Re: pthread_mutex_unlock (was Re: sched_yield() makes OpenLDAP slow)

[Kyle Moffett]

Haven't you OpenLDAP guys realized that the pthread model you're  
actually looking for is this?  POSIX mutexes are not designed to  
mandate scheduling requirements *precisely* because this achieves  
your scheduling goals by explicitly stating what they are.

s: pthread_mutex_lock(&mutex);
s: pthread_cond_wait(&wake_slave, &mutex);

m: [do some work]
m: pthread_cond_signal(&wake_slave);
m: pthread_cond_wait(&wake_master, &mutex);

s: [return from pthread_cond_wait]
s: [do some work]
s: pthread_cond_signal(&wake_master);
s: pthread_cond_wait(&wake_slave, &mutex);

Of course, if that's the model you're looking for, you could always  
do this instead:

void master_func() {
	while (1) {
		[do some work]
		slave_func();
	}
}

void slave_func() {
	[do some work]
}

The semantics are effectively the same.

Cheers,
Kyle Moffett

--
Premature optimization is the root of all evil in programming
  -- C.A.R. Hoare



-
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/