bcachefs: Seralize btree_update operations at btree_update_nodes_written()

Prep work for journalling updates to interior nodes - enforcing ordering
will greatly simplify those changes.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

fs/bcachefs/bcachefs.h

@@ -610,6 +610,7 @@ struct bch_fs {
 
 	mempool_t		btree_interior_update_pool;
 	struct list_head	btree_interior_update_list;
+	struct list_head	btree_interior_updates_unwritten;
 	struct mutex		btree_interior_update_lock;
 	struct closure_waitlist	btree_interior_update_wait;
 

fs/bcachefs/btree_update_interior.c

@@ -666,9 +666,15 @@ static void btree_update_nodes_written(struct closure *cl)
 	 * to child nodes that weren't written yet: now, the child nodes have
 	 * been written so we can write out the update to the interior node.
 	 */
-retry:
 	mutex_lock(&c->btree_interior_update_lock);
 	as->nodes_written = true;
+retry:
+	as = list_first_entry_or_null(&c->btree_interior_updates_unwritten,
+				      struct btree_update, unwritten_list);
+	if (!as || !as->nodes_written) {
+		mutex_unlock(&c->btree_interior_update_lock);
+		return;
+	}
 
 	switch (as->mode) {
 	case BTREE_INTERIOR_NO_UPDATE:
@@ -681,11 +687,12 @@ static void btree_update_nodes_written(struct closure *cl)
 			mutex_unlock(&c->btree_interior_update_lock);
 			btree_node_lock_type(c, b, SIX_LOCK_read);
 			six_unlock_read(&b->c.lock);
+			mutex_lock(&c->btree_interior_update_lock);
 			goto retry;
 		}
 
 		BUG_ON(!btree_node_dirty(b));
-		closure_wait(&btree_current_write(b)->wait, cl);
+		closure_wait(&btree_current_write(b)->wait, &as->cl);
 
 		list_del(&as->write_blocked_list);
 
@@ -694,6 +701,8 @@ static void btree_update_nodes_written(struct closure *cl)
 		 * nodes to be writeable:
 		 */
 		closure_wake_up(&c->btree_interior_update_wait);
+
+		list_del(&as->unwritten_list);
 		mutex_unlock(&c->btree_interior_update_lock);
 
 		/*
@@ -702,6 +711,7 @@ static void btree_update_nodes_written(struct closure *cl)
 		 */
 		bch2_btree_node_write_cond(c, b, true);
 		six_unlock_read(&b->c.lock);
+		continue_at(&as->cl, btree_update_nodes_reachable, system_wq);
 		break;
 
 	case BTREE_INTERIOR_UPDATING_AS:
@@ -716,8 +726,12 @@ static void btree_update_nodes_written(struct closure *cl)
 		/*
 		 * and then we have to wait on that btree_update to finish:
 		 */
-		closure_wait(&as->parent_as->wait, cl);
+		closure_wait(&as->parent_as->wait, &as->cl);
+
+		list_del(&as->unwritten_list);
 		mutex_unlock(&c->btree_interior_update_lock);
+
+		continue_at(&as->cl, btree_update_nodes_reachable, system_wq);
 		break;
 
 	case BTREE_INTERIOR_UPDATING_ROOT:
@@ -728,6 +742,7 @@ static void btree_update_nodes_written(struct closure *cl)
 			mutex_unlock(&c->btree_interior_update_lock);
 			btree_node_lock_type(c, b, SIX_LOCK_read);
 			six_unlock_read(&b->c.lock);
+			mutex_lock(&c->btree_interior_update_lock);
 			goto retry;
 		}
 
@@ -744,6 +759,8 @@ static void btree_update_nodes_written(struct closure *cl)
 		 * can reuse the old nodes it'll have to do a journal commit:
 		 */
 		six_unlock_read(&b->c.lock);
+
+		list_del(&as->unwritten_list);
 		mutex_unlock(&c->btree_interior_update_lock);
 
 		/*
@@ -762,11 +779,12 @@ static void btree_update_nodes_written(struct closure *cl)
 
 		as->journal_seq = bch2_journal_last_unwritten_seq(&c->journal);
 
-		btree_update_wait_on_journal(cl);
-		return;
+		btree_update_wait_on_journal(&as->cl);
+		break;
 	}
 
-	continue_at(cl, btree_update_nodes_reachable, system_wq);
+	mutex_lock(&c->btree_interior_update_lock);
+	goto retry;
 }
 
 /*
@@ -778,6 +796,7 @@ static void btree_update_updated_node(struct btree_update *as, struct btree *b)
 	struct bch_fs *c = as->c;
 
 	mutex_lock(&c->btree_interior_update_lock);
+	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
 
 	BUG_ON(as->mode != BTREE_INTERIOR_NO_UPDATE);
 	BUG_ON(!btree_node_dirty(b));
@@ -858,6 +877,7 @@ static void btree_update_updated_root(struct btree_update *as)
 	struct btree_root *r = &c->btree_roots[as->btree_id];
 
 	mutex_lock(&c->btree_interior_update_lock);
+	list_add_tail(&as->unwritten_list, &c->btree_interior_updates_unwritten);
 
 	BUG_ON(as->mode != BTREE_INTERIOR_NO_UPDATE);
 

fs/bcachefs/btree_update_interior.h

@@ -55,6 +55,7 @@ struct btree_update {
 	struct bch_fs			*c;
 
 	struct list_head		list;
+	struct list_head		unwritten_list;
 
 	/* What kind of update are we doing? */
 	enum {

fs/bcachefs/super.c

@@ -642,6 +642,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts)
 	INIT_LIST_HEAD(&c->list);
 
 	INIT_LIST_HEAD(&c->btree_interior_update_list);
+	INIT_LIST_HEAD(&c->btree_interior_updates_unwritten);
 	mutex_init(&c->btree_reserve_cache_lock);
 	mutex_init(&c->btree_interior_update_lock);