RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-22 04:22:57 +02:00
Code Issues Projects Releases Wiki Activity

[SelectionDAG] remove div-by-zero constant folding restriction

Browse Source 
We don't have this restriction in IR, so it should not be here
either simply out of consistency. Code that wants to handle FP
exceptions is expected to use the 'strict' variants of these
nodes.

We don't get the frem case because frem by 0.0 produces NaN (invalid),
and that's the remaining check here (so the removed check for frem
was dead code AFAIK).

This is the only place in SDAG that uses "HasFPExceptions", so I
think we should remove that entirely as a follow-up patch.

llvm-svn: 359566
This commit is contained in:
Sanjay Patel 2019-04-30 14:37:15 +00:00
parent e3c3046708
commit eb060c9323
3 changed files with 5 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
lib/CodeGen/SelectionDAG/SelectionDAG.cpp
View File

@ -4816,17 +4816,13 @@ SDValue SelectionDAG::foldConstantFPMath(unsigned Opcode, const SDLoc &DL,
break;
case ISD::FDIV:
Status = C1.divide(C2, APFloat::rmNearestTiesToEven);
if (!HasFPExceptions || (Status != APFloat::opInvalidOp &&
Status != APFloat::opDivByZero)) {
if (!HasFPExceptions || Status != APFloat::opInvalidOp)
return getConstantFP(C1, DL, VT);
}
break;
case ISD::FREM:
Status = C1.mod(C2);
if (!HasFPExceptions || (Status != APFloat::opInvalidOp &&
Status != APFloat::opDivByZero)) {
if (!HasFPExceptions || Status != APFloat::opInvalidOp)
return getConstantFP(C1, DL, VT);
}
break;
case ISD::FCOPYSIGN:
C1.copySign(C2);
@ -4840,7 +4836,7 @@ SDValue SelectionDAG::foldConstantFPMath(unsigned Opcode, const SDLoc &DL,
// This can return overflow, underflow, or inexact; we don't care.
// FIXME need to be more flexible about rounding mode.
(void) C1.convert(EVTToAPFloatSemantics(VT), APFloat::rmNearestTiesToEven,
&Unused);
&Unused);
return getConstantFP(C1, DL, VT);
}

6
test/CodeGen/AArch64/fp-const-fold.ll
View File

@ -6,10 +6,8 @@
define double @constant_fold_fdiv_by_zero(double* %p) {
; CHECK-LABEL: constant_fold_fdiv_by_zero:
; CHECK: // %bb.0:
; CHECK-NEXT: mov x8, #1
; CHECK-NEXT: fmov d0, xzr
; CHECK-NEXT: fmov d1, x8
; CHECK-NEXT: fdiv d0, d1, d0
; CHECK-NEXT: mov x8, #9218868437227405312
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: ret
%r = fdiv double 4.940660e-324, 0.0
ret double %r

9
test/CodeGen/X86/fdiv-combine.ll
View File

@ -115,16 +115,7 @@ define float @div_select_constant_fold(i1 zeroext %arg) {
define float @div_select_constant_fold_zero(i1 zeroext %arg) {
; CHECK-LABEL: div_select_constant_fold_zero:
; CHECK: # %bb.0:
; CHECK-NEXT: testl %edi, %edi
; CHECK-NEXT: jne .LBB7_1
; CHECK-NEXT: # %bb.2:
; CHECK-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; CHECK-NEXT: jmp .LBB7_3
; CHECK-NEXT: .LBB7_1:
; CHECK-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; CHECK-NEXT: .LBB7_3:
; CHECK-NEXT: xorps %xmm1, %xmm1
; CHECK-NEXT: divss %xmm1, %xmm0
; CHECK-NEXT: retq
%tmp = select i1 %arg, float 5.000000e+00, float 6.000000e+00
%B2 = fdiv float %tmp, 0.000000e+00