OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 2
Code Issues Packages Projects Releases Wiki Activity

[ValueTracking] fix maxnum miscompile for cannotBeOrderedLessThanZero (PR37776) 

This adds the NAN checks suggested in PR37776:
https://bugs.llvm.org/show_bug.cgi?id=37776

If both operands to maxnum are NAN, that should get constant folded, so we don't 
have to handle that case. This is the same assumption as other FP ops in this
function. Returning 'false' is always conservatively correct.

Copying from the bug report:

Currently, we have this for "when is cannotBeOrderedLessThanZero 
(mustBePositiveOrNaN) true for maxnum":
               L
        -------------------
        | Pos | Neg | NaN |
   ------------------------
   |Pos |  x  |  x  |  x  |
   ------------------------
 R |Neg |  x  |     |  x  |
   ------------------------
   |NaN |  x  |  x  |  x  |
   ------------------------


The cases with (Neg & NaN) are wrong. We should have:

                L
        -------------------
        | Pos | Neg | NaN |
   ------------------------
   |Pos |  x  |  x  |  x  |
   ------------------------
 R |Neg |  x  |     |     |
   ------------------------
   |NaN |  x  |     |  x  |
   ------------------------

Differential Revision: https://reviews.llvm.org/D50081

llvm-svn: 338716
This commit is contained in:
Sanjay Patel 2018-08-02 13:46:20 +00:00
parent 1f3977a856
commit f9a0d593e9
2 changed files with 11 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
llvm/lib/Analysis/ValueTracking.cpp
View File

@ -2817,10 +2817,13 @@ static bool cannotBeOrderedLessThanZeroImpl(const Value *V,
default: default:
break; break;
case Intrinsic::maxnum: case Intrinsic::maxnum:
return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly, return (isKnownNeverNaN(I->getOperand(0)) &&
Depth + 1) || cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI,
cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly, SignBitOnly, Depth + 1)) ||
Depth + 1); (isKnownNeverNaN(I->getOperand(1)) &&
cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI,
SignBitOnly, Depth + 1));
case Intrinsic::minnum: case Intrinsic::minnum:
return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly, return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
Depth + 1) && Depth + 1) &&

6
llvm/test/Transforms/InstSimplify/floating-point-compare.ll
View File

@ -266,13 +266,15 @@ define i1 @orderedLessZeroMinNum(float, float) {
ret i1 %uge ret i1 %uge
} }
; FIXME: This is wrong.
; PR37776: https://bugs.llvm.org/show_bug.cgi?id=37776 ; PR37776: https://bugs.llvm.org/show_bug.cgi?id=37776
; exp() may return nan, leaving %1 as the unknown result, so we can't simplify. ; exp() may return nan, leaving %1 as the unknown result, so we can't simplify.
define i1 @orderedLessZeroMaxNum(float, float) { define i1 @orderedLessZeroMaxNum(float, float) {
; CHECK-LABEL: @orderedLessZeroMaxNum( ; CHECK-LABEL: @orderedLessZeroMaxNum(
; CHECK-NEXT: ret i1 true ; CHECK-NEXT: [[A:%.*]] = call float @llvm.exp.f32(float [[TMP0:%.*]])
; CHECK-NEXT: [[B:%.*]] = call float @llvm.maxnum.f32(float [[A]], float [[TMP1:%.*]])
; CHECK-NEXT: [[UGE:%.*]] = fcmp uge float [[B]], 0.000000e+00
; CHECK-NEXT: ret i1 [[UGE]]
; ;
%a = call float @llvm.exp.f32(float %0) %a = call float @llvm.exp.f32(float %0)
%b = call float @llvm.maxnum.f32(float %a, float %1) %b = call float @llvm.maxnum.f32(float %a, float %1)