rt_timer_list_next_timeout() may return error?

If a timer has a timeout tick 0xFFFFFFFF(RT_TICK_MAX), the timer thread will suspend itself. And if the rt_tick now is 0xFFFFFFFF(RT_TICK_MAX), timer callback should be called immediately.

  /* the fist timer always in the last row */
  static rt_tick_t rt_timer_list_next_timeout(rt_list_t timer_list[])
  {
      struct rt_timer *timer;
      register rt_base_t level;
      rt_tick_t timeout_tick = RT_TICK_MAX;
  
      /* disable interrupt */
      level = rt_hw_interrupt_disable();
  
      if (!rt_list_isempty(&timer_list[RT_TIMER_SKIP_LIST_LEVEL - 1]))
      {
          timer = rt_list_entry(timer_list[RT_TIMER_SKIP_LIST_LEVEL - 1].next,
                                struct rt_timer, row[RT_TIMER_SKIP_LIST_LEVEL - 1]);
          timeout_tick = timer->timeout_tick;
      }
  
      /* enable interrupt */
      rt_hw_interrupt_enable(level);
  
      return timeout_tick;
  }

        /* get the next timeout tick */
        next_timeout = rt_timer_list_next_timeout(rt_soft_timer_list);
        if (next_timeout == RT_TICK_MAX)
        {
            /* no software timer exist, suspend self. */
            rt_thread_suspend(rt_thread_self());
            rt_schedule();
        }

Is that OK?

I suppose that the API should be "rt_err_t rt_timer_list_next_timeout(rt_list_t timer_list[], rt_tick_t *timeout_tick) " instead,

Thanks.

Start by inspecting rt_timer_list_next_timeout() and the software timer thread code shown in the issue. Trace the RT_TICK_MAX cases when a timer exists and when the current tick reaches that value, then determine the expected callback and suspension behavior. Done means the edge case is handled consistently and the API behavior is clearly validated.