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

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (114 commits)
  Btrfs: check for a null fs root when writing to the backup root log
  Btrfs: fix race during transaction joins
  Btrfs: fix a potential btrfs_bio leak on scrub fixups
  Btrfs: rename btrfs_bio multi -> bbio for consistency
  Btrfs: stop leaking btrfs_bios on readahead
  Btrfs: stop the readahead threads on failed mount
  Btrfs: fix extent_buffer leak in the metadata IO error handling
  Btrfs: fix the new inspection ioctls for 32 bit compat
  Btrfs: fix delayed insertion reservation
  Btrfs: ClearPageError during writepage and clean_tree_block
  Btrfs: be smarter about committing the transaction in reserve_metadata_bytes
  Btrfs: make a delayed_block_rsv for the delayed item insertion
  Btrfs: add a log of past tree roots
  btrfs: separate superblock items out of fs_info
  Btrfs: use the global reserve when truncating the free space cache inode
  Btrfs: release metadata from global reserve if we have to fallback for unlink
  Btrfs: make sure to flush queued bios if write_cache_pages waits
  Btrfs: fix extent pinning bugs in the tree log
  Btrfs: make sure btrfs_remove_free_space doesn't leak EAGAIN
  Btrfs: don't wait as long for more batches during SSD log commit
  ...

1 files changed, 518 insertions, 330 deletions

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index c9ee0e18bbdc..9879bd474632 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -23,6 +23,7 @@
 #include <linux/rcupdate.h>
 #include <linux/kthread.h>
 #include <linux/slab.h>
+#include <linux/ratelimit.h>
 #include "compat.h"
 #include "hash.h"
 #include "ctree.h"
@@ -52,6 +53,21 @@ enum {
 	CHUNK_ALLOC_LIMITED = 2,
 };
 
+/*
+ * Control how reservations are dealt with.
+ *
+ * RESERVE_FREE - freeing a reservation.
+ * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for
+ *   ENOSPC accounting
+ * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update
+ *   bytes_may_use as the ENOSPC accounting is done elsewhere
+ */
+enum {
+	RESERVE_FREE = 0,
+	RESERVE_ALLOC = 1,
+	RESERVE_ALLOC_NO_ACCOUNT = 2,
+};
+
 static int update_block_group(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      u64 bytenr, u64 num_bytes, int alloc);
@@ -81,6 +97,8 @@ static int find_next_key(struct btrfs_path *path, int level,
 			 struct btrfs_key *key);
 static void dump_space_info(struct btrfs_space_info *info, u64 bytes,
 			    int dump_block_groups);
+static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache,
+				       u64 num_bytes, int reserve);
 
 static noinline int
 block_group_cache_done(struct btrfs_block_group_cache *cache)
@@ -104,7 +122,6 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
 	if (atomic_dec_and_test(&cache->count)) {
 		WARN_ON(cache->pinned > 0);
 		WARN_ON(cache->reserved > 0);
-		WARN_ON(cache->reserved_pinned > 0);
 		kfree(cache->free_space_ctl);
 		kfree(cache);
 	}
@@ -465,7 +482,8 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
 	 * we likely hold important locks.
 	 */
 	if (trans && (!trans->transaction->in_commit) &&
-	    (root && root != root->fs_info->tree_root)) {
+	    (root && root != root->fs_info->tree_root) &&
+	    btrfs_test_opt(root, SPACE_CACHE)) {
 		spin_lock(&cache->lock);
 		if (cache->cached != BTRFS_CACHE_NO) {
 			spin_unlock(&cache->lock);
@@ -1770,18 +1788,18 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 {
 	int ret;
 	u64 discarded_bytes = 0;
-	struct btrfs_multi_bio *multi = NULL;
+	struct btrfs_bio *bbio = NULL;
 
 
 	/* Tell the block device(s) that the sectors can be discarded */
 	ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD,
-			      bytenr, &num_bytes, &multi, 0);
+			      bytenr, &num_bytes, &bbio, 0);
 	if (!ret) {
-		struct btrfs_bio_stripe *stripe = multi->stripes;
+		struct btrfs_bio_stripe *stripe = bbio->stripes;
 		int i;
 
 
-		for (i = 0; i < multi->num_stripes; i++, stripe++) {
+		for (i = 0; i < bbio->num_stripes; i++, stripe++) {
 			if (!stripe->dev->can_discard)
 				continue;
 
@@ -1800,7 +1818,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 			 */
 			ret = 0;
 		}
-		kfree(multi);
+		kfree(bbio);
 	}
 
 	if (actual_bytes)
@@ -2700,6 +2718,13 @@ again:
 		goto again;
 	}
 
+	/* We've already setup this transaction, go ahead and exit */
+	if (block_group->cache_generation == trans->transid &&
+	    i_size_read(inode)) {
+		dcs = BTRFS_DC_SETUP;
+		goto out_put;
+	}
+
 	/*
 	 * We want to set the generation to 0, that way if anything goes wrong
 	 * from here on out we know not to trust this cache when we load up next
@@ -2749,12 +2774,15 @@ again:
 	if (!ret)
 		dcs = BTRFS_DC_SETUP;
 	btrfs_free_reserved_data_space(inode, num_pages);
+
 out_put:
 	iput(inode);
 out_free:
 	btrfs_release_path(path);
 out:
 	spin_lock(&block_group->lock);
+	if (!ret)
+		block_group->cache_generation = trans->transid;
 	block_group->disk_cache_state = dcs;
 	spin_unlock(&block_group->lock);
 
@@ -3122,16 +3150,13 @@ commit_trans:
 		return -ENOSPC;
 	}
 	data_sinfo->bytes_may_use += bytes;
-	BTRFS_I(inode)->reserved_bytes += bytes;
 	spin_unlock(&data_sinfo->lock);
 
 	return 0;
 }
 
 /*
- * called when we are clearing an delalloc extent from the
- * inode's io_tree or there was an error for whatever reason
- * after calling btrfs_check_data_free_space
+ * Called if we need to clear a data reservation for this inode.
  */
 void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 {
@@ -3144,7 +3169,6 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
 	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);
 }
 
@@ -3165,6 +3189,7 @@ static int should_alloc_chunk(struct btrfs_root *root,
 			      struct btrfs_space_info *sinfo, u64 alloc_bytes,
 			      int force)
 {
+	struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
 	u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly;
 	u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved;
 	u64 thresh;
@@ -3173,11 +3198,18 @@ static int should_alloc_chunk(struct btrfs_root *root,
 		return 1;
 
 	/*
+	 * We need to take into account the global rsv because for all intents
+	 * and purposes it's used space.  Don't worry about locking the
+	 * global_rsv, it doesn't change except when the transaction commits.
+	 */
+	num_allocated += global_rsv->size;
+
+	/*
 	 * in limited mode, we want to have some free space up to
 	 * about 1% of the FS size.
 	 */
 	if (force == CHUNK_ALLOC_LIMITED) {
-		thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+		thresh = btrfs_super_total_bytes(root->fs_info->super_copy);
 		thresh = max_t(u64, 64 * 1024 * 1024,
 			       div_factor_fine(thresh, 1));
 
@@ -3199,7 +3231,7 @@ static int should_alloc_chunk(struct btrfs_root *root,
 	if (num_allocated + alloc_bytes < div_factor(num_bytes, 8))
 		return 0;
 
-	thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+	thresh = btrfs_super_total_bytes(root->fs_info->super_copy);
 
 	/* 256MB or 5% of the FS */
 	thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5));
@@ -3302,24 +3334,26 @@ out:
 /*
  * shrink metadata reservation for delalloc
  */
-static int shrink_delalloc(struct btrfs_trans_handle *trans,
-			   struct btrfs_root *root, u64 to_reclaim, int sync)
+static int shrink_delalloc(struct btrfs_root *root, u64 to_reclaim,
+			   bool wait_ordered)
 {
 	struct btrfs_block_rsv *block_rsv;
 	struct btrfs_space_info *space_info;
+	struct btrfs_trans_handle *trans;
 	u64 reserved;
 	u64 max_reclaim;
 	u64 reclaimed = 0;
 	long time_left;
-	int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT;
+	unsigned long nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT;
 	int loops = 0;
 	unsigned long progress;
 
+	trans = (struct btrfs_trans_handle *)current->journal_info;
 	block_rsv = &root->fs_info->delalloc_block_rsv;
 	space_info = block_rsv->space_info;
 
 	smp_mb();
-	reserved = space_info->bytes_reserved;
+	reserved = space_info->bytes_may_use;
 	progress = space_info->reservation_progress;
 
 	if (reserved == 0)
@@ -3334,7 +3368,8 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
 	}
 
 	max_reclaim = min(reserved, to_reclaim);
-
+	nr_pages = max_t(unsigned long, nr_pages,
+			 max_reclaim >> PAGE_CACHE_SHIFT);
 	while (loops < 1024) {
 		/* have the flusher threads jump in and do some IO */
 		smp_mb();
@@ -3344,9 +3379,9 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
 						WB_REASON_FS_FREE_SPACE);
 
 		spin_lock(&space_info->lock);
-		if (reserved > space_info->bytes_reserved)
-			reclaimed += reserved - space_info->bytes_reserved;
-		reserved = space_info->bytes_reserved;
+		if (reserved > space_info->bytes_may_use)
+			reclaimed += reserved - space_info->bytes_may_use;
+		reserved = space_info->bytes_may_use;
 		spin_unlock(&space_info->lock);
 
 		loops++;
@@ -3357,11 +3392,15 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
 		if (trans && trans->transaction->blocked)
 			return -EAGAIN;
 
-		time_left = schedule_timeout_interruptible(1);
+		if (wait_ordered && !trans) {
+			btrfs_wait_ordered_extents(root, 0, 0);
+		} else {
+			time_left = schedule_timeout_interruptible(1);
 
-		/* We were interrupted, exit */
-		if (time_left)
-			break;
+			/* We were interrupted, exit */
+			if (time_left)
+				break;
+		}
 
 		/* we've kicked the IO a few times, if anything has been freed,
 		 * exit.  There is no sense in looping here for a long time
@@ -3376,34 +3415,90 @@ static int shrink_delalloc(struct btrfs_trans_handle *trans,
 		}
 
 	}
-	if (reclaimed >= to_reclaim && !trans)
-		btrfs_wait_ordered_extents(root, 0, 0);
+
 	return reclaimed >= to_reclaim;
 }
 
-/*
- * Retries tells us how many times we've called reserve_metadata_bytes.  The
- * idea is if this is the first call (retries == 0) then we will add to our
- * reserved count if we can't make the allocation in order to hold our place
- * while we go and try and free up space.  That way for retries > 1 we don't try
- * and add space, we just check to see if the amount of unused space is >= the
- * total space, meaning that our reservation is valid.
+/**
+ * maybe_commit_transaction - possibly commit the transaction if its ok to
+ * @root - the root we're allocating for
+ * @bytes - the number of bytes we want to reserve
+ * @force - force the commit
  *
- * However if we don't intend to retry this reservation, pass -1 as retries so
- * that it short circuits this logic.
+ * This will check to make sure that committing the transaction will actually
+ * get us somewhere and then commit the transaction if it does.  Otherwise it
+ * will return -ENOSPC.
  */
-static int reserve_metadata_bytes(struct btrfs_trans_handle *trans,
-				  struct btrfs_root *root,
+static int may_commit_transaction(struct btrfs_root *root,
+				  struct btrfs_space_info *space_info,
+				  u64 bytes, int force)
+{
+	struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv;
+	struct btrfs_trans_handle *trans;
+
+	trans = (struct btrfs_trans_handle *)current->journal_info;
+	if (trans)
+		return -EAGAIN;
+
+	if (force)
+		goto commit;
+
+	/* See if there is enough pinned space to make this reservation */
+	spin_lock(&space_info->lock);
+	if (space_info->bytes_pinned >= bytes) {
+		spin_unlock(&space_info->lock);
+		goto commit;
+	}
+	spin_unlock(&space_info->lock);
+
+	/*
+	 * See if there is some space in the delayed insertion reservation for
+	 * this reservation.
+	 */
+	if (space_info != delayed_rsv->space_info)
+		return -ENOSPC;
+
+	spin_lock(&delayed_rsv->lock);
+	if (delayed_rsv->size < bytes) {
+		spin_unlock(&delayed_rsv->lock);
+		return -ENOSPC;
+	}
+	spin_unlock(&delayed_rsv->lock);
+
+commit:
+	trans = btrfs_join_transaction(root);
+	if (IS_ERR(trans))
+		return -ENOSPC;
+
+	return btrfs_commit_transaction(trans, root);
+}
+
+/**
+ * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space
+ * @root - the root we're allocating for
+ * @block_rsv - the block_rsv we're allocating for
+ * @orig_bytes - the number of bytes we want
+ * @flush - wether or not we can flush to make our reservation
+ *
+ * This will reserve orgi_bytes number of bytes from the space info associated
+ * with the block_rsv.  If there is not enough space it will make an attempt to
+ * flush out space to make room.  It will do this by flushing delalloc if
+ * possible or committing the transaction.  If flush is 0 then no attempts to
+ * regain reservations will be made and this will fail if there is not enough
+ * space already.
+ */
+static int reserve_metadata_bytes(struct btrfs_root *root,
 				  struct btrfs_block_rsv *block_rsv,
 				  u64 orig_bytes, int flush)
 {
 	struct btrfs_space_info *space_info = block_rsv->space_info;
-	u64 unused;
+	u64 used;
 	u64 num_bytes = orig_bytes;
 	int retries = 0;
 	int ret = 0;
 	bool committed = false;
 	bool flushing = false;
+	bool wait_ordered = false;
 
 again:
 	ret = 0;
@@ -3420,7 +3515,7 @@ again:
 		 * deadlock since we are waiting for the flusher to finish, but
 		 * hold the current transaction open.
 		 */
-		if (trans)
+		if (current->journal_info)
 			return -EAGAIN;
 		ret = wait_event_interruptible(space_info->wait,
 					       !space_info->flush);
@@ -3432,9 +3527,9 @@ again:
 	}
 
 	ret = -ENOSPC;
-	unused = space_info->bytes_used + space_info->bytes_reserved +
-		 space_info->bytes_pinned + space_info->bytes_readonly +
-		 space_info->bytes_may_use;
+	used = space_info->bytes_used + space_info->bytes_reserved +
+		space_info->bytes_pinned + space_info->bytes_readonly +
+		space_info->bytes_may_use;
 
 	/*
 	 * The idea here is that we've not already over-reserved the block group
@@ -3443,10 +3538,9 @@ again:
 	 * lets start flushing stuff first and then come back and try to make
 	 * our reservation.
 	 */
-	if (unused <= space_info->total_bytes) {
-		unused = space_info->total_bytes - unused;
-		if (unused >= num_bytes) {
-			space_info->bytes_reserved += orig_bytes;
+	if (used <= space_info->total_bytes) {
+		if (used + orig_bytes <= space_info->total_bytes) {
+			space_info->bytes_may_use += orig_bytes;
 			ret = 0;
 		} else {
 			/*
@@ -3462,10 +3556,64 @@ again:
 		 * amount plus the amount of bytes that we need for this
 		 * reservation.
 		 */
-		num_bytes = unused - space_info->total_bytes +
+		wait_ordered = true;
+		num_bytes = used - space_info->total_bytes +
 			(orig_bytes * (retries + 1));
 	}
 
+	if (ret) {
+		u64 profile = btrfs_get_alloc_profile(root, 0);
+		u64 avail;
+
+		/*
+		 * If we have a lot of space that's pinned, don't bother doing
+		 * the overcommit dance yet and just commit the transaction.
+		 */
+		avail = (space_info->total_bytes - space_info->bytes_used) * 8;
+		do_div(avail, 10);
+		if (space_info->bytes_pinned >= avail && flush && !committed) {
+			space_info->flush = 1;
+			flushing = true;
+			spin_unlock(&space_info->lock);
+			ret = may_commit_transaction(root, space_info,
+						     orig_bytes, 1);
+			if (ret)
+				goto out;
+			committed = true;
+			goto again;
+		}
+
+		spin_lock(&root->fs_info->free_chunk_lock);
+		avail = root->fs_info->free_chunk_space;
+
+		/*
+		 * If we have dup, raid1 or raid10 then only half of the free
+		 * space is actually useable.
+		 */
+		if (profile & (BTRFS_BLOCK_GROUP_DUP |
+			       BTRFS_BLOCK_GROUP_RAID1 |
+			       BTRFS_BLOCK_GROUP_RAID10))
+			avail >>= 1;
+
+		/*
+		 * If we aren't flushing don't let us overcommit too much, say
+		 * 1/8th of the space.  If we can flush, let it overcommit up to
+		 * 1/2 of the space.
+		 */
+		if (flush)
+			avail >>= 3;
+		else
+			avail >>= 1;
+		 spin_unlock(&root->fs_info->free_chunk_lock);
+
+		if (used + num_bytes < space_info->total_bytes + avail) {
+			space_info->bytes_may_use += orig_bytes;
+			ret = 0;
+		} else {
+			wait_ordered = true;
+		}
+	}
+
 	/*
 	 * Couldn't make our reservation, save our place so while we're trying
 	 * to reclaim space we can actually use it instead of somebody else
@@ -3485,7 +3633,7 @@ again:
 	 * We do synchronous shrinking since we don't actually unreserve
 	 * metadata until after the IO is completed.
 	 */
-	ret = shrink_delalloc(trans, root, num_bytes, 1);
+	ret = shrink_delalloc(root, num_bytes, wait_ordered);
 	if (ret < 0)
 		goto out;
 
@@ -3497,35 +3645,17 @@ again:
 	 * so go back around and try again.
 	 */
 	if (retries < 2) {
+		wait_ordered = true;
 		retries++;
 		goto again;
 	}
 
-	/*
-	 * Not enough space to be reclaimed, don't bother committing the
-	 * transaction.
-	 */
-	spin_lock(&space_info->lock);
-	if (space_info->bytes_pinned < orig_bytes)
-		ret = -ENOSPC;
-	spin_unlock(&space_info->lock);
-	if (ret)
-		goto out;
-
-	ret = -EAGAIN;
-	if (trans)
-		goto out;
-
 	ret = -ENOSPC;
 	if (committed)
 		goto out;
 
-	trans = btrfs_join_transaction(root);
-	if (IS_ERR(trans))
-		goto out;
-	ret = btrfs_commit_transaction(trans, root);
+	ret = may_commit_transaction(root, space_info, orig_bytes, 0);
 	if (!ret) {
-		trans = NULL;
 		committed = true;
 		goto again;
 	}
@@ -3543,10 +3673,12 @@ out:
 static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans,
 					     struct btrfs_root *root)
 {
-	struct btrfs_block_rsv *block_rsv;
-	if (root->ref_cows)
+	struct btrfs_block_rsv *block_rsv = NULL;
+
+	if (root->ref_cows || root == root->fs_info->csum_root)
 		block_rsv = trans->block_rsv;
-	else
+
+	if (!block_rsv)
 		block_rsv = root->block_rsv;
 
 	if (!block_rsv)
@@ -3617,7 +3749,7 @@ static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv,
 		}
 		if (num_bytes) {
 			spin_lock(&space_info->lock);
-			space_info->bytes_reserved -= num_bytes;
+			space_info->bytes_may_use -= num_bytes;
 			space_info->reservation_progress++;
 			spin_unlock(&space_info->lock);
 		}
@@ -3641,9 +3773,6 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv)
 {
 	memset(rsv, 0, sizeof(*rsv));
 	spin_lock_init(&rsv->lock);
-	atomic_set(&rsv->usage, 1);
-	rsv->priority = 6;
-	INIT_LIST_HEAD(&rsv->list);
 }
 
 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
@@ -3664,38 +3793,38 @@ struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root)
 void btrfs_free_block_rsv(struct btrfs_root *root,
 			  struct btrfs_block_rsv *rsv)
 {
-	if (rsv && atomic_dec_and_test(&rsv->usage)) {
-		btrfs_block_rsv_release(root, rsv, (u64)-1);
-		if (!rsv->durable)
-			kfree(rsv);
-	}
+	btrfs_block_rsv_release(root, rsv, (u64)-1);
+	kfree(rsv);
 }
 
-/*
- * make the block_rsv struct be able to capture freed space.
- * the captured space will re-add to the the block_rsv struct
- * after transaction commit
- */
-void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info,
-				 struct btrfs_block_rsv *block_rsv)
+int btrfs_block_rsv_add(struct btrfs_root *root,
+			struct btrfs_block_rsv *block_rsv,
+			u64 num_bytes)
 {
-	block_rsv->durable = 1;
-	mutex_lock(&fs_info->durable_block_rsv_mutex);
-	list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list);
-	mutex_unlock(&fs_info->durable_block_rsv_mutex);
+	int ret;
+
+	if (num_bytes == 0)
+		return 0;
+
+	ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 1);
+	if (!ret) {
+		block_rsv_add_bytes(block_rsv, num_bytes, 1);
+		return 0;
+	}
+
+	return ret;
 }
 
-int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root,
-			struct btrfs_block_rsv *block_rsv,
-			u64 num_bytes)
+int btrfs_block_rsv_add_noflush(struct btrfs_root *root,
+				struct btrfs_block_rsv *block_rsv,
+				u64 num_bytes)
 {
 	int ret;
 
 	if (num_bytes == 0)
 		return 0;
 
-	ret = reserve_metadata_bytes(trans, root, block_rsv, num_bytes, 1);
+	ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 0);
 	if (!ret) {
 		block_rsv_add_bytes(block_rsv, num_bytes, 1);
 		return 0;
@@ -3704,55 +3833,52 @@ int btrfs_block_rsv_add(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-int btrfs_block_rsv_check(struct btrfs_trans_handle *trans,
-			  struct btrfs_root *root,
-			  struct btrfs_block_rsv *block_rsv,
-			  u64 min_reserved, int min_factor)
+int btrfs_block_rsv_check(struct btrfs_root *root,
+			  struct btrfs_block_rsv *block_rsv, int min_factor)
 {
 	u64 num_bytes = 0;
-	int commit_trans = 0;
 	int ret = -ENOSPC;
 
 	if (!block_rsv)
 		return 0;
 
 	spin_lock(&block_rsv->lock);
-	if (min_factor > 0)
-		num_bytes = div_factor(block_rsv->size, min_factor);
-	if (min_reserved > num_bytes)
-		num_bytes = min_reserved;
+	num_bytes = div_factor(block_rsv->size, min_factor);
+	if (block_rsv->reserved >= num_bytes)
+		ret = 0;
+	spin_unlock(&block_rsv->lock);
 
-	if (block_rsv->reserved >= num_bytes) {
+	return ret;
+}
+
+int btrfs_block_rsv_refill(struct btrfs_root *root,
+			  struct btrfs_block_rsv *block_rsv,
+			  u64 min_reserved)
+{
+	u64 num_bytes = 0;
+	int ret = -ENOSPC;
+
+	if (!block_rsv)
+		return 0;
+
+	spin_lock(&block_rsv->lock);
+	num_bytes = min_reserved;
+	if (block_rsv->reserved >= num_bytes)
 		ret = 0;
-	} else {
+	else
 		num_bytes -= block_rsv->reserved;
-		if (block_rsv->durable &&
-		    block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes)
-			commit_trans = 1;
-	}
 	spin_unlock(&block_rsv->lock);
+
 	if (!ret)
 		return 0;
 
-	if (block_rsv->refill_used) {
-		ret = reserve_metadata_bytes(trans, root, block_rsv,
-					     num_bytes, 0);
-		if (!ret) {
-			block_rsv_add_bytes(block_rsv, num_bytes, 0);
-			return 0;
-		}
-	}
-
-	if (commit_trans) {
-		if (trans)
-			return -EAGAIN;
-		trans = btrfs_join_transaction(root);
-		BUG_ON(IS_ERR(trans));
-		ret = btrfs_commit_transaction(trans, root);
+	ret = reserve_metadata_bytes(root, block_rsv, num_bytes, 1);
+	if (!ret) {
+		block_rsv_add_bytes(block_rsv, num_bytes, 0);
 		return 0;
 	}
 
-	return -ENOSPC;
+	return ret;
 }
 
 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv,
@@ -3784,7 +3910,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info)
 	u64 num_bytes;
 	u64 meta_used;
 	u64 data_used;
-	int csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+	int csum_size = btrfs_super_csum_size(fs_info->super_copy);
 
 	sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA);
 	spin_lock(&sinfo->lock);
@@ -3828,12 +3954,12 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info)
 	if (sinfo->total_bytes > num_bytes) {
 		num_bytes = sinfo->total_bytes - num_bytes;
 		block_rsv->reserved += num_bytes;
-		sinfo->bytes_reserved += num_bytes;
+		sinfo->bytes_may_use += num_bytes;
 	}
 
 	if (block_rsv->reserved >= block_rsv->size) {
 		num_bytes = block_rsv->reserved - block_rsv->size;
-		sinfo->bytes_reserved -= num_bytes;
+		sinfo->bytes_may_use -= num_bytes;
 		sinfo->reservation_progress++;
 		block_rsv->reserved = block_rsv->size;
 		block_rsv->full = 1;
@@ -3849,16 +3975,13 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
 
 	space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
 	fs_info->chunk_block_rsv.space_info = space_info;
-	fs_info->chunk_block_rsv.priority = 10;
 
 	space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA);
 	fs_info->global_block_rsv.space_info = space_info;
-	fs_info->global_block_rsv.priority = 10;
-	fs_info->global_block_rsv.refill_used = 1;
 	fs_info->delalloc_block_rsv.space_info = space_info;
 	fs_info->trans_block_rsv.space_info = space_info;
 	fs_info->empty_block_rsv.space_info = space_info;
-	fs_info->empty_block_rsv.priority = 10;
+	fs_info->delayed_block_rsv.space_info = space_info;
 
 	fs_info->extent_root->block_rsv = &fs_info->global_block_rsv;
 	fs_info->csum_root->block_rsv = &fs_info->global_block_rsv;
@@ -3866,10 +3989,6 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info)
 	fs_info->tree_root->block_rsv = &fs_info->global_block_rsv;
 	fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv;
 
-	btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv);
-
-	btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv);
-
 	update_global_block_rsv(fs_info);
 }
 
@@ -3882,37 +4001,8 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info)
 	WARN_ON(fs_info->trans_block_rsv.reserved > 0);
 	WARN_ON(fs_info->chunk_block_rsv.size > 0);
 	WARN_ON(fs_info->chunk_block_rsv.reserved > 0);
-}
-
-int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans,
-				    struct btrfs_root *root,
-				    struct btrfs_block_rsv *rsv)
-{
-	struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv;
-	u64 num_bytes;
-	int ret;
-
-	/*
-	 * Truncate should be freeing data, but give us 2 items just in case it
-	 * needs to use some space.  We may want to be smarter about this in the
-	 * future.
-	 */
-	num_bytes = btrfs_calc_trans_metadata_size(root, 2);
-
-	/* We already have enough bytes, just return */
-	if (rsv->reserved >= num_bytes)
-		return 0;
-
-	num_bytes -= rsv->reserved;
-
-	/*
-	 * You should have reserved enough space before hand to do this, so this
-	 * should not fail.
-	 */
-	ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes);
-	BUG_ON(ret);
-
-	return 0;
+	WARN_ON(fs_info->delayed_block_rsv.size > 0);
+	WARN_ON(fs_info->delayed_block_rsv.reserved > 0);
 }
 
 void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
@@ -3921,9 +4011,7 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
 	if (!trans->bytes_reserved)
 		return;
 
-	BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv);
-	btrfs_block_rsv_release(root, trans->block_rsv,
-				trans->bytes_reserved);
+	btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved);
 	trans->bytes_reserved = 0;
 }
 
@@ -3965,11 +4053,19 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
 	return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes);
 }
 
+/**
+ * drop_outstanding_extent - drop an outstanding extent
+ * @inode: the inode we're dropping the extent for
+ *
+ * This is called when we are freeing up an outstanding extent, either called
+ * after an error or after an extent is written.  This will return the number of
+ * reserved extents that need to be freed.  This must be called with
+ * BTRFS_I(inode)->lock held.
+ */
 static unsigned drop_outstanding_extent(struct inode *inode)
 {
 	unsigned dropped_extents = 0;
 
-	spin_lock(&BTRFS_I(inode)->lock);
 	BUG_ON(!BTRFS_I(inode)->outstanding_extents);
 	BTRFS_I(inode)->outstanding_extents--;
 
@@ -3979,19 +4075,70 @@ static unsigned drop_outstanding_extent(struct inode *inode)
 	 */
 	if (BTRFS_I(inode)->outstanding_extents >=
 	    BTRFS_I(inode)->reserved_extents)
-		goto out;
+		return 0;
 
 	dropped_extents = BTRFS_I(inode)->reserved_extents -
 		BTRFS_I(inode)->outstanding_extents;
 	BTRFS_I(inode)->reserved_extents -= dropped_extents;
-out:
-	spin_unlock(&BTRFS_I(inode)->lock);
 	return dropped_extents;
 }
 
-static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes)
+/**
+ * calc_csum_metadata_size - return the amount of metada space that must be
+ *	reserved/free'd for the given bytes.
+ * @inode: the inode we're manipulating
+ * @num_bytes: the number of bytes in question
+ * @reserve: 1 if we are reserving space, 0 if we are freeing space
+ *
+ * This adjusts the number of csum_bytes in the inode and then returns the
+ * correct amount of metadata that must either be reserved or freed.  We
+ * calculate how many checksums we can fit into one leaf and then divide the
+ * number of bytes that will need to be checksumed by this value to figure out
+ * how many checksums will be required.  If we are adding bytes then the number
+ * may go up and we will return the number of additional bytes that must be
+ * reserved.  If it is going down we will return the number of bytes that must
+ * be freed.
+ *
+ * This must be called with BTRFS_I(inode)->lock held.
+ */
+static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes,
+				   int reserve)
 {
-	return num_bytes >>= 3;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u64 csum_size;
+	int num_csums_per_leaf;
+	int num_csums;
+	int old_csums;
+
+	if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM &&
+	    BTRFS_I(inode)->csum_bytes == 0)
+		return 0;
+
+	old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize);
+	if (reserve)
+		BTRFS_I(inode)->csum_bytes += num_bytes;
+	else
+		BTRFS_I(inode)->csum_bytes -= num_bytes;
+	csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item);
+	num_csums_per_leaf = (int)div64_u64(csum_size,
+					    sizeof(struct btrfs_csum_item) +
+					    sizeof(struct btrfs_disk_key));
+	num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize);
+	num_csums = num_csums + num_csums_per_leaf - 1;
+	num_csums = num_csums / num_csums_per_leaf;
+
+	old_csums = old_csums + num_csums_per_leaf - 1;
+	old_csums = old_csums / num_csums_per_leaf;
+
+	/* No change, no need to reserve more */
+	if (old_csums == num_csums)
+		return 0;
+
+	if (reserve)
+		return btrfs_calc_trans_metadata_size(root,
+						      num_csums - old_csums);
+
+	return btrfs_calc_trans_metadata_size(root, old_csums - num_csums);
 }
 
 int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
@@ -4000,9 +4147,13 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
 	struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv;
 	u64 to_reserve = 0;
 	unsigned nr_extents = 0;
+	int flush = 1;
 	int ret;
 
-	if (btrfs_transaction_in_commit(root->fs_info))
+	if (btrfs_is_free_space_inode(root, inode))
+		flush = 0;
+
+	if (flush && btrfs_transaction_in_commit(root->fs_info))
 		schedule_timeout(1);
 
 	num_bytes = ALIGN(num_bytes, root->sectorsize);
@@ -4018,18 +4169,29 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
 
 		to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents);
 	}
+	to_reserve += calc_csum_metadata_size(inode, num_bytes, 1);
 	spin_unlock(&BTRFS_I(inode)->lock);
 
-	to_reserve += calc_csum_metadata_size(inode, num_bytes);
-	ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1);
+	ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush);
 	if (ret) {
+		u64 to_free = 0;
 		unsigned dropped;
+
+		spin_lock(&BTRFS_I(inode)->lock);
+		dropped = drop_outstanding_extent(inode);
+		to_free = calc_csum_metadata_size(inode, num_bytes, 0);
+		spin_unlock(&BTRFS_I(inode)->lock);
+		to_free += btrfs_calc_trans_metadata_size(root, dropped);
+
 		/*
-		 * We don't need the return value since our reservation failed,
-		 * we just need to clean up our counter.
+		 * Somebody could have come in and twiddled with the
+		 * reservation, so if we have to free more than we would have
+		 * reserved from this reservation go ahead and release those
+		 * bytes.
 		 */
-		dropped = drop_outstanding_extent(inode);
-		WARN_ON(dropped > 1);
+		to_free -= to_reserve;
+		if (to_free)
+			btrfs_block_rsv_release(root, block_rsv, to_free);
 		return ret;
 	}
 
@@ -4038,6 +4200,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
 	return 0;
 }
 
+/**
+ * btrfs_delalloc_release_metadata - release a metadata reservation for an inode
+ * @inode: the inode to release the reservation for
+ * @num_bytes: the number of bytes we're releasing
+ *
+ * This will release the metadata reservation for an inode.  This can be called
+ * once we complete IO for a given set of bytes to release their metadata
+ * reservations.
+ */
 void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -4045,9 +4216,11 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
 	unsigned dropped;
 
 	num_bytes = ALIGN(num_bytes, root->sectorsize);
+	spin_lock(&BTRFS_I(inode)->lock);
 	dropped = drop_outstanding_extent(inode);
 
-	to_free = calc_csum_metadata_size(inode, num_bytes);
+	to_free = calc_csum_metadata_size(inode, num_bytes, 0);
+	spin_unlock(&BTRFS_I(inode)->lock);
 	if (dropped > 0)
 		to_free += btrfs_calc_trans_metadata_size(root, dropped);
 
@@ -4055,6 +4228,21 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
 				to_free);
 }
 
+/**
+ * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc
+ * @inode: inode we're writing to
+ * @num_bytes: the number of bytes we want to allocate
+ *
+ * This will do the following things
+ *
+ * o reserve space in the data space info for num_bytes
+ * o reserve space in the metadata space info based on number of outstanding
+ *   extents and how much csums will be needed
+ * o add to the inodes ->delalloc_bytes
+ * o add it to the fs_info's delalloc inodes list.
+ *
+ * This will return 0 for success and -ENOSPC if there is no space left.
+ */
 int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
 {
 	int ret;
@@ -4072,6 +4260,19 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
 	return 0;
 }
 
+/**
+ * btrfs_delalloc_release_space - release data and metadata space for delalloc
+ * @inode: inode we're releasing space for
+ * @num_bytes: the number of bytes we want to free up
+ *
+ * This must be matched with a call to btrfs_delalloc_reserve_space.  This is
+ * called in the case that we don't need the metadata AND data reservations
+ * anymore.  So if there is an error or we insert an inline extent.
+ *
+ * This function will release the metadata space that was not used and will
+ * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes
+ * list if there are no delalloc bytes left.
+ */
 void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
 {
 	btrfs_delalloc_release_metadata(inode, num_bytes);
@@ -4091,12 +4292,12 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 
 	/* block accounting for super block */
 	spin_lock(&info->delalloc_lock);
-	old_val = btrfs_super_bytes_used(&info->super_copy);
+	old_val = btrfs_super_bytes_used(info->super_copy);
 	if (alloc)
 		old_val += num_bytes;
 	else
 		old_val -= num_bytes;
-	btrfs_set_super_bytes_used(&info->super_copy, old_val);
+	btrfs_set_super_bytes_used(info->super_copy, old_val);
 	spin_unlock(&info->delalloc_lock);
 
 	while (tot