Fix a latent bug in the asm code.

I think that technically, using "*m0" instead of "*0" allows LLVM to pick
between using the same memory as output 0 and using a new memory location.
(So far I haven't been able to construct a testcase that actually breaks
because of this, though)

gen/asmstmt.cpp

@@ -395,7 +395,7 @@ assert(0);
             
             // Add input operand with same value, with original as "matching output".
             std::ostringstream ss;
-            ss << m_cns << (n + asmblock->outputcount);
+            ss << '*' << (n + asmblock->outputcount);
             input_constraints.push_front(ss.str());
             input_values.push_front(output_values[n]);
         }