Skip to content

[MSAN] handle AVX vpermi2var #143463

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 4 commits into from
Jun 13, 2025
Merged

[MSAN] handle AVX vpermi2var #143463

merged 4 commits into from
Jun 13, 2025

Conversation

fmayer
Copy link
Contributor

@fmayer fmayer commented Jun 10, 2025

No description provided.

@fmayer
[𝘀𝗽𝗿] initial version
4ed04b4 
Created using spr 1.3.4
@llvmbot
Copy link
Member

llvmbot commented Jun 10, 2025 •
edited
Loading

@llvm/pr-subscribers-llvm-transforms

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

Author: Florian Mayer (fmayer)

Changes

Patch is 71.59 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/143463.diff

3 Files Affected:

  • (modified) llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp (+39)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics-upgrade.ll (+148-112)
  • (modified) llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics.ll (+147-111)
diff --git a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
index 4d8dda7098aea..e05ea8b759e0b 100644
--- a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
+++ b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
@@ -4175,6 +4175,25 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
     setShadow(&I, IRB.CreateBitCast(CI, getShadowTy(&I)));
     setOriginForNaryOp(I);
   }
+  // Instrument AVX permutation intrinsic.
+  // We apply the same permutation (argument index 1) to the shadows.
+  void handleAVXVpermil2var(IntrinsicInst &I) {
+    IRBuilder<> IRB(&I);
+    Value *AShadow = getShadow(&I, 0);
+    Value *Idx = I.getArgOperand(1);
+    Value *BShadow = getShadow(&I, 2);
+    insertShadowCheck(Idx, &I);
+
+    // Shadows are integer-ish types but some intrinsics require a
+    // different (e.g., floating-point) type.
+    AShadow = IRB.CreateBitCast(AShadow, I.getArgOperand(0)->getType());
+    BShadow = IRB.CreateBitCast(BShadow, I.getArgOperand(2)->getType());
+    CallInst *CI = IRB.CreateIntrinsic(I.getType(), I.getIntrinsicID(),
+                                       {AShadow, Idx, BShadow});
+
+    setShadow(&I, IRB.CreateBitCast(CI, getShadowTy(&I)));
+    setOriginForNaryOp(I);
+  }
 
   // Instrument BMI / BMI2 intrinsics.
   // All of these intrinsics are Z = I(X, Y)
@@ -5130,6 +5149,26 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
       handleAVXVpermilvar(I);
       break;
     }
+    case Intrinsic::x86_avx512_vpermi2var_d_128:
+    case Intrinsic::x86_avx512_vpermi2var_d_256:
+    case Intrinsic::x86_avx512_vpermi2var_d_512:
+    case Intrinsic::x86_avx512_vpermi2var_hi_128:
+    case Intrinsic::x86_avx512_vpermi2var_hi_256:
+    case Intrinsic::x86_avx512_vpermi2var_hi_512:
+    case Intrinsic::x86_avx512_vpermi2var_pd_128:
+    case Intrinsic::x86_avx512_vpermi2var_pd_256:
+    case Intrinsic::x86_avx512_vpermi2var_pd_512:
+    case Intrinsic::x86_avx512_vpermi2var_ps_128:
+    case Intrinsic::x86_avx512_vpermi2var_ps_256:
+    case Intrinsic::x86_avx512_vpermi2var_ps_512:
+    case Intrinsic::x86_avx512_vpermi2var_q_128:
+    case Intrinsic::x86_avx512_vpermi2var_q_256:
+    case Intrinsic::x86_avx512_vpermi2var_q_512:
+    case Intrinsic::x86_avx512_vpermi2var_qi_128:
+    case Intrinsic::x86_avx512_vpermi2var_qi_256:
+    case Intrinsic::x86_avx512_vpermi2var_qi_512:
+      handleAVXVpermil2var(I);
+      break;
 
     case Intrinsic::x86_avx512fp16_mask_add_sh_round:
     case Intrinsic::x86_avx512fp16_mask_sub_sh_round:
diff --git a/llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics-upgrade.ll b/llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics-upgrade.ll
index 5aeaa1221cd21..96f82c4d49a0a 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics-upgrade.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/X86/avx512-intrinsics-upgrade.ll
@@ -13700,8 +13700,8 @@ define <16 x i32>@test_int_x86_avx512_vpermi2var_d_512(<16 x i32> %x0, <16 x i32
 ; CHECK-LABEL: @test_int_x86_avx512_vpermi2var_d_512(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), align 8
 ; CHECK-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @__msan_param_tls, align 8
-; CHECK-NEXT:    [[TMP3:%.*]] = load <16 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load <16 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 136) to ptr), align 8
+; CHECK-NEXT:    [[TMP14:%.*]] = 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:    [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
 ; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP5:%.*]], label [[TMP6:%.*]], !prof [[PROF1]]
@@ -13714,9 +13714,15 @@ define <16 x i32>@test_int_x86_avx512_vpermi2var_d_512(<16 x i32> %x0, <16 x i32
 ; CHECK-NEXT:    [[TMP8:%.*]] = xor i64 [[TMP7]], 87960930222080
 ; CHECK-NEXT:    [[TMP9:%.*]] = inttoptr i64 [[TMP8]] to ptr
 ; CHECK-NEXT:    [[_MSLD:%.*]] = load <16 x i32>, ptr [[TMP9]], align 64
-; CHECK-NEXT:    [[_MSPROP:%.*]] = or <16 x i32> [[TMP2]], [[TMP3]]
-; CHECK-NEXT:    [[_MSPROP1:%.*]] = or <16 x i32> [[_MSPROP]], [[TMP4]]
-; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[X0:%.*]], <16 x i32> [[X1:%.*]], <16 x i32> [[X4:%.*]])
+; CHECK-NEXT:    [[_MSPROP1:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[TMP2]], <16 x i32> [[X1:%.*]], <16 x i32> [[TMP4]])
+; CHECK-NEXT:    [[TMP11:%.*]] = bitcast <16 x i32> [[TMP14]] to i512
+; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP11]], 0
+; CHECK-NEXT:    br i1 [[_MSCMP1]], label [[TMP12:%.*]], label [[TMP13:%.*]], !prof [[PROF1]]
+; CHECK:       12:
+; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
+; CHECK-NEXT:    unreachable
+; CHECK:       13:
+; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[X0:%.*]], <16 x i32> [[X1]], <16 x i32> [[X4:%.*]])
 ; CHECK-NEXT:    store <16 x i32> [[_MSPROP1]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <16 x i32> [[TMP10]]
 ;
@@ -13744,9 +13750,15 @@ define <16 x i32>@test_int_x86_avx512_mask_vpermi2var_d_512(<16 x i32> %x0, <16
 ; CHECK-NEXT:    [[TMP8:%.*]] = xor i64 [[TMP7]], 87960930222080
 ; CHECK-NEXT:    [[TMP9:%.*]] = inttoptr i64 [[TMP8]] to ptr
 ; CHECK-NEXT:    [[_MSLD:%.*]] = load <16 x i32>, ptr [[TMP9]], align 64
-; CHECK-NEXT:    [[_MSPROP:%.*]] = or <16 x i32> [[TMP2]], [[TMP3]]
-; CHECK-NEXT:    [[_MSPROP1:%.*]] = or <16 x i32> [[_MSPROP]], [[_MSLD]]
-; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[X0:%.*]], <16 x i32> [[X1:%.*]], <16 x i32> [[X2]])
+; CHECK-NEXT:    [[_MSPROP1:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[TMP2]], <16 x i32> [[X1:%.*]], <16 x i32> [[_MSLD]])
+; CHECK-NEXT:    [[TMP18:%.*]] = bitcast <16 x i32> [[TMP3]] to i512
+; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP18]], 0
+; CHECK-NEXT:    br i1 [[_MSCMP1]], label [[TMP19:%.*]], label [[TMP20:%.*]], !prof [[PROF1]]
+; CHECK:       12:
+; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
+; CHECK-NEXT:    unreachable
+; CHECK:       13:
+; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32> [[X0:%.*]], <16 x i32> [[X1]], <16 x i32> [[X2]])
 ; CHECK-NEXT:    [[TMP11:%.*]] = bitcast i16 [[TMP4]] to <16 x i1>
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i16 [[X3:%.*]] to <16 x i1>
 ; CHECK-NEXT:    [[TMP13:%.*]] = select <16 x i1> [[TMP12]], <16 x i32> [[_MSPROP1]], <16 x i32> [[TMP3]]
@@ -13768,25 +13780,23 @@ declare <8 x double> @llvm.x86.avx512.mask.vpermi2var.pd.512(<8 x double>, <8 x
 define <8 x double>@test_int_x86_avx512_vpermi2var_pd_512(<8 x double> %x0, <8 x i64> %x1, <8 x double> %x2)  #0 {
 ; CHECK-LABEL: @test_int_x86_avx512_vpermi2var_pd_512(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @__msan_param_tls, align 8
-; CHECK-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
 ; CHECK-NEXT:    [[TMP3:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), align 8
+; CHECK-NEXT:    [[TMP8:%.*]] = load <8 x i64>, 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:    [[TMP4:%.*]] = bitcast <8 x i64> [[TMP1]] to i512
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i512 [[TMP4]], 0
-; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <8 x i64> [[TMP2]] to i512
-; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP5]], 0
-; CHECK-NEXT:    [[_MSOR:%.*]] = or i1 [[_MSCMP]], [[_MSCMP1]]
-; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <8 x i64> [[TMP3]] to i512
+; CHECK-NEXT:    [[TMP4:%.*]] = bitcast <8 x i64> [[TMP1]] to <8 x double>
+; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <8 x i64> [[TMP3]] to <8 x double>
+; CHECK-NEXT:    [[TMP11:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[TMP4]], <8 x i64> [[X1:%.*]], <8 x double> [[TMP5]])
+; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <8 x double> [[TMP11]] to <8 x i64>
+; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <8 x i64> [[TMP8]] to i512
 ; CHECK-NEXT:    [[_MSCMP2:%.*]] = icmp ne i512 [[TMP6]], 0
-; CHECK-NEXT:    [[_MSOR3:%.*]] = or i1 [[_MSOR]], [[_MSCMP2]]
-; CHECK-NEXT:    br i1 [[_MSOR3]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF1]]
-; CHECK:       7:
+; CHECK-NEXT:    br i1 [[_MSCMP2]], label [[TMP12:%.*]], label [[TMP13:%.*]], !prof [[PROF1]]
+; CHECK:       9:
 ; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
 ; CHECK-NEXT:    unreachable
-; CHECK:       8:
-; CHECK-NEXT:    [[TMP9:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[X0:%.*]], <8 x i64> [[X1:%.*]], <8 x double> [[X2:%.*]])
+; CHECK:       10:
+; CHECK-NEXT:    [[TMP9:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[X0:%.*]], <8 x i64> [[X1]], <8 x double> [[X2:%.*]])
 ; CHECK-NEXT:    [[TMP10:%.*]] = bitcast <8 x i64> [[X1]] to <8 x double>
-; CHECK-NEXT:    store <8 x i64> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <8 x i64> [[TMP7]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <8 x double> [[TMP9]]
 ;
   %res = call <8 x double> @llvm.x86.avx512.mask.vpermi2var.pd.512(<8 x double> %x0, <8 x i64> %x1, <8 x double> %x2, i8 -1)
@@ -13797,32 +13807,30 @@ define <8 x double>@test_int_x86_avx512_mask_vpermi2var_pd_512(<8 x double> %x0,
 ;
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_pd_512(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @__msan_param_tls, align 8
-; CHECK-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
 ; CHECK-NEXT:    [[TMP3:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), align 8
+; CHECK-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
 ; CHECK-NEXT:    [[TMP4:%.*]] = load i8, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 192) to ptr), align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <8 x i64> [[TMP1]] to i512
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i512 [[TMP5]], 0
-; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <8 x i64> [[TMP2]] to i512
-; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP6]], 0
-; CHECK-NEXT:    [[_MSOR:%.*]] = or i1 [[_MSCMP]], [[_MSCMP1]]
-; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <8 x i64> [[TMP3]] to i512
+; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <8 x i64> [[TMP1]] to <8 x double>
+; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <8 x i64> [[TMP3]] to <8 x double>
+; CHECK-NEXT:    [[TMP9:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[TMP5]], <8 x i64> [[X1:%.*]], <8 x double> [[TMP6]])
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast <8 x double> [[TMP9]] to <8 x i64>
+; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <8 x i64> [[TMP2]] to i512
 ; CHECK-NEXT:    [[_MSCMP2:%.*]] = icmp ne i512 [[TMP7]], 0
-; CHECK-NEXT:    [[_MSOR3:%.*]] = or i1 [[_MSOR]], [[_MSCMP2]]
-; CHECK-NEXT:    br i1 [[_MSOR3]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF1]]
-; CHECK:       8:
+; CHECK-NEXT:    br i1 [[_MSCMP2]], label [[TMP21:%.*]], label [[TMP22:%.*]], !prof [[PROF1]]
+; CHECK:       10:
 ; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
 ; CHECK-NEXT:    unreachable
-; CHECK:       9:
-; CHECK-NEXT:    [[TMP10:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[X0:%.*]], <8 x i64> [[X1:%.*]], <8 x double> [[X2:%.*]])
+; CHECK:       11:
+; CHECK-NEXT:    [[TMP10:%.*]] = call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> [[X0:%.*]], <8 x i64> [[X1]], <8 x double> [[X2:%.*]])
 ; CHECK-NEXT:    [[TMP11:%.*]] = bitcast <8 x i64> [[X1]] to <8 x double>
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i8 [[TMP4]] to <8 x i1>
 ; CHECK-NEXT:    [[TMP13:%.*]] = bitcast i8 [[X3:%.*]] to <8 x i1>
-; CHECK-NEXT:    [[TMP14:%.*]] = select <8 x i1> [[TMP13]], <8 x i64> zeroinitializer, <8 x i64> [[TMP2]]
+; CHECK-NEXT:    [[TMP14:%.*]] = select <8 x i1> [[TMP13]], <8 x i64> [[TMP8]], <8 x i64> [[TMP2]]
 ; CHECK-NEXT:    [[TMP15:%.*]] = bitcast <8 x double> [[TMP10]] to <8 x i64>
 ; CHECK-NEXT:    [[TMP16:%.*]] = bitcast <8 x double> [[TMP11]] to <8 x i64>
 ; CHECK-NEXT:    [[TMP17:%.*]] = xor <8 x i64> [[TMP15]], [[TMP16]]
-; CHECK-NEXT:    [[TMP18:%.*]] = or <8 x i64> [[TMP17]], zeroinitializer
+; CHECK-NEXT:    [[TMP18:%.*]] = or <8 x i64> [[TMP17]], [[TMP8]]
 ; CHECK-NEXT:    [[TMP19:%.*]] = or <8 x i64> [[TMP18]], [[TMP2]]
 ; CHECK-NEXT:    [[_MSPROP_SELECT:%.*]] = select <8 x i1> [[TMP12]], <8 x i64> [[TMP19]], <8 x i64> [[TMP14]]
 ; CHECK-NEXT:    [[TMP20:%.*]] = select <8 x i1> [[TMP13]], <8 x double> [[TMP10]], <8 x double> [[TMP11]]
@@ -13838,25 +13846,23 @@ declare <16 x float> @llvm.x86.avx512.mask.vpermi2var.ps.512(<16 x float>, <16 x
 define <16 x float>@test_int_x86_avx512_vpermi2var_ps_512(<16 x float> %x0, <16 x i32> %x1, <16 x float> %x2)  #0 {
 ; CHECK-LABEL: @test_int_x86_avx512_vpermi2var_ps_512(
 ; 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:    [[TMP3:%.*]] = load <16 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), align 8
+; CHECK-NEXT:    [[TMP8:%.*]] = 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:    [[TMP4:%.*]] = bitcast <16 x i32> [[TMP1]] to i512
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i512 [[TMP4]], 0
-; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <16 x i32> [[TMP2]] to i512
-; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP5]], 0
-; CHECK-NEXT:    [[_MSOR:%.*]] = or i1 [[_MSCMP]], [[_MSCMP1]]
-; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <16 x i32> [[TMP3]] to i512
+; CHECK-NEXT:    [[TMP4:%.*]] = bitcast <16 x i32> [[TMP1]] to <16 x float>
+; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <16 x i32> [[TMP3]] to <16 x float>
+; CHECK-NEXT:    [[TMP11:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[TMP4]], <16 x i32> [[X1:%.*]], <16 x float> [[TMP5]])
+; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <16 x float> [[TMP11]] to <16 x i32>
+; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <16 x i32> [[TMP8]] to i512
 ; CHECK-NEXT:    [[_MSCMP2:%.*]] = icmp ne i512 [[TMP6]], 0
-; CHECK-NEXT:    [[_MSOR3:%.*]] = or i1 [[_MSOR]], [[_MSCMP2]]
-; CHECK-NEXT:    br i1 [[_MSOR3]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF1]]
-; CHECK:       7:
+; CHECK-NEXT:    br i1 [[_MSCMP2]], label [[TMP12:%.*]], label [[TMP13:%.*]], !prof [[PROF1]]
+; CHECK:       9:
 ; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
 ; CHECK-NEXT:    unreachable
-; CHECK:       8:
-; CHECK-NEXT:    [[TMP9:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[X0:%.*]], <16 x i32> [[X1:%.*]], <16 x float> [[X2:%.*]])
+; CHECK:       10:
+; CHECK-NEXT:    [[TMP9:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[X0:%.*]], <16 x i32> [[X1]], <16 x float> [[X2:%.*]])
 ; CHECK-NEXT:    [[TMP10:%.*]] = bitcast <16 x i32> [[X1]] to <16 x float>
-; CHECK-NEXT:    store <16 x i32> zeroinitializer, ptr @__msan_retval_tls, align 8
+; CHECK-NEXT:    store <16 x i32> [[TMP7]], ptr @__msan_retval_tls, align 8
 ; CHECK-NEXT:    ret <16 x float> [[TMP9]]
 ;
   %res = call <16 x float> @llvm.x86.avx512.mask.vpermi2var.ps.512(<16 x float> %x0, <16 x i32> %x1, <16 x float> %x2, i16 -1)
@@ -13867,32 +13873,30 @@ define <16 x float>@test_int_x86_avx512_mask_vpermi2var_ps_512(<16 x float> %x0,
 ;
 ; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_ps_512(
 ; 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:    [[TMP3:%.*]] = load <16 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), 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:    [[TMP4:%.*]] = load i16, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 192) to ptr), align 8
 ; CHECK-NEXT:    call void @llvm.donothing()
-; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <16 x i32> [[TMP1]] to i512
-; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i512 [[TMP5]], 0
-; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <16 x i32> [[TMP2]] to i512
-; CHECK-NEXT:    [[_MSCMP1:%.*]] = icmp ne i512 [[TMP6]], 0
-; CHECK-NEXT:    [[_MSOR:%.*]] = or i1 [[_MSCMP]], [[_MSCMP1]]
-; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <16 x i32> [[TMP3]] to i512
+; CHECK-NEXT:    [[TMP5:%.*]] = bitcast <16 x i32> [[TMP1]] to <16 x float>
+; CHECK-NEXT:    [[TMP6:%.*]] = bitcast <16 x i32> [[TMP3]] to <16 x float>
+; CHECK-NEXT:    [[TMP9:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[TMP5]], <16 x i32> [[X1:%.*]], <16 x float> [[TMP6]])
+; CHECK-NEXT:    [[TMP8:%.*]] = bitcast <16 x float> [[TMP9]] to <16 x i32>
+; CHECK-NEXT:    [[TMP7:%.*]] = bitcast <16 x i32> [[TMP2]] to i512
 ; CHECK-NEXT:    [[_MSCMP2:%.*]] = icmp ne i512 [[TMP7]], 0
-; CHECK-NEXT:    [[_MSOR3:%.*]] = or i1 [[_MSOR]], [[_MSCMP2]]
-; CHECK-NEXT:    br i1 [[_MSOR3]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF1]]
-; CHECK:       8:
+; CHECK-NEXT:    br i1 [[_MSCMP2]], label [[TMP21:%.*]], label [[TMP22:%.*]], !prof [[PROF1]]
+; CHECK:       10:
 ; CHECK-NEXT:    call void @__msan_warning_noreturn() #[[ATTR8]]
 ; CHECK-NEXT:    unreachable
-; CHECK:       9:
-; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[X0:%.*]], <16 x i32> [[X1:%.*]], <16 x float> [[X2:%.*]])
+; CHECK:       11:
+; CHECK-NEXT:    [[TMP10:%.*]] = call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> [[X0:%.*]], <16 x i32> [[X1]], <16 x float> [[X2:%.*]])
 ; CHECK-NEXT:    [[TMP11:%.*]] = bitcast <16 x i32> [[X1]] to <16 x float>
 ; CHECK-NEXT:    [[TMP12:%.*]] = bitcast i16 [[TMP4]] to <16 x i1>
 ; CHECK-NEXT:    [[TMP13:%.*]] = bitcast i16 [[X3:%.*]] to <16 x i1>
-; CHECK-NEXT:    [[TMP14:%.*]] = select <16 x i1> [[TMP13]], <16 x i32> zeroinitializer, <16 x i32> [[TMP2]]
+; CHECK-NEXT:    [[TMP14:%.*]] = select <16 x i1> [[TMP13]], <16 x i32> [[TMP8]], <16 x i32> [[TMP2]]
 ; CHECK-NEXT:    [[TMP15:%.*]] = bitcast <16 x float> [[TMP10]] to <16 x i32>
 ; CHECK-NEXT:    [[TMP16:%.*]] = bitcast <16 x float> [[TMP11]] to <16 x i32>
 ; CHECK-NEXT:    [[TMP17:%.*]] = xor <16 x i32> [[TMP15]], [[TMP16]]
-; CHECK-NEXT:    [[TMP18:%.*]] = or <16 x i32> [[TMP17]], zeroinitializer
+; CHECK-NEXT:    [[TMP18:%.*]] = or <16 x i32> [[TMP17]], [[TMP8]]
 ; CHECK-NEXT:    [[TMP19:%.*]] = or <16 x i32> [[TMP18]], [[TMP2]]
 ; CHECK-NEXT:    [[_MSPROP_SELECT:%.*]] = select <16 x i1> [[TMP12]], <16 x i32> [[TMP19]], <16 x i32> [[TMP14]]
 ; CHECK-NEXT:    [[TMP20:%.*]] = select <16 x i1> [[TMP13]], <16 x float> [[TMP10]], <16 x float> [[TMP11]]
@@ -13908,12 +13912,18 @@ declare <8 x i64> @llvm.x86.avx512.mask.vpermi2var.q.512(<8 x i64>, <8 x i64>, <
 define <8 x i64>@test_int_x86_avx512_vpermi2var_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)  #0 {
 ; CHECK-LABEL: @test_int_x86_avx512_vpermi2var_q_512(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @__msan_param_tls, align 8
-; CHECK-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
 ; CHECK-NEXT:    [[TMP3:%.*]] = load <8 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 128) to ptr), a...
[truncated]

@fmayer fmayer requested a review from thurstond June 10, 2025 00:58
thurstond
thurstond requested changes Jun 10, 2025
View reviewed changes
Copy link
Contributor

@thurstond thurstond left a comment

Choose a reason for hiding this comment

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

Please add assertions and more tests (comments inline).

@fmayer
comments
99beb83 
Created using spr 1.3.4
@fmayer fmayer requested a review from thurstond June 12, 2025 21:11
@github-actions GitHub Actions
Copy link

github-actions bot commented Jun 12, 2025 •
edited
Loading

✅ With the latest revision this PR passed the C/C++ code formatter.

@fmayer
fmt
dd6d9e4 
Created using spr 1.3.4
thurstond
thurstond reviewed Jun 12, 2025
View reviewed changes
llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
void handleAVXVpermil2var(IntrinsicInst &I) {
assert(I.arg_size() == 3);
assert(I.getArgOperand(0)->getType() == I.getArgOperand(2)->getType());
assert(I.getType() == I.getArgOperand(0)->getType());
Copy link
Contributor

Choose a reason for hiding this comment

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

Please also check that I.getType() is a vector type, and that I.getArgOperand(0) has the same number of elements as I.getArgOperand(1)

@fmayer
assert
4a6142a 
Created using spr 1.3.4
thurstond
thurstond approved these changes Jun 13, 2025
View reviewed changes
llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
assert(cast<FixedVectorType>(I.getArgOperand(1)->getType())
->getNumElements() == ArgVectorSize);
assert(cast<FixedVectorType>(I.getArgOperand(2)->getType())
->getNumElements() == ArgVectorSize);
Copy link
Contributor

Choose a reason for hiding this comment

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

Some of the assertions are redundant (e.g., if operand 0 is a vector, and operand 0's type is the same as operand 2's type, then operand 2 must be a vector with the same number of elements as operand 0), but that's fine.

@fmayer fmayer merged commit b9d7d0a into users/fmayer/spr/main.msan-handle-avx-vpermi2var Jun 13, 2025
7 checks passed
@fmayer fmayer deleted the users/fmayer/spr/msan-handle-avx-vpermi2var branch June 13, 2025 22:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@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.
3 participants
@fmayer @llvmbot @thurstond