Btrfs: Index extent buffers in an rbtree

Before, extent buffers were a temporary object, meant to map a number of pages
at once and collect operations on them.

But, a few extra fields have crept in, and they are also the best place to
store a per-tree block lock field as well.  This commit puts the extent
buffers into an rbtree, and ensures a single extent buffer for each
tree block.
Signed-off-by: Chris Mason <chris.mason@oracle.com>

fs/btrfs/disk-io.c

@@ -381,7 +381,6 @@ int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 
 	end = min_t(u64, eb->len, PAGE_CACHE_SIZE);
 	end = eb->start + end - 1;
-	release_extent_buffer_tail_pages(eb);
 err:
 	free_extent_buffer(eb);
 out:
@@ -563,21 +562,21 @@ static int btree_releasepage(struct page *page, gfp_t gfp_flags)
 	struct extent_map_tree *map;
 	int ret;
 
-	if (page_count(page) > 3) {
-		/* once for page->private, once for the caller, once
-		 * once for the page cache
-		 */
-		return 0;
-	}
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
 	map = &BTRFS_I(page->mapping->host)->extent_tree;
+
 	ret = try_release_extent_state(map, tree, page, gfp_flags);
+	if (!ret) {
+		return 0;
+	}
+
+	ret = try_release_extent_buffer(tree, page);
 	if (ret == 1) {
-		invalidate_extent_lru(tree, page_offset(page), PAGE_CACHE_SIZE);
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
 		page_cache_release(page);
 	}
+
 	return ret;
 }
 
@@ -588,7 +587,8 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	extent_invalidatepage(tree, page, offset);
 	btree_releasepage(page, GFP_NOFS);
 	if (PagePrivate(page)) {
-		invalidate_extent_lru(tree, page_offset(page), PAGE_CACHE_SIZE);
+		printk("warning page private not zero on page %Lu\n",
+		       page_offset(page));
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
 		page_cache_release(page);
@@ -1456,7 +1456,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	free_extent_buffer(tree_root->node);
 fail_sys_array:
 fail_sb_buffer:
-	extent_io_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->io_tree);
 	btrfs_stop_workers(&fs_info->fixup_workers);
 	btrfs_stop_workers(&fs_info->workers);
 	btrfs_stop_workers(&fs_info->endio_workers);
@@ -1705,13 +1704,6 @@ int close_ctree(struct btrfs_root *root)
 
 	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
 
-	extent_io_tree_empty_lru(&fs_info->free_space_cache);
-	extent_io_tree_empty_lru(&fs_info->block_group_cache);
-	extent_io_tree_empty_lru(&fs_info->pinned_extents);
-	extent_io_tree_empty_lru(&fs_info->pending_del);
-	extent_io_tree_empty_lru(&fs_info->extent_ins);
-	extent_io_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->io_tree);
-
 	truncate_inode_pages(fs_info->btree_inode->i_mapping, 0);
 
 	btrfs_stop_workers(&fs_info->fixup_workers);

fs/btrfs/extent_io.h

@@ -54,13 +54,12 @@ struct extent_io_ops {
 
 struct extent_io_tree {
 	struct rb_root state;
+	struct rb_root buffer;
 	struct address_space *mapping;
 	u64 dirty_bytes;
 	spinlock_t lock;
+	spinlock_t buffer_lock;
 	struct extent_io_ops *ops;
-	spinlock_t lru_lock;
-	struct list_head buffer_lru;
-	int lru_size;
 	struct extent_state *last;
 };
 
@@ -87,10 +86,10 @@ struct extent_buffer {
 	unsigned long map_start;
 	unsigned long map_len;
 	struct page *first_page;
-	struct list_head lru;
 	atomic_t refs;
 	int flags;
 	struct list_head leak_list;
+	struct rb_node rb_node;
 };
 
 struct extent_map_tree;
@@ -112,10 +111,10 @@ typedef struct extent_map *(get_extent_t)(struct inode *inode,
 
 void extent_io_tree_init(struct extent_io_tree *tree,
 			  struct address_space *mapping, gfp_t mask);
-void extent_io_tree_empty_lru(struct extent_io_tree *tree);
 int try_release_extent_mapping(struct extent_map_tree *map,
 			       struct extent_io_tree *tree, struct page *page,
 			       gfp_t mask);
+int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page);
 int try_release_extent_state(struct extent_map_tree *map,
 			     struct extent_io_tree *tree, struct page *page,
 			     gfp_t mask);
@@ -241,8 +240,6 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset,
 		      unsigned long *map_start,
 		      unsigned long *map_len, int km);
 void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km);
-int invalidate_extent_lru(struct extent_io_tree *tree, u64 start,
-			  unsigned long len);
 int release_extent_buffer_tail_pages(struct extent_buffer *eb);
 int extent_range_uptodate(struct extent_io_tree *tree,
 			  u64 start, u64 end);

fs/btrfs/inode.c

@@ -2670,7 +2670,6 @@ static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags)
 	map = &BTRFS_I(page->mapping->host)->extent_tree;
 	ret = try_release_extent_mapping(map, tree, page, gfp_flags);
 	if (ret == 1) {
-		invalidate_extent_lru(tree, page_offset(page), PAGE_CACHE_SIZE);
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
 		page_cache_release(page);
@@ -2721,8 +2720,6 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset)
 
 	ClearPageChecked(page);
 	if (PagePrivate(page)) {
-		invalidate_extent_lru(tree, page_offset(page),
-				      PAGE_CACHE_SIZE);
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
 		page_cache_release(page);