Index: compiler-rt/lib/builtins/avr/mulqi3.S
===================================================================
--- compiler-rt/lib/builtins/avr/mulqi3.S
+++ compiler-rt/lib/builtins/avr/mulqi3.S
@@ -8,24 +8,46 @@
 //
 // The corresponding C code is something like:
 //
-// int __mulqi3(char A, char B) {
-//   return __mulhi3((int) A, (int) B);
+// char __mulqi3(char A, char B) {
+//   int S = 0;
+//   while (A != 0) {
+//     if (A & 1)
+//       S += B;
+//     B <<= 1;
+//     A = ((unsigned char) A) >> 1;
+//   }
+//   return S;
 // }
 //
+// __mulqi3 has special ABI, as the implementation of libgcc, the result is
+// returned via R24, while Rtmp and R22 are clobbered.
+//
 //===----------------------------------------------------------------------===//
 
 	.text
 	.align 2
 
+#ifdef __AVR_TINY__
+	.set __tmp_reg__, 16
+#else
+	.set __tmp_reg__, 0
+#endif
+
 	.globl __mulqi3
 	.type  __mulqi3, @function
 
 __mulqi3:
-	mov    r25, r24
-	lsl    r25
-	sbc    r25, r25         ; Promote A from char to int: `(int) A`.
-	mov    r23, r22
-	lsl    r23
-	sbc    r23, r23         ; Promote B from char to int: `(int) B`.
-	rcall  __mulhi3         ; `__mulhi3((int) A, (int) B);`.
-	ret
+	eor   __tmp_reg__, __tmp_reg__ ; S = 0;
+
+__mulqi3_loop:
+	cpi   r24, 0
+	breq  __mulqi3_end             ; while (A != 0) {
+	sbrc  r24, 0                   ;   if (A & 1)
+	add   __tmp_reg__, r22         ;     S += B;
+	add   r22, r22                 ;   B <<= 1;
+	lsr   r24                      ;   A = ((unsigned char) A) >> 1;
+	rjmp  __mulqi3_loop            ; }
+
+__mulqi3_end:
+	mov   r24, __tmp_reg__
+	ret                            ; return S;