----- Joel Sherrill <joel.sherr...@oarcorp.com> schrieb:
>
>
> On 3/18/2015 3:20 PM, Sebastian Huber wrote:
[...]
> > diff --git a/cpukit/score/src/threadchangepriority.c
> > b/cpukit/score/src/threadchangepriority.c
> > index d61dfb8..6ee65f5 100644
> > --- a/cpukit/score/src/threadchangepriority.c
> > +++ b/cpukit/score/src/threadchangepriority.c
> > @@ -29,29 +29,36 @@ void _Thread_Change_priority(
> > bool prepend_it
> > )
> > {
> > + ISR_Level level;
> > +
> > + _ISR_Disable( level );
> > +
> > /*
> > * Do not bother recomputing all the priority related information if
> > * we are not REALLY changing priority.
> > */
> > if ( the_thread->current_priority != new_priority ) {
> > - ISR_Level level;
> > -
> > - _ISR_Disable( level );
> > + uint32_t my_generation = the_thread->priority_generation + 1;
> >
> > the_thread->current_priority = new_priority;
> > + the_thread->priority_generation = my_generation;
> >
> > - if ( _States_Is_ready( the_thread->current_state ) ) {
> > - _Scheduler_Change_priority(
> > - the_thread,
> > - new_priority,
> > - prepend_it
> > - );
> > - } else {
> > - _Scheduler_Update_priority( the_thread, new_priority );
> > - }
> > + _Thread_queue_Requeue( the_thread->Wait.queue, the_thread );
> Can we get a comment above this call that _Thread_queue_Requeue
> checks that the thread is actually blocked on the specified queue so
> we don't need to check the thread state?
I want to get rid of this _Thread_queue_Requeue() in a follow up patch. Things
are getting much more compilicated on SMP since we now have to take the lock
order into account to avoid deadlocks.
_______________________________________________
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel
