RE: FUSYN and RT

[Steven Rostedt <rostedt@goodmis.org>]

On Wed, 2005-04-13 at 10:33 -0700, Daniel Walker wrote:
> On Wed, 2005-04-13 at 08:46, Steven Rostedt wrote:
> > How hard would it be to use the RT mutex PI for the priority inheritance
> > for fusyn?  I only work with the RT mutex now and haven't looked at the
> > fusyn.  Maybe Ingo can make a separate PI system with its own API that
> > both the fusyn and RT mutex can use. This way the fusyn locks can still
> > be separate from the RT mutex locks but still work together. 
> > 
> > Basically can the fusyn work with the rt_mutex_waiter?  That's what I
> > would pull into its own subsystem.  Have another structure that would
> > reside in both the fusyn and RT mutex that would take over for the
> > current rt_mutex that is used in pi_setprio and task_blocks_on_lock in
> > rt.c.  So if both locks used the same PI system, then this should all be
> > cleared up. 
> > 
> > If this doesn't makes sense, or just confusing, I'll explain more :-)  
> 
> I've thought about this as an option, but when I first started this
> thread It seemed like the two could work independently, and safely which
> doesn't appear to be the case any more.
> 
> The problems with pulling out the PI in the RT mutex are that
> pi_setprio() does a walk over lock->owner and we're got two different
> lock structures now . I was thinking we could add something like
> lock_ops (get_owner(), wait_list_add(), wait_list_del(), ?? ) to
> rt_mutex_waiter, or abstract rt_lock. Then pi_setprio would just use the
> lock_ops instead of accessing a structure .. 

Yeah, I was thinking of another structure within rt_mutex and what fusyn
uses. Like a pi_struct of some sort. And this would be the common
structure that holds the owner and prio or what ever that the pi_setprio
and friends need.   This would probably be the easier approach, but for
the long run, I think I like your idea of the ops better.

-- Steve