Skip to content

[msan] Add handling for sse41_round_pd/sse41_round_ps #118441

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

Jump to bottom
Merged
merged 1 commit into from
Dec 4, 2024
Merged

[msan] Add handling for sse41_round_pd/sse41_round_ps #118441

merged 1 commit into from
Dec 4, 2024

Conversation

alexander-shaposhnikov
Copy link
Collaborator

@alexander-shaposhnikov alexander-shaposhnikov commented Dec 3, 2024 •
edited
Loading

Add handling for sse41_round_pd/sse41_round_ps similarly to maybeHandleSimpleNomemIntrinsic.

Test plan: ninja check-all

@llvmbot
Copy link
Member

llvmbot commented Dec 3, 2024

@llvm/pr-subscribers-compiler-rt-sanitizer

Author: Alexander Shaposhnikov (alexander-shaposhnikov)

Changes

Add handling for sse41_round_pd/sse41_round_ps similarly to maybeHandleSimpleNomemIntrinsic.
Previously these intrinsics were processed as function calls and the inserted checks could produce false-positives.

Test plan: ninja check-all

Full diff: https://github.com/llvm/llvm-project/pull/118441.diff

3 Files Affected:

  • (modified) llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp (+21)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll (+2-16)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll (+2-16)
diff --git a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
index dca6bf1adfde8a..0663dbd90b4150 100644
--- a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
+++ b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
@@ -3832,6 +3832,21 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     setOriginForNaryOp(I);
   }
 
+  // _mm_round_ps / _mm_round_ps.
+  // Similar to maybeHandleSimpleNomemIntrinsic except
+  // the second argument is guranteed to be a constant integer.
+  void handleRoundPdPsIntrinsic(IntrinsicInst &I) {
+    assert(I.getArgOperand(0)->getType() == I.getType());
+    unsigned NumArgOperands = I.arg_size();
+    assert(NumArgOperands == 2);
+    assert(isa<ConstantInt>(I.getArgOperand(1)));
+
+    IRBuilder<> IRB(&I);
+    ShadowAndOriginCombiner SC(this, IRB);
+    SC.Add(I.getArgOperand(0));
+    SC.Done(&I);
+  }
+
   // Instrument abs intrinsic.
   // handleUnknownIntrinsic can't handle it because of the last
   // is_int_min_poison argument which does not match the result type.
@@ -4327,10 +4342,16 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
       handlePclmulIntrinsic(I);
       break;
 
+    case Intrinsic::x86_sse41_round_pd:
+    case Intrinsic::x86_sse41_round_ps:
+      handleRoundPdPsIntrinsic(I);
+      break;
+
     case Intrinsic::x86_sse41_round_sd:
     case Intrinsic::x86_sse41_round_ss:
       handleUnarySdSsIntrinsic(I);
       break;
+
     case Intrinsic::x86_sse2_max_sd:
     case Intrinsic::x86_sse_max_ss:
     case Intrinsic::x86_sse2_min_sd:
diff --git a/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll b/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
index f5e8452fd0d585..1fcab72d571ea7 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
@@ -320,15 +320,8 @@ define <2 x double> @test_x86_sse41_round_pd(<2 x double> %a0) #0 {
 ; CHECK-LABEL: @test_x86_sse41_round_pd(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP3:%.*]], label [[TMP4:%.*]], !prof [[PROF1]]
-; CHECK:       3:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       4:
 ; CHECK-NEXT:    [[RES:%.*]] = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <2 x i64> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <2 x i64> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <2 x double> [[RES]]
 ;
   %res = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %a0, i32 7) ; <<2 x double>> [#uses=1]
@@ -341,15 +334,8 @@ define <4 x float> @test_x86_sse41_round_ps(<4 x float> %a0) #0 {
 ; CHECK-LABEL: @test_x86_sse41_round_ps(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP3:%.*]], label [[TMP4:%.*]], !prof [[PROF1]]
-; CHECK:       3:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       4:
 ; CHECK-NEXT:    [[RES:%.*]] = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <4 x i32> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <4 x float> [[RES]]
 ;
   %res = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %a0, i32 7) ; <<4 x float>> [#uses=1]
diff --git a/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll b/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
index a3bda12e395e6c..e51c53375d2b5e 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
@@ -335,15 +335,8 @@ define <2 x double> @test_x86_sse41_round_pd(<2 x double> %a0) #0 {
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr @__msan_va_arg_overflow_size_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP6:%.*]], label [[TMP5:%.*]], !prof [[PROF1]]
-; CHECK:       4:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       5:
 ; CHECK-NEXT:    [[RES:%.*]] = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <2 x i64> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <2 x i64> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <2 x double> [[RES]]
 ;
   %res = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %a0, i32 7) ; <<2 x double>> [#uses=1]
@@ -357,15 +350,8 @@ define <4 x float> @test_x86_sse41_round_ps(<4 x float> %a0) #0 {
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr @__msan_va_arg_overflow_size_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP6:%.*]], label [[TMP5:%.*]], !prof [[PROF1]]
-; CHECK:       4:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       5:
 ; CHECK-NEXT:    [[RES:%.*]] = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <4 x i32> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <4 x float> [[RES]]
 ;
   %res = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %a0, i32 7) ; <<4 x float>> [#uses=1]

@llvmbot
Copy link
Member

llvmbot commented Dec 3, 2024

@llvm/pr-subscribers-llvm-transforms

Author: Alexander Shaposhnikov (alexander-shaposhnikov)

Changes

Add handling for sse41_round_pd/sse41_round_ps similarly to maybeHandleSimpleNomemIntrinsic.
Previously these intrinsics were processed as function calls and the inserted checks could produce false-positives.

Test plan: ninja check-all

Full diff: https://github.com/llvm/llvm-project/pull/118441.diff

3 Files Affected:

  • (modified) llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp (+21)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll (+2-16)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll (+2-16)
diff --git a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
index dca6bf1adfde8a..0663dbd90b4150 100644
--- a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
+++ b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
@@ -3832,6 +3832,21 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     setOriginForNaryOp(I);
   }
 
+  // _mm_round_ps / _mm_round_ps.
+  // Similar to maybeHandleSimpleNomemIntrinsic except
+  // the second argument is guranteed to be a constant integer.
+  void handleRoundPdPsIntrinsic(IntrinsicInst &I) {
+    assert(I.getArgOperand(0)->getType() == I.getType());
+    unsigned NumArgOperands = I.arg_size();
+    assert(NumArgOperands == 2);
+    assert(isa<ConstantInt>(I.getArgOperand(1)));
+
+    IRBuilder<> IRB(&I);
+    ShadowAndOriginCombiner SC(this, IRB);
+    SC.Add(I.getArgOperand(0));
+    SC.Done(&I);
+  }
+
   // Instrument abs intrinsic.
   // handleUnknownIntrinsic can't handle it because of the last
   // is_int_min_poison argument which does not match the result type.
@@ -4327,10 +4342,16 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
       handlePclmulIntrinsic(I);
       break;
 
+    case Intrinsic::x86_sse41_round_pd:
+    case Intrinsic::x86_sse41_round_ps:
+      handleRoundPdPsIntrinsic(I);
+      break;
+
     case Intrinsic::x86_sse41_round_sd:
     case Intrinsic::x86_sse41_round_ss:
       handleUnarySdSsIntrinsic(I);
       break;
+
     case Intrinsic::x86_sse2_max_sd:
     case Intrinsic::x86_sse_max_ss:
     case Intrinsic::x86_sse2_min_sd:
diff --git a/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll b/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
index f5e8452fd0d585..1fcab72d571ea7 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/X86/sse41-intrinsics-x86.ll
@@ -320,15 +320,8 @@ define <2 x double> @test_x86_sse41_round_pd(<2 x double> %a0) #0 {
 ; CHECK-LABEL: @test_x86_sse41_round_pd(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP3:%.*]], label [[TMP4:%.*]], !prof [[PROF1]]
-; CHECK:       3:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       4:
 ; CHECK-NEXT:    [[RES:%.*]] = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <2 x i64> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <2 x i64> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <2 x double> [[RES]]
 ;
   %res = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %a0, i32 7) ; <<2 x double>> [#uses=1]
@@ -341,15 +334,8 @@ define <4 x float> @test_x86_sse41_round_ps(<4 x float> %a0) #0 {
 ; CHECK-LABEL: @test_x86_sse41_round_ps(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP3:%.*]], label [[TMP4:%.*]], !prof [[PROF1]]
-; CHECK:       3:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       4:
 ; CHECK-NEXT:    [[RES:%.*]] = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <4 x i32> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <4 x float> [[RES]]
 ;
   %res = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %a0, i32 7) ; <<4 x float>> [#uses=1]
diff --git a/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll b/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
index a3bda12e395e6c..e51c53375d2b5e 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/i386/sse41-intrinsics-i386.ll
@@ -335,15 +335,8 @@ define <2 x double> @test_x86_sse41_round_pd(<2 x double> %a0) #0 {
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr @__msan_va_arg_overflow_size_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <2 x i64> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP6:%.*]], label [[TMP5:%.*]], !prof [[PROF1]]
-; CHECK:       4:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       5:
 ; CHECK-NEXT:    [[RES:%.*]] = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <2 x i64> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <2 x i64> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <2 x double> [[RES]]
 ;
   %res = call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %a0, i32 7) ; <<2 x double>> [#uses=1]
@@ -357,15 +350,8 @@ define <4 x float> @test_x86_sse41_round_ps(<4 x float> %a0) #0 {
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr @__msan_va_arg_overflow_size_tls, align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to i128
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i128 [[TMP2]], 0
-; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP6:%.*]], label [[TMP5:%.*]], !prof [[PROF1]]
-; CHECK:       4:
-; CHECK-NEXT:    call void @__msan_warning_noreturn()
-; CHECK-NEXT:    unreachable
-; CHECK:       5:
 ; CHECK-NEXT:    [[RES:%.*]] = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> [[A0:%.*]], i32 7)
-; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <4 x i32> [[TMP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <4 x float> [[RES]]
 ;
   %res = call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %a0, i32 7) ; <<4 x float>> [#uses=1]

@vitalybuka vitalybuka requested a review from thurstond December 4, 2024 06:36
vitalybuka
vitalybuka approved these changes Dec 4, 2024
View reviewed changes
Copy link
Collaborator

@vitalybuka vitalybuka left a comment

Choose a reason for hiding this comment

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

LGTM

@vitalybuka
Copy link
Collaborator

@thurstond FYI

I guess it's OK to merge

@alexander-shaposhnikov alexander-shaposhnikov merged commit 95e44d3 into llvm:main Dec 4, 2024
11 checks passed
@alexander-shaposhnikov alexander-shaposhnikov mentioned this pull request Dec 10, 2024
Merged
alexander-shaposhnikov added a commit that referenced this pull request Dec 10, 2024
@alexander-shaposhnikov
[msan] Add support for avx_round_pd_256/avx_round_ps_256 (#119334)
3791323 
Add support for avx_round_pd_256/avx_round_ps_256.
This is a follow-up to #118441

Test plan:
ninja check-all
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@vitalybuka vitalybuka vitalybuka approved these changes

@thurstond thurstond thurstond approved these changes

Assignees
No one assigned
Labels
compiler-rt:sanitizer llvm:transforms
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@alexander-shaposhnikov @llvmbot @vitalybuka @thurstond