Re: [PATCH/RFC 1/2] jump label: make enable/disable o(1)

[Jason Baron <jbaron@redhat.com>]

On Fri, Dec 17, 2010 at 04:12:21PM -0500, Steven Rostedt wrote:
> > On 12/17/2010 12:07 PM, Jason Baron wrote:
> 
> > Not acceptable I would think.
> > 
> > How about:
> > 
> > union fubar {
> >     int key_as_non_atomic;
> >     atomic_t key_as_atomic;
> > };
> 
> I don't even like this union.
> 
> > 
> > Now explain the exact semantics of this thing including how you 
> > guarantee no conflicting accesses *ever* occur.
> 
> I don't like the mixed semantics at all.
> 
> > 
> > 
> > > So for when jump labels are disabled case we could have
> > > one struct:
> > >
> > > struct jump_label_key {
> 
> 	atomic_t state;
> 
> > > }
> > >
> > > and then we could then have (rough c code):
> > >
> > > jump_label_enable(struct jump_label_key *key)
> > > {
> 
> 	if (atomic_read(&key->state))
> 		return;
> 	atomic_inc(&key->state);
> 
> > > }
> > >
> > > jump_label_disable(struct jump_label_key *key)
> > > {
> 
> 	if (!atomic_read(&key->state))
> 		return;
> 	atomic_dec(&key->state);
> 	WARN_ON(atomic_read(&key->state);
> 
> > > }
> > >
> > > jump_label_inc(struct jump_label_key *key)
> > > {
> 
>  	atomic_inc(&key->state)
> 
> > > }
> > >
> > > jump_label_dec(struct jump_label_key *key)
> > > {
> 
>  	atomic_dec((&key->state)
> 
> > > }
> > >
> > > bool unlikely_switch(struct jump_label_key *key)
> > > {
> 
>  	if (atomic_read(&key->state))
> 
> > > 		return true;
> > > 	return false;
> > > }
> > >
> 

hmmm...we were trying to avoid having the atomic_read() for tracepoints
b/c of potential extra cost that Mathieu was concerned about.

> There, now you are guaranteed that you have proper semantics.
> 
> > >

The other issue here was that jump_label.h gets included by
asm/atomic.h, so there a dependency issue to be addressed here as
well....

thanks,

-Jason