simplify _loop so that the "prologe" is shorter and the jump is back into

the first < 5 bytes

test_cases/backwards.h

@@ -1,12 +1,11 @@
 #pragma once
 extern "C" {
 	/**
-	* Raises @num @cnt times
+	* $$ x*x * (x-1)! $$
 	*
-	* @param num
 	* @param cnt
 	*/
-	uint32_t _loop(uint32_t num, uint32_t cnt);
+	uint32_t _loop(uint32_t x);
 
 	/**
 	* Computes factorial