MIPS64: [turbofan] Implemented the TruncateFloat64ToUint64 TurboFan operator.

src/mips64/macro-assembler-mips64.cc

Index: src/mips64/macro-assembler-mips64.cc
diff --git a/src/mips64/macro-assembler-mips64.cc b/src/mips64/macro-assembler-mips64.cc
index c8ae983d502fa04f3723dc6dc195a960fdc68bd5..6307e526c83865d30cacfe6fd4248e48ad7337c0 100644
--- a/src/mips64/macro-assembler-mips64.cc
+++ b/src/mips64/macro-assembler-mips64.cc
@@ -1623,6 +1623,12 @@ void MacroAssembler::Trunc_uw_d(FPURegister fd,
   mtc1(t8, fd);
 }
 
+void MacroAssembler::Trunc_ul_d(FPURegister fd, FPURegister fs,
+                                FPURegister scratch) {
+  Trunc_ul_d(fs, t8, scratch);
+  dmtc1(t8, fd);
+}
+
 
 void MacroAssembler::Trunc_w_d(FPURegister fd, FPURegister fs) {
   trunc_w_d(fd, fs);
@@ -1677,6 +1683,37 @@ void MacroAssembler::Trunc_uw_d(FPURegister fd,
 }
 
 
+void MacroAssembler::Trunc_ul_d(FPURegister fd, Register rs,
+                                FPURegister scratch) {
+  DCHECK(!fd.is(scratch));
+  DCHECK(!rs.is(at));
+
+  // Load 2^63 into scratch as its float representation.
+  li(at, 0x43e0000000000000);
+  dmtc1(at, scratch);
+
+  // Test if scratch > fd.
+  // If fd < 2^63 we can convert it normally.
+  Label simple_convert, done;
+  BranchF(&simple_convert, NULL, lt, fd, scratch);
+
+  // First we subtract 2^63 from fd, then trunc it to rs
+  // and add 2^63 to rs.
+  sub_d(scratch, fd, scratch);
+  trunc_l_d(scratch, scratch);
+  dmfc1(rs, scratch);
+  Or(rs, rs, Operand(1UL << 63));

[Alan Li, 2015/11/20 13:35:22]

maybe splitting the Or into LI and ORI, and put ORI in the delay slot is better?

+  Branch(&done);
+
+  // Simple conversion.
+  bind(&simple_convert);
+  trunc_l_d(scratch, fd);
+  dmfc1(rs, scratch);
+
+  bind(&done);
+}
+
+
 void MacroAssembler::Madd_d(FPURegister fd, FPURegister fr, FPURegister fs,
     FPURegister ft, FPURegister scratch) {
   if (0) {  // TODO(plind): find reasonable arch-variant symbol names.