mirror of
https://github.com/RPCS3/llvm.git
synced 2024-12-23 20:45:06 +00:00
Re-commit: [mips] abs.[ds], and neg.[ds] should be allowed regardless of -enable-no-nans-fp-math
Summary: They behave in accordance with the Has2008 and ABS2008 configuration bits of the processor which are used to select between the 1985 and 2008 versions of IEEE 754. In 1985 mode, these instructions are arithmetic (i.e. they raise invalid operation exceptions when given NaN), in 2008 mode they are non-arithmetic (i.e. they are copies). nmadd.[ds], and nmsub.[ds] are still subject to -enable-no-nans-fp-math because the ISA spec does not explicitly state that they obey Has2008 and ABS2008. Fixed the issue with the previous version of this patch (r205628). A pre-existing 'let Predicate =' statement was removing some predicates that were necessary for FP64 to behave correctly. Reviewers: matheusalmeida Reviewed By: matheusalmeida Differential Revision: http://llvm-reviews.chandlerc.com/D3274 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@205844 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Daniel Sanders
parent
7184f49f40
commit
e777fb4725
@ -245,11 +245,6 @@ MipsTargetLowering::MipsTargetLowering(MipsTargetMachine &TM)
|
||||
setOperationAction(ISD::FCOPYSIGN, MVT::f64, Custom);
|
||||
setOperationAction(ISD::FP_TO_SINT, MVT::i32, Custom);
|
||||
|
||||
if (!TM.Options.NoNaNsFPMath) {
|
||||
setOperationAction(ISD::FABS, MVT::f32, Custom);
|
||||
setOperationAction(ISD::FABS, MVT::f64, Custom);
|
||||
}
|
||||
|
||||
if (hasMips64()) {
|
||||
setOperationAction(ISD::GlobalAddress, MVT::i64, Custom);
|
||||
setOperationAction(ISD::BlockAddress, MVT::i64, Custom);
|
||||
@ -334,11 +329,6 @@ MipsTargetLowering::MipsTargetLowering(MipsTargetMachine &TM)
|
||||
setOperationAction(ISD::FREM, MVT::f32, Expand);
|
||||
setOperationAction(ISD::FREM, MVT::f64, Expand);
|
||||
|
||||
if (!TM.Options.NoNaNsFPMath) {
|
||||
setOperationAction(ISD::FNEG, MVT::f32, Expand);
|
||||
setOperationAction(ISD::FNEG, MVT::f64, Expand);
|
||||
}
|
||||
|
||||
setOperationAction(ISD::EH_RETURN, MVT::Other, Custom);
|
||||
|
||||
setOperationAction(ISD::VAARG, MVT::Other, Expand);
|
||||
@ -779,7 +769,6 @@ LowerOperation(SDValue Op, SelectionDAG &DAG) const
|
||||
case ISD::SETCC: return lowerSETCC(Op, DAG);
|
||||
case ISD::VASTART: return lowerVASTART(Op, DAG);
|
||||
case ISD::FCOPYSIGN: return lowerFCOPYSIGN(Op, DAG);
|
||||
case ISD::FABS: return lowerFABS(Op, DAG);
|
||||
case ISD::FRAMEADDR: return lowerFRAMEADDR(Op, DAG);
|
||||
case ISD::RETURNADDR: return lowerRETURNADDR(Op, DAG);
|
||||
case ISD::EH_RETURN: return lowerEH_RETURN(Op, DAG);
|
||||
@ -1771,65 +1760,6 @@ MipsTargetLowering::lowerFCOPYSIGN(SDValue Op, SelectionDAG &DAG) const {
|
||||
return lowerFCOPYSIGN32(Op, DAG, Subtarget->hasExtractInsert());
|
||||
}
|
||||
|
||||
static SDValue lowerFABS32(SDValue Op, SelectionDAG &DAG,
|
||||
bool HasExtractInsert) {
|
||||
SDValue Res, Const1 = DAG.getConstant(1, MVT::i32);
|
||||
SDLoc DL(Op);
|
||||
|
||||
// If operand is of type f64, extract the upper 32-bit. Otherwise, bitcast it
|
||||
// to i32.
|
||||
SDValue X = (Op.getValueType() == MVT::f32) ?
|
||||
DAG.getNode(ISD::BITCAST, DL, MVT::i32, Op.getOperand(0)) :
|
||||
DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32, Op.getOperand(0),
|
||||
Const1);
|
||||
|
||||
// Clear MSB.
|
||||
if (HasExtractInsert)
|
||||
Res = DAG.getNode(MipsISD::Ins, DL, MVT::i32,
|
||||
DAG.getRegister(Mips::ZERO, MVT::i32),
|
||||
DAG.getConstant(31, MVT::i32), Const1, X);
|
||||
else {
|
||||
SDValue SllX = DAG.getNode(ISD::SHL, DL, MVT::i32, X, Const1);
|
||||
Res = DAG.getNode(ISD::SRL, DL, MVT::i32, SllX, Const1);
|
||||
}
|
||||
|
||||
if (Op.getValueType() == MVT::f32)
|
||||
return DAG.getNode(ISD::BITCAST, DL, MVT::f32, Res);
|
||||
|
||||
SDValue LowX = DAG.getNode(MipsISD::ExtractElementF64, DL, MVT::i32,
|
||||
Op.getOperand(0), DAG.getConstant(0, MVT::i32));
|
||||
return DAG.getNode(MipsISD::BuildPairF64, DL, MVT::f64, LowX, Res);
|
||||
}
|
||||
|
||||
static SDValue lowerFABS64(SDValue Op, SelectionDAG &DAG,
|
||||
bool HasExtractInsert) {
|
||||
SDValue Res, Const1 = DAG.getConstant(1, MVT::i32);
|
||||
SDLoc DL(Op);
|
||||
|
||||
// Bitcast to integer node.
|
||||
SDValue X = DAG.getNode(ISD::BITCAST, DL, MVT::i64, Op.getOperand(0));
|
||||
|
||||
// Clear MSB.
|
||||
if (HasExtractInsert)
|
||||
Res = DAG.getNode(MipsISD::Ins, DL, MVT::i64,
|
||||
DAG.getRegister(Mips::ZERO_64, MVT::i64),
|
||||
DAG.getConstant(63, MVT::i32), Const1, X);
|
||||
else {
|
||||
SDValue SllX = DAG.getNode(ISD::SHL, DL, MVT::i64, X, Const1);
|
||||
Res = DAG.getNode(ISD::SRL, DL, MVT::i64, SllX, Const1);
|
||||
}
|
||||
|
||||
return DAG.getNode(ISD::BITCAST, DL, MVT::f64, Res);
|
||||
}
|
||||
|
||||
SDValue
|
||||
MipsTargetLowering::lowerFABS(SDValue Op, SelectionDAG &DAG) const {
|
||||
if (Subtarget->hasMips64() && (Op.getValueType() == MVT::f64))
|
||||
return lowerFABS64(Op, DAG, Subtarget->hasExtractInsert());
|
||||
|
||||
return lowerFABS32(Op, DAG, Subtarget->hasExtractInsert());
|
||||
}
|
||||
|
||||
SDValue MipsTargetLowering::
|
||||
lowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const {
|
||||
// check the depth
|
||||
|
||||
@ -322,14 +322,12 @@ let isPseudo = 1, isCodeGenOnly = 1 in {
|
||||
def PseudoCVT_D64_L : ABSS_FT<"", FGR64Opnd, GPR64Opnd, II_CVT>;
|
||||
}
|
||||
|
||||
let Predicates = [NoNaNsFPMath, HasStdEnc] in {
|
||||
def FABS_S : MMRel, ABSS_FT<"abs.s", FGR32Opnd, FGR32Opnd, II_ABS, fabs>,
|
||||
ABSS_FM<0x5, 16>;
|
||||
def FNEG_S : MMRel, ABSS_FT<"neg.s", FGR32Opnd, FGR32Opnd, II_NEG, fneg>,
|
||||
ABSS_FM<0x7, 16>;
|
||||
defm FABS : ABSS_M<"abs.d", II_ABS, fabs>, ABSS_FM<0x5, 17>;
|
||||
defm FNEG : ABSS_M<"neg.d", II_NEG, fneg>, ABSS_FM<0x7, 17>;
|
||||
}
|
||||
def FABS_S : MMRel, ABSS_FT<"abs.s", FGR32Opnd, FGR32Opnd, II_ABS, fabs>,
|
||||
ABSS_FM<0x5, 16>;
|
||||
def FNEG_S : MMRel, ABSS_FT<"neg.s", FGR32Opnd, FGR32Opnd, II_NEG, fneg>,
|
||||
ABSS_FM<0x7, 16>;
|
||||
defm FABS : ABSS_M<"abs.d", II_ABS, fabs>, ABSS_FM<0x5, 17>;
|
||||
defm FNEG : ABSS_M<"neg.d", II_NEG, fneg>, ABSS_FM<0x7, 17>;
|
||||
|
||||
def FSQRT_S : MMRel, ABSS_FT<"sqrt.s", FGR32Opnd, FGR32Opnd, II_SQRT_S, fsqrt>,
|
||||
ABSS_FM<0x4, 16>;
|
||||
|
||||
@ -176,8 +176,7 @@ def RelocStatic : Predicate<"TM.getRelocationModel() == Reloc::Static">,
|
||||
AssemblerPredicate<"FeatureMips32">;
|
||||
def RelocPIC : Predicate<"TM.getRelocationModel() == Reloc::PIC_">,
|
||||
AssemblerPredicate<"FeatureMips32">;
|
||||
def NoNaNsFPMath : Predicate<"TM.Options.NoNaNsFPMath">,
|
||||
AssemblerPredicate<"FeatureMips32">;
|
||||
def NoNaNsFPMath : Predicate<"TM.Options.NoNaNsFPMath">;
|
||||
def HasStdEnc : Predicate<"Subtarget.hasStandardEncoding()">,
|
||||
AssemblerPredicate<"!FeatureMips16">;
|
||||
def NotDSP : Predicate<"!Subtarget.hasDSP()">;
|
||||
|
||||
@ -1,21 +1,23 @@
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 | FileCheck %s -check-prefix=32
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32r2 | FileCheck %s -check-prefix=32R2
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64 -mattr=n64 | FileCheck %s -check-prefix=64
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64r2 -mattr=n64 | FileCheck %s -check-prefix=64R2
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 -enable-no-nans-fp-math | FileCheck %s -check-prefix=NO-NAN
|
||||
; Check that abs.[ds] is selected and does not depend on -enable-no-nans-fp-math
|
||||
; They obey the Has2008 and ABS2008 configuration bits which govern the
|
||||
; conformance to IEEE 754 (1985) and IEEE 754 (2008). When these bits are not
|
||||
; present, they confirm to 1985.
|
||||
; In 1985 mode, abs.[ds] are arithmetic (i.e. they raise invalid operation
|
||||
; exceptions when given NaN's). In 2008 mode, they are non-arithmetic (i.e.
|
||||
; they are copies and don't raise any exceptions).
|
||||
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32r2 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 -enable-no-nans-fp-math | FileCheck %s
|
||||
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64 -enable-no-nans-fp-math | FileCheck %s
|
||||
|
||||
define float @foo0(float %a) nounwind readnone {
|
||||
entry:
|
||||
|
||||
; 32: lui $[[T0:[0-9]+]], 32767
|
||||
; 32: ori $[[MSK0:[0-9]+]], $[[T0]], 65535
|
||||
; 32: and $[[AND:[0-9]+]], ${{[0-9]+}}, $[[MSK0]]
|
||||
; 32: mtc1 $[[AND]], $f0
|
||||
|
||||
; 32R2: ins $[[INS:[0-9]+]], $zero, 31, 1
|
||||
; 32R2: mtc1 $[[INS]], $f0
|
||||
|
||||
; NO-NAN: abs.s
|
||||
; CHECK-LABEL: foo0
|
||||
; CHECK: abs.s
|
||||
|
||||
%call = tail call float @fabsf(float %a) nounwind readnone
|
||||
ret float %call
|
||||
@ -26,24 +28,8 @@ declare float @fabsf(float) nounwind readnone
|
||||
define double @foo1(double %a) nounwind readnone {
|
||||
entry:
|
||||
|
||||
; 32: lui $[[T0:[0-9]+]], 32767
|
||||
; 32: ori $[[MSK0:[0-9]+]], $[[T0]], 65535
|
||||
; 32: and $[[AND:[0-9]+]], ${{[0-9]+}}, $[[MSK0]]
|
||||
; 32: mtc1 $[[AND]], $f1
|
||||
|
||||
; 32R2: ins $[[INS:[0-9]+]], $zero, 31, 1
|
||||
; 32R2: mtc1 $[[INS]], $f1
|
||||
|
||||
; 64: daddiu $[[T0:[0-9]+]], $zero, 1
|
||||
; 64: dsll $[[T1:[0-9]+]], ${{[0-9]+}}, 63
|
||||
; 64: daddiu $[[MSK0:[0-9]+]], $[[T1]], -1
|
||||
; 64: and $[[AND:[0-9]+]], ${{[0-9]+}}, $[[MSK0]]
|
||||
; 64: dmtc1 $[[AND]], $f0
|
||||
|
||||
; 64R2: dins $[[INS:[0-9]+]], $zero, 63, 1
|
||||
; 64R2: dmtc1 $[[INS]], $f0
|
||||
|
||||
; NO-NAN: abs.d
|
||||
; CHECK-LABEL: foo1:
|
||||
; CHECK: abs.d
|
||||
|
||||
%call = tail call double @fabs(double %a) nounwind readnone
|
||||
ret double %call
|
||||
|
||||
@ -1,3 +1,10 @@
|
||||
; Check that madd.[ds], msub.[ds], nmadd.[ds], and nmsub.[ds] are supported
|
||||
; correctly.
|
||||
; The spec for nmadd.[ds], and nmsub.[ds] does not state that they obey the
|
||||
; the Has2008 and ABS2008 configuration bits which govern the conformance to
|
||||
; IEEE 754 (1985) and IEEE 754 (2008). These instructions are therefore only
|
||||
; available when -enable-no-nans-fp-math is given.
|
||||
|
||||
; RUN: llc < %s -march=mipsel -mcpu=mips32r2 -enable-no-nans-fp-math | FileCheck %s -check-prefix=32R2 -check-prefix=CHECK
|
||||
; RUN: llc < %s -march=mips64el -mcpu=mips64r2 -mattr=n64 -enable-no-nans-fp-math | FileCheck %s -check-prefix=64R2 -check-prefix=CHECK
|
||||
; RUN: llc < %s -march=mipsel -mcpu=mips32r2 | FileCheck %s -check-prefix=32R2NAN -check-prefix=CHECK
|
||||
@ -5,6 +12,7 @@
|
||||
|
||||
define float @FOO0float(float %a, float %b, float %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO0float:
|
||||
; CHECK: madd.s
|
||||
%mul = fmul float %a, %b
|
||||
%add = fadd float %mul, %c
|
||||
@ -14,6 +22,7 @@ entry:
|
||||
|
||||
define float @FOO1float(float %a, float %b, float %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO1float:
|
||||
; CHECK: msub.s
|
||||
%mul = fmul float %a, %b
|
||||
%sub = fsub float %mul, %c
|
||||
@ -23,6 +32,7 @@ entry:
|
||||
|
||||
define float @FOO2float(float %a, float %b, float %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO2float:
|
||||
; 32R2: nmadd.s
|
||||
; 64R2: nmadd.s
|
||||
; 32R2NAN: madd.s
|
||||
@ -35,6 +45,7 @@ entry:
|
||||
|
||||
define float @FOO3float(float %a, float %b, float %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO3float:
|
||||
; 32R2: nmsub.s
|
||||
; 64R2: nmsub.s
|
||||
; 32R2NAN: msub.s
|
||||
@ -47,6 +58,7 @@ entry:
|
||||
|
||||
define double @FOO10double(double %a, double %b, double %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO10double:
|
||||
; CHECK: madd.d
|
||||
%mul = fmul double %a, %b
|
||||
%add = fadd double %mul, %c
|
||||
@ -56,6 +68,7 @@ entry:
|
||||
|
||||
define double @FOO11double(double %a, double %b, double %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO11double:
|
||||
; CHECK: msub.d
|
||||
%mul = fmul double %a, %b
|
||||
%sub = fsub double %mul, %c
|
||||
@ -65,6 +78,7 @@ entry:
|
||||
|
||||
define double @FOO12double(double %a, double %b, double %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO12double:
|
||||
; 32R2: nmadd.d
|
||||
; 64R2: nmadd.d
|
||||
; 32R2NAN: madd.d
|
||||
@ -77,6 +91,7 @@ entry:
|
||||
|
||||
define double @FOO13double(double %a, double %b, double %c) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-LABEL: FOO13double:
|
||||
; 32R2: nmsub.d
|
||||
; 64R2: nmsub.d
|
||||
; 32R2NAN: msub.d
|
||||
|
||||
@ -1,17 +1,30 @@
|
||||
; RUN: llc < %s -march=mipsel -mcpu=mips32 | FileCheck %s
|
||||
; Check that abs.[ds] is selected and does not depend on -enable-no-nans-fp-math
|
||||
; They obey the Has2008 and ABS2008 configuration bits which govern the
|
||||
; conformance to IEEE 754 (1985) and IEEE 754 (2008). When these bits are not
|
||||
; present, they confirm to 1985.
|
||||
; In 1985 mode, abs.[ds] are arithmetic (i.e. they raise invalid operation
|
||||
; exceptions when given NaN's). In 2008 mode, they are non-arithmetic (i.e.
|
||||
; they are copies and don't raise any exceptions).
|
||||
|
||||
define float @foo0(i32 %a, float %d) nounwind readnone {
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32r2 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mipsel-linux-gnu -mcpu=mips32 -enable-no-nans-fp-math | FileCheck %s
|
||||
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64 | FileCheck %s
|
||||
; RUN: llc < %s -mtriple=mips64el-linux-gnu -mcpu=mips64 -enable-no-nans-fp-math | FileCheck %s
|
||||
|
||||
define float @foo0(float %d) nounwind readnone {
|
||||
entry:
|
||||
; CHECK-NOT: neg.s
|
||||
; CHECK-LABEL: foo0:
|
||||
; CHECK: neg.s
|
||||
%sub = fsub float -0.000000e+00, %d
|
||||
ret float %sub
|
||||
}
|
||||
|
||||
define double @foo1(i32 %a, double %d) nounwind readnone {
|
||||
define double @foo1(double %d) nounwind readnone {
|
||||
entry:
|
||||
; CHECK: foo1
|
||||
; CHECK-NOT: neg.d
|
||||
; CHECK: jr
|
||||
; CHECK-LABEL: foo1:
|
||||
; CHECK: neg.d
|
||||
%sub = fsub double -0.000000e+00, %d
|
||||
ret double %sub
|
||||
}
|
||||
|
||||
Loading…
Reference in New Issue
Block a user