Merge git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable

* git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable:
  Btrfs: add check for changed leaves in setup_leaf_for_split
  Btrfs: create snapshot references in same commit as snapshot
  Btrfs: fix small race with delalloc flushing waitqueue's
  Btrfs: use add_to_page_cache_lru, use __page_cache_alloc
  Btrfs: fix chunk allocate size calculation
  Btrfs: kill max_extent mount option
  Btrfs: fail to mount if we have problems reading the block groups
  Btrfs: check btrfs_get_extent return for IS_ERR()
  Btrfs: handle kmalloc() failure in inode lookup ioctl
  Btrfs: dereferencing freed memory
  Btrfs: Simplify num_stripes's calculation logical for __btrfs_alloc_chunk()
  Btrfs: Add error handle for btrfs_search_slot() in btrfs_read_chunk_tree()
  Btrfs: Remove unnecessary finish_wait() in wait_current_trans()
  Btrfs: add NULL check for do_walk_down()
  Btrfs: remove duplicate include in ioctl.c

Fix trivial conflict in fs/btrfs/compression.c due to slab.h include
cleanups.

fs/btrfs/compression.c

@@ -31,7 +31,6 @@
 #include <linux/swap.h>
 #include <linux/writeback.h>
 #include <linux/bit_spinlock.h>
-#include <linux/pagevec.h>
 #include <linux/slab.h>
 #include "compat.h"
 #include "ctree.h"
@@ -446,7 +445,6 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 	unsigned long nr_pages = 0;
 	struct extent_map *em;
 	struct address_space *mapping = inode->i_mapping;
-	struct pagevec pvec;
 	struct extent_map_tree *em_tree;
 	struct extent_io_tree *tree;
 	u64 end;
@@ -462,7 +460,6 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 
 	end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
 
-	pagevec_init(&pvec, 0);
 	while (last_offset < compressed_end) {
 		page_index = last_offset >> PAGE_CACHE_SHIFT;
 
@@ -479,26 +476,17 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			goto next;
 		}
 
-		page = alloc_page(mapping_gfp_mask(mapping) & ~__GFP_FS);
+		page = __page_cache_alloc(mapping_gfp_mask(mapping) &
+								~__GFP_FS);
 		if (!page)
 			break;
 
-		page->index = page_index;
-		/*
-		 * what we want to do here is call add_to_page_cache_lru,
-		 * but that isn't exported, so we reproduce it here
-		 */
-		if (add_to_page_cache(page, mapping,
-				      page->index, GFP_NOFS)) {
+		if (add_to_page_cache_lru(page, mapping, page_index,
+								GFP_NOFS)) {
 			page_cache_release(page);
 			goto next;
 		}
 
-		/* open coding of lru_cache_add, also not exported */
-		page_cache_get(page);
-		if (!pagevec_add(&pvec, page))
-			__pagevec_lru_add_file(&pvec);
-
 		end = last_offset + PAGE_CACHE_SIZE - 1;
 		/*
 		 * at this point, we have a locked page in the page cache
@@ -552,8 +540,6 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 next:
 		last_offset += PAGE_CACHE_SIZE;
 	}
-	if (pagevec_count(&pvec))
-		__pagevec_lru_add_file(&pvec);
 	return 0;
 }
 

fs/btrfs/ctree.c

@@ -3041,6 +3041,10 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
 	if (ret > 0 || item_size != btrfs_item_size_nr(leaf, path->slots[0]))
 		goto err;
 
+	/* the leaf has  changed, it now has room.  return now */
+	if (btrfs_leaf_free_space(root, path->nodes[0]) >= ins_len)
+		goto err;
+
 	if (key.type == BTRFS_EXTENT_DATA_KEY) {
 		fi = btrfs_item_ptr(leaf, path->slots[0],
 				    struct btrfs_file_extent_item);

fs/btrfs/ctree.h

@@ -835,7 +835,6 @@ struct btrfs_fs_info {
 	u64 last_trans_log_full_commit;
 	u64 open_ioctl_trans;
 	unsigned long mount_opt;
-	u64 max_extent;
 	u64 max_inline;
 	u64 alloc_start;
 	struct btrfs_transaction *running_transaction;

fs/btrfs/disk-io.c

@@ -1635,7 +1635,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	atomic_set(&fs_info->async_submit_draining, 0);
 	atomic_set(&fs_info->nr_async_bios, 0);
 	fs_info->sb = sb;
-	fs_info->max_extent = (u64)-1;
 	fs_info->max_inline = 8192 * 1024;
 	fs_info->metadata_ratio = 0;
 
@@ -1923,7 +1922,11 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	csum_root->track_dirty = 1;
 
-	btrfs_read_block_groups(extent_root);
+	ret = btrfs_read_block_groups(extent_root);
+	if (ret) {
+		printk(KERN_ERR "Failed to read block groups: %d\n", ret);
+		goto fail_block_groups;
+	}
 
 	fs_info->generation = generation;
 	fs_info->last_trans_committed = generation;
@@ -1933,7 +1936,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
 					       "btrfs-cleaner");
 	if (IS_ERR(fs_info->cleaner_kthread))
-		goto fail_csum_root;
+		goto fail_block_groups;
 
 	fs_info->transaction_kthread = kthread_run(transaction_kthread,
 						   tree_root,
@@ -2021,7 +2024,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 
-fail_csum_root:
+fail_block_groups:
+	btrfs_free_block_groups(fs_info);
 	free_extent_buffer(csum_root->node);
 	free_extent_buffer(csum_root->commit_root);
 fail_dev_root:

fs/btrfs/extent-tree.c

@@ -2677,6 +2677,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 
 	INIT_LIST_HEAD(&found->block_groups);
 	init_rwsem(&found->groups_sem);
+	init_waitqueue_head(&found->flush_wait);
+	init_waitqueue_head(&found->allocate_wait);
 	spin_lock_init(&found->lock);
 	found->flags = flags;
 	found->total_bytes = total_bytes;
@@ -2847,7 +2849,7 @@ int btrfs_unreserve_metadata_for_delalloc(struct btrfs_root *root,
 	}
 	spin_unlock(&BTRFS_I(inode)->accounting_lock);
 
-	BTRFS_I(inode)->reserved_extents--;
+	BTRFS_I(inode)->reserved_extents -= num_items;
 	BUG_ON(BTRFS_I(inode)->reserved_extents < 0);
 
 	if (meta_sinfo->bytes_delalloc < num_bytes) {
@@ -2945,12 +2947,10 @@ static void flush_delalloc(struct btrfs_root *root,
 
 	spin_lock(&info->lock);
 
-	if (!info->flushing) {
+	if (!info->flushing)
 		info->flushing = 1;
-		init_waitqueue_head(&info->flush_wait);
-	} else {
+	else
 		wait = true;
-	}
 
 	spin_unlock(&info->lock);
 
@@ -3012,7 +3012,6 @@ static int maybe_allocate_chunk(struct btrfs_root *root,
 	if (!info->allocating_chunk) {
 		info->force_alloc = 1;
 		info->allocating_chunk = 1;
-		init_waitqueue_head(&info->allocate_wait);
 	} else {
 		wait = true;
 	}
@@ -3112,7 +3111,7 @@ int btrfs_reserve_metadata_for_delalloc(struct btrfs_root *root,
 		return -ENOSPC;
 	}
 
-	BTRFS_I(inode)->reserved_extents++;
+	BTRFS_I(inode)->reserved_extents += num_items;
 	check_force_delalloc(meta_sinfo);
 	spin_unlock(&meta_sinfo->lock);
 
@@ -4171,6 +4170,10 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 	ins->offset = 0;
 
 	space_info = __find_space_info(root->fs_info, data);
+	if (!space_info) {
+		printk(KERN_ERR "No space info for %d\n", data);
+		return -ENOSPC;
+	}
 
 	if (orig_root->ref_cows || empty_size)
 		allowed_chunk_alloc = 1;
@@ -5206,6 +5209,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 	next = btrfs_find_tree_block(root, bytenr, blocksize);
 	if (!next) {
 		next = btrfs_find_create_tree_block(root, bytenr, blocksize);
+		if (!next)
+			return -ENOMEM;
 		reada = 1;
 	}
 	btrfs_tree_lock(next);
@@ -5418,7 +5423,8 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
 		if (ret > 0) {
 			path->slots[level]++;
 			continue;
-		}
+		} else if (ret < 0)
+			return ret;
 		level = wc->level;
 	}
 	return 0;
@@ -7370,7 +7376,6 @@ static int find_first_block_group(struct btrfs_root *root,
 		}
 		path->slots[0]++;
 	}
-	ret = -ENOENT;
 out:
 	return ret;
 }

fs/btrfs/extent_io.c

@@ -2678,33 +2678,20 @@ int extent_readpages(struct extent_io_tree *tree,
 {
 	struct bio *bio = NULL;
 	unsigned page_idx;
-	struct pagevec pvec;
 	unsigned long bio_flags = 0;
 
-	pagevec_init(&pvec, 0);
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		struct page *page = list_entry(pages->prev, struct page, lru);
 
 		prefetchw(&page->flags);
 		list_del(&page->lru);
-		/*
-		 * what we want to do here is call add_to_page_cache_lru,
-		 * but that isn't exported, so we reproduce it here
-		 */
-		if (!add_to_page_cache(page, mapping,
+		if (!add_to_page_cache_lru(page, mapping,
 					page->index, GFP_KERNEL)) {
-
-			/* open coding of lru_cache_add, also not exported */
-			page_cache_get(page);
-			if (!pagevec_add(&pvec, page))
-				__pagevec_lru_add_file(&pvec);
 			__extent_read_full_page(tree, page, get_extent,
 						&bio, 0, &bio_flags);
 		}
 		page_cache_release(page);
 	}
-	if (pagevec_count(&pvec))
-		__pagevec_lru_add_file(&pvec);
 	BUG_ON(!list_empty(pages));
 	if (bio)
 		submit_one_bio(READ, bio, 0, bio_flags);

fs/btrfs/inode.c

@@ -797,7 +797,7 @@ static noinline int cow_file_range(struct inode *inode,
 	while (disk_num_bytes > 0) {
 		unsigned long op;
 
-		cur_alloc_size = min(disk_num_bytes, root->fs_info->max_extent);
+		cur_alloc_size = disk_num_bytes;
 		ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
 					   root->sectorsize, 0, alloc_hint,
 					   (u64)-1, &ins, 1);
@@ -1228,30 +1228,9 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page,
 static int btrfs_split_extent_hook(struct inode *inode,
 				    struct extent_state *orig, u64 split)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	u64 size;
-
 	if (!(orig->state & EXTENT_DELALLOC))
 		return 0;
 
-	size = orig->end - orig->start + 1;
-	if (size > root->fs_info->max_extent) {
-		u64 num_extents;
-		u64 new_size;
-
-		new_size = orig->end - split + 1;
-		num_extents = div64_u64(size + root->fs_info->max_extent - 1,
-					root->fs_info->max_extent);
-
-		/*
-		 * if we break a large extent up then leave oustanding_extents
-		 * be, since we've already accounted for the large extent.
-		 */
-		if (div64_u64(new_size + root->fs_info->max_extent - 1,
-			      root->fs_info->max_extent) < num_extents)
-			return 0;
-	}
-
 	spin_lock(&BTRFS_I(inode)->accounting_lock);
 	BTRFS_I(inode)->outstanding_extents++;
 	spin_unlock(&BTRFS_I(inode)->accounting_lock);
@@ -1269,38 +1248,10 @@ static int btrfs_merge_extent_hook(struct inode *inode,
 				   struct extent_state *new,
 				   struct extent_state *other)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	u64 new_size, old_size;
-	u64 num_extents;
-
 	/* not delalloc, ignore it */
 	if (!(other->state & EXTENT_DELALLOC))
 		return 0;
 
-	old_size = other->end - other->start + 1;
-	if (new->start < other->start)
-		new_size = other->end - new->start + 1;
-	else
-		new_size = new->end - other->start + 1;
-
-	/* we're not bigger than the max, unreserve the space and go */
-	if (new_size <= root->fs_info->max_extent) {
-		spin_lock(&BTRFS_I(inode)->accounting_lock);
-		BTRFS_I(inode)->outstanding_extents--;
-		spin_unlock(&BTRFS_I(inode)->accounting_lock);
-		return 0;
-	}
-
-	/*
-	 * If we grew by another max_extent, just return, we want to keep that
-	 * reserved amount.
-	 */
-	num_extents = div64_u64(old_size + root->fs_info->max_extent - 1,
-				root->fs_info->max_extent);
-	if (div64_u64(new_size + root->fs_info->max_extent - 1,
-		      root->fs_info->max_extent) > num_extents)
-		return 0;
-
 	spin_lock(&BTRFS_I(inode)->accounting_lock);
 	BTRFS_I(inode)->outstanding_extents--;
 	spin_unlock(&BTRFS_I(inode)->accounting_lock);
@@ -1329,6 +1280,7 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
 		BTRFS_I(inode)->outstanding_extents++;
 		spin_unlock(&BTRFS_I(inode)->accounting_lock);
 		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;
@@ -1357,6 +1309,7 @@ static int btrfs_clear_bit_hook(struct inode *inode,
 
 		if (bits & EXTENT_DO_ACCOUNTING) {
 			spin_lock(&BTRFS_I(inode)->accounting_lock);
+			WARN_ON(!BTRFS_I(inode)->outstanding_extents);
 			BTRFS_I(inode)->outstanding_extents--;
 			spin_unlock(&BTRFS_I(inode)->accounting_lock);
 			btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
@@ -5385,7 +5338,6 @@ void btrfs_destroy_inode(struct inode *inode)
 void btrfs_drop_inode(struct inode *inode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-
 	if (inode->i_nlink > 0 && btrfs_root_refs(&root->root_item) == 0)
 		generic_delete_inode(inode);
 	else
@@ -5789,18 +5741,15 @@ static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key ins;
-	u64 alloc_size;
 	u64 cur_offset = start;
 	u64 num_bytes = end - start;
 	int ret = 0;
 	u64 i_size;
 
 	while (num_bytes > 0) {
-		alloc_size = min(num_bytes, root->fs_info->max_extent);
-
 		trans = btrfs_start_transaction(root, 1);
 
-		ret = btrfs_reserve_extent(trans, root, alloc_size,
+		ret = btrfs_reserve_extent(trans, root, num_bytes,
 					   root->sectorsize, 0, alloc_hint,
 					   (u64)-1, &ins, 1);
 		if (ret) {