait for us to finish
@@ -184,6 +86,14 @@ static int start_log_trans(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root)
 {
 	int ret;
+
+	mutex_lock(&root->log_mutex);
+	if (root->log_root) {
+		root->log_batch++;
+		atomic_inc(&root->log_writers);
+		mutex_unlock(&root->log_mutex);
+		return 0;
+	}
 	mutex_lock(&root->fs_info->tree_log_mutex);
 	if (!root->fs_info->log_root_tree) {
 		ret = btrfs_init_log_root_tree(trans, root->fs_info);
@@ -193,9 +103,10 @@ static int start_log_trans(struct btrfs_trans_handle *trans,
 		ret = btrfs_add_log_tree(trans, root);
 		BUG_ON(ret);
 	}
-	atomic_inc(&root->fs_info->tree_log_writers);
-	root->fs_info->tree_log_batch++;
 	mutex_unlock(&root->fs_info->tree_log_mutex);
+	root->log_batch++;
+	atomic_inc(&root->log_writers);
+	mutex_unlock(&root->log_mutex);
 	return 0;
 }
 
@@ -212,13 +123,12 @@ static int join_running_log_trans(struct btrfs_root *root)
 	if (!root->log_root)
 		return -ENOENT;
 
-	mutex_lock(&root->fs_info->tree_log_mutex);
+	mutex_lock(&root->log_mutex);
 	if (root->log_root) {
 		ret = 0;
-		atomic_inc(&root->fs_info->tree_log_writers);
-		root->fs_info->tree_log_batch++;
+		atomic_inc(&root->log_writers);
 	}
-	mutex_unlock(&root->fs_info->tree_log_mutex);
+	mutex_unlock(&root->log_mutex);
 	return ret;
 }
 
@@ -228,10 +138,11 @@ static int join_running_log_trans(struct btrfs_root *root)
  */
 static int end_log_trans(struct btrfs_root *root)
 {
-	atomic_dec(&root->fs_info->tree_log_writers);
-	smp_mb();
-	if (waitqueue_active(&root->fs_info->tree_log_wait))
-		wake_up(&root->fs_info->tree_log_wait);
+	if (atomic_dec_and_test(&root->log_writers)) {
+		smp_mb();
+		if (waitqueue_active(&root->log_writer_wait))
+			wake_up(&root->log_writer_wait);
+	}
 	return 0;
 }
 
@@ -1704,6 +1615,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 
 				btrfs_tree_lock(next);
 				clean_tree_block(trans, root, next);
+				btrfs_set_lock_blocking(next);
 				btrfs_wait_tree_block_writeback(next);
 				btrfs_tree_unlock(next);
 
@@ -1750,6 +1662,7 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
 		next = path->nodes[*level];
 		btrfs_tree_lock(next);
 		clean_tree_block(trans, root, next);
+		btrfs_set_lock_blocking(next);
 		btrfs_wait_tree_block_writeback(next);
 		btrfs_tree_unlock(next);
 
@@ -1807,6 +1720,7 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
 
 				btrfs_tree_lock(next);
 				clean_tree_block(trans, root, next);
+				btrfs_set_lock_blocking(next);
 				btrfs_wait_tree_block_writeback(next);
 				btrfs_tree_unlock(next);
 
@@ -1879,6 +1793,7 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
 
 			btrfs_tree_lock(next);
 			clean_tree_block(trans, log, next);
+			btrfs_set_lock_blocking(next);
 			btrfs_wait_tree_block_writeback(next);
 			btrfs_tree_unlock(next);
 
@@ -1902,26 +1817,65 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
 		}
 	}
 	btrfs_free_path(path);
-	if (wc->free)
-		free_extent_buffer(log->node);
 	return ret;
 }
 
-static int wait_log_commit(struct btrfs_root *log)
+/*
+ * helper function to update the item for a given subvolumes log root
+ * in the tree of log roots
+ */
+static int update_log_root(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *log)
+{
+	int ret;
+
+	if (log->log_transid == 1) {
+		/* insert root item on the first sync */
+		ret = btrfs_insert_root(trans, log->fs_info->log_root_tree,
+				&log->root_key, &log->root_item);
+	} else {
+		ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
+				&log->root_key, &log->root_item);
+	}
+	return ret;
+}
+
+static int wait_log_commit(struct btrfs_root *root, unsigned long transid)
 {
 	DEFINE_WAIT(wait);
-	u64 transid = log->fs_info->tree_log_transid;
+	int index = transid % 2;
 
+	/*
+	 * we only allow two pending log transactions at a time,
+	 * so we know that if ours is more than 2 older than the
+	 * current transaction, we're done
+	 */
 	do {
-		prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
-				TASK_UNINTERRUPTIBLE);
-		mutex_unlock(&log->fs_info->tree_log_mutex);
-		if (atomic_read(&log->fs_info->tree_log_commit))
+		prepare_to_wait(&root->log_commit_wait[index],
+				&wait, TASK_UNINTERRUPTIBLE);
+		mutex_unlock(&root->log_mutex);
+		if (root->log_transid < transid + 2 &&
+		    atomic_read(&root->log_commit[index]))
 			schedule();
-		finish_wait(&log->fs_info->tree_log_wait, &wait);
-		mutex_lock(&log->fs_info->tree_log_mutex);
-	} while (transid == log->fs_info->tree_log_transid &&
-		atomic_read(&log->fs_info->tree_log_commit));
+		finish_wait(&root->log_commit_wait[index], &wait);
+		mutex_lock(&root->log_mutex);
+	} while (root->log_transid < transid + 2 &&
+		 atomic_read(&root->log_commit[index]));
+	return 0;
+}
+
+static int wait_for_writer(struct btrfs_root *root)
+{
+	DEFINE_WAIT(wait);
+	while (atomic_read(&root->log_writers)) {
+		prepare_to_wait(&root->log_writer_wait,
+				&wait, TASK_UNINTERRUPTIBLE);
+		mutex_unlock(&root->log_mutex);
+		if (atomic_read(&root->log_writers))
+			schedule();
+		mutex_lock(&root->log_mutex);
+		finish_wait(&root->log_writer_wait, &wait);
+	}
 	return 0;
 }
 
@@ -1933,57 +1887,114 @@ static int wait_log_commit(struct btrfs_root *log)
 int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		   struct btrfs_root *root)
 {
+	int index1;
+	int index2;
 	int ret;
-	unsigned long batch;
 	struct btrfs_root *log = root->log_root;
+	struct btrfs_root *log_root_tree = root->fs_info->log_root_tree;
 
-	mutex_lock(&log->fs_info->tree_log_mutex);
-	if (atomic_read(&log->fs_info->tree_log_commit)) {
-		wait_log_commit(log);
-		goto out;
+	mutex_lock(&root->log_mutex);
+	index1 = root->log_transid % 2;
+	if (atomic_read(&root->log_commit[index1])) {
+		wait_log_commit(root, root->log_transid);
+		mutex_unlock(&root->log_mutex);
+		return 0;
 	}
-	atomic_set(&log->fs_info->tree_log_commit, 1);
+	atomic_set(&root->log_commit[index1], 1);
+
+	/* wait for previous tree log sync to complete */
+	if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
+		wait_log_commit(root, root->log_transid - 1);
 
 	while (1) {
-		batch = log->fs_info->tree_log_batch;
-		mutex_unlock(&log->fs_info->tree_log_mutex);
+		unsigned long batch = root->log_batch;
+		mutex_unlock(&root->log_mutex);
 		schedule_timeout_uninterruptible(1);
-		mutex_lock(&log->fs_info->tree_log_mutex);
-
-		while (atomic_read(&log->fs_info->tree_log_writers)) {
-			DEFINE_WAIT(wait);
-			prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
-					TASK_UNINTERRUPTIBLE);
-			mutex_unlock(&log->fs_info->tree_log_mutex);
-			if (atomic_read(&log->fs_info->tree_log_writers))
-				schedule();
-			mutex_lock(&log->fs_info->tree_log_mutex);
-			finish_wait(&log->fs_info->tree_log_wait, &wait);
-		}
-		if (batch == log->fs_info->tree_log_batch)
+		mutex_lock(&root->log_mutex);
+		wait_for_writer(root);
+		if (batch == root->log_batch)
 			break;
 	}
 
 	ret = btrfs_write_and_wait_marked_extents(log, &log->dirty_log_pages);
 	BUG_ON(ret);
-	ret = btrfs_write_and_wait_marked_extents(root->fs_info->log_root_tree,
-			       &root->fs_info->log_root_tree->dirty_log_pages);
+
+	btrfs_set_root_bytenr(&log->root_item, log->node->start);
+	btrfs_set_root_generation(&log->root_item, trans->transid);
+	btrfs_set_root_level(&log->root_item, btrfs_header_level(log->node));
+
+	root->log_batch = 0;
+	root->log_transid++;
+	log->log_transid = root->log_transid;
+	smp_mb();
+	/*
+	 * log tree has been flushed to disk, new modifications of
+	 * the log will be written to new positions. so it's safe to
+	 * allow log writers to go in.
+	 */
+	mutex_unlock(&root->log_mutex);
+
+	mutex_lock(&log_root_tree->log_mutex);
+	log_root_tree->log_batch++;
+	atomic_inc(&log_root_tree->log_writers);
+	mutex_unlock(&log_root_tree->log_mutex);
+
+	ret = update_log_root(trans, log);
+	BUG_ON(ret);
+
+	mutex_lock(&log_root_tree->log_mutex);
+	if (atomic_dec_and_test(&log_root_tree->log_writers)) {
+		smp_mb();
+		if (waitqueue_active(&log_root_tree->log_writer_wait))
+			wake_up(&log_root_tree->log_writer_wait);
+	}
+
+	index2 = log_root_tree->log_transid % 2;
+	if (atomic_read(&log_root_tree->log_commit[index2])) {
+		wait_log_commit(log_root_tree, log_root_tree->log_transid);
+		mutex_unlock(&log_root_tree->log_mutex);
+		goto out;
+	}
+	atomic_set(&log_root_tree->log_commit[index2], 1);
+
+	if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2]))
+		wait_log_commit(log_root_tree, log_root_tree->log_transid - 1);
+
+	wait_for_writer(log_root_tree);
+
+	ret = btrfs_write_and_wait_marked_extents(log_root_tree,
+				&log_root_tree->dirty_log_pages);
 	BUG_ON(ret);
 
 	btrfs_set_super_log_root(&root->fs_info->super_for_commit,
-				 log->fs_info->log_root_tree->node->start);
+				log_root_tree->node->start);
 	btrfs_set_super_log_root_level(&root->fs_info->super_for_commit,
-		       btrfs_header_level(log->fs_info->log_root_tree->node));
+				btrfs_header_level(log_root_tree->node));
+
+	log_root_tree->log_batch = 0;
+	log_root_tree->log_transid++;
+	smp_mb();
+
+	mutex_unlock(&log_root_tree->log_mutex);
+
+	/*
+	 * nobody else is going to jump in and write the the ctree
+	 * super here because the log_commit atomic below is protecting
+	 * us.  We must be called with a transaction handle pinning
+	 * the running transaction open, so a full commit can't hop
+	 * in and cause problems either.
+	 */
+	write_ctree_super(trans, root->fs_info->tree_root, 2);
 
-	write_ctree_super(trans, log->fs_info->tree_root, 2);
-	log->fs_info->tree_log_transid++;
-	log->fs_info->tree_log_batch = 0;
-	atomic_set(&log->fs_info->tree_log_commit, 0);
+	atomic_set(&log_root_tree->log_commit[index2], 0);
 	smp_mb();
-	if (waitqueue_active(&log->fs_info->tree_log_wait))
-		wake_up(&log->fs_info->tree_log_wait);
+	if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
+		wake_up(&log_root_tree->log_commit_wait[index2]);
 out:
-	mutex_unlock(&log->fs_info->tree_log_mutex);
+	atomic_set(&root->log_commit[index1], 0);
+	smp_mb();
+	if (waitqueue_active(&root->log_commit_wait[index1]))
+		wake_up(&root->log_commit_wait[index1]);
 	return 0;
 }
 
@@ -2019,38 +2030,18 @@ int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
 				   start, end, GFP_NOFS);
 	}
 
-	log = root->log_root;
-	ret = btrfs_del_root(trans, root->fs_info->log_root_tree,
-			     &log->root_key);
-	BUG_ON(ret);
+	if (log->log_transid > 0) {
+		ret = btrfs_del_root(trans, root->fs_info->log_root_tree,
+				     &log->root_key);
+		BUG_ON(ret);
+	}
 	root->log_root = NULL;
-	kfree(root->log_root);
+	free_extent_buffer(log->node);
+	kfree(log);
 	return 0;
 }
 
 /*
- * helper function to update the item for a given subvolumes log root
- * in the tree of log roots
- */
-static int update_log_root(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *log)
-{
-	u64 bytenr = btrfs_root_bytenr(&log->root_item);
-	int ret;
-
-	if (log->node->start == bytenr)
-		return 0;
-
-	btrfs_set_root_bytenr(&log->root_item, log->node->start);
-	btrfs_set_root_generation(&log->root_item, trans->transid);
-	btrfs_set_root_level(&log->root_item, btrfs_header_level(log->node));
-	ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
-				&log->root_key, &log->root_item);
-	BUG_ON(ret);
-	return ret;
-}
-
-/*
  * If both a file and directory are logged, and unlinks or renames are
  * mixed in, we have a few interesting corners:
  *
@@ -2711,11 +2702,6 @@ next_slot:
 
 	btrfs_free_path(path);
 	btrfs_free_path(dst_path);
-
-	mutex_lock(&root->fs_info->tree_log_mutex);
-	ret = update_log_root(trans, log);
-	BUG_ON(ret);
-	mutex_unlock(&root->fs_info->tree_log_mutex);
 out:
 	return 0;
 }
@@ -2846,7 +2832,9 @@ again:
 		BUG_ON(!wc.replay_dest);
 
 		wc.replay_dest->log_root = log;
+		mutex_lock(&fs_info->trans_mutex);
 		btrfs_record_root_in_trans(wc.replay_dest);
+		mutex_unlock(&fs_info->trans_mutex);
 		ret = walk_log_tree(trans, log, &wc);
 		BUG_ON(ret);
 
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 3451e1cca2b..dd06e18e5aa 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -20,7 +20,6 @@
 #include <linux/buffer_head.h>
 #include <linux/blkdev.h>
 #include <linux/random.h>
-#include <linux/version.h>
 #include <asm/div64.h>
 #include "compat.h"
 #include "ctree.h"
@@ -104,10 +103,8 @@ static noinline struct btrfs_device *__find_device(struct list_head *head,
 						   u64 devid, u8 *uuid)
 {
 	struct btrfs_device *dev;
-	struct list_head *cur;
 
-	list_for_each(cur, head) {
-		dev = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(dev, head, dev_list) {
 		if (dev->devid == devid &&
 		    (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
 			return dev;
@@ -118,11 +115,9 @@ static noinline struct btrfs_device *__find_device(struct list_head *head,
 
 static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
 {
-	struct list_head *cur;
 	struct btrfs_fs_devices *fs_devices;
 
-	list_for_each(cur, &fs_uuids) {
-		fs_devices = list_entry(cur, struct btrfs_fs_devices, list);
+	list_for_each_entry(fs_devices, &fs_uuids, list) {
 		if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
 			return fs_devices;
 	}
@@ -159,6 +154,7 @@ static noinline int run_scheduled_bios(struct btrfs_device *device)
 loop:
 	spin_lock(&device->io_lock);
 
+loop_lock:
 	/* take all the bios off the list at once and process them
 	 * later on (without the lock held).  But, remember the
 	 * tail and other pointers so the bios can be properly reinserted
@@ -208,7 +204,7 @@ loop:
 		 * is now congested.  Back off and let other work structs
 		 * run instead
 		 */
-		if (pending && bdi_write_congested(bdi) &&
+		if (pending && bdi_write_congested(bdi) && num_run > 16 &&
 		    fs_info->fs_devices->open_devices > 1) {
 			struct bio *old_head;
 
@@ -220,7 +216,8 @@ loop:
 				tail->bi_next = old_head;
 			else
 				device->pending_bio_tail = tail;
-			device->running_pending = 0;
+
+			device->running_pending = 1;
 
 			spin_unlock(&device->io_lock);
 			btrfs_requeue_work(&device->work);
@@ -229,6 +226,11 @@ loop:
 	}
 	if (again)
 		goto loop;
+
+	spin_lock(&device->io_lock);
+	if (device->pending_bios)
+		goto loop_lock;
+	spin_unlock(&device->io_lock);
 done:
 	return 0;
 }
@@ -345,14 +347,11 @@ error:
 
 int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
 {
-	struct list_head *tmp;
-	struct list_head *cur;
-	struct btrfs_device *device;
+	struct btrfs_device *device, *next;
 
 	mutex_lock(&uuid_mutex);
 again:
-	list_for_each_safe(cur, tmp, &fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
 		if (device->in_fs_metadata)
 			continue;
 
@@ -383,14 +382,12 @@ again:
 
 static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 {
-	struct list_head *cur;
 	struct btrfs_device *device;
 
 	if (--fs_devices->opened > 0)
 		return 0;
 
-	list_for_each(cur, &fs_devices->devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, &fs_devices->devices, dev_list) {
 		if (device->bdev) {
 			close_bdev_exclusive(device->bdev, device->mode);
 			fs_devices->open_devices--;
@@ -439,7 +436,6 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 {
 	struct block_device *bdev;
 	struct list_head *head = &fs_devices->devices;
-	struct list_head *cur;
 	struct btrfs_device *device;
 	struct block_device *latest_bdev = NULL;
 	struct buffer_head *bh;
@@ -450,8 +446,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 	int seeding = 1;
 	int ret = 0;
 
-	list_for_each(cur, head) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, head, dev_list) {
 		if (device->bdev)
 			continue;
 		if (!device->name)
@@ -578,7 +573,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 		       *(unsigned long long *)disk_super->fsid,
 		       *(unsigned long long *)(disk_super->fsid + 8));
 	}
-	printk(KERN_INFO "devid %llu transid %llu %s\n",
+	printk(KERN_CONT "devid %llu transid %llu %s\n",
 	       (unsigned long long)devid, (unsigned long long)transid, path);
 	ret = device_list_add(path, disk_super, devid, fs_devices_ret);
 
@@ -1017,14 +1012,12 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	}
 
 	if (strcmp(device_path, "missing") == 0) {
-		struct list_head *cur;
 		struct list_head *devices;
 		struct btrfs_device *tmp;
 
 		device = NULL;
 		devices = &root->fs_info->fs_devices->devices;
-		list_for_each(cur, devices) {
-			tmp = list_entry(cur, struct btrfs_device, dev_list);
+		list_for_each_entry(tmp, devices, dev_list) {
 			if (tmp->in_fs_metadata && !tmp->bdev) {
 				device = tmp;
 				break;
@@ -1280,7 +1273,6 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device;
 	struct block_device *bdev;
-	struct list_head *cur;
 	struct list_head *devices;
 	struct super_block *sb = root->fs_info->sb;
 	u64 total_bytes;
@@ -1304,8 +1296,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	mutex_lock(&root->fs_info->volume_mutex);
 
 	devices = &root->fs_info->fs_devices->devices;
-	list_for_each(cur, devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, devices, dev_list) {
 		if (device->bdev == bdev) {
 			ret = -EEXIST;
 			goto error;
@@ -1383,6 +1374,12 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 		ret = btrfs_add_device(trans, root, device);
 	}
 
+	/*
+	 * we've got more storage, clear any full flags on the space
+	 * infos
+	 */
+	btrfs_clear_space_info_full(root->fs_info);
+
 	unlock_chunks(root);
 	btrfs_commit_transaction(trans, root);
 
@@ -1468,6 +1465,8 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
 	device->fs_devices->total_rw_bytes += diff;
 
 	device->total_bytes = new_size;
+	btrfs_clear_space_info_full(device->dev_root->fs_info);
+
 	return btrfs_update_device(trans, device);
 }
 
@@ -1704,7 +1703,6 @@ static u64 div_factor(u64 num, int factor)
 int btrfs_balance(struct btrfs_root *dev_root)
 {
 	int ret;
-	struct list_head *cur;
 	struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
 	struct btrfs_device *device;
 	u64 old_size;
@@ -1723,8 +1721,7 @@ int btrfs_balance(struct btrfs_root *dev_root)
 	dev_root = dev_root->fs_info->dev_root;
 
 	/* step one make some room on all the devices */
-	list_for_each(cur, devices) {
-		device = list_entry(cur, struct btrfs_device, dev_list);
+	list_for_each_entry(device, devices, dev_list) {
 		old_size = device->total_bytes;
 		size_to_free = div_factor(old_size, 1);
 		size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
@@ -2905,10 +2902,6 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
 		free_extent_map(em);
 	}
 
-	map = kzalloc(sizeof(*map), GFP_NOFS);
-	if (!map)
-		return -ENOMEM;
-
 	em = alloc_extent_map(GFP_NOFS);
 	if (!em)
 		return -ENOMEM;
@@ -3117,6 +3110,8 @@ int btrfs_read_sys_array(struct btrfs_root *root)
 	if (!sb)
 		return -ENOMEM;
 	btrfs_set_buffer_uptodate(sb);
+	btrfs_set_buffer_lockdep_class(sb, 0);
+
 	write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
 	array_size = btrfs_super_sys_array_size(super_copy);
 
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index 7f332e27089..a9d3bf4d268 100644
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -21,6 +21,7 @@
 #include <linux/slab.h>
 #include <linux/rwsem.h>
 #include <linux/xattr.h>
+#include <linux/security.h>
 #include "ctree.h"
 #include "btrfs_inode.h"
 #include "transaction.h"
@@ -45,9 +46,12 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
 	/* lookup the xattr by name */
 	di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name,
 				strlen(name), 0);
-	if (!di || IS_ERR(di)) {
+	if (!di) {
 		ret = -ENODATA;
 		goto out;
+	} else if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto out;
 	}
 
 	leaf = path->nodes[0];
@@ -62,6 +66,14 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
 		ret = -ERANGE;
 		goto out;
 	}
+
+	/*
+	 * The way things are packed into the leaf is like this
+	 * |struct btrfs_dir_item|name|data|
+	 * where name is the xattr name, so security.foo, and data is the
+	 * content of the xattr.  data_ptr points to the location in memory
+	 * where the data starts in the in memory leaf
+	 */
 	data_ptr = (unsigned long)((char *)(di + 1) +
 				   btrfs_dir_name_len(leaf, di));
 	read_extent_buffer(leaf, buffer, data_ptr,
@@ -86,7 +98,7 @@ int __btrfs_setxattr(struct inode *inode, const char *name,
 	if (!path)
 		return -ENOMEM;
 
-	trans = btrfs_start_transaction(root, 1);
+	trans = btrfs_join_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, inode);
 
 	/* first lets see if we already have this xattr */
@@ -176,7 +188,6 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
-	ret = 0;
 	advance = 0;
 	while (1) {
 		leaf = path->nodes[0];
@@ -320,3 +331,34 @@ int btrfs_removexattr(struct dentry *dentry, const char *name)
 		return -EOPNOTSUPP;
 	return __btrfs_setxattr(dentry->d_inode, name, NULL, 0, XATTR_REPLACE);
 }
+
+int btrfs_xattr_security_init(struct inode *inode, struct inode *dir)
+{
+	int err;
+	size_t len;
+	void *value;
+	char *suffix;
+	char *name;
+
+	err = security_inode_init_security(inode, dir, &suffix, &value, &len);
+	if (err) {
+		if (err == -EOPNOTSUPP)
+			return 0;
+		return err;
+	}
+
+	name = kmalloc(XATTR_SECURITY_PREFIX_LEN + strlen(suffix) + 1,
+		       GFP_NOFS);
+	if (!name) {
+		err = -ENOMEM;
+	} else {
+		strcpy(name, XATTR_SECURITY_PREFIX);
+		strcpy(name + XATTR_SECURITY_PREFIX_LEN, suffix);
+		err = __btrfs_setxattr(inode, name, value, len, 0);
+		kfree(name);
+	}
+
+	kfree(suffix);
+	kfree(value);
+	return err;
+}
diff --git a/fs/btrfs/xattr.h b/fs/btrfs/xattr.h
index 5b1d08f8e68..c71e9c3cf3f 100644
--- a/fs/btrfs/xattr.h
+++ b/fs/btrfs/xattr.h
@@ -36,4 +36,6 @@ extern int btrfs_setxattr(struct dentry *dentry, const char *name,
 		const void *value, size_t size, int flags);
 extern int btrfs_removexattr(struct dentry *dentry, const char *name);
 
+extern int btrfs_xattr_security_init(struct inode *inode, struct inode *dir);
+
 #endif /* __XATTR__ */
diff --git a/fs/buffer.c b/fs/buffer.c
index b58208f1640..891e1c78e4f 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -760,15 +760,9 @@ EXPORT_SYMBOL(mark_buffer_dirty_inode);
  * If warn is true, then emit a warning if the page is not uptodate and has
  * not been truncated.
  */
-static int __set_page_dirty(struct page *page,
+static void __set_page_dirty(struct page *page,
 		struct address_space *mapping, int warn)
 {
-	if (unlikely(!mapping))
-		return !TestSetPageDirty(page);
-
-	if (TestSetPageDirty(page))
-		return 0;
-
 	spin_lock_irq(&mapping->tree_lock);
 	if (page->mapping) {	/* Race with truncate? */
 		WARN_ON_ONCE(warn && !PageUptodate(page));
@@ -777,6 +771,7 @@ static int __set_page_dirty(struct page *page,
 			__inc_zone_page_state(page, NR_FILE_DIRTY);
 			__inc_bdi_stat(mapping->backing_dev_info,
 					BDI_RECLAIMABLE);
+			task_dirty_inc(current);
 			task_io_account_write(PAGE_CACHE_SIZE);
 		}
 		radix_tree_tag_set(&mapping->page_tree,
@@ -784,8 +779,6 @@ static int __set_page_dirty(struct page *page,
 	}
 	spin_unlock_irq(&mapping->tree_lock);
 	__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
-
-	return 1;
 }
 
 /*
@@ -815,6 +808,7 @@ static int __set_page_dirty(struct page *page,
  */
 int __set_page_dirty_buffers(struct page *page)
 {
+	int newly_dirty;
 	struct address_space *mapping = page_mapping(page);
 
 	if (unlikely(!mapping))
@@ -830,9 +824,12 @@ int __set_page_dirty_buffers(struct page *page)
 			bh = bh->b_this_page;
 		} while (bh != head);
 	}
+	newly_dirty = !TestSetPageDirty(page);
 	spin_unlock(&mapping->private_lock);
 
-	return __set_page_dirty(page, mapping, 1);
+	if (newly_dirty)
+		__set_page_dirty(page, mapping, 1);
+	return newly_dirty;
 }
 EXPORT_SYMBOL(__set_page_dirty_buffers);
 
@@ -1261,8 +1258,11 @@ void mark_buffer_dirty(struct buffer_head *bh)
 			return;
 	}
 
-	if (!test_set_buffer_dirty(bh))
-		__set_page_dirty(bh->b_page, page_mapping(bh->b_page), 0);
+	if (!test_set_buffer_dirty(bh)) {
+		struct page *page = bh->b_page;
+		if (!TestSetPageDirty(page))
+			__set_page_dirty(page, page_mapping(page), 0);
+	}
 }
 
 /*
@@ -2688,7 +2688,7 @@ int nobh_write_end(struct file *file, struct address_space *mapping,
 	struct buffer_head *bh;
 	BUG_ON(fsdata != NULL && page_has_buffers(page));
 
-	if (unlikely(copied < len) && !page_has_buffers(page))
+	if (unlikely(copied < len) && head)
 		attach_nobh_buffers(page, head);
 	if (page_has_buffers(page))
 		return generic_write_end(file, mapping, pos, len,
@@ -3108,7 +3108,7 @@ int sync_dirty_buffer(struct buffer_head *bh)
 	if (test_clear_buffer_dirty(bh)) {
 		get_bh(bh);
 		bh->b_end_io = end_buffer_write_sync;
-		ret = submit_bh(WRITE_SYNC, bh);
+		ret = submit_bh(WRITE, bh);
 		wait_on_buffer(bh);
 		if (buffer_eopnotsupp(bh)) {
 			clear_buffer_eopnotsupp(bh);
diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index 73ac7ebd1df..851388fafc7 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -1,3 +1,13 @@
+Version 1.57
+------------
+Improve support for multiple security contexts to the same server. We
+used to use the same "vcnumber" for all connections which could cause
+the server to treat subsequent connections, especially those that
+are authenticated as guest, as reconnections, invalidating the earlier
+user's smb session.  This fix allows cifs to mount multiple times to the
+same server with different userids without risking invalidating earlier
+established security contexts.
+
 Version 1.56
 ------------
 Add "forcemandatorylock" mount option to allow user to use mandatory
@@ -7,7 +17,10 @@ specified and user does not have access to query information about the
 top of the share.  Fix problem in 2.6.28 resolving DFS paths to
 Samba servers (worked to Windows).  Fix rmdir so that pending search
 (readdir) requests do not get invalid results which include the now
-removed directory.
+removed directory.  Fix oops in cifs_dfs_ref.c when prefixpath is not reachable
+when using DFS.  Add better file create support to servers which support
+the CIFS POSIX protocol extensions (this adds support for new flags
+on create, and improves semantics for write of locked ranges).
 
 Version 1.55
 ------------
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index 7ac481841f8..2b1d28a9ee2 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -100,5 +100,5 @@ extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg);
 extern const struct export_operations cifs_export_ops;
 #endif /* EXPERIMENTAL */
 
-#define CIFS_VERSION   "1.56"
+#define CIFS_VERSION   "1.57"
 #endif				/* _CIFSFS_H */
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 94c1ca0ec95..e004f6db5fc 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -164,9 +164,12 @@ struct TCP_Server_Info {
 	/* multiplexed reads or writes */
 	unsigned int maxBuf;	/* maxBuf specifies the maximum */
 	/* message size the server can send or receive for non-raw SMBs */
-	unsigned int maxRw;	/* maxRw specifies the maximum */
+	unsigned int max_rw;	/* maxRw specifies the maximum */
 	/* message size the server can send or receive for */
 	/* SMB_COM_WRITE_RAW or SMB_COM_READ_RAW. */
+	unsigned int max_vcs;	/* maximum number of smb sessions, at least
+				   those that can be specified uniquely with
+				   vcnumbers */
 	char sessid[4];		/* unique token id for this session */
 	/* (returned on Negotiate */
 	int capabilities; /* allow selective disabling of caps by smb sess */
@@ -210,6 +213,7 @@ struct cifsSesInfo {
 	unsigned overrideSecFlg;  /* if non-zero override global sec flags */
 	__u16 ipc_tid;		/* special tid for connection to IPC share */
 	__u16 flags;
+	__u16 vcnum;
 	char *serverOS;		/* name of operating system underlying server */
 	char *serverNOS;	/* name of network operating system of server */
 	char *serverDomain;	/* security realm of server */
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index 382ba629880..083dfc57c7a 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -42,6 +42,7 @@ extern void _FreeXid(unsigned int);
 #define GetXid() (int)_GetXid(); cFYI(1,("CIFS VFS: in %s as Xid: %d with uid: %d",__func__, xid,current_fsuid()));
 #define FreeXid(curr_xid) {_FreeXid(curr_xid); cFYI(1,("CIFS VFS: leaving %s (xid = %d) rc = %d",__func__,curr_xid,(int)rc));}
 extern char *build_path_from_dentry(struct dentry *);
+extern char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb);
 extern char *build_wildcard_path_from_dentry(struct dentry *direntry);
 /* extern void renew_parental_timestamps(struct dentry *direntry);*/
 extern int SendReceive(const unsigned int /* xid */ , struct cifsSesInfo *,
@@ -91,6 +92,9 @@ extern u64 cifs_UnixTimeToNT(struct timespec);
 extern __le64 cnvrtDosCifsTm(__u16 date, __u16 time);
 extern struct timespec cnvrtDosUnixTm(__u16 date, __u16 time);
 
+extern void posix_fill_in_inode(struct inode *tmp_inode,
+				FILE_UNIX_BASIC_INFO *pData, int isNewInode);
+extern struct inode *cifs_new_inode(struct super_block *sb, __u64 *inum);
 extern int cifs_get_inode_info(struct inode **pinode,
 			const unsigned char *search_path,
 			FILE_ALL_INFO *pfile_info,
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index 552642a507c..939e2f76b95 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -528,14 +528,15 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses)
 		server->maxReq = le16_to_cpu(rsp->MaxMpxCount);
 		server->maxBuf = min((__u32)le16_to_cpu(rsp->MaxBufSize),
 				(__u32)CIFSMaxBufSize + MAX_CIFS_HDR_SIZE);
+		server->max_vcs = le16_to_cpu(rsp->MaxNumberVcs);
 		GETU32(server->sessid) = le32_to_cpu(rsp->SessionKey);
 		/* even though we do not use raw we might as well set this
 		accurately, in case we ever find a need for it */
 		if ((le16_to_cpu(rsp->RawMode) & RAW_ENABLE) == RAW_ENABLE) {
-			server->maxRw = 0xFF00;
+			server->max_rw = 0xFF00;
 			server->capabilities = CAP_MPX_MODE | CAP_RAW_MODE;
 		} else {
-			server->maxRw = 0;/* we do not need to use raw anyway */
+			server->max_rw = 0;/* do not need to use raw anyway */
 			server->capabilities = CAP_MPX_MODE;
 		}
 		tmp = (__s16)le16_to_cpu(rsp->ServerTimeZone);
@@ -638,7 +639,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses)
 	/* probably no need to store and check maxvcs */
 	server->maxBuf = min(le32_to_cpu(pSMBr->MaxBufferSize),
 			(__u32) CIFSMaxBufSize + MAX_CIFS_HDR_SIZE);
-	server->maxRw = le32_to_cpu(pSMBr->MaxRawSize);
+	server->max_rw = le32_to_cpu(pSMBr->MaxRawSize);
 	cFYI(DBG2, ("Max buf = %d", ses->server->maxBuf));
 	GETU32(ses->server->sessid) = le32_to_cpu(pSMBr->SessionKey);
 	server->capabilities = le32_to_cpu(pSMBr->Capabilities);
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 2209be94305..da0f4ffa061 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -23,7 +23,6 @@
 #include <linux/string.h>
 #include <linux/list.h>
 #include <linux/wait.h>
-#include <linux/ipv6.h>
 #include <linux/pagemap.h>
 #include <linux/ctype.h>
 #include <linux/utsname.h>
@@ -35,6 +34,7 @@
 #include <linux/freezer.h>
 #include <asm/uaccess.h>
 #include <asm/processor.h>
+#include <net/ipv6.h>
 #include "cifspdu.h"
 #include "cifsglob.h"
 #include "cifsproto.h"
@@ -1379,8 +1379,8 @@ cifs_find_tcp_session(struct sockaddr_storage *addr)
 		     server->addr.sockAddr.sin_addr.s_addr))
 			continue;
 		else if (addr->ss_family == AF_INET6 &&
-			 memcmp(&server->addr.sockAddr6.sin6_addr,
-				&addr6->sin6_addr, sizeof(addr6->sin6_addr)))
+			 !ipv6_addr_equal(&server->addr.sockAddr6.sin6_addr,
+					  &addr6->sin6_addr))
 			continue;
 
 		++server->srv_count;
@@ -2180,6 +2180,33 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info,
 			   "mount option supported"));
 }
 
+static int
+is_path_accessible(int xid, struct cifsTconInfo *tcon,
+		   struct cifs_sb_info *cifs_sb, const char *full_path)
+{
+	int rc;
+	__u64 inode_num;
+	FILE_ALL_INFO *pfile_info;
+
+	rc = CIFSGetSrvInodeNumber(xid, tcon, full_path, &inode_num,
+				   cifs_sb->local_nls,
+				   cifs_sb->mnt_cifs_flags &
+						CIFS_MOUNT_MAP_SPECIAL_CHR);
+	if (rc != -EOPNOTSUPP)
+		return rc;
+
+	pfile_info = kmalloc(sizeof(FILE_ALL_INFO), GFP_KERNEL);
+	if (pfile_info == NULL)
+		return -ENOMEM;
+
+	rc = CIFSSMBQPathInfo(xid, tcon, full_path, pfile_info,
+			      0 /* not legacy */, cifs_sb->local_nls,
+			      cifs_sb->mnt_cifs_flags &
+				CIFS_MOUNT_MAP_SPECIAL_CHR);
+	kfree(pfile_info);
+	return rc;
+}
+
 int
 cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb,
 	   char *mount_data, const char *devname)
@@ -2190,6 +2217,7 @@ cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb,
 	struct cifsSesInfo *pSesInfo = NULL;
 	struct cifsTconInfo *tcon = NULL;
 	struct TCP_Server_Info *srvTcp = NULL;
+	char   *full_path;
 
 	xid = GetXid();
 
@@ -2426,6 +2454,23 @@ mount_fail_check:
 		cifs_sb->rsize = min(cifs_sb->rsize,
 			       (tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE));
 
+	if (!rc && cifs_sb->prepathlen) {
+		/* build_path_to_root works only when we have a valid tcon */
+		full_path = cifs_build_path_to_root(cifs_sb);
+		if (full_path == NULL) {
+			rc = -ENOMEM;
+			goto mount_fail_check;
+		}
+		rc = is_path_accessible(xid, tcon, cifs_sb, full_path);
+		if (rc) {
+			cERROR(1, ("Path %s in not accessible: %d",
+						full_path, rc));
+			kfree(full_path);
+			goto mount_fail_check;
+		}
+		kfree(full_path);
+	}
+
 	/* volume_info->password is freed above when existing session found
 	(in which case it is not needed anymore) but when new sesion is created
 	the password ptr is put in the new session structure (in which case the
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 964aad03c5a..89fb7283265 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -3,7 +3,7 @@
  *
  *   vfs operations that deal with dentries
  *
- *   Copyright (C) International Business Machines  Corp., 2002,2008
+ *   Copyright (C) International Business Machines  Corp., 2002,2009
  *   Author(s): Steve French (sfrench@us.ibm.com)
  *
  *   This library is free software; you can redistribute it and/or modify
@@ -129,6 +129,78 @@ cifs_bp_rename_retry:
 	return full_path;
 }
 
+static int cifs_posix_open(char *full_path, struct inode **pinode,
+		    struct super_block *sb, int mode, int oflags,
+		    int *poplock, __u16 *pnetfid, int xid)
+{
+	int rc;
+	__u32 oplock;
+	FILE_UNIX_BASIC_INFO *presp_data;
+	__u32 posix_flags = 0;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
+
+	cFYI(1, ("posix open %s", full_path));
+
+	presp_data = kzalloc(sizeof(FILE_UNIX_BASIC_INFO), GFP_KERNEL);
+	if (presp_data == NULL)
+		return -ENOMEM;
+
+/* So far cifs posix extensions can only map the following flags.
+   There are other valid fmode oflags such as FMODE_LSEEK, FMODE_PREAD, but
+   so far we do not seem to need them, and we can treat them as local only */
+	if ((oflags & (FMODE_READ | FMODE_WRITE)) ==
+		(FMODE_READ | FMODE_WRITE))
+		posix_flags = SMB_O_RDWR;
+	else if (oflags & FMODE_READ)
+		posix_flags = SMB_O_RDONLY;
+	else if (oflags & FMODE_WRITE)
+		posix_flags = SMB_O_WRONLY;
+	if (oflags & O_CREAT)
+		posix_flags |= SMB_O_CREAT;
+	if (oflags & O_EXCL)
+		posix_flags |= SMB_O_EXCL;
+	if (oflags & O_TRUNC)
+		posix_flags |= SMB_O_TRUNC;
+	if (oflags & O_APPEND)
+		posix_flags |= SMB_O_APPEND;
+	if (oflags & O_SYNC)
+		posix_flags |= SMB_O_SYNC;
+	if (oflags & O_DIRECTORY)
+		posix_flags |= SMB_O_DIRECTORY;
+	if (oflags & O_NOFOLLOW)
+		posix_flags |= SMB_O_NOFOLLOW;
+	if (oflags & O_DIRECT)
+		posix_flags |= SMB_O_DIRECT;
+
+
+	rc = CIFSPOSIXCreate(xid, cifs_sb->tcon, posix_flags, mode,
+			pnetfid, presp_data, &oplock, full_path,
+			cifs_sb->local_nls, cifs_sb->mnt_cifs_flags &
+					CIFS_MOUNT_MAP_SPECIAL_CHR);
+	if (rc)
+		goto posix_open_ret;
+
+	if (presp_data->Type == cpu_to_le32(-1))
+		goto posix_open_ret; /* open ok, caller does qpathinfo */
+
+	/* get new inode and set it up */
+	if (!pinode)
+		goto posix_open_ret; /* caller does not need info */
+
+	*pinode = cifs_new_inode(sb, &presp_data->UniqueId);
+
+	/* We do not need to close the file if new_inode fails since
+	   the caller will retry qpathinfo as long as inode is null */
+	if (*pinode == NULL)
+		goto posix_open_ret;
+
+	posix_fill_in_inode(*pinode, presp_data, 1);
+
+posix_open_ret:
+	kfree(presp_data);
+	return rc;
+}
+
 static void setup_cifs_dentry(struct cifsTconInfo *tcon,
 			      struct dentry *direntry,
 			      struct inode *newinode)
@@ -150,7 +222,14 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	int xid;
 	int create_options = CREATE_NOT_DIR;
 	int oplock = 0;
-	/* BB below access is too much for the mknod to request */
+	int oflags;
+	/*
+	 * BB below access is probably too much for mknod to request
+	 *    but we have to do query and setpathinfo so requesting
+	 *    less could fail (unless we want to request getatr and setatr
+	 *    permissions (only).  At least for POSIX we do not have to
+	 *    request so much.
+	 */
 	int desiredAccess = GENERIC_READ | GENERIC_WRITE;
 	__u16 fileHandle;
 	struct cifs_sb_info *cifs_sb;
@@ -174,13 +253,43 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	}
 
 	mode &= ~current->fs->umask;
+	if (oplockEnabled)
+		oplock = REQ_OPLOCK;
 
-	if (nd && (nd->flags & LOOKUP_OPEN)) {
-		int oflags = nd->intent.open.flags;
+	if (nd && (nd->flags & LOOKUP_OPEN))
+		oflags = nd->intent.open.flags;
+	else
+		oflags = FMODE_READ;
+
+	if (tcon->unix_ext && (tcon->ses->capabilities & CAP_UNIX) &&
+	    (CIFS_UNIX_POSIX_PATH_OPS_CAP &
+			le64_to_cpu(tcon->fsUnixInfo.Capability))) {
+		rc = cifs_posix_open(full_path, &newinode, inode->i_sb,
+				     mode, oflags, &oplock, &fileHandle, xid);
+		/* EIO could indicate that (posix open) operation is not
+		   supported, despite what server claimed in capability
+		   negotation.  EREMOTE indicates DFS junction, which is not
+		   handled in posix open */
+
+		if ((rc == 0) && (newinode == NULL))
+			goto cifs_create_get_file_info; /* query inode info */
+		else if (rc == 0) /* success, no need to query */
+			goto cifs_create_set_dentry;
+		else if ((rc != -EIO) && (rc != -EREMOTE) &&
+			 (rc != -EOPNOTSUPP)) /* path not found or net err */
+			goto cifs_create_out;
+		/* else fallthrough to retry, using older open call, this is
+		   case where server does not support this SMB level, and
+		   falsely claims capability (also get here for DFS case
+		   which should be rare for path not covered on files) */
+	}
 
+	if (nd && (nd->flags & LOOKUP_OPEN)) {
+		/* if the file is going to stay open, then we
+		   need to set the desired access properly */
 		desiredAccess = 0;
 		if (oflags & FMODE_READ)
-			desiredAccess |= GENERIC_READ;
+			desiredAccess |= GENERIC_READ; /* is this too little? */
 		if (oflags & FMODE_WRITE) {
 			desiredAccess |= GENERIC_WRITE;
 			if (!(oflags & FMODE_READ))
@@ -199,8 +308,6 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 
 	/* BB add processing to set equivalent of mode - e.g. via CreateX with
 	   ACLs */
-	if (oplockEnabled)
-		oplock = REQ_OPLOCK;
 
 	buf = kmalloc(sizeof(FILE_ALL_INFO), GFP_KERNEL);
 	if (buf == NULL) {
@@ -233,116 +340,112 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	}
 	if (rc) {
 		cFYI(1, ("cifs_create returned 0x%x", rc));
-	} else {
-		/* If Open reported that we actually created a file
-		then we now have to set the mode if possible */
-		if ((tcon->unix_ext) && (oplock & CIFS_CREATE_ACTION)) {
-			struct cifs_unix_set_info_args args = {
+		goto cifs_create_out;
+	}
+
+	/* If Open reported that we actually created a file
+	   then we now have to set the mode if possible */
+	if ((tcon->unix_ext) && (oplock & CIFS_CREATE_ACTION)) {
+		struct cifs_unix_set_info_args args = {
 				.mode	= mode,
 				.ctime	= NO_CHANGE_64,
 				.atime	= NO_CHANGE_64,
 				.mtime	= NO_CHANGE_64,
 				.device	= 0,
-			};
+		};
 
-			if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID) {
-				args.uid = (__u64) current_fsuid();
-				if (inode->i_mode & S_ISGID)
-					args.gid = (__u64) inode->i_gid;
-				else
-					args.gid = (__u64) current_fsgid();
-			} else {
-				args.uid = NO_CHANGE_64;
-				args.gid = NO_CHANGE_64;
-			}
-			CIFSSMBUnixSetInfo(xid, tcon, full_path, &args,
-				cifs_sb->local_nls,
-				cifs_sb->mnt_cifs_flags &
-					CIFS_MOUNT_MAP_SPECIAL_CHR);
+		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID) {
+			args.uid = (__u64) current_fsuid();
+			if (inode->i_mode & S_ISGID)
+				args.gid = (__u64) inode->i_gid;
+			else
+				args.gid = (__u64) current_fsgid();
 		} else {
-			/* BB implement mode setting via Windows security
-			   descriptors e.g. */
-			/* CIFSSMBWinSetPerms(xid,tcon,path,mode,-1,-1,nls);*/
-
-			/* Could set r/o dos attribute if mode & 0222 == 0 */
+			args.uid = NO_CHANGE_64;
+			args.gid = NO_CHANGE_64;
 		}
+		CIFSSMBUnixSetInfo(xid, tcon, full_path, &args,
+			cifs_sb->local_nls,
+			cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
+	} else {
+		/* BB implement mode setting via Windows security
+		   descriptors e.g. */
+		/* CIFSSMBWinSetPerms(xid,tcon,path,mode,-1,-1,nls);*/
 
-		/* server might mask mode so we have to query for it */
-		if (tcon->unix_ext)
-			rc = cifs_get_inode_info_unix(&newinode, full_path,
-						 inode->i_sb, xid);
-		else {
-			rc = cifs_get_inode_info(&newinode, full_path,
-						 buf, inode->i_sb, xid,
-						 &fileHandle);
-			if (newinode) {
-				if (cifs_sb->mnt_cifs_flags &
-				    CIFS_MOUNT_DYNPERM)
-					newinode->i_mode = mode;
-				if ((oplock & CIFS_CREATE_ACTION) &&
-				    (cifs_sb->mnt_cifs_flags &
-				     CIFS_MOUNT_SET_UID)) {
-					newinode->i_uid = current_fsuid();
-					if (inode->i_mode & S_ISGID)
-						newinode->i_gid =
-							inode->i_gid;
-					else
-						newinode->i_gid =
-							current_fsgid();
-				}
+		/* Could set r/o dos attribute if mode & 0222 == 0 */
+	}
+
+cifs_create_get_file_info:
+	/* server might mask mode so we have to query for it */
+	if (tcon->unix_ext)
+		rc = cifs_get_inode_info_unix(&newinode, full_path,
+					      inode->i_sb, xid);
+	else {
+		rc = cifs_get_inode_info(&newinode, full_path, buf,
+					 inode->i_sb, xid, &fileHandle);
+		if (newinode) {
+			if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DYNPERM)
+				newinode->i_mode = mode;
+			if ((oplock & CIFS_CREATE_ACTION) &&
+			    (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID)) {
+				newinode->i_uid = current_fsuid();
+				if (inode->i_mode & S_ISGID)
+					newinode->i_gid = inode->i_gid;
+				else
+					newinode->i_gid = current_fsgid();
 			}
 		}
+	}
 
-		if (rc != 0) {
-			cFYI(1, ("Create worked, get_inode_info failed rc = %d",
-				 rc));
-		} else
-			setup_cifs_dentry(tcon, direntry, newinode);
-
-		if ((nd == NULL /* nfsd case - nfs srv does not set nd */) ||
-			(!(nd->flags & LOOKUP_OPEN))) {
-			/* mknod case - do not leave file open */
-			CIFSSMBClose(xid, tcon, fileHandle);
-		} else if (newinode) {
-			struct cifsFileInfo *pCifsFile =
-			   kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
-
-			if (pCifsFile == NULL)
-				goto cifs_create_out;
-			pCifsFile->netfid = fileHandle;
-			pCifsFile->pid = current->tgid;
-			pCifsFile->pInode = newinode;
-			pCifsFile->invalidHandle = false;
-			pCifsFile->closePend     = false;
-			init_MUTEX(&pCifsFile->fh_sem);
-			mutex_init(&pCifsFile->lock_mutex);
-			INIT_LIST_HEAD(&pCifsFile->llist);
-			atomic_set(&pCifsFile->wrtPending, 0);
-
-			/* set the following in open now
+cifs_create_set_dentry:
+	if (rc == 0)
+		setup_cifs_dentry(tcon, direntry, newinode);
+	else
+		cFYI(1, ("Create worked, get_inode_info failed rc = %d", rc));
+
+	/* nfsd case - nfs srv does not set nd */
+	if ((nd == NULL) || (!(nd->flags & LOOKUP_OPEN))) {
+		/* mknod case - do not leave file open */
+		CIFSSMBClose(xid, tcon, fileHandle);
+	} else if (newinode) {
+		struct cifsFileInfo *pCifsFile =
+			kzalloc(sizeof(struct cifsFileInfo), GFP_KERNEL);
+
+		if (pCifsFile == NULL)
+			goto cifs_create_out;
+		pCifsFile->netfid = fileHandle;
+		pCifsFile->pid = current->tgid;
+		pCifsFile->pInode = newinode;
+		pCifsFile->invalidHandle = false;
+		pCifsFile->closePend     = false;
+		init_MUTEX(&pCifsFile->fh_sem);
+		mutex_init(&pCifsFile->lock_mutex);
+		INIT_LIST_HEAD(&pCifsFile->llist);
+		atomic_set(&pCifsFile->wrtPending, 0);
+
+		/* set the following in open now
 				pCifsFile->pfile = file; */
-			write_lock(&GlobalSMBSeslock);
-			list_add(&pCifsFile->tlist, &tcon->openFileList);
-			pCifsInode = CIFS_I(newinode);
-			if (pCifsInode) {
-				/* if readable file instance put first in list*/
-				if (write_only) {
-					list_add_tail(&pCifsFile->flist,
-						&pCifsInode->openFileList);
-				} else {
-					list_add(&pCifsFile->flist,
-						&pCifsInode->openFileList);
-				}
-				if ((oplock & 0xF) == OPLOCK_EXCLUSIVE) {
-					pCifsInode->clientCanCacheAll = true;
-					pCifsInode->clientCanCacheRead = true;
-					cFYI(1, ("Exclusive Oplock inode %p",
-						newinode));
-				} else if ((oplock & 0xF) == OPLOCK_READ)
-					pCifsInode->clientCanCacheRead = true;
+		write_lock(&GlobalSMBSeslock);
+		list_add(&pCifsFile->tlist, &tcon->openFileList);
+		pCifsInode = CIFS_I(newinode);
+		if (pCifsInode) {
+			/* if readable file instance put first in list*/
+			if (write_only) {
+				list_add_tail(&pCifsFile->flist,
+					      &pCifsInode->openFileList);
+			} else {
+				list_add(&pCifsFile->flist,
+					 &pCifsInode->openFileList);
 			}
-			write_unlock(&GlobalSMBSeslock);
+			if ((oplock & 0xF) == OPLOCK_EXCLUSIVE) {
+				pCifsInode->clientCanCacheAll = true;
+				pCifsInode->clientCanCacheRead = true;
+				cFYI(1, ("Exclusive Oplock inode %p",
+					newinode));
+			} else if ((oplock & 0xF) == OPLOCK_READ)
+				pCifsInode->clientCanCacheRead = true;
 		}
+		write_unlock(&GlobalSMBSeslock);
 	}
 cifs_create_out:
 	kfree(buf);
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index bcf7b518466..4690a360c85 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -199,6 +199,49 @@ static void fill_fake_finddataunix(FILE_UNIX_BASIC_INFO *pfnd_dat,
 	pfnd_dat->Gid = cpu_to_le64(pinode->i_gid);
 }
 
+/**
+ * cifs_new inode - create new inode, initialize, and hash it
+ * @sb - pointer to superblock
+ * @inum - if valid pointer and serverino is enabled, replace i_ino with val
+ *
+ * Create a new inode, initialize it for CIFS and hash it. Returns the new
+ * inode or NULL if one couldn't be allocated.
+ *
+ * If the share isn't mounted with "serverino" or inum is a NULL pointer then
+ * we'll just use the inode number assigned by new_inode(). Note that this can
+ * mean i_ino collisions since the i_ino assigned by new_inode is not
+ * guaranteed to be unique.
+ */
+struct inode *
+cifs_new_inode(struct super_block *sb, __u64 *inum)
+{
+	struct inode *inode;
+
+	inode = new_inode(sb);
+	if (inode == NULL)
+		return NULL;
+
+	/*
+	 * BB: Is i_ino == 0 legal? Here, we assume that it is. If it isn't we
+	 *     stop passing inum as ptr. Are there sanity checks we can use to
+	 *     ensure that the server is really filling in that field? Also,
+	 *     if serverino is disabled, perhaps we should be using iunique()?
+	 */
+	if (inum && (CIFS_SB(sb)->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM))
+		inode->i_ino = (unsigned long) *inum;
+
+	/*
+	 * must set this here instead of cifs_alloc_inode since VFS will
+	 * clobber i_flags
+	 */
+	if (sb->s_flags & MS_NOATIME)
+		inode->i_flags |= S_NOATIME | S_NOCMTIME;
+
+	insert_inode_hash(inode);
+
+	return inode;
+}
+
 int cifs_get_inode_info_unix(struct inode **pinode,
 	const unsigned char *full_path, struct super_block *sb, int xid)
 {
@@ -233,22 +276,11 @@ int cifs_get_inode_info_unix(struct inode **pinode,
 
 	/* get new inode */
 	if (*pinode == NULL) {
-		*pinode = new_inode(sb);
+		*pinode = cifs_new_inode(sb, &find_data.UniqueId);
 		if (*pinode == NULL) {
 			rc = -ENOMEM;
 			goto cgiiu_exit;
 		}
-		/* Is an i_ino of zero legal? */
-		/* note ino incremented to unique num in new_inode */
-		/* Are there sanity checks we can use to ensure that
-		   the server is really filling in that field? */
-		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM)
-			(*pinode)->i_ino = (unsigned long)find_data.UniqueId;
-
-		if (sb->s_flags & MS_NOATIME)
-			(*pinode)->i_flags |= S_NOATIME | S_NOCMTIME;
-
-		insert_inode_hash(*pinode);
 	}
 
 	inode = *pinode;
@@ -465,11 +497,9 @@ int cifs_get_inode_info(struct inode **pinode,
 
 	/* get new inode */
 	if (*pinode == NULL) {
-		*pinode = new_inode(sb);
-		if (*pinode == NULL) {
-			rc = -ENOMEM;
-			goto cgii_exit;
-		}
+		__u64 inode_num;
+		__u64 *pinum = &inode_num;
+
 		/* Is an i_ino of zero legal? Can we use that to check
 		   if the server supports returning inode numbers?  Are
 		   there other sanity checks we can use to ensure that
@@ -486,22 +516,26 @@ int cifs_get_inode_info(struct inode **pinode,
 
 		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) {
 			int rc1 = 0;
-			__u64 inode_num;
 
 			rc1 = CIFSGetSrvInodeNumber(xid, pTcon,
-					full_path, &inode_num,
+					full_path, pinum,
 					cifs_sb->local_nls,
 					cifs_sb->mnt_cifs_flags &
 						CIFS_MOUNT_MAP_SPECIAL_CHR);
 			if (rc1) {
 				cFYI(1, ("GetSrvInodeNum rc %d", rc1));
+				pinum = NULL;
 				/* BB EOPNOSUPP disable SERVER_INUM? */
-			} else /* do we need cast or hash to ino? */
-				(*pinode)->i_ino = inode_num;
-		} /* else ino incremented to unique num in new_inode*/
-		if (sb->s_flags & MS_NOATIME)
-			(*pinode)->i_flags |= S_NOATIME | S_NOCMTIME;
-		insert_inode_hash(*pinode);
+			}
+		} else {
+			pinum = NULL;
+		}
+
+		*pinode = cifs_new_inode(sb, pinum);
+		if (*pinode == NULL) {
+			rc = -ENOMEM;
+			goto cgii_exit;
+		}
 	}
 	inode = *pinode;
 	cifsInfo = CIFS_I(inode);
@@ -621,7 +655,7 @@ static const struct inode_operations cifs_ipc_inode_ops = {
 	.lookup = cifs_lookup,
 };
 
-static char *build_path_to_root(struct cifs_sb_info *cifs_sb)
+char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb)
 {
 	int pplen = cifs_sb->prepathlen;
 	int dfsplen;
@@ -678,7 +712,7 @@ struct inode *cifs_iget(struct super_block *sb, unsigned long ino)
 		return inode;
 
 	cifs_sb = CIFS_SB(inode->i_sb);
-	full_path = build_path_to_root(cifs_sb);
+	full_path = cifs_build_path_to_root(cifs_sb);
 	if (full_path == NULL)
 		return ERR_PTR(-ENOMEM);
 
@@ -1017,7 +1051,7 @@ out_reval:
 	return rc;
 }
 
-static void posix_fill_in_inode(struct inode *tmp_inode,
+void posix_fill_in_inode(struct inode *tmp_inode,
 	FILE_UNIX_BASIC_INFO *pData, int isNewInode)
 {
 	struct cifsInodeInfo *cifsInfo = CIFS_I(tmp_inode);
@@ -1114,24 +1148,14 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 			else
 				direntry->d_op = &cifs_dentry_ops;
 
-			newinode = new_inode(inode->i_sb);
+			newinode = cifs_new_inode(inode->i_sb,
+						  &pInfo->UniqueId);
 			if (newinode == NULL) {
 				kfree(pInfo);
 				goto mkdir_get_info;
 			}
 
-			/* Is an i_ino of zero legal? */
-			/* Are there sanity checks we can use to ensure that
-			   the server is really filling in that field? */
-			if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) {
-				newinode->i_ino =
-					(unsigned long)pInfo->UniqueId;
-			} /* note ino incremented to unique num in new_inode */
-			if (inode->i_sb->s_flags & MS_NOATIME)
-				newinode->i_flags |= S_NOATIME | S_NOCMTIME;
 			newinode->i_nlink = 2;
-
-			insert_inode_hash(newinode);
 			d_instantiate(direntry, newinode);
 
 			/* we already checked in POSIXCreate whether
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 9f51f9bf029..c2c01ff4c32 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -56,35 +56,34 @@ static inline void dump_cifs_file_struct(struct file *file, char *label)
 }
 #endif /* DEBUG2 */
 
-/* Returns one if new inode created (which therefore needs to be hashed) */
+/* Returns 1 if new inode created, 2 if both dentry and inode were */
 /* Might check in the future if inode number changed so we can rehash inode */
-static int construct_dentry(struct qstr *qstring, struct file *file,
-	struct inode **ptmp_inode, struct dentry **pnew_dentry)
+static int
+construct_dentry(struct qstr *qstring, struct file *file,
+		 struct inode **ptmp_inode, struct dentry **pnew_dentry,
+		 __u64 *inum)
 {
-	struct dentry *tmp_dentry;
-	struct cifs_sb_info *cifs_sb;
-	struct cifsTconInfo *pTcon;
+	struct dentry *tmp_dentry = NULL;
+	struct super_block *sb = file->f_path.dentry->d_sb;
 	int rc = 0;
 
 	cFYI(1, ("For %s", qstring->name));
-	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
 
 	qstring->hash = full_name_hash(qstring->name, qstring->len);
 	tmp_dentry = d_lookup(file->f_path.dentry, qstring);
 	if (tmp_dentry) {
+		/* BB: overwrite old name? i.e. tmp_dentry->d_name and
+		 * tmp_dentry->d_name.len??
+		 */
 		cFYI(0, ("existing dentry with inode 0x%p",
 			 tmp_dentry->d_inode));
 		*ptmp_inode = tmp_dentry->d_inode;
-/* BB overwrite old name? i.e. tmp_dentry->d_name and tmp_dentry->d_name.len??*/
 		if (*ptmp_inode == NULL) {
-			*ptmp_inode = new_inode(file->f_path.dentry->d_sb);
+			*ptmp_inode = cifs_new_inode(sb, inum);
 			if (*ptmp_inode == NULL)
 				return rc;
 			rc = 1;
 		}
-		if (file->f_path.dentry->d_sb->s_flags & MS_NOATIME)
-			(*ptmp_inode)->i_flags |= S_NOATIME | S_NOCMTIME;
 	} else {
 		tmp_dentry = d_alloc(file->f_path.dentry, qstring);
 		if (tmp_dentry == NULL) {
@@ -93,15 +92,14 @@ static int construct_dentry(struct qstr *qstring, struct file *file,
 			return rc;
 		}
 
-		*ptmp_inode = new_inode(file->f_path.dentry->d_sb);
-		if (pTcon->nocase)
+		if (CIFS_SB(sb)->tcon->nocase)
 			tmp_dentry->d_op = &cifs_ci_dentry_ops;
 		else
 			tmp_dentry->d_op = &cifs_dentry_ops;
+
+		*ptmp_inode = cifs_new_inode(sb, inum);
 		if (*ptmp_inode == NULL)
 			return rc;
-		if (file->f_path.dentry->d_sb->s_flags & MS_NOATIME)
-			(*ptmp_inode)->i_flags |= S_NOATIME | S_NOCMTIME;
 		rc = 2;
 	}
 
@@ -822,7 +820,7 @@ static int find_cifs_entry(const int xid, struct cifsTconInfo *pTcon,
 /* inode num, inode type and filename returned */
 static int cifs_get_name_from_search_buf(struct qstr *pqst,
 	char *current_entry, __u16 level, unsigned int unicode,
-	struct cifs_sb_info *cifs_sb, int max_len, ino_t *pinum)
+	struct cifs_sb_info *cifs_sb, int max_len, __u64 *pinum)
 {
 	int rc = 0;
 	unsigned int len = 0;
@@ -842,9 +840,7 @@ static int cifs_get_name_from_search_buf(struct qstr *pqst,
 			len = strnlen(filename, PATH_MAX);
 		}
 
-		/* BB fixme - hash low and high 32 bits if not 64 bit arch BB */
-		if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM)
-			*pinum = pFindData->UniqueId;
+		*pinum = pFindData->UniqueId;
 	} else if (level == SMB_FIND_FILE_DIRECTORY_INFO) {
 		FILE_DIRECTORY_INFO *pFindData =
 			(FILE_DIRECTORY_INFO *)current_entry;
@@ -907,7 +903,7 @@ static int cifs_filldir(char *pfindEntry, struct file *file,
 	struct qstr qstring;
 	struct cifsFileInfo *pCifsF;
 	unsigned int obj_type;
-	ino_t  inum;
+	__u64  inum;
 	struct cifs_sb_info *cifs_sb;
 	struct inode *tmp_inode;
 	struct dentry *tmp_dentry;
@@ -940,20 +936,18 @@ static int cifs_filldir(char *pfindEntry, struct file *file,
 	if (rc)
 		return rc;
 
-	rc = construct_dentry(&qstring, file, &tmp_inode, &tmp_dentry);
+	/* only these two infolevels return valid inode numbers */
+	if (pCifsF->srch_inf.info_level == SMB_FIND_FILE_UNIX ||
+	    pCifsF->srch_inf.info_level == SMB_FIND_FILE_ID_FULL_DIR_INFO)
+		rc = construct_dentry(&qstring, file, &tmp_inode, &tmp_dentry,
+					&inum);
+	else
+		rc = construct_dentry(&qstring, file, &tmp_inode, &tmp_dentry,
+					NULL);
+
 	if ((tmp_inode == NULL) || (tmp_dentry == NULL))
 		return -ENOMEM;
 
-	if (rc) {
-		/* inode created, we need to hash it with right inode number */
-		if (inum != 0) {
-			/* BB fixme - hash the 2 32 quantities bits together if
-			 *  necessary BB */
-			tmp_inode->i_ino = inum;
-		}
-		insert_inode_hash(tmp_inode);
-	}
-
 	/* we pass in rc below, indicating whether it is a new inode,
 	   so we can figure out whether to invalidate the inode cached
 	   data if the file has changed */
diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c
index 5f22de7b79a..5c68b4282be 100644
--- a/fs/cifs/sess.c
+++ b/fs/cifs/sess.c
@@ -34,15 +34,99 @@
 extern void SMBNTencrypt(unsigned char *passwd, unsigned char *c8,
 			 unsigned char *p24);
 
+/* Checks if this is the first smb session to be reconnected after
+   the socket has been reestablished (so we know whether to use vc 0).
+   Called while holding the cifs_tcp_ses_lock, so do not block */
+static bool is_first_ses_reconnect(struct cifsSesInfo *ses)
+{
+	struct list_head *tmp;
+	struct cifsSesInfo *tmp_ses;
+
+	list_for_each(tmp, &ses->server->smb_ses_list) {
+		tmp_ses = list_entry(tmp, struct cifsSesInfo,
+				     smb_ses_list);
+		if (tmp_ses->need_reconnect == false)
+			return false;
+	}
+	/* could not find a session that was already connected,
+	   this must be the first one we are reconnecting */
+	return true;
+}
+
+/*
+ *	vc number 0 is treated specially by some servers, and should be the
+ *      first one we request.  After that we can use vcnumbers up to maxvcs,
+ *	one for each smb session (some Windows versions set maxvcs incorrectly
+ *	so maxvc=1 can be ignored).  If we have too many vcs, we can reuse
+ *	any vc but zero (some servers reset the connection on vcnum zero)
+ *
+ */
+static __le16 get_next_vcnum(struct cifsSesInfo *ses)
+{
+	__u16 vcnum = 0;
+	struct list_head *tmp;
+	struct cifsSesInfo *tmp_ses;
+	__u16 max_vcs = ses->server->max_vcs;
+	__u16 i;
+	int free_vc_found = 0;
+
+	/* Quoting the MS-SMB specification: "Windows-based SMB servers set this
+	field to one but do not enforce this limit, which allows an SMB client
+	to establish more virtual circuits than allowed by this value ... but
+	other server implementations can enforce this limit." */
+	if (max_vcs < 2)
+		max_vcs = 0xFFFF;
+
+	write_lock(&cifs_tcp_ses_lock);
+	if ((ses->need_reconnect) && is_first_ses_reconnect(ses))
+			goto get_vc_num_exit;  /* vcnum will be zero */
+	for (i = ses->server->srv_count - 1; i < max_vcs; i++) {
+		if (i == 0) /* this is the only connection, use vc 0 */
+			break;
+
+		free_vc_found = 1;
+
+		list_for_each(tmp, &ses->server->smb_ses_list) {
+			tmp_ses = list_entry(tmp, struct cifsSesInfo,
+					     smb_ses_list);
+			if (tmp_ses->vcnum == i) {
+				free_vc_found = 0;
+				break; /* found duplicate, try next vcnum */
+			}
+		}
+		if (free_vc_found)
+			break; /* we found a vcnumber that will work - use it */
+	}
+
+	if (i == 0)
+		vcnum = 0; /* for most common case, ie if one smb session, use
+			      vc zero.  Also for case when no free vcnum, zero
+			      is safest to send (some clients only send zero) */
+	else if (free_vc_found == 0)
+		vcnum = 1;  /* we can not reuse vc=0 safely, since some servers
+				reset all uids on that, but 1 is ok. */
+	else
+		vcnum = i;
+	ses->vcnum = vcnum;
+get_vc_num_exit:
+	write_unlock(&cifs_tcp_ses_lock);
+
+	return le16_to_cpu(vcnum);
+}
+
 static __u32 cifs_ssetup_hdr(struct cifsSesInfo *ses, SESSION_SETUP_ANDX *pSMB)
 {
 	__u32 capabilities = 0;
 
 	/* init fields common to all four types of SessSetup */
-	/* note that header is initialized to zero in header_assemble */
+	/* Note that offsets for first seven fields in req struct are same  */
+	/*	in CIFS Specs so does not matter which of 3 forms of struct */
+	/*	that we use in next few lines                               */
+	/* Note that header is initialized to zero in header_assemble */
 	pSMB->req.AndXCommand = 0xFF;
 	pSMB->req.MaxBufferSize = cpu_to_le16(ses->server->maxBuf);
 	pSMB->req.MaxMpxCount = cpu_to_le16(ses->server->maxReq);
+	pSMB->req.VcNumber = get_next_vcnum(ses);
 
 	/* Now no need to set SMBFLG_CASELESS or obsolete CANONICAL PATH */
 
@@ -71,7 +155,6 @@ static __u32 cifs_ssetup_hdr(struct cifsSesInfo *ses, SESSION_SETUP_ANDX *pSMB)
 	if (ses->capabilities & CAP_UNIX)
 		capabilities |= CAP_UNIX;
 
-	/* BB check whether to init vcnum BB */
 	return capabilities;
 }
 
@@ -228,7 +311,7 @@ static int decode_unicode_ssetup(char **pbcc_area, int bleft,
 
 	kfree(ses->serverOS);
 	/* UTF-8 string will not grow more than four times as big as UCS-16 */
-	ses->serverOS = kzalloc(4 * len, GFP_KERNEL);
+	ses->serverOS = kzalloc((4 * len) + 2 /* trailing null */, GFP_KERNEL);
 	if (ses->serverOS != NULL)
 		cifs_strfromUCS_le(ses->serverOS, (__le16 *)data, len, nls_cp);
 	data += 2 * (len + 1);
@@ -241,7 +324,7 @@ static int decode_unicode_ssetup(char **pbcc_area, int bleft,
 		return rc;
 
 	kfree(ses->serverNOS);
-	ses->serverNOS = kzalloc(4 * len, GFP_KERNEL); /* BB this is wrong length FIXME BB */
+	ses->serverNOS = kzalloc((4 * len) + 2 /* trailing null */, GFP_KERNEL);
 	if (ses->serverNOS != NULL) {
 		cifs_strfromUCS_le(ses->serverNOS, (__le16 *)data, len,
 				   nls_cp);
diff --git a/fs/compat.c b/fs/compat.c
index 25589f8322f..0949b43794a 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1408,7 +1408,7 @@ int compat_do_execve(char * filename,
 	bprm->cred = prepare_exec_creds();
 	if (!bprm->cred)
 		goto out_unlock;
-	check_unsafe_exec(bprm);
+	check_unsafe_exec(bprm, current->files);
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
index c8f8d5904f5..45e59d3c7f1 100644
--- a/fs/compat_ioctl.c
+++ b/fs/compat_ioctl.c
@@ -785,7 +785,7 @@ static int sg_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
 
 	if (copy_in_user(&sgio->status, &sgio32->status,
 			 (4 * sizeof(unsigned char)) +
-			 (2 * sizeof(unsigned (short))) +
+			 (2 * sizeof(unsigned short)) +
 			 (3 * sizeof(int))))
 		return -EFAULT;
 
@@ -1913,6 +1913,9 @@ COMPATIBLE_IOCTL(FIONREAD)  /* This is also TIOCINQ */
 /* 0x00 */
 COMPATIBLE_IOCTL(FIBMAP)
 COMPATIBLE_IOCTL(FIGETBSZ)
+/* 'X' - originally XFS but some now in the VFS */
+COMPATIBLE_IOCTL(FIFREEZE)
+COMPATIBLE_IOCTL(FITHAW)
 /* RAID */
 COMPATIBLE_IOCTL(RAID_VERSION)
 COMPATIBLE_IOCTL(GET_ARRAY_INFO)
@@ -1938,6 +1941,8 @@ ULONG_IOCTL(SET_BITMAP_FILE)
 /* Big K */
 COMPATIBLE_IOCTL(PIO_FONT)
 COMPATIBLE_IOCTL(GIO_FONT)
+COMPATIBLE_IOCTL(PIO_CMAP)
+COMPATIBLE_IOCTL(GIO_CMAP)
 ULONG_IOCTL(KDSIGACCEPT)
 COMPATIBLE_IOCTL(KDGETKEYCODE)
 COMPATIBLE_IOCTL(KDSETKEYCODE)
diff --git a/fs/configfs/dir.c b/fs/configfs/dir.c
index 9c235839114..8e93341f3e8 100644
--- a/fs/configfs/dir.c
+++ b/fs/configfs/dir.c
@@ -553,24 +553,12 @@ static void detach_groups(struct config_group *group)
 
 		child = sd->s_dentry;
 
-		/*
-		 * Note: we hide this from lockdep since we have no way
-		 * to teach lockdep about recursive
-		 * I_MUTEX_PARENT -> I_MUTEX_CHILD patterns along a path
-		 * in an inode tree, which are valid as soon as
-		 * I_MUTEX_PARENT -> I_MUTEX_CHILD is valid from a
-		 * parent inode to one of its children.
-		 */
-		lockdep_off();
 		mutex_lock(&child->d_inode->i_mutex);
-		lockdep_on();
 
 		configfs_detach_group(sd->s_element);
 		child->d_inode->i_flags |= S_DEAD;
 
-		lockdep_off();
 		mutex_unlock(&child->d_inode->i_mutex);
-		lockdep_on();
 
 		d_delete(child);
 		dput(child);
@@ -760,22 +748,11 @@ static int configfs_attach_item(struct config_item *parent_item,
 			 * We are going to remove an inode and its dentry but
 			 * the VFS may already have hit and used them. Thus,
 			 * we must lock them as rmdir() would.
-			 *
-			 * Note: we hide this from lockdep since we have no way
-			 * to teach lockdep about recursive
-			 * I_MUTEX_PARENT -> I_MUTEX_CHILD patterns along a path
-			 * in an inode tree, which are valid as soon as
-			 * I_MUTEX_PARENT -> I_MUTEX_CHILD is valid from a
-			 * parent inode to one of its children.
 			 */
-			lockdep_off();
 			mutex_lock(&dentry->d_inode->i_mutex);
-			lockdep_on();
 			configfs_remove_dir(item);
 			dentry->d_inode->i_flags |= S_DEAD;
-			lockdep_off();
 			mutex_unlock(&dentry->d_inode->i_mutex);
-			lockdep_on();
 			d_delete(dentry);
 		}
 	}
@@ -810,25 +787,14 @@ static int configfs_attach_group(struct config_item *parent_item,
 		 *
 		 * We must also lock the inode to remove it safely in case of
 		 * error, as rmdir() would.
-		 *
-		 * Note: we hide this from lockdep since we have no way
-		 * to teach lockdep about recursive
-		 * I_MUTEX_PARENT -> I_MUTEX_CHILD patterns along a path
-		 * in an inode tree, which are valid as soon as
-		 * I_MUTEX_PARENT -> I_MUTEX_CHILD is valid from a
-		 * parent inode to one of its children.
 		 */
-		lockdep_off();
 		mutex_lock_nested(&dentry->d_inode->i_mutex, I_MUTEX_CHILD);
-		lockdep_on();
 		ret = populate_groups(to_config_group(item));
 		if (ret) {
 			configfs_detach_item(item);
 			dentry->d_inode->i_flags |= S_DEAD;
 		}
-		lockdep_off();
 		mutex_unlock(&dentry->d_inode->i_mutex);
-		lockdep_on();
 		if (ret)
 			d_delete(dentry);
 	}
@@ -990,17 +956,7 @@ static int configfs_depend_prep(struct dentry *origin,
 	BUG_ON(!origin || !sd);
 
 	/* Lock this guy on the way down */
-	/*
-	 * Note: we hide this from lockdep since we have no way
-	 * to teach lockdep about recursive
-	 * I_MUTEX_PARENT -> I_MUTEX_CHILD patterns along a path
-	 * in an inode tree, which are valid as soon as
-	 * I_MUTEX_PARENT -> I_MUTEX_CHILD is valid from a
-	 * parent inode to one of its children.
-	 */
-	lockdep_off();
 	mutex_lock(&sd->s_dentry->d_inode->i_mutex);
-	lockdep_on();
 	if (sd->s_element == target)  /* Boo-yah */
 		goto out;
 
@@ -1014,9 +970,7 @@ static int configfs_depend_prep(struct dentry *origin,
 	}
 
 	/* We looped all our children and didn't find target */
-	lockdep_off();
 	mutex_unlock(&sd->s_dentry->d_inode->i_mutex);
-	lockdep_on();
 	ret = -ENOENT;
 
 out:
@@ -1036,16 +990,11 @@ static void configfs_depend_rollback(struct dentry *origin,
 	struct dentry *dentry = item->ci_dentry;
 
 	while (dentry != origin) {
-		/* See comments in configfs_depend_prep() */
-		lockdep_off();
 		mutex_unlock(&dentry->d_inode->i_mutex);
-		lockdep_on();
 		dentry = dentry->d_parent;
 	}
 
-	lockdep_off();
 	mutex_unlock(&origin->d_inode->i_mutex);
-	lockdep_on();
 }
 
 int configfs_depend_item(struct configfs_subsystem *subsys,
@@ -1380,16 +1329,8 @@ static int configfs_rmdir(struct inode *dir, struct dentry *dentry)
 			}
 
 			/* Wait until the racing operation terminates */
-			/*
-			 * Note: we hide this from lockdep since we are locked
-			 * with subclass I_MUTEX_NORMAL from vfs_rmdir() (why
-			 * not I_MUTEX_CHILD?), and I_MUTEX_XATTR or
-			 * I_MUTEX_QUOTA are not relevant for the locked inode.
-			 */
-			lockdep_off();
 			mutex_lock(wait_mutex);
 			mutex_unlock(wait_mutex);
-			lockdep_on();
 		}
 	} while (ret == -EAGAIN);
 
diff --git a/fs/dcache.c b/fs/dcache.c
index 937df0fb0da..07e2d4a44bd 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1180,7 +1180,7 @@ struct dentry *d_obtain_alias(struct inode *inode)
 	iput(inode);
 	return res;
 }
-EXPORT_SYMBOL_GPL(d_obtain_alias);
+EXPORT_SYMBOL(d_obtain_alias);
 
 /**
  * d_splice_alias - splice a disconnected dentry into the tree if one exists
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index 5f3231b9633..bff4052b05e 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -198,9 +198,6 @@ static int mknod_ptmx(struct super_block *sb)
 
 	fsi->ptmx_dentry = dentry;
 	rc = 0;
-
-	printk(KERN_DEBUG "Created ptmx node in devpts ino %lu\n",
-			inode->i_ino);
 out:
 	mutex_unlock(&root->d_inode->i_mutex);
 	return rc;
@@ -369,8 +366,6 @@ static int new_pts_mount(struct file_system_type *fs_type, int flags,
 	struct pts_fs_info *fsi;
 	struct pts_mount_opts *opts;
 
-	printk(KERN_NOTICE "devpts: newinstance mount\n");
-
 	err = get_sb_nodev(fs_type, flags, data, devpts_fill_super, mnt);
 	if (err)
 		return err;
diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c
index c01e043670e..8b65f289ee0 100644
--- a/fs/ecryptfs/crypto.c
+++ b/fs/ecryptfs/crypto.c
@@ -946,6 +946,8 @@ static int ecryptfs_copy_mount_wide_sigs_to_inode_sigs(
 	list_for_each_entry(global_auth_tok,
 			    &mount_crypt_stat->global_auth_tok_list,
 			    mount_crypt_stat_list) {
+		if (global_auth_tok->flags & ECRYPTFS_AUTH_TOK_FNEK)
+			continue;
 		rc = ecryptfs_add_keysig(crypt_stat, global_auth_tok->sig);
 		if (rc) {
 			printk(KERN_ERR "Error adding keysig; rc = [%d]\n", rc);
@@ -1322,14 +1324,13 @@ static int ecryptfs_write_headers_virt(char *page_virt, size_t max,
 }
 
 static int
-ecryptfs_write_metadata_to_contents(struct ecryptfs_crypt_stat *crypt_stat,
-				    struct dentry *ecryptfs_dentry,
-				    char *virt)
+ecryptfs_write_metadata_to_contents(struct dentry *ecryptfs_dentry,
+				    char *virt, size_t virt_len)
 {
 	int rc;
 
 	rc = ecryptfs_write_lower(ecryptfs_dentry->d_inode, virt,
-				  0, crypt_stat->num_header_bytes_at_front);
+				  0, virt_len);
 	if (rc)
 		printk(KERN_ERR "%s: Error attempting to write header "
 		       "information to lower file; rc = [%d]\n", __func__,
@@ -1339,7 +1340,6 @@ ecryptfs_write_metadata_to_contents(struct ecryptfs_crypt_stat *crypt_stat,
 
 static int
 ecryptfs_write_metadata_to_xattr(struct dentry *ecryptfs_dentry,
-				 struct ecryptfs_crypt_stat *crypt_stat,
 				 char *page_virt, size_t size)
 {
 	int rc;
@@ -1349,6 +1349,17 @@ ecryptfs_write_metadata_to_xattr(struct dentry *ecryptfs_dentry,
 	return rc;
 }
 
+static unsigned long ecryptfs_get_zeroed_pages(gfp_t gfp_mask,
+					       unsigned int order)
+{
+	struct page *page;
+
+	page = alloc_pages(gfp_mask | __GFP_ZERO, order);
+	if (page)
+		return (unsigned long) page_address(page);
+	return 0;
+}
+
 /**
  * ecryptfs_write_metadata
  * @ecryptfs_dentry: The eCryptfs dentry
@@ -1365,7 +1376,9 @@ int ecryptfs_write_metadata(struct dentry *ecryptfs_dentry)
 {
 	struct ecryptfs_crypt_stat *crypt_stat =
 		&ecryptfs_inode_to_private(ecryptfs_dentry->d_inode)->crypt_stat;
+	unsigned int order;
 	char *virt;
+	size_t virt_len;
 	size_t size = 0;
 	int rc = 0;
 
@@ -1381,33 +1394,35 @@ int ecryptfs_write_metadata(struct dentry *ecryptfs_dentry)
 		rc = -EINVAL;
 		goto out;
 	}
+	virt_len = crypt_stat->num_header_bytes_at_front;
+	order = get_order(virt_len);
 	/* Released in this function */
-	virt = (char *)get_zeroed_page(GFP_KERNEL);
+	virt = (char *)ecryptfs_get_zeroed_pages(GFP_KERNEL, order);
 	if (!virt) {
 		printk(KERN_ERR "%s: Out of memory\n", __func__);
 		rc = -ENOMEM;
 		goto out;
 	}
-	rc = ecryptfs_write_headers_virt(virt, PAGE_CACHE_SIZE, &size,
-					 crypt_stat, ecryptfs_dentry);
+	rc = ecryptfs_write_headers_virt(virt, virt_len, &size, crypt_stat,
+					 ecryptfs_dentry);
 	if (unlikely(rc)) {
 		printk(KERN_ERR "%s: Error whilst writing headers; rc = [%d]\n",
 		       __func__, rc);
 		goto out_free;
 	}
 	if (crypt_stat->flags & ECRYPTFS_METADATA_IN_XATTR)
-		rc = ecryptfs_write_metadata_to_xattr(ecryptfs_dentry,
-						      crypt_stat, virt, size);
+		rc = ecryptfs_write_metadata_to_xattr(ecryptfs_dentry, virt,
+						      size);
 	else
-		rc = ecryptfs_write_metadata_to_contents(crypt_stat,
-							 ecryptfs_dentry, virt);
+		rc = ecryptfs_write_metadata_to_contents(ecryptfs_dentry, virt,
+							 virt_len);
 	if (rc) {
 		printk(KERN_ERR "%s: Error writing metadata out to lower file; "
 		       "rc = [%d]\n", __func__, rc);
 		goto out_free;
 	}
 out_free:
-	free_page((unsigned long)virt);
+	free_pages((unsigned long)virt, order);
 out:
 	return rc;
 }
@@ -1716,7 +1731,7 @@ static int ecryptfs_copy_filename(char **copied_name, size_t *copied_name_size,
 {
 	int rc = 0;
 
-	(*copied_name) = kmalloc((name_size + 2), GFP_KERNEL);
+	(*copied_name) = kmalloc((name_size + 1), GFP_KERNEL);
 	if (!(*copied_name)) {
 		rc = -ENOMEM;
 		goto out;
@@ -1726,7 +1741,7 @@ static int ecryptfs_copy_filename(char **copied_name, size_t *copied_name_size,
 						 * in printing out the
 						 * string in debug
 						 * messages */
-	(*copied_name_size) = (name_size + 1);
+	(*copied_name_size) = name_size;
 out:
 	return rc;
 }
@@ -2206,17 +2221,19 @@ int ecryptfs_decode_and_decrypt_filename(char **plaintext_name,
 					 struct dentry *ecryptfs_dir_dentry,
 					 const char *name, size_t name_size)
 {
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat =
+		&ecryptfs_superblock_to_private(
+			ecryptfs_dir_dentry->d_sb)->mount_crypt_stat;
 	char *decoded_name;
 	size_t decoded_name_size;
 	size_t packet_size;
 	int rc = 0;
 
-	if ((name_size > ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE)
+	if ((mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)
+	    && !(mount_crypt_stat->flags & ECRYPTFS_ENCRYPTED_VIEW_ENABLED)
+	    && (name_size > ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE)
 	    && (strncmp(name, ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX,
 			ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE) == 0)) {
-		struct ecryptfs_mount_crypt_stat *mount_crypt_stat =
-			&ecryptfs_superblock_to_private(
-				ecryptfs_dir_dentry->d_sb)->mount_crypt_stat;
 		const char *orig_name = name;
 		size_t orig_name_size = name_size;
 
diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h
index c11fc95714a..ac749d4d644 100644
--- a/fs/ecryptfs/ecryptfs_kernel.h
+++ b/fs/ecryptfs/ecryptfs_kernel.h
@@ -328,6 +328,7 @@ struct ecryptfs_dentry_info {
  */
 struct ecryptfs_global_auth_tok {
 #define ECRYPTFS_AUTH_TOK_INVALID 0x00000001
+#define ECRYPTFS_AUTH_TOK_FNEK    0x00000002
 	u32 flags;
 	struct list_head mount_crypt_stat_list;
 	struct key *global_auth_tok_key;
@@ -619,7 +620,6 @@ int ecryptfs_interpose(struct dentry *hidden_dentry,
 		       u32 flags);
 int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
 					struct dentry *lower_dentry,
-					struct ecryptfs_crypt_stat *crypt_stat,
 					struct inode *ecryptfs_dir_inode,
 					struct nameidata *ecryptfs_nd);
 int ecryptfs_decode_and_decrypt_filename(char **decrypted_name,
@@ -696,7 +696,7 @@ ecryptfs_write_header_metadata(char *virt,
 int ecryptfs_add_keysig(struct ecryptfs_crypt_stat *crypt_stat, char *sig);
 int
 ecryptfs_add_global_auth_tok(struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
-			   char *sig);
+			   char *sig, u32 global_auth_tok_flags);
 int ecryptfs_get_global_auth_tok_for_sig(
 	struct ecryptfs_global_auth_tok **global_auth_tok,
 	struct ecryptfs_mount_crypt_stat *mount_crypt_stat, char *sig);
diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c
index 5697899a168..55b3145b807 100644
--- a/fs/ecryptfs/inode.c
+++ b/fs/ecryptfs/inode.c
@@ -246,7 +246,6 @@ out:
  */
 int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
 					struct dentry *lower_dentry,
-					struct ecryptfs_crypt_stat *crypt_stat,
 					struct inode *ecryptfs_dir_inode,
 					struct nameidata *ecryptfs_nd)
 {
@@ -254,6 +253,7 @@ int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
 	struct vfsmount *lower_mnt;
 	struct inode *lower_inode;
 	struct ecryptfs_mount_crypt_stat *mount_crypt_stat;
+	struct ecryptfs_crypt_stat *crypt_stat;
 	char *page_virt = NULL;
 	u64 file_size;
 	int rc = 0;
@@ -314,6 +314,11 @@ int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
 			goto out_free_kmem;
 		}
 	}
+	crypt_stat = &ecryptfs_inode_to_private(
+					ecryptfs_dentry->d_inode)->crypt_stat;
+	/* TODO: lock for crypt_stat comparison */
+	if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED))
+			ecryptfs_set_default_sizes(crypt_stat);
 	rc = ecryptfs_read_and_validate_header_region(page_virt,
 						      ecryptfs_dentry->d_inode);
 	if (rc) {
@@ -362,9 +367,7 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode,
 {
 	char *encrypted_and_encoded_name = NULL;
 	size_t encrypted_and_encoded_name_size;
-	struct ecryptfs_crypt_stat *crypt_stat = NULL;
 	struct ecryptfs_mount_crypt_stat *mount_crypt_stat = NULL;
-	struct ecryptfs_inode_info *inode_info;
 	struct dentry *lower_dir_dentry, *lower_dentry;
 	int rc = 0;
 
@@ -388,26 +391,15 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode,
 	}
 	if (lower_dentry->d_inode)
 		goto lookup_and_interpose;
-	inode_info =  ecryptfs_inode_to_private(ecryptfs_dentry->d_inode);
-	if (inode_info) {
-		crypt_stat = &inode_info->crypt_stat;
-		/* TODO: lock for crypt_stat comparison */
-		if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED))
-			ecryptfs_set_default_sizes(crypt_stat);
-	}
-	if (crypt_stat)
-		mount_crypt_stat = crypt_stat->mount_crypt_stat;
-	else
-		mount_crypt_stat = &ecryptfs_superblock_to_private(
-			ecryptfs_dentry->d_sb)->mount_crypt_stat;
-	if (!(crypt_stat && (crypt_stat->flags & ECRYPTFS_ENCRYPT_FILENAMES))
-	    && !(mount_crypt_stat && (mount_crypt_stat->flags
-				     & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)))
+	mount_crypt_stat = &ecryptfs_superblock_to_private(
+				ecryptfs_dentry->d_sb)->mount_crypt_stat;
+	if (!(mount_crypt_stat
+	    && (mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)))
 		goto lookup_and_interpose;
 	dput(lower_dentry);
 	rc = ecryptfs_encrypt_and_encode_filename(
 		&encrypted_and_encoded_name, &encrypted_and_encoded_name_size,
-		crypt_stat, mount_crypt_stat, ecryptfs_dentry->d_name.name,
+		NULL, mount_crypt_stat, ecryptfs_dentry->d_name.name,
 		ecryptfs_dentry->d_name.len);
 	if (rc) {
 		printk(KERN_ERR "%s: Error attempting to encrypt and encode "
@@ -426,7 +418,7 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode,
 	}
 lookup_and_interpose:
 	rc = ecryptfs_lookup_and_interpose_lower(ecryptfs_dentry, lower_dentry,
-						 crypt_stat, ecryptfs_dir_inode,
+						 ecryptfs_dir_inode,
 						 ecryptfs_nd);
 	goto out;
 out_d_drop:
diff --git a/fs/ecryptfs/keystore.c b/fs/ecryptfs/keystore.c
index ff539420cc6..e4a6223c314 100644
--- a/fs/ecryptfs/keystore.c
+++ b/fs/ecryptfs/keystore.c
@@ -2375,7 +2375,7 @@ struct kmem_cache *ecryptfs_global_auth_tok_cache;
 
 int
 ecryptfs_add_global_auth_tok(struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
-			     char *sig)
+			     char *sig, u32 global_auth_tok_flags)
 {
 	struct ecryptfs_global_auth_tok *new_auth_tok;
 	int rc = 0;
@@ -2389,6 +2389,7 @@ ecryptfs_add_global_auth_tok(struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
 		goto out;
 	}
 	memcpy(new_auth_tok->sig, sig, ECRYPTFS_SIG_SIZE_HEX);
+	new_auth_tok->flags = global_auth_tok_flags;
 	new_auth_tok->sig[ECRYPTFS_SIG_SIZE_HEX] = '\0';
 	mutex_lock(&mount_crypt_stat->global_auth_tok_list_mutex);
 	list_add(&new_auth_tok->mount_crypt_stat_list,
diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c
index 789cf2e1be1..aed56c25539 100644
--- a/fs/ecryptfs/main.c
+++ b/fs/ecryptfs/main.c
@@ -319,7 +319,7 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 		case ecryptfs_opt_ecryptfs_sig:
 			sig_src = args[0].from;
 			rc = ecryptfs_add_global_auth_tok(mount_crypt_stat,
-							  sig_src);
+							  sig_src, 0);
 			if (rc) {
 				printk(KERN_ERR "Error attempting to register "
 				       "global sig; rc = [%d]\n", rc);
@@ -370,7 +370,8 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 				ECRYPTFS_SIG_SIZE_HEX] = '\0';
 			rc = ecryptfs_add_global_auth_tok(
 				mount_crypt_stat,
-				mount_crypt_stat->global_default_fnek_sig);
+				mount_crypt_stat->global_default_fnek_sig,
+				ECRYPTFS_AUTH_TOK_FNEK);
 			if (rc) {
 				printk(KERN_ERR "Error attempting to register "
 				       "global fnek sig [%s]; rc = [%d]\n",
diff --git a/fs/exec.c b/fs/exec.c
index 9881dc3bb48..b9f1c144b7a 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -1056,16 +1056,32 @@ EXPORT_SYMBOL(install_exec_creds);
  * - the caller must hold current->cred_exec_mutex to protect against
  *   PTRACE_ATTACH
  */
-void check_unsafe_exec(struct linux_binprm *bprm)
+void check_unsafe_exec(struct linux_binprm *bprm, struct files_struct *files)
 {
-	struct task_struct *p = current;
+	struct task_struct *p = current, *t;
+	unsigned long flags;
+	unsigned n_fs, n_files, n_sighand;
 
 	bprm->unsafe = tracehook_unsafe_exec(p);
 
-	if (atomic_read(&p->fs->count) > 1 ||
-	    atomic_read(&p->files->count) > 1 ||
-	    atomic_read(&p->sighand->count) > 1)
+	n_fs = 1;
+	n_files = 1;
+	n_sighand = 1;
+	lock_task_sighand(p, &flags);
+	for (t = next_thread(p); t != p; t = next_thread(t)) {
+		if (t->fs == p->fs)
+			n_fs++;
+		if (t->files == files)
+			n_files++;
+		n_sighand++;
+	}
+
+	if (atomic_read(&p->fs->count) > n_fs ||
+	    atomic_read(&p->files->count) > n_files ||
+	    atomic_read(&p->sighand->count) > n_sighand)
 		bprm->unsafe |= LSM_UNSAFE_SHARE;
+
+	unlock_task_sighand(p, &flags);
 }
 
 /* 
@@ -1284,7 +1300,7 @@ int do_execve(char * filename,
 	bprm->cred = prepare_exec_creds();
 	if (!bprm->cred)
 		goto out_unlock;
-	check_unsafe_exec(bprm);
+	check_unsafe_exec(bprm, displaced);
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index da8bdeaa2e6..7c6e3606f0e 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -1185,9 +1185,12 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data)
 	es = sbi->s_es;
 	if (((sbi->s_mount_opt & EXT2_MOUNT_XIP) !=
 	    (old_mount_opt & EXT2_MOUNT_XIP)) &&
-	    invalidate_inodes(sb))
-		ext2_warning(sb, __func__, "busy inodes while remounting "\
-			     "xip remain in cache (no functional problem)");
+	    invalidate_inodes(sb)) {
+		ext2_warning(sb, __func__, "refusing change of xip flag "
+			     "with busy inodes while remounting");
+		sbi->s_mount_opt &= ~EXT2_MOUNT_XIP;
+		sbi->s_mount_opt |= old_mount_opt & EXT2_MOUNT_XIP;
+	}
 	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
 		return 0;
 	if (*flags & MS_RDONLY) {
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index b70d90e08a3..4a970411a45 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -2428,12 +2428,13 @@ static void ext3_write_super (struct super_block * sb)
 
 static int ext3_sync_fs(struct super_block *sb, int wait)
 {
-	sb->s_dirt = 0;
-	if (wait)
-		ext3_force_commit(sb);
-	else
-		journal_start_commit(EXT3_SB(sb)->s_journal, NULL);
+	tid_t target;
 
+	sb->s_dirt = 0;
+	if (journal_start_commit(EXT3_SB(sb)->s_journal, &target)) {
+		if (wait)
+			log_wait_commit(EXT3_SB(sb)->s_journal, target);
+	}
 	return 0;
 }
 
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 9a50b8052dc..de9459b4cb9 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -609,7 +609,9 @@ int ext4_claim_free_blocks(struct ext4_sb_info *sbi,
  */
 int ext4_should_retry_alloc(struct super_block *sb, int *retries)
 {
-	if (!ext4_has_free_blocks(EXT4_SB(sb), 1) || (*retries)++ > 3)
+	if (!ext4_has_free_blocks(EXT4_SB(sb), 1) ||
+	    (*retries)++ > 3 ||
+	    !EXT4_SB(sb)->s_journal)
 		return 0;
 
 	jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id);
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index aafc9eba1c2..b0c87dce66a 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -868,7 +868,7 @@ static inline unsigned ext4_rec_len_from_disk(__le16 dlen)
 {
 	unsigned len = le16_to_cpu(dlen);
 
-	if (len == EXT4_MAX_REC_LEN)
+	if (len == EXT4_MAX_REC_LEN || len == 0)
 		return 1 << 16;
 	return len;
 }
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index e2eab196875..e0aa4fe4f59 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -1122,7 +1122,8 @@ ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path,
 	struct ext4_extent_idx *ix;
 	struct ext4_extent *ex;
 	ext4_fsblk_t block;
-	int depth, ee_len;
+	int depth;	/* Note, NOT eh_depth; depth from top of tree */
+	int ee_len;
 
 	BUG_ON(path == NULL);
 	depth = path->p_depth;
@@ -1179,7 +1180,8 @@ got_index:
 		if (bh == NULL)
 			return -EIO;
 		eh = ext_block_hdr(bh);
-		if (ext4_ext_check_header(inode, eh, depth)) {
+		/* subtract from p_depth to get proper eh_depth */
+		if (ext4_ext_check_header(inode, eh, path->p_depth - depth)) {
 			put_bh(bh);
 			return -EIO;
 		}
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 4fb86a0061d..2d2b3585ee9 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -188,7 +188,7 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 	struct ext4_group_desc *gdp;
 	struct ext4_super_block *es;
 	struct ext4_sb_info *sbi;
-	int fatal = 0, err, count;
+	int fatal = 0, err, count, cleared;
 	ext4_group_t flex_group;
 
 	if (atomic_read(&inode->i_count) > 1) {
@@ -248,8 +248,10 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 		goto error_return;
 
 	/* Ok, now we can actually update the inode bitmaps.. */
-	if (!ext4_clear_bit_atomic(sb_bgl_lock(sbi, block_group),
-					bit, bitmap_bh->b_data))
+	spin_lock(sb_bgl_lock(sbi, block_group));
+	cleared = ext4_clear_bit(bit, bitmap_bh->b_data);
+	spin_unlock(sb_bgl_lock(sbi, block_group));
+	if (!cleared)
 		ext4_error(sb, "ext4_free_inode",
 			   "bit already cleared for inode %lu", ino);
 	else {
@@ -696,6 +698,7 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode)
 	struct inode *ret;
 	ext4_group_t i;
 	int free = 0;
+	static int once = 1;
 	ext4_group_t flex_group;
 
 	/* Cannot create files in a deleted directory */
@@ -715,6 +718,14 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode)
 
 	if (sbi->s_log_groups_per_flex) {
 		ret2 = find_group_flex(sb, dir, &group);
+		if (ret2 == -1) {
+			ret2 = find_group_other(sb, dir, &group);
+			if (ret2 == 0 && once)
+				once = 0;
+				printk(KERN_NOTICE "ext4: find_group_flex "
+				       "failed, fallback succeeded dir %lu\n",
+				       dir->i_ino);
+		}
 		goto got_group;
 	}
 
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 03ba20be132..c7fed5b1874 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -47,8 +47,10 @@
 static inline int ext4_begin_ordered_truncate(struct inode *inode,
 					      loff_t new_size)
 {
-	return jbd2_journal_begin_ordered_truncate(&EXT4_I(inode)->jinode,
-						   new_size);
+	return jbd2_journal_begin_ordered_truncate(
+					EXT4_SB(inode->i_sb)->s_journal,
+					&EXT4_I(inode)->jinode,
+					new_size);
 }
 
 static void ext4_invalidatepage(struct page *page, unsigned long offset);
@@ -1366,6 +1368,10 @@ retry:
 		goto out;
 	}
 
+	/* We cannot recurse into the filesystem as the transaction is already
+	 * started */
+	flags |= AOP_FLAG_NOFS;
+
 	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		ext4_journal_stop(handle);
@@ -1375,7 +1381,7 @@ retry:
 	*pagep = page;
 
 	ret = block_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
-							ext4_get_block);
+				ext4_get_block);
 
 	if (!ret && ext4_should_journal_data(inode)) {
 		ret = walk_page_buffers(handle, page_buffers(page),
@@ -2437,6 +2443,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 	int no_nrwrite_index_update;
 	int pages_written = 0;
 	long pages_skipped;
+	int range_cyclic, cycled = 1, io_done = 0;
 	int needed_blocks, ret = 0, nr_to_writebump = 0;
 	struct ext4_sb_info *sbi = EXT4_SB(mapping->host->i_sb);
 
@@ -2488,9 +2495,15 @@ static int ext4_da_writepages(struct address_space *mapping,
 	if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
 		range_whole = 1;
 
-	if (wbc->range_cyclic)
+	range_cyclic = wbc->range_cyclic;
+	if (wbc->range_cyclic) {
 		index = mapping->writeback_index;
-	else
+		if (index)
+			cycled = 0;
+		wbc->range_start = index << PAGE_CACHE_SHIFT;
+		wbc->range_end  = LLONG_MAX;
+		wbc->range_cyclic = 0;
+	} else
 		index = wbc->range_start >> PAGE_CACHE_SHIFT;
 
 	mpd.wbc = wbc;
@@ -2504,6 +2517,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 	wbc->no_nrwrite_index_update = 1;
 	pages_skipped = wbc->pages_skipped;
 
+retry:
 	while (!ret && wbc->nr_to_write > 0) {
 
 		/*
@@ -2530,7 +2544,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 
 		ext4_journal_stop(handle);
 
-		if (mpd.retval == -ENOSPC) {
+		if ((mpd.retval == -ENOSPC) && sbi->s_journal) {
 			/* commit the transaction which would
 			 * free blocks released in the transaction
 			 * and try again
@@ -2546,6 +2560,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 			pages_written += mpd.pages_written;
 			wbc->pages_skipped = pages_skipped;
 			ret = 0;
+			io_done = 1;
 		} else if (wbc->nr_to_write)
 			/*
 			 * There is no more writeout needed
@@ -2554,6 +2569,13 @@ static int ext4_da_writepages(struct address_space *mapping,
 			 */
 			break;
 	}
+	if (!io_done && !cycled) {
+		cycled = 1;
+		index = 0;
+		wbc->range_start = index << PAGE_CACHE_SHIFT;
+		wbc->range_end  = mapping->writeback_index - 1;
+		goto retry;
+	}
 	if (pages_skipped != wbc->pages_skipped)
 		printk(KERN_EMERG "This should not happen leaving %s "
 				"with nr_to_write = %ld ret = %d\n",
@@ -2561,6 +2583,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 
 	/* Update index */
 	index += pages_written;
+	wbc->range_cyclic = range_cyclic;
 	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
 		/*
 		 * set the writeback_index so that range_cyclic
@@ -2648,6 +2671,9 @@ retry:
 		ret = PTR_ERR(handle);
 		goto out;
 	}
+	/* We cannot recurse into the filesystem as the transaction is already
+	 * started */
+	flags |= AOP_FLAG_NOFS;
 
 	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index deba54f6cbe..9f61e62f435 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -1447,7 +1447,7 @@ static void ext4_mb_measure_extent(struct ext4_allocation_context *ac,
 	struct ext4_free_extent *gex = &ac->ac_g_ex;
 
 	BUG_ON(ex->fe_len <= 0);
-	BUG_ON(ex->fe_len >= EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
+	BUG_ON(ex->fe_len > EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
 	BUG_ON(ex->fe_start >= EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
 	BUG_ON(ac->ac_status != AC_STATUS_CONTINUE);
 
@@ -3292,7 +3292,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
 	}
 	BUG_ON(start + size <= ac->ac_o_ex.fe_logical &&
 			start > ac->ac_o_ex.fe_logical);
-	BUG_ON(size <= 0 || size >= EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
+	BUG_ON(size <= 0 || size > EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
 
 	/* now prepare goal request */
 
@@ -3589,6 +3589,7 @@ static void ext4_mb_put_pa(struct ext4_allocation_context *ac,
 			struct super_block *sb, struct ext4_prealloc_space *pa)
 {
 	ext4_group_t grp;
+	ext4_fsblk_t grp_blk;
 
 	if (!atomic_dec_and_test(&pa->pa_count) || pa->pa_free != 0)
 		return;
@@ -3603,8 +3604,12 @@ static void ext4_mb_put_pa(struct ext4_allocation_context *ac,
 	pa->pa_deleted = 1;
 	spin_unlock(&pa->pa_lock);
 
-	/* -1 is to protect from crossing allocation group */
-	ext4_get_group_no_and_offset(sb, pa->pa_pstart - 1, &grp, NULL);
+	grp_blk = pa->pa_pstart;
+	/* If linear, pa_pstart may be in the next group when pa is used up */
+	if (pa->pa_linear)
+		grp_blk--;
+
+	ext4_get_group_no_and_offset(sb, grp_blk, &grp, NULL);
 
 	/*
 	 * possible race:
@@ -3693,6 +3698,8 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
 	pa->pa_free = pa->pa_len;
 	atomic_set(&pa->pa_count, 1);
 	spin_lock_init(&pa->pa_lock);
+	INIT_LIST_HEAD(&pa->pa_inode_list);
+	INIT_LIST_HEAD(&pa->pa_group_list);
 	pa->pa_deleted = 0;
 	pa->pa_linear = 0;
 
@@ -3755,6 +3762,7 @@ ext4_mb_new_group_pa(struct ext4_allocation_context *ac)
 	atomic_set(&pa->pa_count, 1);
 	spin_lock_init(&pa->pa_lock);
 	INIT_LIST_HEAD(&pa->pa_inode_list);
+	INIT_LIST_HEAD(&pa->pa_group_list);
 	pa->pa_deleted = 0;
 	pa->pa_linear = 1;
 
@@ -4476,23 +4484,26 @@ static int ext4_mb_release_context(struct ext4_allocation_context *ac)
 			pa->pa_free -= ac->ac_b_ex.fe_len;
 			pa->pa_len -= ac->ac_b_ex.fe_len;
 			spin_unlock(&pa->pa_lock);
-			/*
-			 * We want to add the pa to the right bucket.
-			 * Remove it from the list and while adding
-			 * make sure the list to which we are adding
-			 * doesn't grow big.
-			 */
-			if (likely(pa->pa_free)) {
-				spin_lock(pa->pa_obj_lock);
-				list_del_rcu(&pa->pa_inode_list);
-				spin_unlock(pa->pa_obj_lock);
-				ext4_mb_add_n_trim(ac);
-			}
 		}
-		ext4_mb_put_pa(ac, ac->ac_sb, pa);
 	}
 	if (ac->alloc_semp)
 		up_read(ac->alloc_semp);
+	if (pa) {
+		/*
+		 * We want to add the pa to the right bucket.
+		 * Remove it from the list and while adding
+		 * make sure the list to which we are adding
+		 * doesn't grow big.  We need to release
+		 * alloc_semp before calling ext4_mb_add_n_trim()
+		 */
+		if (pa->pa_linear && likely(pa->pa_free)) {
+			spin_lock(pa->pa_obj_lock);
+			list_del_rcu(&pa->pa_inode_list);
+			spin_unlock(pa->pa_obj_lock);
+			ext4_mb_add_n_trim(ac);
+		}
+		ext4_mb_put_pa(ac, ac->ac_sb, pa);
+	}
 	if (ac->ac_bitmap_page)
 		page_cache_release(ac->ac_bitmap_page);
 	if (ac->ac_buddy_page)
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index 734abca25e3..fe64d9f7985 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -481,7 +481,7 @@ int ext4_ext_migrate(struct inode *inode)
 					+ 1);
 	if (IS_ERR(handle)) {
 		retval = PTR_ERR(handle);
-		goto err_out;
+		return retval;
 	}
 	tmp_inode = ext4_new_inode(handle,
 				inode->i_sb->s_root->d_inode,
@@ -489,8 +489,7 @@ int ext4_ext_migrate(struct inode *inode)
 	if (IS_ERR(tmp_inode)) {
 		retval = -ENOMEM;
 		ext4_journal_stop(handle);
-		tmp_inode = NULL;
-		goto err_out;
+		return retval;
 	}
 	i_size_write(tmp_inode, i_size_read(inode));
 	/*
@@ -618,8 +617,7 @@ err_out:
 
 	ext4_journal_stop(handle);
 
-	if (tmp_inode)
-		iput(tmp_inode);
+	iput(tmp_inode);
 
 	return retval;
 }
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index e5f06a5f045..39d1993cfa1 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -3046,14 +3046,17 @@ static void ext4_write_super(struct super_block *sb)
 static int ext4_sync_fs(struct super_block *sb, int wait)
 {
 	int ret = 0;
+	tid_t target;
 
 	trace_mark(ext4_sync_fs, "dev %s wait %d", sb->s_id, wait);
 	sb->s_dirt = 0;
 	if (EXT4_SB(sb)->s_journal) {
-		if (wait)
-			ret = ext4_force_commit(sb);
-		else
- 			jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
+		if (jbd2_journal_start_commit(EXT4_SB(sb)->s_journal,
+					      &target)) {
+			if (wait)
+				jbd2_log_wait_commit(EXT4_SB(sb)->s_journal,
+						     target);
+		}
 	} else {
 		ext4_commit_super(sb, EXT4_SB(sb)->s_es, wait);
 	}
@@ -3088,7 +3091,6 @@ static int ext4_freeze(struct super_block *sb)
 
 		/* Journal blocked and flushed, clear needs_recovery flag. */
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
-		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
 		error = ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
 		if (error)
 			goto out;
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index 6b74d09adbe..de0004fe6e0 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -202,9 +202,9 @@ static sector_t _fat_bmap(struct address_space *mapping, sector_t block)
 	sector_t blocknr;
 
 	/* fat_get_cluster() assumes the requested blocknr isn't truncated. */
-	mutex_lock(&mapping->host->i_mutex);
+	down_read(&mapping->host->i_alloc_sem);
 	blocknr = generic_block_bmap(mapping, block, fat_get_block);
-	mutex_unlock(&mapping->host->i_mutex);
+	up_read(&mapping->host->i_alloc_sem);
 
 	return blocknr;
 }
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index e5eaa62fd17..e3fe9918faa 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -274,6 +274,7 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc)
 	int ret;
 
 	BUG_ON(inode->i_state & I_SYNC);
+	WARN_ON(inode->i_state & I_NEW);
 
 	/* Set I_SYNC, reset I_DIRTY */
 	dirty = inode->i_state & I_DIRTY;
@@ -298,6 +299,7 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc)
 	}
 
 	spin_lock(&inode_lock);
+	WARN_ON(inode->i_state & I_NEW);
 	inode->i_state &= ~I_SYNC;
 	if (!(inode->i_state & I_FREEING)) {
 		if (!(inode->i_state & I_DIRTY) &&
@@ -470,6 +472,11 @@ void generic_sync_sb_inodes(struct super_block *sb,
 			break;
 		}
 
+		if (inode->i_state & I_NEW) {
+			requeue_io(inode);
+			continue;
+		}
+
 		if (wbc->nonblocking && bdi_write_congested(bdi)) {
 			wbc->encountered_congestion = 1;
 			if (!sb_is_blkdev_sb(sb))
@@ -531,7 +538,7 @@ void generic_sync_sb_inodes(struct super_block *sb,
 		list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 			struct address_space *mapping;
 
-			if (inode->i_state & (I_FREEING|I_WILL_FREE))
+			if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW))
 				continue;
 			mapping = inode->i_mapping;
 			if (mapping->nrpages == 0)
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index 6903d37af03..9b800d97a68 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -108,7 +108,8 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 
 	if (hugetlb_reserve_pages(inode,
 				vma->vm_pgoff >> huge_page_order(h),
-				len >> huge_page_shift(h), vma))
+				len >> huge_page_shift(h), vma,
+				vma->vm_flags))
 		goto out;
 
 	ret = 0;
@@ -947,7 +948,7 @@ static int can_do_hugetlb_shm(void)
 			can_do_mlock());
 }
 
-struct file *hugetlb_file_setup(const char *name, size_t size)
+struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag)
 {
 	int error = -ENOMEM;
 	struct file *file;
@@ -981,7 +982,8 @@ struct file *hugetlb_file_setup(const char *name, size_t size)
 
 	error = -ENOMEM;
 	if (hugetlb_reserve_pages(inode, 0,
-			size >> huge_page_shift(hstate_inode(inode)), NULL))
+			size >> huge_page_shift(hstate_inode(inode)), NULL,
+			acctflag))
 		goto out_inode;
 
 	d_instantiate(dentry, inode);
diff --git a/fs/inode.c b/fs/inode.c
index 40e37c02656..8a47ccd9fc3 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -369,6 +369,7 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose)
 		invalidate_inode_buffers(inode);
 		if (!atomic_read(&inode->i_count)) {
 			list_move(&inode->i_list, dispose);
+			WARN_ON(inode->i_state & I_NEW);
 			inode->i_state |= I_FREEING;
 			count++;
 			continue;
@@ -470,6 +471,7 @@ static void prune_icache(int nr_to_scan)
 				continue;
 		}
 		list_move(&inode->i_list, &freeable);
+		WARN_ON(inode->i_state & I_NEW);
 		inode->i_state |= I_FREEING;
 		nr_pruned++;
 	}
@@ -666,6 +668,7 @@ void unlock_new_inode(struct inode *inode)
 	 * just created it (so there can be no old holders
 	 * that haven't tested I_LOCK).
 	 */
+	WARN_ON((inode->i_state & (I_LOCK|I_NEW)) != (I_LOCK|I_NEW));
 	inode->i_state &= ~(I_LOCK|I_NEW);
 	wake_up_inode(inode);
 }
@@ -1155,6 +1158,7 @@ void generic_delete_inode(struct inode *inode)
 
 	list_del_init(&inode->i_list);
 	list_del_init(&inode->i_sb_list);
+	WARN_ON(inode->i_state & I_NEW);
 	inode->i_state |= I_FREEING;
 	inodes_stat.nr_inodes--;
 	spin_unlock(&inode_lock);
@@ -1196,16 +1200,19 @@ static void generic_forget_inode(struct inode *inode)
 			spin_unlock(&inode_lock);
 			return;
 		}
+		WARN_ON(inode->i_state & I_NEW);
 		inode->i_state |= I_WILL_FREE;
 		spin_unlock(&inode_lock);
 		write_inode_now(inode, 1);
 		spin_lock(&inode_lock);
+		WARN_ON(inode->i_state & I_NEW);
 		inode->i_state &= ~I_WILL_FREE;
 		inodes_stat.nr_unused--;
 		hlist_del_init(&inode->i_hash);
 	}
 	list_del_init(&inode->i_list);
 	list_del_init(&inode->i_sb_list);
+	WARN_ON(inode->i_state & I_NEW);
 	inode->i_state |= I_FREEING;
 	inodes_stat.nr_inodes--;
 	spin_unlock(&inode_lock);
diff --git a/fs/internal.h b/fs/internal.h
index 53af885f173..0d8ac497b3d 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -43,7 +43,7 @@ extern void __init chrdev_init(void);
 /*
  * exec.c
  */
-extern void check_unsafe_exec(struct linux_binprm *);
+extern void check_unsafe_exec(struct linux_binprm *, struct files_struct *);
 
 /*
  * namespace.c
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index 9e4fa52d7dc..e79c07812af 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -427,7 +427,7 @@ int __log_space_left(journal_t *journal)
 }
 
 /*
- * Called under j_state_lock.  Returns true if a transaction was started.
+ * Called under j_state_lock.  Returns true if a transaction commit was started.
  */
 int __log_start_commit(journal_t *journal, tid_t target)
 {
@@ -495,7 +495,8 @@ int journal_force_commit_nested(journal_t *journal)
 
 /*
  * Start a commit of the current running transaction (if any).  Returns true
- * if a transaction was started, and fills its tid in at *ptid
+ * if a transaction is going to be committed (or is currently already
+ * committing), and fills its tid in at *ptid
  */
 int journal_start_commit(journal_t *journal, tid_t *ptid)
 {
@@ -505,15 +506,19 @@ int journal_start_commit(journal_t *journal, tid_t *ptid)
 	if (journal->j_running_transaction) {
 		tid_t tid = journal->j_running_transaction->t_tid;
 
-		ret = __log_start_commit(journal, tid);
-		if (ret && ptid)
+		__log_start_commit(journal, tid);
+		/* There's a running transaction and we've just made sure
+		 * it's commit has been scheduled. */
+		if (ptid)
 			*ptid = tid;
-	} else if (journal->j_committing_transaction && ptid) {
+		ret = 1;
+	} else if (journal->j_committing_transaction) {
 		/*
 		 * If ext3_write_super() recently started a commit, then we
 		 * have to wait for completion of that transaction
 		 */
-		*ptid = journal->j_committing_transaction->t_tid;
+		if (ptid)
+			*ptid = journal->j_committing_transaction->t_tid;
 		ret = 1;
 	}
 	spin_unlock(&journal->j_state_lock);
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index eb343008ede..58144102bf2 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -450,7 +450,7 @@ int __jbd2_log_space_left(journal_t *journal)
 }
 
 /*
- * Called under j_state_lock.  Returns true if a transaction was started.
+ * Called under j_state_lock.  Returns true if a transaction commit was started.
  */
 int __jbd2_log_start_commit(journal_t *journal, tid_t target)
 {
@@ -518,7 +518,8 @@ int jbd2_journal_force_commit_nested(journal_t *journal)
 
 /*
  * Start a commit of the current running transaction (if any).  Returns true
- * if a transaction was started, and fills its tid in at *ptid
+ * if a transaction is going to be committed (or is currently already
+ * committing), and fills its tid in at *ptid
  */
 int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
 {
@@ -528,15 +529,19 @@ int jbd2_journal_start_commit(journal_t *journal, tid_t *ptid)
 	if (journal->j_running_transaction) {
 		tid_t tid = journal->j_running_transaction->t_tid;
 
-		ret = __jbd2_log_start_commit(journal, tid);
-		if (ret && ptid)
+		__jbd2_log_start_commit(journal, tid);
+		/* There's a running transaction and we've just made sure
+		 * it's commit has been scheduled. */
+		if (ptid)
 			*ptid = tid;
-	} else if (journal->j_committing_transaction && ptid) {
+		ret = 1;
+	} else if (journal->j_committing_transaction) {
 		/*
 		 * If ext3_write_super() recently started a commit, then we
 		 * have to wait for completion of that transaction
 		 */
-		*ptid = journal->j_committing_transaction->t_tid;
+		if (ptid)
+			*ptid = journal->j_committing_transaction->t_tid;
 		ret = 1;
 	}
 	spin_unlock(&journal->j_state_lock);
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 46b4e347ed7..28ce21d8598 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -2129,26 +2129,46 @@ done:
 }
 
 /*
- * This function must be called when inode is journaled in ordered mode
- * before truncation happens. It starts writeout of truncated part in
- * case it is in the committing transaction so that we stand to ordered
- * mode consistency guarantees.
+ * File truncate and transaction commit interact with each other in a
+ * non-trivial way.  If a transaction writing data block A is
+ * committing, we cannot discard the data by truncate until we have
+ * written them.  Otherwise if we crashed after the transaction with
+ * write has committed but before the transaction with truncate has
+ * committed, we could see stale data in block A.  This function is a
+ * helper to solve this problem.  It starts writeout of the truncated
+ * part in case it is in the committing transaction.
+ *
+ * Filesystem code must call this function when inode is journaled in
+ * ordered mode before truncation happens and after the inode has been
+ * placed on orphan list with the new inode size. The second condition
+ * avoids the race that someone writes new data and we start
+ * committing the transaction after this function has been called but
+ * before a transaction for truncate is started (and furthermore it
+ * allows us to optimize the case where the addition to orphan list
+ * happens in the same transaction as write --- we don't have to write
+ * any data in such case).
  */
-int jbd2_journal_begin_ordered_truncate(struct jbd2_inode *inode,
+int jbd2_journal_begin_ordered_truncate(journal_t *journal,
+					struct jbd2_inode *jinode,
 					loff_t new_size)
 {
-	journal_t *journal;
-	transaction_t *commit_trans;
+	transaction_t *inode_trans, *commit_trans;
 	int ret = 0;
 
-	if (!inode->i_transaction && !inode->i_next_transaction)
+	/* This is a quick check to avoid locking if not necessary */
+	if (!jinode->i_transaction)
 		goto out;
-	journal = inode->i_transaction->t_journal;
+	/* Locks are here just to force reading of recent values, it is
+	 * enough that the transaction was not committing before we started
+	 * a transaction adding the inode to orphan list */
 	spin_lock(&journal->j_state_lock);
 	commit_trans = journal->j_committing_transaction;
 	spin_unlock(&journal->j_state_lock);
-	if (inode->i_transaction == commit_trans) {
-		ret = filemap_fdatawrite_range(inode->i_vfs_inode->i_mapping,
+	spin_lock(&journal->j_list_lock);
+	inode_trans = jinode->i_transaction;
+	spin_unlock(&journal->j_list_lock);
+	if (inode_trans == commit_trans) {
+		ret = filemap_fdatawrite_range(jinode->i_vfs_inode->i_mapping,
 			new_size, LLONG_MAX);
 		if (ret)
 			jbd2_journal_abort(journal, ret);
diff --git a/fs/jffs2/background.c b/fs/jffs2/background.c
index 3cceef4ad2b..e9580104b6b 100644
--- a/fs/jffs2/background.c
+++ b/fs/jffs2/background.c
@@ -95,13 +95,17 @@ static int jffs2_garbage_collect_thread(void *_c)
 			spin_unlock(&c->erase_completion_lock);
 			
 
-		/* This thread is purely an optimisation. But if it runs when
-		   other things could be running, it actually makes things a
-		   lot worse. Use yield() and put it at the back of the runqueue
-		   every time. Especially during boot, pulling an inode in
-		   with read_inode() is much preferable to having the GC thread
-		   get there first. */
-		yield();
+		/* Problem - immediately after bootup, the GCD spends a lot
+		 * of time in places like jffs2_kill_fragtree(); so much so
+		 * that userspace processes (like gdm and X) are starved
+		 * despite plenty of cond_resched()s and renicing.  Yield()
+		 * doesn't help, either (presumably because userspace and GCD
+		 * are generally competing for a higher latency resource -
+		 * disk).
+		 * This forces the GCD to slow the hell down.   Pulling an
+		 * inode in with read_inode() is much preferable to having
+		 * the GC thread get there first. */
+		schedule_timeout_interruptible(msecs_to_jiffies(50));
 
 		/* Put_super will send a SIGKILL and then wait on the sem.
 		 */
diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c
index 6ca08ad887c..1fc1e92356e 100644
--- a/fs/jffs2/readinode.c
+++ b/fs/jffs2/readinode.c
@@ -220,7 +220,7 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 				struct jffs2_tmp_dnode_info *tn)
 {
 	uint32_t fn_end = tn->fn->ofs + tn->fn->size;
-	struct jffs2_tmp_dnode_info *this;
+	struct jffs2_tmp_dnode_info *this, *ptn;
 
 	dbg_readinode("insert fragment %#04x-%#04x, ver %u at %08x\n", tn->fn->ofs, fn_end, tn->version, ref_offset(tn->fn->raw));
 
@@ -251,11 +251,18 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 	if (this) {
 		/* If the node is coincident with another at a lower address,
 		   back up until the other node is found. It may be relevant */
-		while (this->overlapped)
-			this = tn_prev(this);
-
-		/* First node should never be marked overlapped */
-		BUG_ON(!this);
+		while (this->overlapped) {
+			ptn = tn_prev(this);
+			if (!ptn) {
+				/*
+				 * We killed a node which set the overlapped
+				 * flags during the scan. Fix it up.
+				 */
+				this->overlapped = 0;
+				break;
+			}
+			this = ptn;
+		}
 		dbg_readinode("'this' found %#04x-%#04x (%s)\n", this->fn->ofs, this->fn->ofs + this->fn->size, this->fn ? "data" : "hole");
 	}
 
@@ -360,7 +367,17 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 			}
 			if (!this->overlapped)
 				break;
-			this = tn_prev(this);
+
+			ptn = tn_prev(this);
+			if (!ptn) {
+				/*
+				 * We killed a node which set the overlapped
+				 * flags during the scan. Fix it up.
+				 */
+				this->overlapped = 0;
+				break;
+			}
+			this = ptn;
 		}
 	}
 
@@ -456,8 +473,15 @@ static int jffs2_build_inode_fragtree(struct jffs2_sb_info *c,
 		eat_last(&rii->tn_root, &last->rb);
 		ver_insert(&ver_root, last);
 
-		if (unlikely(last->overlapped))
-			continue;
+		if (unlikely(last->overlapped)) {
+			if (pen)
+				continue;
+			/*
+			 * We killed a node which set the overlapped
+			 * flags during the scan. Fix it up.
+			 */
+			last->overlapped = 0;
+		}
 
 		/* Now we have a bunch of nodes in reverse version
 		   order, in the tree at ver_root. Most of the time,
diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index 1f3b0fc0d35..aedc47a264c 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -139,6 +139,55 @@ int nlmclnt_block(struct nlm_wait *block, struct nlm_rqst *req, long timeout)
 	return 0;
 }
 
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+static const struct in6_addr *nlmclnt_map_v4addr(const struct sockaddr *sap,
+						 struct in6_addr *addr_mapped)
+{
+	const struct sockaddr_in *sin = (const struct sockaddr_in *)sap;
+
+	switch (sap->sa_family) {
+	case AF_INET6:
+		return &((const struct sockaddr_in6 *)sap)->sin6_addr;
+	case AF_INET:
+		ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, addr_mapped);
+		return addr_mapped;
+	}
+
+	return NULL;
+}
+
+/*
+ * If lockd is using a PF_INET6 listener, all incoming requests appear
+ * to come from AF_INET6 remotes.  The address of AF_INET remotes are
+ * mapped to AF_INET6 automatically by the network layer.  In case the
+ * user passed an AF_INET server address at mount time, ensure both
+ * addresses are AF_INET6 before comparing them.
+ */
+static int nlmclnt_cmp_addr(const struct nlm_host *host,
+			    const struct sockaddr *sap)
+{
+	const struct in6_addr *addr1;
+	const struct in6_addr *addr2;
+	struct in6_addr addr1_mapped;
+	struct in6_addr addr2_mapped;
+
+	addr1 = nlmclnt_map_v4addr(nlm_addr(host), &addr1_mapped);
+	if (likely(addr1 != NULL)) {
+		addr2 = nlmclnt_map_v4addr(sap, &addr2_mapped);
+		if (likely(addr2 != NULL))
+			return ipv6_addr_equal(addr1, addr2);
+	}
+
+	return 0;
+}
+#else	/* !(CONFIG_IPV6 || CONFIG_IPV6_MODULE) */
+static int nlmclnt_cmp_addr(const struct nlm_host *host,
+			    const struct sockaddr *sap)
+{
+	return nlm_cmp_addr(nlm_addr(host), sap);
+}
+#endif	/* !(CONFIG_IPV6 || CONFIG_IPV6_MODULE) */
+
 /*
  * The server lockd has called us back to tell us the lock was granted
  */
@@ -166,7 +215,7 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 		 */
 		if (fl_blocked->fl_u.nfs_fl.owner->pid != lock->svid)
 			continue;
-		if (!nlm_cmp_addr(nlm_addr(block->b_host), addr))
+		if (!nlmclnt_cmp_addr(block->b_host, addr))
 			continue;
 		if (nfs_compare_fh(NFS_FH(fl_blocked->fl_file->f_path.dentry->d_inode) ,fh) != 0)
 			continue;
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 6063a8e4b9f..763b78a6e9d 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -427,7 +427,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 			goto out;
 		case -EAGAIN:
 			ret = nlm_lck_denied;
-			goto out;
+			break;
 		case FILE_LOCK_DEFERRED:
 			if (wait)
 				break;
@@ -443,6 +443,10 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 			goto out;
 	}
 
+	ret = nlm_lck_denied;
+	if (!wait)
+		goto out;
+
 	ret = nlm_lck_blocked;
 
 	/* Append to list of blocked */
diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index d1d1eb84679..618865b3128 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -3,7 +3,7 @@
  *
  *  Copyright (C) 1991, 1992  Linus Torvalds
  *
- *  Copyright (C) 1996  Gertjan van Wingerde    (gertjan@cs.vu.nl)
+ *  Copyright (C) 1996  Gertjan van Wingerde
  *	Minix V2 fs support.
  *
  *  Modified for 680x0 by Andreas Schwab
diff --git a/fs/namespace.c b/fs/namespace.c
index 228d8c4bfd1..06f8e63f6cb 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -614,9 +614,11 @@ static inline void __mntput(struct vfsmount *mnt)
 	 */
 	for_each_possible_cpu(cpu) {
 		struct mnt_writer *cpu_writer = &per_cpu(mnt_writers, cpu);
-		if (cpu_writer->mnt != mnt)
-			continue;
 		spin_lock(&cpu_writer->lock);
+		if (cpu_writer->mnt != mnt) {
+			spin_unlock(&cpu_writer->lock);
+			continue;
+		}
 		atomic_add(cpu_writer->count, &mnt->__mnt_writers);
 		cpu_writer->count = 0;
 		/*
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 9b728f3565a..574158ae239 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -255,6 +255,32 @@ static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
 	}
 	return 0;
 }
+
+/*
+ * Test if two ip6 socket addresses refer to the same socket by
+ * comparing relevant fields. The padding bytes specifically, are not
+ * compared. sin6_flowinfo is not compared because it only affects QoS
+ * and sin6_scope_id is only compared if the address is "link local"
+ * because "link local" addresses need only be unique to a specific
+ * link. Conversely, ordinary unicast addresses might have different
+ * sin6_scope_id.
+ *
+ * The caller should ensure both socket addresses are AF_INET6.
+ */
+static int nfs_sockaddr_cmp_ip6(const struct sockaddr *sa1,
+				const struct sockaddr *sa2)
+{
+	const struct sockaddr_in6 *saddr1 = (const struct sockaddr_in6 *)sa1;
+	const struct sockaddr_in6 *saddr2 = (const struct sockaddr_in6 *)sa2;
+
+	if (!ipv6_addr_equal(&saddr1->sin6_addr,
+			     &saddr1->sin6_addr))
+		return 0;
+	if (ipv6_addr_scope(&saddr1->sin6_addr) == IPV6_ADDR_SCOPE_LINKLOCAL &&
+	    saddr1->sin6_scope_id != saddr2->sin6_scope_id)
+		return 0;
+	return saddr1->sin6_port == saddr2->sin6_port;
+}
 #else
 static int nfs_sockaddr_match_ipaddr4(const struct sockaddr_in *sa1,
 				 const struct sockaddr_in *sa2)
@@ -270,9 +296,52 @@ static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
 	return nfs_sockaddr_match_ipaddr4((const struct sockaddr_in *)sa1,
 			(const struct sockaddr_in *)sa2);
 }
+
+static int nfs_sockaddr_cmp_ip6(const struct sockaddr * sa1,
+				const struct sockaddr * sa2)
+{
+	return 0;
+}
 #endif
 
 /*
+ * Test if two ip4 socket addresses refer to the same socket, by
+ * comparing relevant fields. The padding bytes specifically, are
+ * not compared.
+ *
+ * The caller should ensure both socket addresses are AF_INET.
+ */
+static int nfs_sockaddr_cmp_ip4(const struct sockaddr *sa1,
+				const struct sockaddr *sa2)
+{
+	const struct sockaddr_in *saddr1 = (const struct sockaddr_in *)sa1;
+	const struct sockaddr_in *saddr2 = (const struct sockaddr_in *)sa2;
+
+	if (saddr1->sin_addr.s_addr != saddr2->sin_addr.s_addr)
+		return 0;
+	return saddr1->sin_port == saddr2->sin_port;
+}
+
+/*
+ * Test if two socket addresses represent the same actual socket,
+ * by comparing (only) relevant fields.
+ */
+static int nfs_sockaddr_cmp(const struct sockaddr *sa1,
+			    const struct sockaddr *sa2)
+{
+	if (sa1->sa_family != sa2->sa_family)
+		return 0;
+
+	switch (sa1->sa_family) {
+	case AF_INET:
+		return nfs_sockaddr_cmp_ip4(sa1, sa2);
+	case AF_INET6:
+		return nfs_sockaddr_cmp_ip6(sa1, sa2);
+	}
+	return 0;
+}
+
+/*
  * Find a client by IP address and protocol version
  * - returns NULL if no such client
  */
@@ -344,8 +413,10 @@ struct nfs_client *nfs_find_client_next(struct nfs_client *clp)
 static struct nfs_client *nfs_match_client(const struct nfs_client_initdata *data)
 {
 	struct nfs_client *clp;
+	const struct sockaddr *sap = data->addr;
 
 	list_for_each_entry(clp, &nfs_client_list, cl_share_link) {
+	        const struct sockaddr *clap = (struct sockaddr *)&clp->cl_addr;
 		/* Don't match clients that failed to initialise properly */
 		if (clp->cl_cons_state < 0)
 			continue;
@@ -358,7 +429,7 @@ static struct nfs_client *nfs_match_client(const struct nfs_client_initdata *dat
 			continue;
 
 		/* Match the full socket address */
-		if (memcmp(&clp->cl_addr, data->addr, sizeof(clp->cl_addr)) != 0)
+		if (!nfs_sockaddr_cmp(sap, clap))
 			continue;
 
 		atomic_inc(&clp->cl_count);
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index e35c8199f82..672368f865c 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1892,8 +1892,14 @@ static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask)
 	cache.cred = cred;
 	cache.jiffies = jiffies;
 	status = NFS_PROTO(inode)->access(inode, &cache);
-	if (status != 0)
+	if (status != 0) {
+		if (status == -ESTALE) {
+			nfs_zap_caches(inode);
+			if (!S_ISDIR(inode->i_mode))
+				set_bit(NFS_INO_STALE, &NFS_I(inode)->flags);
+		}
 		return status;
+	}
 	nfs_access_add_cache(inode, &cache);
 out:
 	if ((mask & ~cache.mask & (MAY_READ | MAY_WRITE | MAY_EXEC)) == 0)
diff --git a/fs/nfs/nfs3acl.c b/fs/nfs/nfs3acl.c
index cef62557c87..6bbf0e6daad 100644
--- a/fs/nfs/nfs3acl.c
+++ b/fs/nfs/nfs3acl.c
@@ -292,7 +292,7 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 {
 	struct nfs_server *server = NFS_SERVER(inode);
 	struct nfs_fattr fattr;
-	struct page *pages[NFSACL_MAXPAGES] = { };
+	struct page *pages[NFSACL_MAXPAGES];
 	struct nfs3_setaclargs args = {
 		.inode = inode,
 		.mask = NFS_ACL,
@@ -303,7 +303,7 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 		.rpc_argp	= &args,
 		.rpc_resp	= &fattr,
 	};
-	int status, count;
+	int status;
 
 	status = -EOPNOTSUPP;
 	if (!nfs_server_capable(inode, NFS_CAP_ACLS))
@@ -319,6 +319,20 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 	if (S_ISDIR(inode->i_mode)) {
 		args.mask |= NFS_DFACL;
 		args.acl_default = dfacl;
+		args.len = nfsacl_size(acl, dfacl);
+	} else
+		args.len = nfsacl_size(acl, NULL);
+
+	if (args.len > NFS_ACL_INLINE_BUFSIZE) {
+		unsigned int npages = 1 + ((args.len - 1) >> PAGE_SHIFT);
+
+		status = -ENOMEM;
+		do {
+			args.pages[args.npages] = alloc_page(GFP_KERNEL);
+			if (args.pages[args.npages] == NULL)
+				goto out_freepages;
+			args.npages++;
+		} while (args.npages < npages);
 	}
 
 	dprintk("NFS call setacl\n");
@@ -329,10 +343,6 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 	nfs_zap_acl_cache(inode);
 	dprintk("NFS reply setacl: %d\n", status);
 
-	/* pages may have been allocated at the xdr layer. */
-	for (count = 0; count < NFSACL_MAXPAGES && args.pages[count]; count++)
-		__free_page(args.pages[count]);
-
 	switch (status) {
 		case 0:
 			status = nfs_refresh_inode(inode, &fattr);
@@ -346,6 +356,11 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 		case -ENOTSUPP:
 			status = -EOPNOTSUPP;
 	}
+out_freepages:
+	while (args.npages != 0) {
+		args.npages--;
+		__free_page(args.pages[args.npages]);
+	}
 out:
 	return status;
 }
diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
index 11cdddec143..6cdeacffde4 100644
--- a/fs/nfs/nfs3xdr.c
+++ b/fs/nfs/nfs3xdr.c
@@ -82,8 +82,10 @@
 #define NFS3_commitres_sz	(1+NFS3_wcc_data_sz+2)
 
 #define ACL3_getaclargs_sz	(NFS3_fh_sz+1)
-#define ACL3_setaclargs_sz	(NFS3_fh_sz+1+2*(2+5*3))
-#define ACL3_getaclres_sz	(1+NFS3_post_op_attr_sz+1+2*(2+5*3))
+#define ACL3_setaclargs_sz	(NFS3_fh_sz+1+ \
+				XDR_QUADLEN(NFS_ACL_INLINE_BUFSIZE))
+#define ACL3_getaclres_sz	(1+NFS3_post_op_attr_sz+1+ \
+				XDR_QUADLEN(NFS_ACL_INLINE_BUFSIZE))
 #define ACL3_setaclres_sz	(1+NFS3_post_op_attr_sz)
 
 /*
@@ -703,28 +705,18 @@ nfs3_xdr_setaclargs(struct rpc_rqst *req, __be32 *p,
                    struct nfs3_setaclargs *args)
 {
 	struct xdr_buf *buf = &req->rq_snd_buf;
-	unsigned int base, len_in_head, len = nfsacl_size(
-		(args->mask & NFS_ACL)   ? args->acl_access  : NULL,
-		(args->mask & NFS_DFACL) ? args->acl_default : NULL);
-	int count, err;
+	unsigned int base;
+	int err;
 
 	p = xdr_encode_fhandle(p, NFS_FH(args->inode));
 	*p++ = htonl(args->mask);
-	base = (char *)p - (char *)buf->head->iov_base;
-	/* put as much of the acls into head as possible. */
-	len_in_head = min_t(unsigned int, buf->head->iov_len - base, len);
-	len -= len_in_head;
-	req->rq_slen = xdr_adjust_iovec(req->rq_svec, p + (len_in_head >> 2));
-
-	for (count = 0; (count << PAGE_SHIFT) < len; count++) {
-		args->pages[count] = alloc_page(GFP_KERNEL);
-		if (!args->pages[count]) {
-			while (count)
-				__free_page(args->pages[--count]);
-			return -ENOMEM;
-		}
-	}
-	xdr_encode_pages(buf, args->pages, 0, len);
+	req->rq_slen = xdr_adjust_iovec(req->rq_svec, p);
+	base = req->rq_slen;
+
+	if (args->npages != 0)
+		xdr_encode_pages(buf, args->pages, 0, args->len);
+	else
+		req->rq_slen += args->len;
 
 	err = nfsacl_encode(buf, base, args->inode,
 			    (args->mask & NFS_ACL) ?
diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c
index 30befc39b3c..2a2a0a7143a 100644
--- a/fs/nfs/nfs4namespace.c
+++ b/fs/nfs/nfs4namespace.c
@@ -21,7 +21,9 @@
 #define NFSDBG_FACILITY		NFSDBG_VFS
 
 /*
- * Check if fs_root is valid
+ * Convert the NFSv4 pathname components into a standard posix path.
+ *
+ * Note that the resulting string will be placed at the end of the buffer
  */
 static inline char *nfs4_pathname_string(const struct nfs4_pathname *pathname,
 					 char *buffer, ssize_t buflen)
@@ -99,21 +101,20 @@ static struct vfsmount *try_location(struct nfs_clone_mount *mountdata,
 {
 	struct vfsmount *mnt = ERR_PTR(-ENOENT);
 	char *mnt_path;
-	int page2len;
+	unsigned int maxbuflen;
 	unsigned int s;
 
 	mnt_path = nfs4_pathname_string(&location->rootpath, page2, PAGE_SIZE);
 	if (IS_ERR(mnt_path))
 		return mnt;
 	mountdata->mnt_path = mnt_path;
-	page2 += strlen(mnt_path) + 1;
-	page2len = PAGE_SIZE - strlen(mnt_path) - 1;
+	maxbuflen = mnt_path - 1 - page2;
 
 	for (s = 0; s < location->nservers; s++) {
 		const struct nfs4_string *buf = &location->servers[s];
 		struct sockaddr_storage addr;
 
-		if (buf->len <= 0 || buf->len >= PAGE_SIZE)
+		if (buf->len <= 0 || buf->len >= maxbuflen)
 			continue;
 
 		mountdata->addr = (struct sockaddr *)&addr;
@@ -126,8 +127,8 @@ static struct vfsmount *try_location(struct nfs_clone_mount *mountdata,
 			continue;
 		nfs_set_port(mountdata->addr, NFS_PORT);
 
-		strncpy(page2, buf->data, page2len);
-		page2[page2len] = '\0';
+		memcpy(page2, buf->data, buf->len);
+		page2[buf->len] = '\0';
 		mountdata->hostname = page2;
 
 		snprintf(page, PAGE_SIZE, "%s:%s",
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index f65953be39c..9250067943d 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -2596,6 +2596,7 @@ static nfsd4_enc nfsd4_enc_ops[] = {
 	[OP_LOOKUPP]		= (nfsd4_enc)nfsd4_encode_noop,
 	[OP_NVERIFY]		= (nfsd4_enc)nfsd4_encode_noop,
 	[OP_OPEN]		= (nfsd4_enc)nfsd4_encode_open,
+	[OP_OPENATTR]		= (nfsd4_enc)nfsd4_encode_noop,
 	[OP_OPEN_CONFIRM]	= (nfsd4_enc)nfsd4_encode_open_confirm,
 	[OP_OPEN_DOWNGRADE]	= (nfsd4_enc)nfsd4_encode_open_downgrade,
 	[OP_PUTFH]		= (nfsd4_enc)nfsd4_encode_noop,
diff --git a/fs/notify/inotify/inotify.c b/fs/notify/inotify/inotify.c
index dae3f28f30d..331f2e88e28 100644
--- a/fs/notify/inotify/inotify.c
+++ b/fs/notify/inotify/inotify.c
@@ -156,7 +156,7 @@ static int inotify_handle_get_wd(struct inotify_handle *ih,
 	int ret;
 
 	do {
-		if (unlikely(!idr_pre_get(&ih->idr, GFP_KERNEL)))
+		if (unlikely(!idr_pre_get(&ih->idr, GFP_NOFS)))
 			return -ENOSPC;
 		ret = idr_get_new_above(&ih->idr, watch, ih->last_wd+1, &watch->wd);
 	} while (ret == -EAGAIN);
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 60fe74035db..19e3a96aa02 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -176,7 +176,8 @@ static int ocfs2_dinode_insert_check(struct inode *inode,
 
 	BUG_ON(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL);
 	mlog_bug_on_msg(!ocfs2_sparse_alloc(osb) &&
-			(OCFS2_I(inode)->ip_clusters != rec->e_cpos),
+			(OCFS2_I(inode)->ip_clusters !=
+			 le32_to_cpu(rec->e_cpos)),
 			"Device %s, asking for sparse allocation: inode %llu, "
 			"cpos %u, clusters %u\n",
 			osb->dev_str,
@@ -4796,6 +4797,29 @@ out:
 	return ret;
 }
 
+static int ocfs2_replace_extent_rec(struct inode *inode,
+				    handle_t *handle,
+				    struct ocfs2_path *path,
+				    struct ocfs2_extent_list *el,
+				    int split_index,
+				    struct ocfs2_extent_rec *split_rec)
+{
+	int ret;
+
+	ret = ocfs2_path_bh_journal_access(handle, inode, path,
+					   path_num_items(path) - 1);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	el->l_recs[split_index] = *split_rec;
+
+	ocfs2_journal_dirty(handle, path_leaf_bh(path));
+out:
+	return ret;
+}
+
 /*
  * Mark part or all of the extent record at split_index in the leaf
  * pointed to by path as written. This removes the unwritten
@@ -4885,7 +4909,9 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 
 	if (ctxt.c_contig_type == CONTIG_NONE) {
 		if (ctxt.c_split_covers_rec)
-			el->l_recs[split_index] = *split_rec;
+			ret = ocfs2_replace_extent_rec(inode, handle,
+						       path, el,
+						       split_index, split_rec);
 		else
 			ret = ocfs2_split_and_insert(inode, handle, path, et,
 						     &last_eb_bh, split_index,
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index a067a6cffb0..8e1709a679b 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -227,7 +227,7 @@ int ocfs2_read_inline_data(struct inode *inode, struct page *page,
 	size = i_size_read(inode);
 
 	if (size > PAGE_CACHE_SIZE ||
-	    size > ocfs2_max_inline_data(inode->i_sb)) {
+	    size > ocfs2_max_inline_data_with_xattr(inode->i_sb, di)) {
 		ocfs2_error(inode->i_sb,
 			    "Inode %llu has with inline data has bad size: %Lu",
 			    (unsigned long long)OCFS2_I(inode)->ip_blkno,
@@ -1555,6 +1555,7 @@ static int ocfs2_try_to_write_inline_data(struct address_space *mapping,
 	int ret, written = 0;
 	loff_t end = pos + len;
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_dinode *di = NULL;
 
 	mlog(0, "Inode %llu, write of %u bytes at off %llu. features: 0x%x\n",
 	     (unsigned long long)oi->ip_blkno, len, (unsigned long long)pos,
@@ -1587,7 +1588,9 @@ static int ocfs2_try_to_write_inline_data(struct address_space *mapping,
 	/*
 	 * Check whether the write can fit.
 	 */
-	if (mmap_page || end > ocfs2_max_inline_data(inode->i_sb))
+	di = (struct ocfs2_dinode *)wc->w_di_bh->b_data;
+	if (mmap_page ||
+	    end > ocfs2_max_inline_data_with_xattr(inode->i_sb, di))
 		return 0;
 
 do_inline_write:
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 54e182a27ca..0a281394785 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -1849,12 +1849,12 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data,
 		if (!mle) {
 			if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN &&
 			    res->owner != assert->node_idx) {
-				mlog(ML_ERROR, "assert_master from "
-					  "%u, but current owner is "
-					  "%u! (%.*s)\n",
-				       assert->node_idx, res->owner,
-				       namelen, name);
-				goto kill;
+				mlog(ML_ERROR, "DIE! Mastery assert from %u, "
+				     "but current owner is %u! (%.*s)\n",
+				     assert->node_idx, res->owner, namelen,
+				     name);
+				__dlm_print_one_lock_resource(res);
+				BUG();
 			}
 		} else if (mle->type != DLM_MLE_MIGRATION) {
 			if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN) {
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index d1295203029..4060bb328bc 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -181,8 +181,7 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm,
 
 		spin_lock(&res->spinlock);
 		/* This ensures that clear refmap is sent after the set */
-		__dlm_wait_on_lockres_flags(res, (DLM_LOCK_RES_SETREF_INPROG |
-						  DLM_LOCK_RES_MIGRATING));
+		__dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG);
 		spin_unlock(&res->spinlock);
 
 		/* clear our bit from the master's refmap, ignore errors */
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index 86ca085ef32..fcf879ed693 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -117,11 +117,11 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm,
 	else
 		BUG_ON(res->owner == dlm->node_num);
 
-	spin_lock(&dlm->spinlock);
+	spin_lock(&dlm->ast_lock);
 	/* We want to be sure that we're not freeing a lock
 	 * that still has AST's pending... */
 	in_use = !list_empty(&lock->ast_list);
-	spin_unlock(&dlm->spinlock);
+	spin_unlock(&dlm->ast_lock);
 	if (in_use) {
 	       mlog(ML_ERROR, "lockres %.*s: Someone is calling dlmunlock "
 		    "while waiting for an ast!", res->lockname.len,
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 206a2370876..7219a86d34c 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -320,9 +320,14 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
 					struct ocfs2_lock_res *lockres);
 static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
 						int convert);
-#define ocfs2_log_dlm_error(_func, _err, _lockres) do {			\
-	mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \
-	     _err, _func, _lockres->l_name);				\
+#define ocfs2_log_dlm_error(_func, _err, _lockres) do {					\
+	if ((_lockres)->l_type != OCFS2_LOCK_TYPE_DENTRY)				\
+		mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n",	\
+		     _err, _func, _lockres->l_name);					\
+	else										\
+		mlog(ML_ERROR, "DLM error %d while calling %s on resource %.*s%08x\n",	\
+		     _err, _func, OCFS2_DENTRY_LOCK_INO_START - 1, (_lockres)->l_name,	\
+		     (unsigned int)ocfs2_get_dentry_lock_ino(_lockres));		\
 } while (0)
 static int ocfs2_downconvert_thread(void *arg);
 static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index 3c3532e1307..172850a9a12 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)
 static inline int ocfs2_begin_ordered_truncate(struct inode *inode,
 					       loff_t new_size)
 {
-	return jbd2_journal_begin_ordered_truncate(&OCFS2_I(inode)->ip_jinode,
-						   new_size);
+	return jbd2_journal_begin_ordered_truncate(
+				OCFS2_SB(inode->i_sb)->journal->j_journal,
+				&OCFS2_I(inode)->ip_jinode,
+				new_size);
 }
 
 #endif /* OCFS2_JOURNAL_H */
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index 084aba86c3b..4b11762f249 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -532,7 +532,8 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 
 		fe->i_dyn_features = cpu_to_le16(feat | OCFS2_INLINE_DATA_FL);
 
-		fe->id2.i_data.id_count = cpu_to_le16(ocfs2_max_inline_data(osb->sb));
+		fe->id2.i_data.id_count = cpu_to_le16(
+				ocfs2_max_inline_data_with_xattr(osb->sb, fe));
 	} else {
 		fel = &fe->id2.i_list;
 		fel->l_tree_depth = 0;
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 077384135f4..946d3c34b90 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -341,6 +341,9 @@ struct ocfs2_super
 	struct ocfs2_node_map		osb_recovering_orphan_dirs;
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
+
+	/* used to protect metaecc calculation check of xattr. */
+	spinlock_t osb_xattr_lock;
 };
 
 #define OCFS2_SB(sb)	    ((struct ocfs2_super *)(sb)->s_fs_info)
diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h
index c7ae45aaa36..2332ef740f4 100644
--- a/fs/ocfs2/ocfs2_fs.h
+++ b/fs/ocfs2/ocfs2_fs.h
@@ -1070,12 +1070,6 @@ static inline int ocfs2_fast_symlink_chars(struct super_block *sb)
 		 offsetof(struct ocfs2_dinode, id2.i_symlink);
 }
 
-static inline int ocfs2_max_inline_data(struct super_block *sb)
-{
-	return sb->s_blocksize -
-		offsetof(struct ocfs2_dinode, id2.i_data.id_data);
-}
-
 static inline int ocfs2_max_inline_data_with_xattr(struct super_block *sb,
 						   struct ocfs2_dinode *di)
 {
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index b1cb38fbe80..7ac83a81ee5 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -1537,6 +1537,13 @@ static int ocfs2_get_sector(struct super_block *sb,
 	unlock_buffer(*bh);
 	ll_rw_block(READ, 1, bh);
 	wait_on_buffer(*bh);
+	if (!buffer_uptodate(*bh)) {
+		mlog_errno(-EIO);
+		brelse(*bh);
+		*bh = NULL;
+		return -EIO;
+	}
+
 	return 0;
 }
 
@@ -1747,6 +1754,7 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	INIT_LIST_HEAD(&osb->blocked_lock_list);
 	osb->blocked_lock_count = 0;
 	spin_lock_init(&osb->osb_lock);
+	spin_lock_init(&osb->osb_xattr_lock);
 	ocfs2_init_inode_steal_slot(osb);
 
 	atomic_set(&osb->alloc_stats.moves, 0);
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 915039fffe6..2563df89fc2 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -82,13 +82,14 @@ struct ocfs2_xattr_set_ctxt {
 
 #define OCFS2_XATTR_ROOT_SIZE	(sizeof(struct ocfs2_xattr_def_value_root))
 #define OCFS2_XATTR_INLINE_SIZE	80
+#define OCFS2_XATTR_HEADER_GAP	4
 #define OCFS2_XATTR_FREE_IN_IBODY	(OCFS2_MIN_XATTR_INLINE_SIZE \
 					 - sizeof(struct ocfs2_xattr_header) \
-					 - sizeof(__u32))
+					 - OCFS2_XATTR_HEADER_GAP)
 #define OCFS2_XATTR_FREE_IN_BLOCK(ptr)	((ptr)->i_sb->s_blocksize \
 					 - sizeof(struct ocfs2_xattr_block) \
 					 - sizeof(struct ocfs2_xattr_header) \
-					 - sizeof(__u32))
+					 - OCFS2_XATTR_HEADER_GAP)
 
 static struct ocfs2_xattr_def_value_root def_xv = {
 	.xv.xr_list.l_count = cpu_to_le16(1),
@@ -274,10 +275,12 @@ static int ocfs2_read_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
 			       bucket->bu_blocks, bucket->bu_bhs, 0,
 			       NULL);
 	if (!rc) {
+		spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 		rc = ocfs2_validate_meta_ecc_bhs(bucket->bu_inode->i_sb,
 						 bucket->bu_bhs,
 						 bucket->bu_blocks,
 						 &bucket_xh(bucket)->xh_check);
+		spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 		if (rc)
 			mlog_errno(rc);
 	}
@@ -310,9 +313,11 @@ static void ocfs2_xattr_bucket_journal_dirty(handle_t *handle,
 {
 	int i;
 
+	spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 	ocfs2_compute_meta_ecc_bhs(bucket->bu_inode->i_sb,
 				   bucket->bu_bhs, bucket->bu_blocks,
 				   &bucket_xh(bucket)->xh_check);
+	spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 
 	for (i = 0; i < bucket->bu_blocks; i++)
 		ocfs2_journal_dirty(handle, bucket->bu_bhs[i]);
@@ -542,8 +547,12 @@ int ocfs2_calc_xattr_init(struct inode *dir,
 	 * when blocksize = 512, may reserve one more cluser for
 	 * xattr bucket, otherwise reserve one metadata block
 	 * for them is ok.
+	 * If this is a new directory with inline data,
+	 * we choose to reserve the entire inline area for
+	 * directory contents and force an external xattr block.
 	 */
 	if (dir->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE ||
+	    (S_ISDIR(mode) && ocfs2_supports_inline_data(osb)) ||
 	    (s_size + a_size) > OCFS2_XATTR_FREE_IN_IBODY) {
 		ret = ocfs2_reserve_new_metadata_blocks(osb, 1, xattr_ac);
 		if (ret) {
@@ -1507,7 +1516,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		last += 1;
 	}
 
-	free = min_offs - ((void *)last - xs->base) - sizeof(__u32);
+	free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP;
 	if (free < 0)
 		return -EIO;
 
@@ -2190,7 +2199,7 @@ static int ocfs2_xattr_can_be_in_inode(struct inode *inode,
 		last += 1;
 	}
 
-	free = min_offs - ((void *)last - xs->base) - sizeof(__u32);
+	free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP;
 	if (free < 0)
 		return 0;
 
@@ -2592,8 +2601,9 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
 
 	if (!ret) {
 		/* Update inode ctime. */
-		ret = ocfs2_journal_access(ctxt->handle, inode, xis->inode_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_journal_access_di(ctxt->handle, inode,
+					      xis->inode_bh,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -4785,19 +4795,33 @@ static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
 						char *val,
 						int value_len)
 {
-	int offset;
+	int ret, offset, block_off;
 	struct ocfs2_xattr_value_root *xv;
 	struct ocfs2_xattr_entry *xe = xs->here;
+	struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket);
+	void *base;
 
 	BUG_ON(!xs->base || !xe || ocfs2_xattr_is_local(xe));
 
-	offset = le16_to_cpu(xe->xe_name_offset) +
-		 OCFS2_XATTR_SIZE(xe->xe_name_len);
+	ret = ocfs2_xattr_bucket_get_name_value(inode, xh,
+						xe - xh->xh_entries,
+						&block_off,
+						&offset);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
 
-	xv = (struct ocfs2_xattr_value_root *)(xs->base + offset);
+	base = bucket_block(xs->bucket, block_off);
+	xv = (struct ocfs2_xattr_value_root *)(base + offset +
+		 OCFS2_XATTR_SIZE(xe->xe_name_len));
 
-	return __ocfs2_xattr_set_value_outside(inode, handle,
-					       xv, val, value_len);
+	ret = __ocfs2_xattr_set_value_outside(inode, handle,
+					      xv, val, value_len);
+	if (ret)
+		mlog_errno(ret);
+out:
+	return ret;
 }
 
 static int ocfs2_rm_xattr_cluster(struct inode *inode,
@@ -5060,8 +5084,8 @@ try_again:
 	xh_free_start = le16_to_cpu(xh->xh_free_start);
 	header_size = sizeof(struct ocfs2_xattr_header) +
 			count * sizeof(struct ocfs2_xattr_entry);
-	max_free = OCFS2_XATTR_BUCKET_SIZE -
-		le16_to_cpu(xh->xh_name_value_len) - header_size;
+	max_free = OCFS2_XATTR_BUCKET_SIZE - header_size -
+		le16_to_cpu(xh->xh_name_value_len) - OCFS2_XATTR_HEADER_GAP;
 
 	mlog_bug_on_msg(header_size > blocksize, "bucket %llu has header size "
 			"of %u which exceed block size\n",
@@ -5094,7 +5118,7 @@ try_again:
 			need = 0;
 	}
 
-	free = xh_free_start - header_size;
+	free = xh_free_start - header_size - OCFS2_XATTR_HEADER_GAP;
 	/*
 	 * We need to make sure the new name/value pair
 	 * can exist in the same block.
@@ -5127,7 +5151,8 @@ try_again:
 			}
 
 			xh_free_start = le16_to_cpu(xh->xh_free_start);
-			free = xh_free_start - header_size;
+			free = xh_free_start - header_size
+				- OCFS2_XATTR_HEADER_GAP;
 			if (xh_free_start % blocksize < need)
 				free -= xh_free_start % blocksize;
 
diff --git a/fs/pipe.c b/fs/pipe.c
index 3a48ba5179d..14f502b89cf 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -699,12 +699,12 @@ pipe_rdwr_fasync(int fd, struct file *filp, int on)
 	int retval;
 
 	mutex_lock(&inode->i_mutex);
-
 	retval = fasync_helper(fd, filp, on, &pipe->fasync_readers);
-
-	if (retval >= 0)
+	if (retval >= 0) {
 		retval = fasync_helper(fd, filp, on, &pipe->fasync_writers);
-
+		if (retval < 0) /* this can happen only if on == T */
+			fasync_helper(-1, filp, 0, &pipe->fasync_readers);
+	}
 	mutex_unlock(&inode->i_mutex);
 
 	if (retval < 0)
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 0c9de19a163..beaa0ce3b82 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3066,7 +3066,6 @@ static int proc_task_readdir(struct file * filp, void * dirent, filldir_t filldi
 	int retval = -ENOENT;
 	ino_t ino;
 	int tid;
-	unsigned long pos = filp->f_pos;  /* avoiding "long long" filp->f_pos */
 	struct pid_namespace *ns;
 
 	task = get_proc_task(inode);
@@ -3083,18 +3082,18 @@ static int proc_task_readdir(struct file * filp, void * dirent, filldir_t filldi
 		goto out_no_task;
 	retval = 0;
 
-	switch (pos) {
+	switch ((unsigned long)filp->f_pos) {
 	case 0:
 		ino = inode->i_ino;
-		if (filldir(dirent, ".", 1, pos, ino, DT_DIR) < 0)
+		if (filldir(dirent, ".", 1, filp->f_pos, ino, DT_DIR) < 0)
 			goto out;
-		pos++;
+		filp->f_pos++;
 		/* fall through */
 	case 1:
 		ino = parent_ino(dentry);
-		if (filldir(dirent, "..", 2, pos, ino, DT_DIR) < 0)
+		if (filldir(dirent, "..", 2, filp->f_pos, ino, DT_DIR) < 0)
 			goto out;
-		pos++;
+		filp->f_pos++;
 		/* fall through */
 	}
 
@@ -3104,9 +3103,9 @@ static int proc_task_readdir(struct file * filp, void * dirent, filldir_t filldi
 	ns = filp->f_dentry->d_sb->s_fs_info;
 	tid = (int)filp->f_version;
 	filp->f_version = 0;
-	for (task = first_tid(leader, tid, pos - 2, ns);
+	for (task = first_tid(leader, tid, filp->f_pos - 2, ns);
 	     task;
-	     task = next_tid(task), pos++) {
+	     task = next_tid(task), filp->f_pos++) {
 		tid = task_pid_nr_ns(task, ns);
 		if (proc_task_fill_cache(filp, dirent, filldir, task, tid) < 0) {
 			/* returning this tgid failed, save it as the first
@@ -3117,7 +3116,6 @@ static int proc_task_readdir(struct file * filp, void * dirent, filldir_t filldi
 		}
 	}
 out:
-	filp->f_pos = pos;
 	put_task_struct(leader);
 out_no_task:
 	return retval;
diff --git a/fs/proc/inode.c b/fs/proc/inode.c
index 3e76bb9b3ad..d8bb5c671f4 100644
--- a/fs/proc/inode.c
+++ b/fs/proc/inode.c
@@ -485,8 +485,10 @@ struct inode *proc_get_inode(struct super_block *sb, unsigned int ino,
 			}
 		}
 		unlock_new_inode(inode);
-	} else
+	} else {
 	       module_put(de->owner);
+	       de_put(de);
+	}
 	return inode;
 
 out_ino:
diff --git a/fs/proc/page.c b/fs/proc/page.c
index 767d95a6d1b..e9983837d08 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -80,7 +80,7 @@ static const struct file_operations proc_kpagecount_operations = {
 #define KPF_RECLAIM    9
 #define KPF_BUDDY     10
 
-#define kpf_copy_bit(flags, srcpos, dstpos) (((flags >> srcpos) & 1) << dstpos)
+#define kpf_copy_bit(flags, dstpos, srcpos) (((flags >> srcpos) & 1) << dstpos)
 
 static ssize_t kpageflags_read(struct file *file, char __user *buf,
 			     size_t count, loff_t *ppos)
@@ -107,7 +107,7 @@ static ssize_t kpageflags_read(struct file *file, char __user *buf,
 		else
 			kflags = ppage->flags;
 
-		uflags = kpf_copy_bit(KPF_LOCKED, PG_locked, kflags) |
+		uflags = kpf_copy_bit(kflags, KPF_LOCKED, PG_locked) |
 			kpf_copy_bit(kflags, KPF_ERROR, PG_error) |
 			kpf_copy_bit(kflags, KPF_REFERENCED, PG_referenced) |
 			kpf_copy_bit(kflags, KPF_UPTODATE, PG_uptodate) |
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index b9b567a2837..5d7c7ececa6 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -114,6 +114,9 @@ int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize)
 		if (!pagevec_add(&lru_pvec, page))
 			__pagevec_lru_add_file(&lru_pvec);
 
+		/* prevent the page from being discarded on memory pressure */
+		SetPageDirty(page);
+
 		unlock_page(page);
 	}
 
@@ -126,6 +129,7 @@ int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize)
 	return -EFBIG;
 
  add_error:
+	pagevec_lru_add_file(&lru_pvec);
 	page_cache_release(pages + loop);
 	for (loop++; loop < npages; loop++)
 		__free_page(pages + loop);
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 5267098532b..a1a4cfe1921 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -48,8 +48,16 @@ int seq_open(struct file *file, const struct seq_operations *op)
 	 */
 	file->f_version = 0;
 
-	/* SEQ files support lseek, but not pread/pwrite */
-	file->f_mode &= ~(FMODE_PREAD | FMODE_PWRITE);
+	/*
+	 * seq_files support lseek() and pread().  They do not implement
+	 * write() at all, but we clear FMODE_PWRITE here for historical
+	 * reasons.
+	 *
+	 * If a client of seq_files a) implements file.write() and b) wishes to
+	 * support pwrite() then that client will need to implement its own
+	 * file.open() which calls seq_open() and then sets FMODE_PWRITE.
+	 */
+	file->f_mode &= ~FMODE_PWRITE;
 	return 0;
 }
 EXPORT_SYMBOL(seq_open);
@@ -131,6 +139,22 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
 	int err = 0;
 
 	mutex_lock(&m->lock);
+
+	/* Don't assume *ppos is where we left it */
+	if (unlikely(*ppos != m->read_pos)) {
+		m->read_pos = *ppos;
+		while ((err = traverse(m, *ppos)) == -EAGAIN)
+			;
+		if (err) {
+			/* With prejudice... */
+			m->read_pos = 0;
+			m->version = 0;
+			m->index = 0;
+			m->count = 0;
+			goto Done;
+		}
+	}
+
 	/*
 	 * seq_file->op->..m_start/m_stop/m_next may do special actions
 	 * or optimisations based on the file->f_version, so we want to
@@ -230,8 +254,10 @@ Fill:
 Done:
 	if (!copied)
 		copied = err;
-	else
+	else {
 		*ppos += copied;
+		m->read_pos += copied;
+	}
 	file->f_version = m->version;
 	mutex_unlock(&m->lock);
 	return copied;
@@ -266,16 +292,18 @@ loff_t seq_lseek(struct file *file, loff_t offset, int origin)
 			if (offset < 0)
 				break;
 			retval = offset;
-			if (offset != file->f_pos) {
+			if (offset != m->read_pos) {
 				while ((retval=traverse(m, offset)) == -EAGAIN)
 					;
 				if (retval) {
 					/* with extreme prejudice... */
 					file->f_pos = 0;
+					m->read_pos = 0;
 					m->version = 0;
 					m->index = 0;
 					m->count = 0;
 				} else {
+					m->read_pos = offset;
 					retval = file->f_pos = offset;
 				}
 			}
diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c
index c837dfc2b3c..2a796031034 100644
--- a/fs/squashfs/block.c
+++ b/fs/squashfs/block.c
@@ -80,7 +80,7 @@ static struct buffer_head *get_block_length(struct super_block *sb,
  * generated a larger block - this does occasionally happen with zlib).
  */
 int squashfs_read_data(struct super_block *sb, void **buffer, u64 index,
-			int length, u64 *next_index, int srclength)
+			int length, u64 *next_index, int srclength, int pages)
 {
 	struct squashfs_sb_info *msblk = sb->s_fs_info;
 	struct buffer_head **bh;
@@ -184,7 +184,7 @@ int squashfs_read_data(struct super_block *sb, void **buffer, u64 index,
 				offset = 0;
 			}
 
-			if (msblk->stream.avail_out == 0) {
+			if (msblk->stream.avail_out == 0 && page < pages) {
 				msblk->stream.next_out = buffer[page++];
 				msblk->stream.avail_out = PAGE_CACHE_SIZE;
 			}
@@ -201,25 +201,20 @@ int squashfs_read_data(struct super_block *sb, void **buffer, u64 index,
 				zlib_init = 1;
 			}
 
-			zlib_err = zlib_inflate(&msblk->stream, Z_NO_FLUSH);
+			zlib_err = zlib_inflate(&msblk->stream, Z_SYNC_FLUSH);
 
 			if (msblk->stream.avail_in == 0 && k < b)
 				put_bh(bh[k++]);
 		} while (zlib_err == Z_OK);
 
 		if (zlib_err != Z_STREAM_END) {
-			ERROR("zlib_inflate returned unexpected result"
-				" 0x%x, srclength %d, avail_in %d,"
-				" avail_out %d\n", zlib_err, srclength,
-				msblk->stream.avail_in,
-				msblk->stream.avail_out);
+			ERROR("zlib_inflate error, data probably corrupt\n");
 			goto release_mutex;
 		}
 
 		zlib_err = zlib_inflateEnd(&msblk->stream);
 		if (zlib_err != Z_OK) {
-			ERROR("zlib_inflateEnd returned unexpected result 0x%x,"
-				" srclength %d\n", zlib_err, srclength);
+			ERROR("zlib_inflate error, data probably corrupt\n");
 			goto release_mutex;
 		}
 		length = msblk->stream.total_out;
@@ -268,7 +263,8 @@ block_release:
 		put_bh(bh[k]);
 
 read_failure:
-	ERROR("sb_bread failed reading block 0x%llx\n", cur_index);
+	ERROR("squashfs_read_data failed to read block 0x%llx\n",
+					(unsigned long long) index);
 	kfree(bh);
 	return -EIO;
 }
diff --git a/fs/squashfs/cache.c b/fs/squashfs/cache.c
index f29eda16d25..1c4739e33af 100644
--- a/fs/squashfs/cache.c
+++ b/fs/squashfs/cache.c
@@ -119,7 +119,7 @@ struct squashfs_cache_entry *squashfs_cache_get(struct super_block *sb,
 
 			entry->length = squashfs_read_data(sb, entry->data,
 				block, length, &entry->next_index,
-				cache->block_size);
+				cache->block_size, cache->pages);
 
 			spin_lock(&cache->lock);
 
@@ -406,7 +406,7 @@ int squashfs_read_table(struct super_block *sb, void *buffer, u64 block,
 	for (i = 0; i < pages; i++, buffer += PAGE_CACHE_SIZE)
 		data[i] = buffer;
 	res = squashfs_read_data(sb, data, block, length |
-		SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, length);
+		SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, length, pages);
 	kfree(data);
 	return res;
 }
diff --git a/fs/squashfs/inode.c b/fs/squashfs/inode.c
index 7a63398bb85..9101dbde39e 100644
--- a/fs/squashfs/inode.c
+++ b/fs/squashfs/inode.c
@@ -133,7 +133,8 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 	type = le16_to_cpu(sqshb_ino->inode_type);
 	switch (type) {
 	case SQUASHFS_REG_TYPE: {
-		unsigned int frag_offset, frag_size, frag;
+		unsigned int frag_offset, frag;
+		int frag_size;
 		u64 frag_blk;
 		struct squashfs_reg_inode *sqsh_ino = &squashfs_ino.reg;
 
@@ -175,7 +176,8 @@ int squashfs_read_inode(struct inode *inode, long long ino)
 		break;
 	}
 	case SQUASHFS_LREG_TYPE: {
-		unsigned int frag_offset, frag_size, frag;
+		unsigned int frag_offset, frag;
+		int frag_size;
 		u64 frag_blk;
 		struct squashfs_lreg_inode *sqsh_ino = &squashfs_ino.lreg;
 
diff --git a/fs/squashfs/squashfs.h b/fs/squashfs/squashfs.h
index 6b2515d027d..0e9feb6adf7 100644
--- a/fs/squashfs/squashfs.h
+++ b/fs/squashfs/squashfs.h
@@ -34,7 +34,7 @@ static inline struct squashfs_inode_info *squashfs_i(struct inode *inode)
 
 /* block.c */
 extern int squashfs_read_data(struct super_block *, void **, u64, int, u64 *,
-				int);
+				int, int);
 
 /* cache.c */
 extern struct squashfs_cache *squashfs_cache_init(char *, int, int);
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
index 071df5b5b49..681ec0d8379 100644
--- a/fs/squashfs/super.c
+++ b/fs/squashfs/super.c
@@ -389,7 +389,7 @@ static int __init init_squashfs_fs(void)
 		return err;
 	}
 
-	printk(KERN_INFO "squashfs: version 4.0 (2009/01/03) "
+	printk(KERN_INFO "squashfs: version 4.0 (2009/01/31) "
 		"Phillip Lougher\n");
 
 	return 0;
diff --git a/fs/super.c b/fs/super.c
index 645e5403f2a..6ce501447ad 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -82,7 +82,22 @@ static struct super_block *alloc_super(struct file_system_type *type)
 		 * lock ordering than usbfs:
 		 */
 		lockdep_set_class(&s->s_lock, &type->s_lock_key);
-		down_write(&s->s_umount);
+		/*
+		 * sget() can have s_umount recursion.
+		 *
+		 * When it cannot find a suitable sb, it allocates a new
+		 * one (this one), and tries again to find a suitable old
+		 * one.
+		 *
+		 * In case that succeeds, it will acquire the s_umount
+		 * lock of the old one. Since these are clearly distrinct
+		 * locks, and this object isn't exposed yet, there's no
+		 * risk of deadlocks.
+		 *
+		 * Annotate this by putting this lock in a different
+		 * subclass.
+		 */
+		down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);
 		s->s_count = S_BIAS;
 		atomic_set(&s->s_active, 1);
 		mutex_init(&s->s_vfs_rename_mutex);
@@ -301,7 +316,7 @@ void generic_shutdown_super(struct super_block *sb)
 		/*
 		 * wait for asynchronous fs operations to finish before going further
 		 */
-		async_synchronize_full_special(&sb->s_async_list);
+		async_synchronize_full_domain(&sb->s_async_list);
 
 		/* bad name - it should be evict_inodes() */
 		invalidate_inodes(sb);
@@ -356,8 +371,10 @@ retry:
 				continue;
 			if (!grab_super(old))
 				goto retry;
-			if (s)
+			if (s) {
+				up_write(&s->s_umount);
 				destroy_super(s);
+			}
 			return old;
 		}
 	}
@@ -372,6 +389,7 @@ retry:
 	err = set(s, data);
 	if (err) {
 		spin_unlock(&sb_lock);
+		up_write(&s->s_umount);
 		destroy_super(s);
 		return ERR_PTR(err);
 	}
@@ -470,7 +488,7 @@ restart:
 		sb->s_count++;
 		spin_unlock(&sb_lock);
 		down_read(&sb->s_umount);
-		async_synchronize_full_special(&sb->s_async_list);
+		async_synchronize_full_domain(&sb->s_async_list);
 		if (sb->s_root && (wait || sb->s_dirt))
 			sb->s_op->sync_fs(sb, wait);
 		up_read(&sb->s_umount);
diff --git a/fs/timerfd.c b/fs/timerfd.c
index 6a123b8ff3f..b042bd7034b 100644
--- a/fs/timerfd.c
+++ b/fs/timerfd.c
@@ -186,10 +186,9 @@ SYSCALL_DEFINE2(timerfd_create, int, clockid, int, flags)
 	BUILD_BUG_ON(TFD_CLOEXEC != O_CLOEXEC);
 	BUILD_BUG_ON(TFD_NONBLOCK != O_NONBLOCK);
 
-	if (flags & ~(TFD_CLOEXEC | TFD_NONBLOCK))
-		return -EINVAL;
-	if (clockid != CLOCK_MONOTONIC &&
-	    clockid != CLOCK_REALTIME)
+	if ((flags & ~TFD_CREATE_FLAGS) ||
+	    (clockid != CLOCK_MONOTONIC &&
+	     clockid != CLOCK_REALTIME))
 		return -EINVAL;
 
 	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
@@ -201,7 +200,7 @@ SYSCALL_DEFINE2(timerfd_create, int, clockid, int, flags)
 	hrtimer_init(&ctx->tmr, clockid, HRTIMER_MODE_ABS);
 
 	ufd = anon_inode_getfd("[timerfd]", &timerfd_fops, ctx,
-			       flags & (O_CLOEXEC | O_NONBLOCK));
+			       flags & TFD_SHARED_FCNTL_FLAGS);
 	if (ufd < 0)
 		kfree(ctx);
 
@@ -219,7 +218,8 @@ SYSCALL_DEFINE4(timerfd_settime, int, ufd, int, flags,
 	if (copy_from_user(&ktmr, utmr, sizeof(ktmr)))
 		return -EFAULT;
 
-	if (!timespec_valid(&ktmr.it_value) ||
+	if ((flags & ~TFD_SETTIME_FLAGS) ||
+	    !timespec_valid(&ktmr.it_value) ||
 	    !timespec_valid(&ktmr.it_interval))
 		return -EINVAL;
 
diff --git a/fs/ufs/super.c b/fs/ufs/super.c
index e65212dfb60..261a1c2f22d 100644
--- a/fs/ufs/super.c
+++ b/fs/ufs/super.c
@@ -41,7 +41,7 @@
  * Stefan Reinauer <stepan@home.culture.mipt.ru>
  *
  * Module usage counts added on 96/04/29 by
- * Gertjan van Wingerde <gertjan@cs.vu.nl>
+ * Gertjan van Wingerde <gwingerde@gmail.com>
  *
  * Clean swab support on 19970406 by
  * Francois-Rene Rideau <fare@tunes.org>
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index d71dc44e21e..aa1016bb913 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -34,6 +34,12 @@
 #include <linux/backing-dev.h>
 #include <linux/freezer.h>
 
+#include "xfs_sb.h"
+#include "xfs_inum.h"
+#include "xfs_ag.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
+
 static kmem_zone_t *xfs_buf_zone;
 STATIC int xfsbufd(void *);
 STATIC int xfsbufd_wakeup(int, gfp_t);
@@ -166,6 +172,75 @@ test_page_region(
 }
 
 /*
+ *	Mapping of multi-page buffers into contiguous virtual space
+ */
+
+typedef struct a_list {
+	void		*vm_addr;
+	struct a_list	*next;
+} a_list_t;
+
+static a_list_t		*as_free_head;
+static int		as_list_len;
+static DEFINE_SPINLOCK(as_lock);
+
+/*
+ *	Try to batch vunmaps because they are costly.
+ */
+STATIC void
+free_address(
+	void		*addr)
+{
+	a_list_t	*aentry;
+
+#ifdef CONFIG_XEN
+	/*
+	 * Xen needs to be able to make sure it can get an exclusive
+	 * RO mapping of pages it wants to turn into a pagetable.  If
+	 * a newly allocated page is also still being vmap()ed by xfs,
+	 * it will cause pagetable construction to fail.  This is a
+	 * quick workaround to always eagerly unmap pages so that Xen
+	 * is happy.
+	 */
+	vunmap(addr);
+	return;
+#endif
+
+	aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT);
+	if (likely(aentry)) {
+		spin_lock(&as_lock);
+		aentry->next = as_free_head;
+		aentry->vm_addr = addr;
+		as_free_head = aentry;
+		as_list_len++;
+		spin_unlock(&as_lock);
+	} else {
+		vunmap(addr);
+	}
+}
+
+STATIC void
+purge_addresses(void)
+{
+	a_list_t	*aentry, *old;
+
+	if (as_free_head == NULL)
+		return;
+
+	spin_lock(&as_lock);
+	aentry = as_free_head;
+	as_free_head = NULL;
+	as_list_len = 0;
+	spin_unlock(&as_lock);
+
+	while ((old = aentry) != NULL) {
+		vunmap(aentry->vm_addr);
+		aentry = aentry->next;
+		kfree(old);
+	}
+}
+
+/*
  *	Internal xfs_buf_t object manipulation
  */
 
@@ -264,7 +339,7 @@ xfs_buf_free(
 		uint		i;
 
 		if ((bp->b_flags & XBF_MAPPED) && (bp->b_page_count > 1))
-                       vm_unmap_ram(bp->b_addr - bp->b_offset, bp->b_page_count);
+			free_address(bp->b_addr - bp->b_offset);
 
 		for (i = 0; i < bp->b_page_count; i++) {
 			struct page	*page = bp->b_pages[i];
@@ -386,8 +461,10 @@ _xfs_buf_map_pages(
 		bp->b_addr = page_address(bp->b_pages[0]) + bp->b_offset;
 		bp->b_flags |= XBF_MAPPED;
 	} else if (flags & XBF_MAPPED) {
-               bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count,
-                                       -1, PAGE_KERNEL);
+		if (as_list_len > 64)
+			purge_addresses();
+		bp->b_addr = vmap(bp->b_pages, bp->b_page_count,
+					VM_MAP, PAGE_KERNEL);
 		if (unlikely(bp->b_addr == NULL))
 			return -ENOMEM;
 		bp->b_addr += bp->b_offset;
@@ -1364,10 +1441,12 @@ xfs_unregister_buftarg(
 
 void
 xfs_free_buftarg(
-	xfs_buftarg_t		*btp)
+	struct xfs_mount	*mp,
+	struct xfs_buftarg	*btp)
 {
 	xfs_flush_buftarg(btp, 1);
-	xfs_blkdev_issue_flush(btp);
+	if (mp->m_flags & XFS_MOUNT_BARRIER)
+		xfs_blkdev_issue_flush(btp);
 	xfs_free_bufhash(btp);
 	iput(btp->bt_mapping->host);
 
@@ -1672,6 +1751,8 @@ xfsbufd(
 			count++;
 		}
 
+		if (as_list_len > 0)
+			purge_addresses();
 		if (count)
 			blk_run_address_space(target->bt_mapping);
 
diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h
index 288ae7c4c80..9b4d666ad31 100644
--- a/fs/xfs/linux-2.6/xfs_buf.h
+++ b/fs/xfs/linux-2.6/xfs_buf.h
@@ -413,7 +413,7 @@ static inline int XFS_bwrite(xfs_buf_t *bp)
  *	Handling of buftargs.
  */
 extern xfs_buftarg_t *xfs_alloc_buftarg(struct block_device *, int);
-extern void xfs_free_buftarg(xfs_buftarg_t *);
+extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *);
 extern void xfs_wait_buftarg(xfs_buftarg_t *);
 extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int);
 extern int xfs_flush_buftarg(xfs_buftarg_t *, int);
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index c71e226da7f..32ae5028e96 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -734,15 +734,15 @@ xfs_close_devices(
 {
 	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
 		struct block_device *logdev = mp->m_logdev_targp->bt_bdev;
-		xfs_free_buftarg(mp->m_logdev_targp);
+		xfs_free_buftarg(mp, mp->m_logdev_targp);
 		xfs_blkdev_put(logdev);
 	}
 	if (mp->m_rtdev_targp) {
 		struct block_device *rtdev = mp->m_rtdev_targp->bt_bdev;
-		xfs_free_buftarg(mp->m_rtdev_targp);
+		xfs_free_buftarg(mp, mp->m_rtdev_targp);
 		xfs_blkdev_put(rtdev);
 	}
-	xfs_free_buftarg(mp->m_ddev_targp);
+	xfs_free_buftarg(mp, mp->m_ddev_targp);
 }
 
 /*
@@ -811,9 +811,9 @@ xfs_open_devices(
 
  out_free_rtdev_targ:
 	if (mp->m_rtdev_targp)
-		xfs_free_buftarg(mp->m_rtdev_targp);
+		xfs_free_buftarg(mp, mp->m_rtdev_targp);
  out_free_ddev_targ:
-	xfs_free_buftarg(mp->m_ddev_targp);
+	xfs_free_buftarg(mp, mp->m_ddev_targp);
  out_close_rtdev:
 	if (rtdev)
 		xfs_blkdev_put(rtdev);
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index e2fb6210d4c..478e587087f 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -246,9 +246,6 @@ xfs_iget_cache_miss(
 		goto out_destroy;
 	}
 
-	if (lock_flags)
-		xfs_ilock(ip, lock_flags);
-
 	/*
 	 * Preload the radix tree so we can insert safely under the
 	 * write spinlock. Note that we cannot sleep inside the preload
@@ -256,7 +253,16 @@ xfs_iget_cache_miss(
 	 */
 	if (radix_tree_preload(GFP_KERNEL)) {
 		error = EAGAIN;
-		goto out_unlock;
+		goto out_destroy;
+	}
+
+	/*
+	 * Because the inode hasn't been added to the radix-tree yet it can't
+	 * be found by another thread, so we can do the non-sleeping lock here.
+	 */
+	if (lock_flags) {
+		if (!xfs_ilock_nowait(ip, lock_flags))
+			BUG();
 	}
 
 	mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1);
@@ -284,7 +290,6 @@ xfs_iget_cache_miss(
 out_preload_end:
 	write_unlock(&pag->pag_ici_lock);
 	radix_tree_preload_end();
-out_unlock:
 	if (lock_flags)
 		xfs_iunlock(ip, lock_flags);
 out_destroy:
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index b1047de2fff..61af610d79b 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -1455,10 +1455,19 @@ xlog_recover_add_to_trans(
 	item = item->ri_prev;
 
 	if (item->ri_total == 0) {		/* first region to be added */
-		item->ri_total	= in_f->ilf_size;
-		ASSERT(item->ri_total <= XLOG_MAX_REGIONS_IN_ITEM);
-		item->ri_buf = kmem_zalloc((item->ri_total *
-					    sizeof(xfs_log_iovec_t)), KM_SLEEP);
+		if (in_f->ilf_size == 0 ||
+		    in_f->ilf_size > XLOG_MAX_REGIONS_IN_ITEM) {
+			xlog_warn(
+	"XFS: bad number of regions (%d) in inode log format",
+				  in_f->ilf_size);
+			ASSERT(0);
+			return XFS_ERROR(EIO);
+		}
+
+		item->ri_total = in_f->ilf_size;
+		item->ri_buf =
+			kmem_zalloc(item->ri_total * sizeof(xfs_log_iovec_t),
+				    KM_SLEEP);
 	}
 	ASSERT(item->ri_total > item->ri_cnt);
 	/* Description region is ri_buf[0] */