Author: hans Date: Wed Aug 8 04:31:39 2018 New Revision: 339234 URL: http://llvm.org/viewvc/llvm-project?rev=339234&view=rev Log: Merging r338716: ------------------------------------------------------------------------ r338716 | spatel | 2018-08-02 15:46:20 +0200 (Thu, 02 Aug 2018) | 41 lines
[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 ------------------------------------------------------------------------ Modified: llvm/branches/release_70/ (props changed) llvm/branches/release_70/lib/Analysis/ValueTracking.cpp llvm/branches/release_70/test/Transforms/InstSimplify/floating-point-compare.ll Propchange: llvm/branches/release_70/ ------------------------------------------------------------------------------ --- svn:mergeinfo (original) +++ svn:mergeinfo Wed Aug 8 04:31:39 2018 @@ -1,3 +1,3 @@ /llvm/branches/Apple/Pertwee:110850,110961 /llvm/branches/type-system-rewrite:133420-134817 -/llvm/trunk:155241,338552,338554,338569,338599,338610,338658,338665,338682,338703,338709,338751,338762,338817,338915,338968 +/llvm/trunk:155241,338552,338554,338569,338599,338610,338658,338665,338682,338703,338709,338716,338751,338762,338817,338915,338968 Modified: llvm/branches/release_70/lib/Analysis/ValueTracking.cpp URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_70/lib/Analysis/ValueTracking.cpp?rev=339234&r1=339233&r2=339234&view=diff ============================================================================== --- llvm/branches/release_70/lib/Analysis/ValueTracking.cpp (original) +++ llvm/branches/release_70/lib/Analysis/ValueTracking.cpp Wed Aug 8 04:31:39 2018 @@ -2817,10 +2817,13 @@ static bool cannotBeOrderedLessThanZeroI default: break; case Intrinsic::maxnum: - return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly, - Depth + 1) || - cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly, - Depth + 1); + return (isKnownNeverNaN(I->getOperand(0)) && + cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, + SignBitOnly, Depth + 1)) || + (isKnownNeverNaN(I->getOperand(1)) && + cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, + SignBitOnly, Depth + 1)); + case Intrinsic::minnum: return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly, Depth + 1) && Modified: llvm/branches/release_70/test/Transforms/InstSimplify/floating-point-compare.ll URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_70/test/Transforms/InstSimplify/floating-point-compare.ll?rev=339234&r1=339233&r2=339234&view=diff ============================================================================== --- llvm/branches/release_70/test/Transforms/InstSimplify/floating-point-compare.ll (original) +++ llvm/branches/release_70/test/Transforms/InstSimplify/floating-point-compare.ll Wed Aug 8 04:31:39 2018 @@ -266,13 +266,15 @@ define i1 @orderedLessZeroMinNum(float, ret i1 %uge } -; FIXME: This is wrong. ; 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. define i1 @orderedLessZeroMaxNum(float, float) { ; 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) %b = call float @llvm.maxnum.f32(float %a, float %1) _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits