CHG: a better micro second (us) spindely function. At average it has 8-10us delay, but its linear. Making error less when you want to wait longer..

SpinDelayCountUs(20) gives a delay of 28us.
SpinDelayCountUs(100) gives a delay of 110us.
SpinDelayCountUs(500) gives a delay of 508us.

armsrc/util.h

@@ -65,6 +65,7 @@ uint32_t RAMFUNC GetTickCount();
 void StartCountUS();
 uint32_t RAMFUNC GetCountUS();
 void ResetUSClock(void);
+void SpinDelayCountUs(uint32_t us);
 //uint32_t RAMFUNC GetDeltaCountUS();
 
 void StartCountSspClk();