Re: [PATCH] RT: Add priority-queuing and priority-inheritance to workqueue infrastructure

[Daniel Walker <dwalker@mvista.com>]

On Thu, 2007-08-02 at 00:32 +0400, Oleg Nesterov wrote:
> On 08/02, Oleg Nesterov wrote:
> >
> > And I don't understand why rt_mutex_setprio() is called just before
> > calling work->func(). This means that a high-priority work could
> > be delayed by the low-priority ->current_work.
> 
> Aha, I missed the rt_mutex_setprio() in insert_work().
> 
> This is not good either. Suppose we have
> 
> 	void work_handler(struct work_struct *self)
> 	{
> 		if (!try_lock()) {
> 			// try again later...
> 			queue_work(wq, self);
> 			return;
> 		}
> 
> 		do_the_work();
> 	}
> 
> What if that work was queued by the low-priority thread, and
> then a high-priority thread inserts a new work when work_handler()
> is running?

You mean the above queue_work(wq, self) would get an arbitrarily higher
priority vs. what it would normally? 

Yeah, I suppose we would want follow the priority inside the workqueue
thread only ..

Daniel