RC_Channel: prevent float exception with bad RCn_MIN/MAX/TRIM

thanks to Michael for finding this

libraries/RC_Channel/RC_Channel.cpp

@@ -370,8 +370,14 @@ RC_Channel::norm_input()
     float ret;
     int16_t reverse_mul = (_reverse==-1?-1:1);
     if (radio_in < radio_trim) {
+        if (radio_min >= radio_trim) {
+            return 0.0f;
+        }
         ret = reverse_mul * (float)(radio_in - radio_trim) / (float)(radio_trim - radio_min);
     } else {
+        if (radio_max <= radio_trim) {
+            return 0.0f;
+        }
         ret = reverse_mul * (float)(radio_in - radio_trim) / (float)(radio_max  - radio_trim);
     }
     return constrain_float(ret, -1.0f, 1.0f);