MIPS: [turbofan] Properly implement Float64/32 Min/Max instructions.

src/compiler/machine-operator.h

Index: src/compiler/machine-operator.h
diff --git a/src/compiler/machine-operator.h b/src/compiler/machine-operator.h
index 2a80feecded6f01574f7c156228af6e41573c115..11f62b2e4b736c1e4f4c303a769200fa28f7fc03 100644
--- a/src/compiler/machine-operator.h
+++ b/src/compiler/machine-operator.h
@@ -236,10 +236,14 @@ class MachineOperatorBuilder final : public ZoneObject {
   // Floating point min/max complying to IEEE 754 (single-precision).
   const OptionalOperator Float32Max();
   const OptionalOperator Float32Min();
+  bool HasFloat32Max() const { return flags_ & kFloat32Max; }

[Benedikt Meurer, 2015/11/03 05:39:00]

You don't need these functions. Just use Float32Max().IsSupported() instead.

[dusan.milosavljevic, 2015/11/03 13:48:34]

Done.

[dusan.milosavljevic, 2015/11/03 13:48:34]

Right, exactly what I need. Thank you.

+  bool HasFloat64Max() const { return flags_ & kFloat64Max; }

[Benedikt Meurer, 2015/11/03 05:39:00]

Same as above.

 
   // Floating point min/max complying to IEEE 754 (double-precision).
   const OptionalOperator Float64Max();
   const OptionalOperator Float64Min();
+  bool HasFloat32Min() const { return flags_ & kFloat32Min; }

[Benedikt Meurer, 2015/11/03 05:39:00]

Same as above.

[dusan.milosavljevic, 2015/11/03 13:48:34]

Done.

+  bool HasFloat64Min() const { return flags_ & kFloat64Min; }
 
   // Floating point abs complying to IEEE 754 (single-precision).
   const Operator* Float32Abs();