release_sem_etc(): Reschedule only when a thread with a priority greater

than that of the current thread has been woken up. I didn't see the
reason why the thread should otherwise relinquish the rest of its
quantum. I noticed for instance that client and app server window
threads were ping-ponging more than seemed necessary. In most cases
when the client sent a port message it would be unscheduled although it
had run only for a few microseconds and had still stuff to do.
I measured a relatively Terminal-heavy "find /boot" (second run), which
does now take 5-10% less time.


git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@27314 a95241bf-73f2-0310-859d-f6bbb57e9c96

src/system/kernel/sem.cpp

@@ -919,7 +919,8 @@ release_sem_etc(sem_id id, int32 count, uint32 flags)
 		flags |= B_RELEASE_IF_WAITING_ONLY;
 	}
 
-	bool unblockedAny = false;
+	struct thread* currentThread = thread_get_current_thread();
+	bool reschedule = false;
 
 	SpinLocker threadLocker(gThreadSpinlock);
 
@@ -948,7 +949,7 @@ release_sem_etc(sem_id id, int32 count, uint32 flags)
 			sSems[slot].u.used.count += delta;
 			sSems[slot].u.used.net_count += delta - entry->count;
 			count -= delta;
-			unblockedAny = true;
+			reschedule |= entry->thread->priority > currentThread->priority;
 		} else {
 			// The thread is no longer waiting, but still queued, which
 			// means acquiration failed and we can just remove it.
@@ -966,7 +967,7 @@ release_sem_etc(sem_id id, int32 count, uint32 flags)
 
 	// If we've unblocked another thread reschedule, if we've not explicitly
 	// been told not to.
-	if (unblockedAny && (flags & B_DO_NOT_RESCHEDULE) == 0) {
+	if (reschedule && (flags & B_DO_NOT_RESCHEDULE) == 0) {
 		semLocker.Unlock();
 		threadLocker.Lock();
 		scheduler_reschedule();