nilfs2: ensure to clear dirty state when deleting metadata file block

This would fix the following failure during GC:

 nilfs_cpfile_delete_checkpoints: cannot delete block
 NILFS: GC failed during preparation: cannot delete checkpoints: err=-2

The problem was caused by a break in state consistency between page
cache and btree; the above block was removed from the btree but the
page buffering the block was remaining in the page cache in dirty
state.

This resolves the inconsistency by ensuring to clear dirty state of
the page buffering the deleted block.

Reported-by: David Arendt <admin@prnet.org>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

fs/nilfs2/mdt.c

@@ -300,7 +300,7 @@ int nilfs_mdt_delete_block(struct inode *inode, unsigned long block)
 	int err;
 
 	err = nilfs_bmap_delete(ii->i_bmap, block);
-	if (likely(!err)) {
+	if (!err || err == -ENOENT) {
 		nilfs_mdt_mark_dirty(inode);
 		nilfs_mdt_forget_block(inode, block);
 	}

fs/nilfs2/page.c

@@ -128,7 +128,8 @@ void nilfs_forget_buffer(struct buffer_head *bh)
 
 	lock_buffer(bh);
 	clear_buffer_nilfs_volatile(bh);
-	if (test_clear_buffer_dirty(bh) && nilfs_page_buffers_clean(page))
+	clear_buffer_dirty(bh);
+	if (nilfs_page_buffers_clean(page))
 		__nilfs_clear_page_dirty(page);
 
 	clear_buffer_uptodate(bh);