Improve mulU8X16toH16 accordingly

src/modules/math.h

@@ -14,19 +14,17 @@ static inline uint16_t mulU8X16toH16(const uint8_t charIn1, const uint16_t intIn
     intRes = ((uint32_t)charIn1 * (uint32_t)intIn2) >> 8;
 #else
     asm volatile(
-        "clr r26        \n\t"
+        "mul %B2, %A1    \n\t"
-        "mul %A1, %B2   \n\t"
+        "movw %0, r0     \n\t"
-        "movw %A0, r0   \n\t"
+        "mul %A2, %A1    \n\t"
-        "mul %A1, %A2   \n\t"
+        "lsl r0          \n\t" //push MSB to carry for rounding
-        "add %A0, r1    \n\t"
+        "adc %A0, r1     \n\t" //add with carry (for rounding)
-        "adc %B0, r26   \n\t"
+        "clr r1          \n\t" //make r1 __zero_reg__ again
-        "lsr r0         \n\t"
+        "adc %B0, r1     \n\t" //propagate carry of addition (add 0 with carry)
-        "adc %A0, r26   \n\t"
-        "adc %B0, r26   \n\t"
-        "clr r1         \n\t"
         : "=&r"(intRes)
-        : "d"(charIn1), "d"(intIn2)
+        : "r"(charIn1), "r"(intIn2)
-        : "r26");
+        : "r0", "r1" //clobbers: Technically these are either scratch registers or always 0 registers, but I'm making sure the compiler knows just in case.
+    );
 #endif
     return intRes;
 }