Skip to content

[ValueTracking] Don't take sign bit from NaN operands #157250

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Sep 16, 2025
Merged

[ValueTracking] Don't take sign bit from NaN operands #157250

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c4fcf16
[ValueTracking] Add pre-commit tests. NFC.
dtcxzyw Sep 6, 2025
6da191e
[ValueTracking] Don't take sign bit from NaN operands
dtcxzyw Sep 6, 2025
e6b8d2a
[ValueTracking] Add commuted tests. NFC.
dtcxzyw Sep 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions llvm/lib/Analysis/ValueTracking.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5058,6 +5058,11 @@ void computeKnownFPClass(const Value *V, const APInt &DemandedElts,
KnownRHS.isKnownNeverPosZero()) &&
(KnownLHS.isKnownNeverPosZero() ||
KnownRHS.isKnownNeverNegZero()))) {
// Don't take sign bit from NaN operands.
if (!KnownLHS.isKnownNeverNaN())
KnownLHS.SignBit = std::nullopt;
if (!KnownRHS.isKnownNeverNaN())
KnownRHS.SignBit = std::nullopt;
if ((IID == Intrinsic::maximum || IID == Intrinsic::maximumnum ||
IID == Intrinsic::maxnum) &&
(KnownLHS.SignBit == false || KnownRHS.SignBit == false))
Expand Down
15 changes: 15 additions & 0 deletions llvm/test/Transforms/InstCombine/is_fpclass.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3922,6 +3922,21 @@ define i1 @test_class_is_not_psub_pnorm_pinf__dynamic(float %arg) #3 {
ret i1 %class
}

; Make sure we don't take sign bit from NaN operands.

define i1 @minnum_qnan(i32 %x) {
; CHECK-LABEL: @minnum_qnan(
; CHECK-NEXT: entry:
; CHECK-NEXT: ret i1 true
;
entry:
%qnan_bits = or i32 %x, -5938
%qnan = bitcast i32 %qnan_bits to float
%min = call float @llvm.minnum.f32(float %qnan, float 0.000000e+00)
%test = call i1 @llvm.is.fpclass.f32(float %min, i32 64)
ret i1 %test
}

declare i1 @llvm.is.fpclass.f32(float, i32 immarg)
declare i1 @llvm.is.fpclass.f64(double, i32 immarg)
declare <2 x i1> @llvm.is.fpclass.v2f32(<2 x float>, i32 immarg)
Expand Down