Re: Multiplexed registers and invalidating the register cache

[Orjan Friberg <orjan.friberg@axis.com>]

Andrew Cagney wrote:
> 
> Oops, yes, pass the current_target (I ment that it shouldn't have a 
> parameter trying to tell the observers anything beyond the fact that the 
> target has changed).  (the sed script generates () instead of (void)).

Ok.

> There are architectures where registers live in memory so the 
> information is simply misleading.

Ok, I'm convinced.  (Thanks for clarifying.)

> Eventually, yes.  The two can co-habitate for a bit, first just:
> 
> - add the observer

Ok.  observer.texi gets the following lines added:

   @deftypefun void target_changed (struct target_ops *@var{current_target})
   The target's register contents has changed.
   @end deftypefun

(Proper patch delayed until I've got all the pieces in place.)

> - add code to frame.c and regcache.c to register themselves

I'm sorry; this is the part I don't get - my skull must be getting really thick. 
  Are you saying that frame.c and regcache.c should register an observer each 
for the target_changed event, like this (frame.c):

   void
   frame_observer_target_changed (struct target_ops *current_target)
   {
     flush_cached_frames ();
   }
   observer_attach_target_changed (frame_observer_target_changed);

and (regcache.c):

   void
   regcache_observer_target_changed (struct target_ops *current_target)
   {
     registers_changed ();
   }
   observer_attach_target_changed (regcache_observer_target_changed);

> - add code to the problem area to trigger the observer

Ok; something like (example from valops.c):

         if (deprecated_register_changed_hook)
           deprecated_register_changed_hook (-1);
         target_changed_event ();
+       observer_notify_target_changed (current_target);
         break;
        }

-- 
Orjan Friberg
Axis Communications