Skip to content

[msan] Check mask and rounding mode in handleAVX512VectorConvertFPToInt #147782

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
thurstond merged 2 commits into from
Jul 9, 2025
Merged

[msan] Check mask and rounding mode in handleAVX512VectorConvertFPToInt #147782

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c47c743
[msan] Check mask and rounding mode in handleAVX512VectorConvertFPToInt
thurstond Jul 9, 2025
b4b203b
Add comment about rounding mode
thurstond Jul 9, 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
4 changes: 3 additions & 1 deletion llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4391,7 +4391,7 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
Value *A = I.getOperand(0);
Value *WriteThrough = I.getOperand(1);
Value *Mask = I.getOperand(2);
[[maybe_unused]] Value *RoundingMode = I.getOperand(3);
Value *RoundingMode = I.getOperand(3);

assert(isa<FixedVectorType>(A->getType()));
assert(A->getType()->isFPOrFPVectorTy());
Expand All @@ -4406,8 +4406,10 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {

assert(Mask->getType()->isIntegerTy());
assert(Mask->getType()->getScalarSizeInBits() == ANumElements);
insertCheckShadowOf(Mask, &I);

assert(RoundingMode->getType()->isIntegerTy());
insertCheckShadowOf(RoundingMode, &I);
Copy link
Contributor

@fmayer fmayer Jul 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just confirming that the implementation doesn't just look at a few bits of this?

Copy link
Contributor Author

@thurstond thurstond Jul 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think technically only 4 out of the 16 bits are used, though I would think mistakes were made if someone was intentionally only initializing the necessary bits. (I expect the rounding mode is more commonly a constant e.g., as seen in avx512-intrinsics.ll. In that case, the shadow check is actually elided.)

Copy link
Contributor

@fmayer fmayer Jul 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please leave a comment to that effect. So we remember in case your expectation turns out to be wrong.

Copy link
Contributor Author

@thurstond thurstond Jul 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done


assert(I.getType() == WriteThrough->getType());

Expand Down
7 changes: 7 additions & 0 deletions llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7941,13 +7941,20 @@ declare <16 x i32> @llvm.x86.avx512.mask.cvtps2dq.512(<16 x float>, <16 x i32>,

define <16 x i32>@test_int_x86_avx512_mask_cvt_ps2dq_512(<16 x float> %x0, <16 x i32> %x1, i16 %x2) #0 {
; CHECK-LABEL: @test_int_x86_avx512_mask_cvt_ps2dq_512(
; CHECK-NEXT: [[TMP10:%.*]] = load i16, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), align 8
; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i32>, ptr @__msan_param_tls, align 8
; CHECK-NEXT: [[TMP2:%.*]] = load <16 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
; CHECK-NEXT: call void @llvm.donothing()
; CHECK-NEXT: [[TMP3:%.*]] = bitcast i16 [[X2:%.*]] to <16 x i1>
; CHECK-NEXT: [[TMP4:%.*]] = icmp ne <16 x i32> [[TMP1]], zeroinitializer
; CHECK-NEXT: [[TMP5:%.*]] = sext <16 x i1> [[TMP4]] to <16 x i32>
; CHECK-NEXT: [[TMP6:%.*]] = select <16 x i1> [[TMP3]], <16 x i32> [[TMP5]], <16 x i32> [[TMP2]]
; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i16 [[TMP10]], 0
; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP11:%.*]], label [[TMP12:%.*]], !prof [[PROF1]]
; CHECK: 8:
; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR10]]
; CHECK-NEXT: unreachable
; CHECK: 9:
; CHECK-NEXT: [[RES:%.*]] = call <16 x i32> @llvm.x86.avx512.mask.cvtps2dq.512(<16 x float> [[X0:%.*]], <16 x i32> [[X1:%.*]], i16 [[X2]], i32 10)
; CHECK-NEXT: [[TMP7:%.*]] = icmp ne <16 x i32> [[TMP1]], zeroinitializer
; CHECK-NEXT: [[TMP8:%.*]] = sext <16 x i1> [[TMP7]] to <16 x i32>
Expand Down