[llvm-commits] [llvm] r137484 - /llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp
Akira Hatanaka
ahatanak at gmail.com
Fri Aug 12 11:09:59 PDT 2011
Author: ahatanak
Date: Fri Aug 12 13:09:59 2011
New Revision: 137484
URL: http://llvm.org/viewvc/llvm-project?rev=137484&view=rev
Log:
When constant double 0.0 is lowered, make sure 0 is copied directly from an
integer register to a floating point register. It is not valid to interpret
the value of a floating pointer register as part of a double precision
floating point value after a single precision floating point computational
or move instruction stores its result to the register.
- In the test case, the following code is generated before this patch is
applied:
mtc1 $zero, $f2 ; unformatted copy to $f2
mov.s $f0, $f2 ; $f0 is in single format
sdc1 $f12, 0($sp)
mov.s $f1, $f2 ; $f1 is in single format
c.eq.d $f12, $f0 ; $f0 cannot be interpreted as double
- The following code is generated after this patch is applied:
mtc1 $zero, $f0 ; unformatted copy to $f0
mtc1 $zero, $f1 ; unformatted copy to $f1
c.eq.d $f12, $f0 ; $f0 can be interpreted as double
Bhanu Chetlapalli and Chris Dearman at MIPS technologies reported this bug and
provided the test case.
Modified:
llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp
Modified: llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp?rev=137484&r1=137483&r2=137484&view=diff
==============================================================================
--- llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp (original)
+++ llvm/trunk/lib/Target/Mips/MipsISelDAGToDAG.cpp Fri Aug 12 13:09:59 2011
@@ -417,15 +417,8 @@
if (Node->getValueType(0) == MVT::f64 && CN->isExactlyValue(+0.0)) {
SDValue Zero = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), dl,
Mips::ZERO, MVT::i32);
- SDValue Undef = SDValue(
- CurDAG->getMachineNode(TargetOpcode::IMPLICIT_DEF, dl, MVT::f64), 0);
- SDNode *MTC = CurDAG->getMachineNode(Mips::MTC1, dl, MVT::f32, Zero);
- SDValue I0 = CurDAG->getTargetInsertSubreg(Mips::sub_fpeven, dl,
- MVT::f64, Undef, SDValue(MTC, 0));
- SDValue I1 = CurDAG->getTargetInsertSubreg(Mips::sub_fpodd, dl,
- MVT::f64, I0, SDValue(MTC, 0));
- ReplaceUses(SDValue(Node, 0), I1);
- return I1.getNode();
+ return CurDAG->getMachineNode(Mips::BuildPairF64, dl, MVT::f64, Zero,
+ Zero);
}
break;
}
More information about the llvm-commits
mailing list