btrfs: remove unused logic of limiting async delalloc pages

A proposed patch in https://marc.info/?l=linux-btrfs&m=147859791003837
pointed out bad limit threshold in cow_file_range_async, but it turned
out that the whole logic is not necessary and is done by writeback. We
agreed to remove it.
Signed-off-by: NDavid Sterba <dsterba@suse.com>

fs/btrfs/inode.c

@@ -1167,7 +1167,6 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page,
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	unsigned long nr_pages;
 	u64 cur_end;
-	int limit = 10 * SZ_1M;
 
 	clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED,
 			 1, 0, NULL, GFP_NOFS);
@@ -1199,12 +1198,6 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page,
 
 		btrfs_queue_work(fs_info->delalloc_workers, &async_cow->work);
 
-		if (atomic_read(&fs_info->async_delalloc_pages) > limit) {
-			wait_event(fs_info->async_submit_wait,
-				   (atomic_read(&fs_info->async_delalloc_pages) <
-				    limit));
-		}
-
 		while (atomic_read(&fs_info->async_submit_draining) &&
 		       atomic_read(&fs_info->async_delalloc_pages)) {
 			wait_event(fs_info->async_submit_wait,