forked from Imagelibrary/rtems
bsps/cache: Optimize smp_cache_broadcast()
Directly call the handler on the executing processor instead of doing this indirectly via a per-CPU job.
This commit is contained in:
3
bsps/shared/cache/cacheimpl.h
vendored
3
bsps/shared/cache/cacheimpl.h
vendored
@@ -299,7 +299,8 @@ static void smp_cache_broadcast( SMP_Action_handler handler, void *arg )
|
||||
cpu_self = _Per_CPU_Get();
|
||||
}
|
||||
|
||||
_SMP_Broadcast_action( handler, arg );
|
||||
( *handler )( arg );
|
||||
_SMP_Othercast_action( handler, arg );
|
||||
|
||||
if ( isr_level == 0 ) {
|
||||
_Thread_Dispatch_enable( cpu_self );
|
||||
|
||||
Reference in New Issue
Block a user