Why does INSERT ... ON DUPLICATE KEY UPDATE affected rows reports 2?

If I execute the following statement:

INSERT INTO users(id, username) VALUES(102, 'test') ON DUPLICATE KEY UPDATE username='test';

And if the value for column id exists in DB, but the value for username is different, then the reported number of affected rows is 2. To be more precise:

• If the new record is an exact duplicate, and nothing is updated/inserted then the affected rows is 0
• If the id does not exist yet, then a new record is inserted, and affected rows is 1
• If the id exists, but the existing record is updated the value for affected rows is set to 2

What is the reason for this? Is this to provide information back of what action took place? Are there actually 2 rows being affected?

Tested on MariaDB 10.3.7 with

CREATE TABLE IF NOT EXISTS `users` (
    `id` int(11) NOT NULL AUTO_INCREMENT,
    `username` varchar(355) NOT NULL,
    PRIMARY KEY (`id`)
  ) ENGINE=InnoDB AUTO_INCREMENT=103 DEFAULT CHARSET=utf8mb4;

INSERT INTO `users` (`id`, `username`) VALUES (102, 'Some name');

INSERT... ON DUPLICATE KEY UPDATE executes two actions. You can demonstrate this with triggers that set session variables as a side-effect:

mysql> create trigger i before insert on users for each row set @i = true;

mysql> create trigger u before update on users for each row set @u = true;

mysql> INSERT INTO users(id, username) VALUES(102, 'test') ON DUPLICATE KEY UPDATE username='test';
Query OK, 2 rows affected (0.01 sec)

mysql> select @i, @u;
+------+------+
| @i   | @u   |
+------+------+
|    1 |    1 |
+------+------+