Re: [TIP][RFC 6/7] futex: add requeue_pi calls

[Thomas Gleixner]

On Thu, 5 Mar 2009, Darren Hart wrote:
> int rt_mutex_start_proxy_lock(struct rt_mutex *lock,
> 			      struct rt_mutex_waiter *waiter,
> 			      struct task_struct *task, int detect_deadlock)
> {
> 	int ret;
> 
> 	spin_lock(&lock->wait_lock);
> 	ret = task_blocks_on_rt_mutex(lock, waiter, task, detect_deadlock);
> 
> 
> I add the following line to fix the bug.  Question is, should I use this
> atomic
> optimization here (under the lock->wait_lock) or should I just do "lock->owner
> |= RT_MUTEX_HAS_WAITERS" ?
> 
> =====>	mark_rt_mutex_waiters(lock);

This is still not enough as I explained in the review of the original
patch. What you need to do is:

	if (try_to_take_rt_mutex(lock, task)) {
	   	spin_unlock(&lock->wait_lock);
		/* The caller needs to wake up task, as it is now the owner */
		return WAKEIT;
	}

	ret = task_blocks_on_rt_mutex(lock, waiter, task, detect_deadlock);
 
> 	if (ret && !waiter->task) {
> 		/*
> 		 * Reset the return value. We might have
> 		 * returned with -EDEADLK and the owner
> 		 * released the lock while we were walking the
> 		 * pi chain.  Let the waiter sort it out.
> 		 */
> 		ret = 0;
> 	}
> 	spin_unlock(&lock->wait_lock);
> 
> 	debug_rt_mutex_print_deadlock(waiter);
> 
> 	return ret;
> }

Thanks,

	tglx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/